text
stringlengths 4
2.52M
| meta
dict |
|---|---|
Background
==========
The relevance of solar power in the renewable energy field is constantly increasing due to its ready availability and to the fact that the available amount exceeds by several orders of magnitude the needs of the human race. The search for new materials with better performances than standard Si-based solar cells is also constantly increasing. Organic materials \[[@B1]\] emerged as a very attractive solution for this scope, their lower efficiencies with respect to inorganic materials being compensated by lower fabrication costs and higher flexibility. Hybrid materials have also been investigated as a way to combine the low production costs of organic materials with the high efficiency of inorganic materials \[[@B2]-[@B8]\].
Among the adopted strategies for new materials, interface geometry often plays a major role in the collection of photogenerated carriers, and bulk heterojunctions \[[@B3],[@B9]-[@B12]\] - intimately mixing the two junction materials while keeping them separate in a 'fractal like' high-surface interface - are a very promising design for solar cells. This concept was introduced in the mid 1990s for organic solar cells \[[@B5],[@B13]-[@B15]\] to shorten the exciton travel distance from the photon absorption site towards the charge-separating interface and then both to reduce the spontaneous recombination and to increase the collection efficiency. The very large interfacial area available for charge separation processes also increases the carrier collection efficiency.
We investigate here a new hybrid material for photovoltaic applications composed by *n*-type porous silicon (PSi) and eumelanin, a natural pigment featuring relatively high electrical conductivity \[[@B16]\] and believed to rely mainly on proton-based conduction \[[@B16]-[@B18]\]. Porous Si is a large specific area material, whose properties depend on Si substrate doping and on fabrication parameters \[[@B19]\]. Its application span ranges from biosensor \[[@B19]\] to drug delivery \[[@B20]\] and optoelectronics \[[@B21]\]. In the photovoltaic field, PSi has been considered up to now mainly as an antireflection coating for crystalline Si \[[@B22]\], and there are very few studies about its photovoltaic properties \[[@B23],[@B24]\]. Porous Si-organic hybrids have recently been considered \[[@B25]\], but literature reports are mainly on amorphous \[[@B26]\] or crystalline \[[@B27],[@B28]\] Si. An exception are the interesting results by Nahor et al. \[[@B25]\] reporting a study of a PSi-organic hybrid material realized using conjugate polymers for solar cells, which highlight the potential of PSi-based bulk heterojunctions.
Melanins are a class of natural pigments responsible for the colorations of human skin and hair \[[@B29]\] in the range from light, yellow-reddish (pheomelanins) and dark, brownish black (eumelanins). Their unique status among natural pigments is due to their socioeconomic and biomedical relevance, encompassing racial pigmentation, skin photoprotection, sun tanning, and pigmentary disorders. Moreover, they display a quite unusual set of physicochemical properties such as broadband monotonic absorption in the ultraviolet-visible \[[@B18]\].
These features have suggested the possible use of synthetic (artificial) eumelanins in the development of a new generation of bioinspired electrically active devices \[[@B30]-[@B32]\]. More recently, eumelanin biopolymers have also been proposed for optoelectronic and photovoltaic applications \[[@B33],[@B34]\]. Due to their wide absorbance covering the whole visible light spectrum, eumelanins behave as very efficient photoreceptors \[[@B31]\].
As a part of a large project aimed to assess the scope of melanins and melanin-like materials for the development of novel hybrid functional architectures \[[@B35],[@B36]\], we investigate here '5,6-dihydroxyindole-melanin immobilized PSi' as a prototypal device for the generation of white light-induced photocurrent.
Methods
=======
Porous silicon
--------------
PSi samples were prepared by electrochemical etch in the dark of (100)-oriented, phosphorous doped, *n*^+^-type monocrystalline Si wafers from Siltronix (Archamps Technopole, Archamps, France) using the procedure described in \[[@B37]\]. The resistivity of the wafers is in the 0.007 to 0.003 ohm/cm range. The constant current etching process \[[@B38],[@B39]\] has been performed in a polyvinylchloride electrochemical cell using HF:H~2~O:ethanol solution in 15:15:70 percent, respectively. The potential source was a PARSTAT 2273 potentiostat from Princeton Applied Research (Oak Ridge, TN, USA). The various components of the fabrication cell are schematically shown in Figure [1](#F1){ref-type="fig"}. The silicon substrate acts as the working electrode and a platinum wire or grid is used as a counter electrode.
{#F1}
All samples considered have a nominal 5 μm thickness. The relation between the samples\' thickness and their respective fabrication time has been studied by means of scanning electron microscopy (SEM). In Figure [2](#F2){ref-type="fig"} we show an example of thickness measurement performed on a 29 μm-thick PSi sample. The scale is shown at the bottom of the image. The porous layer (comprised between the two orange lines) and the bulk Si substrate are indicated, together with the measured thickness value. The thickness measurements proved to be fully reproducible for any thickness below or equal to 29 μm (the thickest sample tested) and directly proportional to the formation time. The samples\' porosity, measured by gravimetry, was 55 % (empty/full ratio) for all samples. A few samples have been electrochemically oxidized, in the same cell used for the formation process, using a 0.1 M aqueous solution of KNO~3~ and a constant current *I* = −5 mA. The optical reflectivity measurements were performed using a PerkinElmer (Perkin Elmer Italia SpA, Monza (MI), Italy) Lambda 950 spectrometer equipped with the Universal Reflectance Accessory using an incidence angle of 8°.
{#F2}
Eumelanin
---------
Eumelanin formation starts from a tyrosinase-catalyzed oxidation of tyrosine (Figure [3](#F3){ref-type="fig"}) in a multistep process up to 5,6-dihydroxyindole (DHI) and 5,6 dihydroxyindole-2-carboxylic acid. Oxidative polymerization of these indoles then gives rise to the black-brown variety of melanin biopolymer: the eumelanin \[[@B29]\].
![**Schematic path of eumelanin formation**\[[@B29]\]**and proposed hierarchical structure of the pigment.** Early DHI oligomers are shown \[[@B40]\].](1556-276X-7-377-3){#F3}
In our experiments, DHI \[[@B41]\] and synthetic eumelanin \[[@B42]\] were prepared according to the literature: a solution of DHI (50 mmol) in phosphate buffer, 0.1 M, pH 7.4, (10 ml) was treated with tyrosinase (400 units) under a stream of oxygen for 4 h at 25 °C, then acidified to pH 4.5 and washed first with distilled water and then in methanol/water 7/3.
Eumelanin immobilization on porous Si was obtained by treating the substrates with the appropriate solution/suspension of DHI or synthetic melanin: methanol, methanol/water, or methanol/phosphate buffer pH 7.0 were used as liquid phase. The substrates were processed with three cycles, 5 min each, in ultrasound bath, followed by 4 h oxygen exposition. When suitable, tyrosinase was also added to the mixture to promote oxidation process.
Photocurrent measuring procedure
--------------------------------
Electrical contacts were realized by deposition of gold spots on top of the empty and impregnated porous layers by sputtering using an Emitech K450 sputter coater (Quorum Technologies Ltd, East Grinstead, West Sussex, UK). On each sample, four contacts were realized to test the reproducibility of the procedure.
The samples\' photoconductivity was tested using a PM8 Analytical prober and a Keithley multimeter (Keithley Instruments Inc., Cleveland, OH, USA). The light source was a tungsten-halogen lamp, whose spectral range at the output of the optical system was in the 400 to 850 nm interval. The active external surface of the samples involved in the photocurrent generation is estimated of the order of a squared millimeter.
Results and discussion
======================
Different approaches for the immobilization of eumelanin in the porous matrix have been explored, including treatment of the porous Si with preformed synthetic eumelanin and *in situ*-induced oxidative polymerization of DHI.
To gain insight into the mode of coupling (Figure [3](#F3){ref-type="fig"}) of DHI during polymer buildup in this procedure, which is important to match the basic structural features of the immobilized DHI-derived polymer \[[@B30],[@B32]\] and to validate the procedure, mother liquors were collected and the reaction was stopped in the early stages by the addition of sodium dithionite to reduce the oxidized species. The crude oligomer-containing mixture was acetylated according to an established protocol \[[@B40]\] and examined by thin layer chromatography (TLC). This treatment of the mixture allowed the isolation of two main eumelanin oligomer intermediates, which were identified as the acetylated 2,7′- and 2,4′-biindolyls.
The oligomer identification has been performed using the protocols reported in \[[@B40]\] and \[[@B43]\]. In detail, after 2 min of the substrate treatment, the liquors were removed and the oxidation was halted by the addition of 5 % *w/w* aqueous solution of sodium dithionite and acidified to pH 4 with 3 M HCl. The reaction mixture was extracted repeatedly with ethyl acetate (3 × 250 mL), and the combined organic layers were dried over sodium sulfate and taken to dryness.
The residue was acetylated with acetic anhydride-pyridine 95:5 (*v/v*) and fractionated by preparative TLC (CHCl~3~/MeOH 98:2) to identify the constituents by comparison with synthetic standards with known structure.
The optical reflectivity measurements were performed on both empty and impregnated PSi matrices to explore the impregnation procedure. In Figure [4](#F4){ref-type="fig"} we show optical reflectivity spectra obtained on two PSi samples, where the thin-layer Fabry-Pérot interference fringes are clearly visible. The red curve corresponds to an impregnated PSi sample and the blue curve to an empty one with equal thickness. The discontinuity for wavelengths longer than 890 nm in the red curve is due to instrumental adjustment routine at the detector change. There are two main differences between the two curves. The first one is in the intensity of the interference fringes. The fact that the impregnated sample shows less intense fringes can be attributed to an increase of the absorption coefficient of the layer \[[@B44]\]. The second difference between the two curves is the higher number of fringes per wavelength range for the impregnated sample. This increase can be correlated to an increase of the optical thickness of the sample \[[@B44]\]. Given that the two samples have the same nominal thickness, this difference is clearly related mainly to the change in the refractive index.
{#F4}
In Figure [5](#F5){ref-type="fig"} we show the dependence of the interference peak maxima number on the inverse of the wavelength for the two samples of Figure [4](#F4){ref-type="fig"}, the color code being also the same. If the refractive index had been constant, this plot would have been a straight line \[[@B45]\]. In our case, however, considering the large spectral range explored, the refractive index is not constant. For this reason, the curves have been fitted separately for the higher (left) and lower (right) energy sides of the spectra. Since the slopes of the linear fits related to the impregnated sample are always higher than those related to the empty sample, the results of Figure [4](#F4){ref-type="fig"} demonstrate that the refractive index of the impregnated PSi layer is higher than that of the empty PSi layer in the whole spectral range examined. From the results of Figures [4](#F4){ref-type="fig"} and [5](#F5){ref-type="fig"}, we can then conclude that the eumelanin impregnation process leads to an increase of both the absorption coefficient and the refractive index of the PSi matrix. Moreover, the absence of beats in the interference fringes of the impregnated layer indicates that the impregnation process is quite homogeneous throughout the whole PSi thickness and that there is no double layer given by a partial, depth-limited, eumelanin pore penetration.
{ref-type="fig"}**.** Blue triangles refer to the empty PSi layer and red circles to the eumelanin-impregnated PSi layer. The linear fits of the experimental data are also shown for the lower (right) and higher (left) energy sides of the explored spectral range.](1556-276X-7-377-5){#F5}
These results are in agreement with the reported literature absorption coefficients of eumelanin and PSi. To show this, in Figure [6](#F6){ref-type="fig"}, we plot the absorption coefficients of 78 % *p*^+^-type and 58 % *p*-type porous Si \[[@B44]\], bulk Si \[[@B46]\], and eumelanin \[[@B16]\]. The absorption coefficients dispersion curves for eumelanin have been obtained from thin (80 nm) and thick (800 nm) film. All symbols are explained in the figure caption.
![**Literature absorption coefficients dispersion curves for PSi, Si and eumelanin.** Comparison of the absorption coefficients dispersion curves reported in literature for several materials: eumelanin film, thin (80 nm, full squares), and thick (800 nm, empty squares), \[[@B16]\]; 58 % porosity *p*-type (inverted triangles) and 78 % *p*^+^-type (upright triangles) porous Si, \[[@B44]\]; and bulk Si (full circles) \[[@B46]\]. The plot shows how the absorption coefficient of eumelanin is significantly larger than that of PSi, becoming more than one order of magnitude larger for wavelengths longer than about 650 nm. It is worth noting that this is also valid with respect to the undoped bulk Si.](1556-276X-7-377-6){#F6}
The data shown in Figure [6](#F6){ref-type="fig"} clearly evidenced that the optical absorption coefficient of both thin and thick eumelanin films is significantly greater than that of PSi and even bulk silicon for photon energies lower than 2.5 eV (500 nm). For wavelengths longer than 600 to 650 nm, the eumelanin absorption coefficient is still significant (well above than 10^4^ cm^−1^) and more than a factor of ten larger than that of PSi whose absolute value becomes less than 10^3^ cm^−1^. This is where the effect of eumelanin may be expected to generate a more significant difference in the photoconductive behavior of empty and impregnated porous layers. It is important to note that from \[[@B16]\], the eumelanin absorption coefficient at 1,400 nm is still more than 6,000 cm^−1^ for thin films and 8,000 cm^−1^ for thick films.
The photoconductive properties of porous Si samples were studied with and without eumelanin for unoxidized layers. A few samples were oxidized at several oxidation levels. This test has been done because, although the partial oxidation process may reduce the PSi conductivity, the oxidation-induced modification of the eumelanin adhesion to the pore walls could, in principle, more than counterbalance the effect. In all cases, however, all the partially oxidized samples showed no measurable photosensitivity. Whatever the impinging light intensity value (up to a maximum of about 200 W/m^2^) is, with and without eumelanin, our results conclusively show that even a thin oxide layer on the pore\'s wall is sufficient to severely limit the layers\' photosensitivity.
All non-oxidized samples, for all eumelanin immobilization approaches considered here, showed a marked photosensitivity. When illuminating with the whole lamp spectrum, the variation in the photocurrent intensity from very low ambient light to the maximum impinging light intensity showed an increase of more than three orders of magnitude, fully reproducible for the same gold plot. However, while the wavelength dependence was highly reproducible, as discussed later in more detail, a variability was observed in the maximum measured photocurrent when using different gold contacts on the same sample and/or on different samples, even if nominally identical. At the same time, the maximum photogenerated voltage showed a much lower variation, being between 100 and 150 mV for impregnated samples and between 120 and 200 mV for empty samples, suggesting a contact-related and non material-related issue. Empty PSi layers showed more stable results with respect to the impregnated samples. The maximum absolute photocurrent values measured with full illumination for impregnated and empty PSi were within about a factor of two, while the photovoltage measurement were much closer (5.1 mA/200 mV for empty PSi and 2.2 mA/150 mV for impregnated samples). The observed variability did not appear to depend on the eumelanin immobilization procedure.
To better evaluate the effect of the eumelanin insertion on the photoconductive properties of the samples and to discriminate possible contact issues from the samples\' behavior, we measured the wavelength dependence of the light response of impregnated and empty porous Si samples using low-pass optical filters in front of the samples. By this approach, using a given filter the light arriving onto the samples\' surface will only have wavelengths longer than the cutoff filter\'s. This means that the photocurrent measured with each filter will be given by all photocarriers generated by the interaction of the samples with the residual spectral range, that is, the part whose photon energy is lower than that characteristic of the cutoff filter. Consequently, if the carrier photogeneration occurs in the whole source spectrum, the plot of the photocurrent as a function of the cutoff wavelengths is expected to be a monotone curve which is increasing when moving towards shorter wavelengths. The results obtained on typical samples in our experiments are shown in Figure [7](#F7){ref-type="fig"}. The samples indicated as 'empty PSi 1' and 'impregnated PSi 1' are the same samples whose optical reflectivity spectra have been shown in Figure [4](#F4){ref-type="fig"}. The absolute current values are comparable (*I*~max~ = 0.8 μA for the two empty PSi layers and *I*~max~ = 0.9 μA for the impregnated sample labeled impregnated PSi 1) for three of the samples shown, while for one of the impregnated samples, the maximum absolute photocurrent is significantly lower (*I*~max~ = 0.09 μA, *V*~max~ = 100 mV with full spectral range illumination).
{#F7}
To separate the contact-related issues from the photogeneration behavior of the samples, the photocurrent intensity and the wavelength dependent photogeneration behavior must be separated. If the behavior of the samples would prove to be the same with respect to the maximum recorded photocurrent, it would be possible to attribute the intensity differences from sample to sample essentially to contact-related issues and not to the hybrid material building procedure adopted. Accordingly, we show in Figure [8](#F8){ref-type="fig"} the photocurrents of the two impregnated PSi samples of Figure [7](#F7){ref-type="fig"}, normalized by dividing each value by the maximum measured photocurrent of that sample. As in Figure [7](#F7){ref-type="fig"}, the normalized photocurrents are plotted vs. the low-pass filters cutoff wavelengths. Please note, however, that the peak at 500 nm of the impregnated PSi 1 sample, being clearly due to signal noise, has not been used as a reference for the normalization.
{ref-type="fig"}**.** It is worth noting the almost complete superposition of the two behaviors even if the absolute photocurrent values are significantly different.](1556-276X-7-377-8){#F8}
The striking feature of Figure [8](#F8){ref-type="fig"} is the almost complete superposition of the two curves, indicating that the relative behavior of the two impregnated samples is the same, despite the factor of 20 within their maximum values. This clearly indicates, as stated above, that the observed photocurrent intensity fluctuation is fundamentally related to the contact fabrication process and does not point to a poor reproducibility of the hybrid material behavior.
From the results shown in Figures [7](#F7){ref-type="fig"} and [8](#F8){ref-type="fig"}, there are several considerations that can be made. First, all curves show a monotonic photocurrent increase when going from longer to shorter wavelengths, indicating that the absorption takes place in the whole available light spectrum for both impregnated and empty PSi layers. The second significant feature is that impregnated samples show a steep photocurrent increase in the lower energy part of the spectrum. This is coherent with what can be expected by the comparison of the optical absorption coefficients of the materials, as discussed earlier in this work. If we compare empty and impregnated samples with similar maximum photocurrent, about 40 % of the total photocurrent is generated, for impregnated samples, when using wavelengths longer than 780 nm; while for empty PSi layers, the same wavelength range gives only about 15 % of the maximum photocurrent.
The latter effect is even more significant if we keep in mind the very low intensity of the impinging light in the 700 to 850 nm range: for the impregnated samples, 40 % of the total photocurrent is obtained with a very limited part of the total spectrum; while to obtain the same amount of photocurrent with the empty PSi samples, we need to increase the wavelength range up to 550 nm, including most of the available lamp spectral range.
These results show how the impregnation of the PSi matrix with eumelanin significantly increases the capability of the layer to efficiently photogenerate carriers from light especially when approaching the infrared region.
Conclusions
===========
We have shown that the photovoltaic properties of PSi may be significantly improved by impregnation with eumelanin. In particular, we showed that introducing the pigment in the porous Si matrix leads to a significantly more efficient photocurrent generation in the lower energy part of the experimental wavelength range explored with respect to empty porous Si layers.
This result not only contributes to expand the scope of heterojunctions in developing a new hybrid material but also provides the first evidence of the eumelanin capability to efficiently collect the photon energy. In a given substrate, eumelanin seems to act as a kind of 'antenna', modifying the range of the useful wavelength range for photoconversion application.
Although further experimental and theoretical studies are needed to improve the electrical contacts reproducibility and to reach a deeper understanding of the observed behavior in view of future development and applications, the proof of principle device presented here opens a new window into the evolving panorama of eumelanin-based devices and contributes to the development of bioinspired and biocompatible optoelectronic devices.
Competing interests
===================
The authors declare that they have no competing interests.
Authors' contributions
======================
GM conceived of the study, participated in its design and coordination, prepared part of the samples and performed the reflectivity measurements and analysis. LM performed the photoconductivity measurements and prepared part of the samples. SS participated in the design of the study. AP participated in the design of the study, carried out the eumelanin preparation and performed the PSi impregnation process with eumelanin. All authors contributed to the data analysis. All authors read and approved the final manuscript.
Acknowledgements
================
One of the authors (AP) acknowledges partial financial support by Ministry of Education, Universities and Research (MIUR, Italy), PRIN 2008 project. Andrea Falqui from the Italian Institute of Technology, Genova, Italy is gratefully acknowledged for the SEM measurements.
| {
"pile_set_name": "PubMed Central"
} |
Background
==========
*Dicistroviridae* is a recently established family of small, non-enveloped viruses, containing a +ssRNA genome, and is classified under the order *Picornavirales*. This family contains 14 members classified within two genera, *Cripavirus* (type species *Cricket paralysis virus*, CrPV), and *Aparavirus* (type species *Acute bee paralysis virus*, ACP) \[[@B1],[@B2]\]. A third genus has been recently proposed, *Triatovirus* (type species *Triatoma virus*, TrV) \[[@B3]\]. All dicistroviruses are pathogenic to arthropods, although primarily to insects, representing significant threats to the health of beneficial arthropods such as *acute bee paralysis virus*, *Black queen cell virus*, *Kashmir bee virus*, which infects honeybees \[[@B4]-[@B6]\], and *Taura syndrome virus*, which is pathogenic to shrimps \[[@B7]\]. Two dicistrovirus members infect the model organism *Drosophila melanogaster*, *Drosophila C virus* (DCV) and CrPV \[[@B8],[@B9]\], and eight of them are pathogenic to insect pests, as it is the case with CrPV, which infects field crickets and the *Olive Fruit Fly* as well \[[@B7]\]. The aforementioned CrPV is quite ubiquitous and replicates in many other insect species spanning five orders \[[@B10]\]. Although dicistroviruses are believed to be arthropod-specific, antibodies against CrPV have been detected in sera from a pig, horse, and cattle \[[@B11]\]. Similarly, high levels of anti-TrV antibodies were detected in chickens used to feed a colony of TrV-infected insects, although the authors of this work concluded that the chickens were apparently refractory to the infection with TrV \[[@B12]\].
Compared to other families of picornaviruses, little is known about dicistrovirus infection, replication mechanism or gene function. Most information about cellular infection and the replication cycle of dicistroviruses comes from studies on DCV, and by comparison with mammalian picornaviruses \[[@B1]\]. Many members of the *Dicistroviridae* family are considered novel candidates to be used as biopesticides \[[@B1],[@B13]-[@B17]\].
TrV and *Solenopsis invicta virus*-1 are the two lone members of the *Dicistroviridae* family that infect insects of medical importance \[[@B1]\]. In fact, TrV is a natural enemy of *Triatoma infestans* (*Hemiptera: Reduviidae*), one of the main vectors of the trypanosomiasis disease called Chagas disease, a severe human illness most prevalent in almost all Latin American countries \[[@B18],[@B19]\]. To date, TrV is the only entomopathogenic virus found in triatomines \[[@B20]\]. The viral particles are spherical, with a diameter of 30 nm \[[@B21]\], and consist of a non-enveloped capsid that encloses a 9010 nt long viral genome \[[@B22]\]. The capsid contains four structural proteins VP1 (29.7 kDa), VP2 (28.4 kDa), VP3 (31.8 kDa) and a minor one VP0 (37.3 kDa). In addition to these four structural proteins, this virus has a low molecular weight protein of 5.5 kDa that appears to be detached from the capsid, lying at the particle interior and presumably in contact with the genome \[[@B3],[@B20],[@B23]\]. TrV can also infect natural populations of *T. sordida* and several experimental populations of triatomines as well \[[@B24]\]. Viral replication takes place in the midgut epithelium cells of triatomines, causing many deleterious sublethal effects in *T. infestans* and *T. patagonica* colonies, such as reduction of the longevity, increase of the developmental time, a decrease in both fecundity and fertility of eggs, and even a cumulative mortality higher than 97% \[[@B16],[@B25]-[@B28]\].
As pointed out previously, the infectivity of TrV in vertebrates remains unclear. Therefore, in this work we studied the infectivity of TrV in mice (*Mus musculus* - BALB/c mice). Our results indicate that inoculations of both infective and non-infective TrV particles elicit the same immune response, and that this insect virus is unable to replicate in this animal model. Based on the bicistronic character of dicistrovirus genomes, it was speculated that these pathogens may not infect vertebrates, and this report constitutes important experimental evidence to support that hypothesis.
Methods
=======
*Triatoma virus* purification
-----------------------------
Virus purification was performed according to \[[@B3]\]. Dry feces obtained from infected insects were collected on Whatman filter papers and excised from them using a standard surgery scalpel until two g of material was collected. A buffer composed of 10 mM NaCl, 1 mM MgCl~2~ and 200 mM citric acid at pH 6 (Lysis buffer) was added at a ratio of 100 mL per gram of dry material. PMSF (N 98.5%, purchased from Sigma) thawed at −20°C in isopropanol (Merck, GR for analysis) was added up to a final concentration of 1 mM. The mixture was homogenized for 5 min using a vortex mixer and then sonicated using a Sanyo MSE Soniprep 150 operating at 10 s ON/10 s OFF for 20 pulses. After centrifugation, the resulting pellet was resuspended in NMT buffer and then loaded on top of a 35 mL continuous 5%--30% sucrose gradient. The gradients were cooled in ice to avoid particle diffusion within them. The individual optical density of each aliquot was then measured at both 260 nm and 280 nm. After plotting the absorbance profile, selected aliquots were loaded onto 12.5% polyacrylamide SDS-PAGE gels to check the purity of the samples. Following gradient fractionation, equivalent samples were pooled together and dialyzed overnight against 2 L of NMT buffer.
Infection with *Triatoma virus* (TrV)
-------------------------------------
Eight groups of female *Mus musculus* BALB/c mice (3 mice per group) between 5 and 8 weeks of age obtained from Instituto de Higiene e Medicina Tropical, Lisbon, Portugal, were used in the experimental infection with TrV. Groups 1-5 were inoculated by intraperitoneal injection. Group 1 is the control group, which was inoculated only with PBS. Group 2 was inoculated with 3.0 μg of empty TrV particles (capsids without genome, therefore non-infective) \[[@B3]\]. Group 3 was inoculated with 3.0 μg of TrV, group 4 and group 5 with 0.3 μg and 0.03 μg of TrV respectively. Group 6 served as a negative control of *per os* inoculation, therefore it was inoculated only with 100 μl of PBS delivered *per os* intraesophageally by using a 1 ml tuberculin syringe and Gavage needle. Group 7 was inoculated with 3.0 μg of empty TrV particles in 100 μl of PBS, delivered *per os* intraesophageally by using a 1 ml tuberculin syringe and Gavage needle. Group 8 was inoculated with 3.0 μg of TrV in 100 μl of PBS, delivered *per os* intraesophageally by using a 1 ml tuberculin syringe and Gavage needle. After inoculation blood and feces samples were collected at different times (3, 7, 30 and 45 days). Forty-five days after inoculation, the animals were sacrificed.
All experiments on animals were conducted with prior approval from the Animal Welfare of General Management of Veterinarians (DGV - Portugal). All manipulations of mice satisfied the requirements of the Regulations of Experimental Animal from the Ethics Committee of IHMT/UNL-Portugal.
Searching for TrV in blood and feces samples
--------------------------------------------
Blood samples from each mouse were collected into individual micro tubes containing 5 μl of Heparin-Sodium (B. Braun, USA). The RNA was extracted according to the product manual supplied with the RNA isolation Kits, (Bioline, UK). RNA concentration was determined by measuring absorbance at 260 nm (Abs~260~) in a spectrophotometer (NanoDrop 1000, Thermo Scientific).
Feces samples were collected into micro tubes. For each group of mice we pooled the samples from all mice of this group. 1,2 mg of fecal samples resuspended in 200 μl PBS were homogenized and centrifuged (5,000 rpm) for 10-15 minutes. 50 μl of supernatant was then homogenized in TRIZOL reagent (Genbiotech, Argentina), and viral RNA-TrV (vRNA) was purified according to the manufacturer's instructions. As a positive control of RNA extraction, we carried out the same procedure simultaneously with feces from triatomines infected with TrV.
Purified vRNA (0.724 μg) was used as a template for the positive control in the RT-PCR reaction. For the other RT-PCR, we used 5 μl of the samples resulting from the purification of the vRNA from blood and feces samples from the mice used in the infectivity experiment. The cDNA synthesis was performed according to the cDNA Synthesis Kit protocol (Bioline, UK). 5 μl of the extraction product was used according to the OneStep RT-PCR protocol (QIAGEN; USA). TrV positive PCR reactions have a 832 bp product according to the protocol established in our previous work \[[@B26]\]. As a positive control of RT-PCR and RNA extraction from blood samples, we performed a PCR with two primer pairs from β-actin gene: β-actin sense: 5´ TGGAATCCTGTGGCATCCATGAAAC 3´ and β-actin anti-sense: 3´ TAAAACGCAGCTCAGTAACAGTCCG 5´, with an expected product of 348 bp \[[@B29]\]. PCR products were visualized on 1.5% agarose gels for TrV and 2% agarose gel for β-actin, stained with ethidium bromide, and their sizes were determined by comparison against DNA markers, HyperLadder I (Bioline, UK) and HyperLadder IV (Bioline, UK) respectively.
Production of mouse polyclonal antibodies to anti-TrV
-----------------------------------------------------
The serum used as a positive control in ELISA reactions were produced in female *Mus musculus* BALB/c mice between 5 and 8 weeks of age. The mice were inoculated (subcutaneous injection) with 100 μg of empty TrV particles mixed with 100 μl of Freund's complete adjuvant (Sigma-Aldrich, USA). Fifty days after the first inoculation, the animals were inoculated with 100 μg of empty TrV particles mixed with 100 μl of Freund's incomplete adjuvant (Sigma-Aldrich, USA). Ten days after the second inoculation, the animals were sacrificed and hyperimmune sera was obtained.
Detection of anti-TrV IgGs antibodies by ELISA
----------------------------------------------
Naturally empty TrV particles were selected as antigens for their protein-only composition. These particles contain almost the same proteins as full particles, but at least 30% of its composition was integrated by 7 additional polypeptides that were determined to be misprocessed products from the cleavage of the structural protein precursor P1 \[[@B3]\]. Nevertheless, since both full and empty particle proteins share the same origin (P1), their antigenic properties remain essentially the same.
Three different concentrations of total proteins (50, 100 and 150 ng/well) from the empty TrV particles were used to optimize the amount of antigen used in this reaction. We chose 100 ng/well because with this value the amount of antigen is no longer the limiting reagent of the reaction. Total protein extract from empty TrV particles (in carbonate buffer pH 8.5) was adsorbed overnight at 4°C onto 96-well micro-plates (Nunc, Denmark) for use in indirect enzyme-linked immunosorbent assays (ELISA), using sera from inoculated mice. Subsequently, horseradish peroxidase, (HRP)-conjugated rat anti-mouse IgG (1:4000, Sigma- Aldrich, USA), rat anti-mouse IgG1 (500 ng/ml, AbD serotec, UK) or rat anti-mouse IgG2a (250 ng/ml, AbD serotec, UK) were used.
Results
=======
Absence of clinical infection signals
-------------------------------------
Mice inoculated with TrV by intraperitoneal and *per os* route did not show any behavioral alteration or clinical signals of viral infection (e.g. leg paralysis, change in food intake rate, weight loss, decrease in motility and death), when compared with mice inoculated with empty TrV particles or saline solution (control groups), this is contrary to what happens in triatomines, where the insects inoculated with the virus show leg paralysis, delayed development and death \[[@B16],[@B22]\].
Absence of TrV in blood and feces samples
-----------------------------------------
We analyzed all blood and feces samples from each group of animals at 0, 3 and 45 days after inoculation. No RT-PCR products corresponding to TrV were detected from blood or feces samples of any of the groups of mice inoculated with TrV or mice inoculated with empty TrV particles (Figure [1](#F1){ref-type="fig"}). All blood samples showed PCR products corresponding to β-actin gene, with expected size (348 bp) (Figure [1](#F1){ref-type="fig"}-D).
{#F1}
Detection anti-TrV IgGs antibodies by ELISA
-------------------------------------------
We performed ELISA to detect antibodies against TrV in sera extracted from inoculated mice. Total protein extract from empty TrV particles was used to detect the presence of anti-TrV antibodies in sera from non-inoculated and inoculated BALB/c mice with TrV. We produced polyclonal antibodies to anti-TrV in mice in order to establish basal and positive signals in the ELISA. The cut-off value was defined using sera from non-inoculated mice, mean OD (optical density) + 3SD (standard deviation) at 490 nm. We chose this weighting (3SD), because all sera from non-inoculated mice had OD values less than the *cut-off* value.
Forty-five days after intraperitoneal inoculations, all inoculated mice showed high levels of IgG anti-TrV antibodies when compared with sera from non-inoculated mice (Figure [2](#F2){ref-type="fig"}) or mice inoculated by *per os* route. Sera reactivity was directly proportional to the amount of antigen used in inoculation. However, there was no significant difference (F = 0.0064; p-value \> 0.05) between the reactivity of sera from mice inoculated with empty TrV particles to that of mice inoculated with the same amount of RNA-full TrV capsid (Additional file [1](#S1){ref-type="supplementary-material"}). All mice inoculated by the *per os* route presented levels of anti-TrV antibodies similar to mice inoculated only with PBS (Figure [2](#F2){ref-type="fig"}), indicating that *per os* inoculation with TrV produces little or no IgG antibody response in mice.
{#F2}
To define which subclasses of IgG anti-TrV were detected in sera from inoculated mice and in order to determine if TrV can replicate in mice, we measured the titers of IgG1 and IgG2a antibodies specific to TrV (Figure [2](#F2){ref-type="fig"}-B and C). The results show that there is no significant difference in the IgG2a/IgG1 ratio between mice inoculated with TrV, mice inoculated with empty TrV particles or non-inoculated mice (Figure [2](#F2){ref-type="fig"}-D).
Discussion
==========
RT-PCR has been used to detect a variety of RNA viruses in different kinds of samples \[[@B9],[@B30]\]. In this study, a RT-PCR assay has been developed to detect TrV from blood and feces samples from mice inoculated with TrV. Two specific pairs of primers for TrV were used and the 832 bp products expected for the TrV-specific assay were produced only when TrV vRNA was present, indicating that this assay is highly specific for TrV.
In this PCR assay, PCR products were detected from samples with low concentrations of plasmid DNA (6.2 pg) encoding specific TrV sequences or TrV vRNA (2.5 ng) (data not shown). Therefore, if TrV were able to replicate in mice, their presence would be detected by RT-PCR assay from blood and feces samples of inoculated mice. However, we were not able to detect any PCR products from blood or fecal samples of mice inoculated with TrV. The absence of PCR products from these samples may indicate that TrV is unable to replicate in mice, this contrasts to what happens in triatomines, where this virus causes a high mortality rate, reduces fecundity and delays the development of infected insects \[[@B16],[@B25]\].
It has been postulated that the immune response triggered by viruses mainly activates the Th1 subset of T helper cells and enhances the production of INF-γ, which is the reason many systemic viral infections in mice result in preferential increases in virus-specific IgG2a antibodies in serum \[[@B31],[@B32]\]. This preferential class switch is dependent on viral replication and is attributed in part to virus-induced production of IFN-γ \[[@B31]\]. In this study we observed that there is no significant difference in the ratio IgG2a/IgG1 in sera from animals inoculated with TrV when compared with non-inoculated animals or mice inoculated only with empty TrV particles. The results reported in this work support the idea that TrV is unable to replicate in vertebrates.
Conclusions
===========
The results of the ELISA for anti-TrV together with the results of a TrV vRNA search by RT-PCR in blood and feces samples from mice inoculated with TrV indicate that this virus is not infective in mice, at least under the conditions explored in this study (infected by *per os* and intraperitoneal inoculations). The results reported in this work support the idea that TrV is a virus that only infects invertebrates.
To date, this study constitutes an important approach to evaluate the response of one vertebrate animal model to the experimental infection with a dicistrovirus. Future studies may lead to a better understanding of the TrV host range, and hopefully may also characterize the effect observed in humans and other vertebrates that are most likely to be exposed to dicistroviruses \[[@B10],[@B22]\]. Certainly, to improve the current knowledge on TrV, and in general, to gain information on the life cycle of any member of the *Dicistroviridae* family, will greatly help in establishing a proof of principle for employing dicistroviruses as biological agents to control insect pests.
Competing interests
===================
The authors declare that they have no competing interests.
Authors' contributions
======================
Conceived and designed the experiments: MSS, DMAG, GAM. Performed the experiments: JBQ, JA, MSS. Analysed the data: JBQ, GAM, MMAG, MSS. Wrote the paper: JBQ, DMAG, MSS. All authors have read and approved the final manuscript.
Supplementary Material
======================
###### Additional file 1
**Statistical analysis (ANOVA, with Microsoft Excel®) between the two groups of mice inoculated with the same concentration of RNA-full TrV capsids and empty TrV particles respectively (3 μg).** Considering that the data is normally distributed, this test helped to identify differences between mice inoculated with TrV and mice inoculated with the same amount of empty TrV particles. The null hypothesis (H~0~) assumes that the means are statistically the same (H~0~: μ~1~ = μ~2~), and the alternate hypothesis (H~A~) assumes that the means are statistically different (H~A~: μ~1~ ≠ μ~2~), at 95% confidence. Since the F statistic is smaller than the critical value, we fail to reject the null hypothesis. Sum of squares (SS); Degrees of freedom (df); Mean square (MS).
######
Click here for file
Acknowledgments
===============
The authors are grateful to Rubén Sánchez-Eugenia and Sonia López (UBF) for their technical assistance in TrV purification, and to Dr. Carlos Robello from the Instituto Pasteur Montevideo for the gift of the complementary cDNA used as positive control in PCR assays. Thanks also to all personnel from the Department of Virology, Faculty of Veterinary Sciences, National University of La Plata. JA was supported by a UPV/EHU contract (UPV-IT-461-07). GAM is a researcher of the CONICET, Argentina. DMAG was partially supported by grant SAIOTEK (Y-CRYSTAL), Gobierno Vasco (GV), and MICINN (BFU2007-62062), Spain. This work was partially supported by Acción Especial AE-2009-1-21, from the GV, Spain, and CYTED 209RT0364 action (RedTrV: <http://www.redtrv.org>), both granted to DMAG.
| {
"pile_set_name": "PubMed Central"
} |
Introduction
============
Around 7% of malignant neoplasms occur in the kidney; these include Wilms tumor (WT; also known as nephroblastoma), clear cell sarcoma of the kidney, malignant rhabdoid tumor, and renal cell carcinoma \[[@b1-crt-2019-313],[@b2-crt-2019-313]\]. WT is the most common primary kidney tumor in children and is treated with a multimodal approach that consists of surgery, radiation therapy, and anticancer drugs \[[@b3-crt-2019-313]\]. Based on tumor stage and pathological findings of upfront nephrectomy, the National Wilms Tumor Study Group (NWTSG) recommends treating WT according to the risk of relapse \[[@b4-crt-2019-313]\]. In contrast, the International Society of Pediatric Oncology advocates upfront chemotherapy before nephrectomy, with the intensity of subsequent therapies adjusted according to tumor response \[[@b5-crt-2019-313]\]. Attempts to treat WT patients according to the risk of tumor recurrence using prognostic markers has led to improvement in outcome over the past 40 years, such that the long-term survival rate is now 90% \[[@b6-crt-2019-313]\].
Tumor suppressor genes contribute to tumorigenesis by disabling the function of both normal alleles of a gene \[[@b7-crt-2019-313],[@b8-crt-2019-313]\]. Loss of heterozygosity (LOH) at a chromosomal locus comprising a tumor suppressor gene results in functional defects that can lead to pediatric WT \[[@b7-crt-2019-313],[@b9-crt-2019-313]\]. It was recently reported that LOH at chromosome 1p and/or 16q is associated with high recurrence and mortality rates in WT patients. The National Wilms Tumor Study (NWTS)-5 reported that tumor-specific LOH at 1p and 16q was associated with a higher risk of relapse and death in favorable histology Wilms tumor (FHWT), and treatment intensity was increased in these patients \[[@b6-crt-2019-313]\].
However, different investigators have reported variable findings regarding the prognostic value of LOH at 1p and 16q \[[@b10-crt-2019-313]-[@b12-crt-2019-313]\]: some have shown that LOH at both loci predicts survival, while others have found that only one or the other affects prognosis \[[@b13-crt-2019-313]\]. These discrepancies could be due to several causes, not only differences in the study populations, but also differences in ethnic variation. The present study investigated the prognostic significance of LOH at 1p and 16q in a Korean pediatric FHWT cohort.
Materials and Methods
=====================
1. Patients and treatment
-------------------------
The study protocol was approved by the WT committee of the Korean Society of Pediatric Hematology Oncology Group (K-PHOG), which analyzed patient information and specimens from the four participating hospitals, each of which passed Institutional Review Board deliberation, filled out the case report form (CRF), and sent five-panel slides for each patient to the committee for retrospective review. All patints included in this study were diagnosed as FHWT between 1996 and 2016; examined by chest X-ray, abdominal sonography, computed tomography, or magnetic resonance imaging; staged according to the NWTSG staging system; and treated according to National Wilms' Tumor Study 4 (NWTS-4) guidelines \[[@b6-crt-2019-313]\]. One institution modified the NWTS-4 treatment guideline; initial fine needle biopsy and delayed nephrectomy was done at 6 to 9 weeks if the largest diameter of the tumor is greater than 8 cm (modified NWTS). The CRF included data on the patient's age, sex, histological findings, staging at the time of diagnosis, surgery, radiation, chemotherapy, relapse, and survival.
Upfront nephrectomy was initially recommended for all patients. However, chemotherapy was given prior to nephrectomy if the tumor was deemed unresectable by the surgeon or if the surgery was life-threatening, the tumor thrombus extended into the inferior vena cava above the level of the hepatic veins, or the disease was bilateral by Doppler sonography. Stage I-II patients received two drugs including vincristine and actinomycin D for 18 weeks without radiotherapy, and stage III-IV patients were treated with a threedrug regimen consisting of vincristine, actinomycin, and doxorubicin for 24 weeks in addition to receiving radiotherapy. Stage V patients were treated with nephron-sparing surgery and chemotherapy or delayed nephrectomy after preoperative chemotherapy and then postoperative chemotherapy.
2. Histopathological examination and DNA extraction for microsatellite analysis
-------------------------------------------------------------------------------
Paraffin-embedded kidney specimens containing normal and tumor tissues that were cut into 5-μm sections and mounted on glass slides (five per patient) were sent from participating institutions to that of the principal investigator. Pathological diagnoses were reviewed at a central location and tumor cells were microdissected under a microscope from hematoxylin and eosin-stained sections. DNA was extracted from tumor and non-tumor sections for microsatellite analysis according to standard procedures. Two tissue sections from each case were transferred using a disposable razor blade to 900 μL of xylene in a 1.5-mL microcentrifuge tube. Equal volumes of 100% and 70% ethanol were added and the samples were pelleted (10 minutes in a microcentrifuge), dried at 56°C, and incubated overnight in cell lysis buffer (cat. No. 1045723, Qiagen, Valencia, CA) with 1.5 μL of proteinase K (41 μg/mg). The proteinase K was inactivated by heating the samples at 56°C for 30-60 minutes; 100 μL of protein precipitation solution (cat. No. 1045701, Qiagen) was added to each sample, followed by centrifugation for 5 minutes. The supernatant was removed and isopropanol was added to precipitate the DNA, which was washed with 70% ethanol and transferred to a new 1.5-mL tube.
3. Detection of LOH at 1p and 16q
---------------------------------
To detect LOH at 1p and 16q, 16 polymorphic microsatellite markers were selected based on a previous study \[[@b6-crt-2019-313]\]. Forward primers were labeled with four different fluorescent dyes (VIC, 6FAM, PET, or NED) ([Table 1](#t1-crt-2019-313){ref-type="table"}). Polymerase chain reaction (PCR) amplification was performed on tumor and normal tissue samples in triplicate using the Qiagen Multiplex PCR Plus Kit (Hilden, Germany) under the following conditions: 95°C for 10 minutes; five cycles of 95°C for 20 seconds, 60°C for 90 seconds (−1°C in each cycle), and 72°C for 1 minute; 30 cycles of 95°C for 20 seconds, 55°C for 90 seconds, and 72°C for 1 minute; and 60°C for 30 minutes. PCR products were examined using a 3500xL Dx Genetic Analyzer (Thermo Fisher Scientific, Waltham, MA) and LOH was detected using GeneMapper v.4.1 software (Thermo Fisher Scientific) according to the manufacturer's instructions.
4. Statistical analysis
-----------------------
Event-free survival (EFS) refers to the time from diagnosis to the first occurrence of progression, relapse after response or death from any cause. Overall survival (OS) was defined as the time from diagnosis to the day of death or last follow up. EFS and OS were calculated with the Kaplan-Meier method, and the log-rank test and the Cox proportional hazards model were used for uni- and multivariate analyses, respectively. The chi-square test was used to compare the contingency tables. Two-sided p-values \< 0.05 were considered statistically significant.
5. Ethical statement
--------------------
This study was performed in accordance with the ethical guidelines of the "World Medical Association Declaration of Helsinki-Ethical Principles for Medical Research Involving Human Subjects." This study was approved by the Institutional Review Board of participating institutions and Ajou University Hospital (IRB No. AJIRB-BMR-OBS-16-143). Written informed consent was waived from the IRB.
Results
=======
A total of 101 FHWT patients were enrolled in the study. The median age was 2.8 years (range, 0.3 to 10.4 years). The male-to-female ratio (48:53) was 1.04. The characteristics of the study population are summarized in [Table 2](#t2-crt-2019-313){ref-type="table"}. Among the patients, 43 (42.6%) and 52 (51.5%) had tumors in the right and left kidneys, respectively, whereas six (5.9%) had tumors in both kidneys. Eleven patients showed distant metastasis to the lung. Upfront nephrectomy was performed in 54 patients and delayed nephrectomy in 47; the latter was performed after preoperative chemotherapy since 37 cases were deemed inoperable by the surgeon, with six cases of bilateral disease and four in which the tumor thrombus extended into the inferior vena cava above the level of the hepatic veins. Forty patients received radiotherapy.
The status of LOH at 1p and 16q in the patients is shown in [Table 3](#t3-crt-2019-313){ref-type="table"}. There were 19 patients (18.8%) with LOH at 1p and 16q respectively, including five exhibiting LOH at both loci. The frequency of LOH at 1p was higher among younger patients (p=0.049), especially those under 4 years of age. There was no difference in LOH prevalence according to tumor stage ([Table 3](#t3-crt-2019-313){ref-type="table"}).
Among the 101 patients, there were 12 recurrences and two deaths. The 3-year OS and EFS of all patients were 98.0% and 87.6%, respectively. The 3-year EFS rate was 89.5% in patients with and 87.2% in those without LOH at 1p (p=0.798 \[log-rank test\]) ([Fig. 1A](#f1-crt-2019-313){ref-type="fig"}). The 3-year EFS rate was lower in patients with LOH than in those without LOH at 16q (73.7% vs. 91.1%, p=0.047 \[log-rank test\]) ([Fig. 1B](#f1-crt-2019-313){ref-type="fig"}). Patients with LOH at both 1p and 16q did not show a lower EFS rate than those in the other groups ([Fig. 1C](#f1-crt-2019-313){ref-type="fig"}).
A multivariate analysis showed that LOH at 16q was a significant negative prognostic factor for EFS (hazard ratio, 3.95; p=0.037), which was not true of LOH at 1p (hazard ratio, 0.83; p=0.817) ([Table 4](#t4-crt-2019-313){ref-type="table"}).
Discussion
==========
There has been significant progress in the treatment of FHWT over the past 30 years, and the survival rate is now about 90% \[[@b14-crt-2019-313]\]. This excellent outcome is the result of a multimodal approach that combines surgery, chemotherapy, and radiation therapy according to tumor stage and histology. Additionally, tumors at high risk of recurrence are treated more aggressively than those at low risk, such that overtreatment can be avoided in the latter cases \[[@b15-crt-2019-313]\]. Recently, efforts have been made to identify new molecular markers that can be used to predict tumor recurrence by stratifying high- and low-risk patients, which can ultimately prolong survival by reducing long-term treatment sequelae \[[@b7-crt-2019-313],[@b9-crt-2019-313]\]. There have been many reports of LOH in relation to the risk of recurrence of WT, for instance at 1p, 11q, 16q, and 22q \[[@b10-crt-2019-313],[@b13-crt-2019-313],[@b16-crt-2019-313]\]. Of these, LOH at 1p and LOH 16q are the most widely investigated and were the focus of the present work.
In our study, 19 patients each out of 101 (18.9%) had LOH at 1p and 16q. We compared these rates to the ones reported in other studies ([Table 5](#t5-crt-2019-313){ref-type="table"}). Of the 232 WT patients enrolled in NWTS-3 and -4, 21 patients (12%) had LOH at 1p and 35 (17%) at 16q \[[@b10-crt-2019-313]\]. In NWTS-5, 196 out of 1,727 FHWT patients (11.3%) had LOH at 1p and 301 (17.4%) had LOH at 16q \[[@b4-crt-2019-313]\]. Of the 426 FHWT patients enrolled in the UKW 1-3 trials, the prevalence of LOH at 1p and 16q was 10.3% and 17.4%, respectively \[[@b11-crt-2019-313]\]; among the 125 patients who participated in the AIEOP-TW-2003 trial, the rates were 19% and 15%, respectively \[[@b13-crt-2019-313]\]; and in an Egyptian FHWT cohort (n=100), the rates were 16% and 25%, respectively \[[@b12-crt-2019-313]\]. Thus, with the exception of one report \[[@b12-crt-2019-313]\], the prevalence of LOH at 1p and 16q is about 10%-19% and 14%-20%, respectively, whereas the proportion of patients with both alterations is between 2.6% and 5% (i.e., 4.6% \[[@b4-crt-2019-313]\], 2.6% \[[@b11-crt-2019-313]\], and 4% \[[@b13-crt-2019-313]\]). Five out of 101 patients in our study had LOH at both 1p and 16q, which is comparable to the rates in these other studies.
The frequencies of LOH at 1p and 16q are higher among FHWT patients over 4 years old \[[@b4-crt-2019-313]\], and lower in patients younger than 2 years old \[[@b12-crt-2019-313]\]. One study found that the frequency LOH at 1p but not at 16q increased after 4 years of age \[[@b11-crt-2019-313]\]. Others have shown a low prevalence of LOH at 1p but not at 16q in patients under 2 years old \[[@b13-crt-2019-313]\]. In our study, the rate of LOH at 1p was higher among younger FHWT patients (p=0.045), especially in those under 4 years old ([Table 3](#t3-crt-2019-313){ref-type="table"}). These same age-group differences in the frequency of LOHs appearing among the reports may represent interracial biological differences in WTs. It was reported that the frequency of LOH at 1p or 16q increases with tumor stage \[[@b12-crt-2019-313]\]; however, in most other studies \[[@b6-crt-2019-313],[@b11-crt-2019-313],[@b13-crt-2019-313]\] including ours, there was no association between the frequency of these alterations and tumor stage.
Several studies have investigated whether LOH at 1p and/ or 16q can predict FHWT recurrence and patient survival ([Table 5](#t5-crt-2019-313){ref-type="table"}). NWTS-5 is the largest of these studies to date; it was prospective in nature and used stage-specific treatment, and showed that LOH at both 1p and 16q loci was associated with worse prognosis \[[@b4-crt-2019-313]\]. In NWTS-3 and -4, which preceded NWTS-5, LOH at 16q was associated with 3.3- and 12-fold higher rates of relapse and mortality, but LOH at 1p tended to increase recurrence and mortality rates without statistically significant \[[@b2-crt-2019-313],[@b10-crt-2019-313]\]. In this study, it is thought that the preceding NWTS-3,4 study was a statistical difference caused by the smaller sample sizes than the NWTS-5 study.
However, it is thought that there would be another reason for different results in determining LOH(s) at 1q and/or 16q as prognostic factor(s). In a study of 125 children with FHWT treated with the AIEOP-TW-2003 protocol which was conducted in Italy from 2003 to 2008, LOH at 1p was associated with worse disease-free survival \[[@b13-crt-2019-313]\]. Meanwhile, in the UKW 1-3 trials conducted in the United Kingdom with a relatively large number of subjects, 426 FHWT patients, LOH at 16q but not at 1p was also associated with a higher risk of relapse and death \[[@b11-crt-2019-313]\]. In a study of Egyptian FHWT patients, LOH at 16q alone or in combination with LOH at 1p had a higher recurrence rate and lower 3-year EFS \[[@b12-crt-2019-313]\]. It can be inferred that genetic variability of the WT may have racial differences to produce different outcomes for each country that conducted the study. In our study, EFS was lower in patients with LOH at 1p only or at both 1p and 16q compared with those with neither alteration; this was not statistically significant, possibly due to our small sample size. Nonetheless, our results show that only LOH 16q negatively affects prognosis, which is consistent with previous findings \[[@b4-crt-2019-313],[@b10-crt-2019-313]-[@b12-crt-2019-313]\]. This result is thought to reflect racial differences, as stated earlier in the study conducted in United Kingdom, Egypt, and Italy.
Only cases of LOH at both 1p and 16q in FHWT stages I-IV are stratified into poor prognosis groups and receive additional intensive chemotherapy, although both types of alteration show a low prevalence of 4% \[[@b6-crt-2019-313]\]. Thus, the absolute number of patients who would benefit from clinical decisions made based on the identification of LOH may be relatively small if they were classified as high risk and given intensive chemotherapy \[[@b13-crt-2019-313]\]. In our study, five patients among the 19 with LOH at 16q experienced recurrence; assuming that this group be received more aggressive treatment in the future study, the remaining 14 will be likely overtreated. Therefore, it is important that patients are stratified with higher statistical power and treated in combination with other risk factors rather than classified as a recurrent risk group based solely on detection of LOH at 16q \[[@b9-crt-2019-313]\].
The limitation of our study is that although it was a nationwide study conducted by K-PHOG the sample size was relatively small, this is especially true for LOH at 1p that is statistically insignificant. Second, the study was retrospective and therefore did not have a uniform staging system and treatment strategy. Therefore, the next study needs a prospective study with a uniformed staging system and treatment strategy.
This study is meaningful in that it is the first to identify LOH at 16q as a significant negative prognostic factor affecting outcome in Korean pediatric FHWT patients. However, this study is limited to its small sample size, large-scale multicenter trials are warranted to investigate the prognostic value of LOH at 1p and 16q in Korean children with FHWT.
Conflict of interest relevant to this article was not reported.
This study was supported by a grant from Ajou University Medical Center (grant No. M-2015-C0460-00110).
{#f1-crt-2019-313}
######
Information on microsatellite markers used in this study
No. Name Forward primer Reverse primer Size (bp) Location Tube no.
----- ---------- --------------------------------- --------------------------- ----------- ---------- ----------
1 D1S244 VIC: GAGCAGCACCGTACAAAT AGCTCCGCTCCCTGTAAT 285-296 1p36.22 1
2 D1S468 6FAM: AATTAACCGTTTTGGTCCT GCGACACACACTTCCC 173-191 1p36.33 1
3 D1S2870 PET: GATCATGCCAATGCACTAT CCAGGGTGACACAGCA 190-212 1p36.31 1
4 D1S214 6FAM: CCGAATGACAAGGTGAGACT AATGTTGTTTCCAAAGTGGC 120-142 1p36.31 1
5 D1S1612 NED: TCCCATGCCAAAATTCTTAG GAAAGAAAGAGAAAGAAGGAAGG 94-130 1p36.23 1
6 D1S243 PET: CACACAGGCTCACATGCC GCTCCAGCGTCATGGACT 142-170 1p36.33 2
7 D1S450 NED: GCTCCAATGTCCAAGGG GGGTACTCAGATGGCTGGT 243-267 1p36.22 2
8 D16S518 6FAM: GGCCTTTTGGCAGTCA ACCTTGGCCTCCCACC 271-290 16q23.1 3
9 D16S3025 PET: TCCATTGGACTTATAACCATG AGCTGAGAGACATCTGGG 90-110 16q22.1 2
10 D16S3043 6FAM: CATTAATATGGAGCCTTATAGATTG AAATGTTGAGCACTTGAATAAAAT 118-150 16q21 2
11 D16S400 6FAM: GTCATCCGACTTCTCACAGG AATATGAACCCTCCATGCTG 192-202 16q21 3
12 D16S421 VIC: ACATGAACCGATTGGACTGA CCGTTCCCTATATTTCCTGG 206-216 16q22.1 2
13 D16S422 NED: CAGTGTAACCTGGGGGC CTTTCGATTAGTTTAGCAGAATGAG 188-212 16q23.3 2
14 D16S2621 PET: GTCATATGGGCCAATTCCC TACCGCGTAGTGAGACTGTG 239-263 16q24.2 3
15 D16S2624 VIC: TGAGGCAATTTGTTACAGAGC TAATGTACCTGGTACCAAAAACA 132-148 16q22.2 3
16 D16S3101 NED: TTCCTGAATGTCATGTAGTTGG TGTCATCGGGGCTTGTAG 158-166 16q23.1 2
######
Characteristics of patients with favorable histology Wilms tumors
Patient characteristic No. of patients (%) (n=101)
--------------------------------------- -----------------------------
**Age at diagnosis (yr)**
Mean (range) 2.8 (0.3-10.4)
\< 2 52 (51.5)
2-10 23 (22.8)
\> 10 26 (25.7)
**Sex**
Male 48 (47.5)
Female 53 (52.5)
**Primary tumor site**
Right 43 (42.6)
Left 52 (51.5)
Both 6 (5.9)
**Stage**
I 25 (24.8)
II 15 (14.9)
III 43 (42.6)
IV 12 (11.9)
V 6 (5.9)
**Treatment strategy**
NWTS-4 53 (52.5)
Modified NWTS-4 48 (47.5)
**Nephrectomy**
Upfront 54 (53.4)
Delayed 47 (46.5)
**Metastasis to lung at diagnosis** 11 (10.9)
**Loss of heterozygosity for 1p/16q**
1p only 14 (13.9)
16q only 14 (13.9)
Both 1p and 16q 5 (5.0)
NWTS-4, National Wilms' Tumor Study 4.
######
Prevalence of loss of heterozygosity for 1p and 16q according to age and disease stage
Patient characteristic No. of patients (%)
--------------------------- --------------------- ---------- ------- ----------- ----------- ------- ---------- ---------- ---------- ----------- -------
**Age at diagnosis (yr)**
\< 2 43 (52.4) 9 (47.4) 0.049 42 (51.2) 10 (52.6) 0.427 8 (57.1) 9 (64.3) 1 (20.0) 34 (50.0) 0.116
2-4 15 (18.3) 8 (42.1) 17 (20.7) 6 (31.6) 5 (35.7) 3 (21.4) 3 (60.0) 12 (17.6)
\> 4 24 (29.3) 2 (10.5) 23 (28.0) 3 (15.8) 1 (7.1) 2 (14.3) 1 (20.0) 22 (32.4)
**Stage**
I 19 (23.2) 6 (31.6) 0.411 24 (29.3) 1 (5.3) 0.086 6 (42.9) 1 (7.1) 0 18 (26.5) 0.226
II 13 (15.9) 2 (10.5) 12 (14.6) 3 (15.8) 2 (14.3) 3 (21.4) 0 10 (14.7)
III 36 (43.9) 7 (36.8) 30 (36.6) 13 (68.4) 3 (21.4) 9 (64.3) 4 (80.0) 27 (39.7)
IV 8 (9.8) 4 (21.1) 11 (13.4) 1 (5.3) 3 (21.4) 0 1 (20.0) 8 (11.8)
V 6 (7.3) 0 5 (6.1) 1 (5.3) 0 1 (7.1) 0 5 (7.4)
######
Event-free survival rates and hazard ratio according to age, stage, and LOH at 1p and 16q status (multivariate analysis)
Event-free survival (3-year) (%) Hazard ratio (95% CI) p-value
------------------------ ---------------------------------- ----------------------- ---------
**Age (yr)**
\< 2 90.0 1
2-4 91.3 0.84 (0.15-4.53) 0.835
\> 4 79.6 2.40 (0.67-8.55) 0.176
**Stage**
I-II 90.6 1
III-V 84.5 1.11 (0.24-5.15) 0.893
**Treatment strategy**
NWTS-4 88.2 1
Modified NWTS-4 86.9 0.73 (0.15-3.51) 0.695
**LOH 1p**
Absent 82.7 1
Present 89.5 0.83 (0.17-4.01) 0.817
**LOH 16q**
Absent 91.1 1
Present 73.7 3.95 (1.08-14.39) 0.037
LOH, loss of heterozygosity; CI, confidence interval; NWTS-4, National Wilms' Tumor Study 4.
######
Summary of previous studies on the prevalence and hazard ratio of relapse risk associated with LOH at 1p or/and 16q
Study Source of study No. of patients Prevalence of LOH 1p/16q/both (%) HR or RR p-value
------------------------------------------ ---------------------------------- ----------------- ----------------------------------- ------------------------------------------------------------------------------------------------------------------------------------------------------------------- ----------
Grundy et al. \[[@b10-crt-2019-313]\] NWTS-3,4 232 12/17/4 RR 3.3 for 16q vs. none 0.01
Grundy et al. \[[@b6-crt-2019-313]\] NWTS-5 1,727 11.3/17.4/4.6 Stage I-II, RR 2.9 (95% CI, 1.51-5.49) for both vs. none, 1p only, and 16q only stage III-IV, RR 2.4 (95% CI, 1.20-4.82) for both vs. none, 1p only, and 16q only 0.001
Messahel et al. \[[@b11-crt-2019-313]\] UKW 1-3 426 10.3/17.4/2.6 HR 2.64 (95% CI, 1.47-4.92) for 16q vs. none \< 0.001
Spreafico et al. \[[@b13-crt-2019-313]\] AIEOP-TW-2003 125 19/14/4 HR 4.1 (95% CI, 1.60-10.8) for 1p vs. none 0.0009
Fawzy et al. \[[@b12-crt-2019-313]\] National Cancer Institute, Egypt 100 26/25/12 53.8%a), 50% for 16q only, and both vs. 1p only 0.007
This study K-PHOG 101 18.8/18.8/5 HR 3.84 (95% CI, 1.31-13.0) for 16q vs. none 0.031
LOH, loss of heterozygosity; HR, hazard ratio; RR, relative risk (Cox model for relapse or events for specific LOH); NWTS-4, National Wilms' Tumor Study 4; CI, confidence interval. ^a)^Event-free survival rate according to the Kaplan-Meier analysis and compared with the log-rank test.
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#S0001}
============
Psoriatic arthritis (PsA) is a chronic, inflammatory musculoskeletal disease that is characterized by peripheral arthritis, enthesitis, dactylitis, and spondylitis with or without nail and skin lesions.[@CIT0001],[@CIT0002] It is a systemic disease that develops in up to 40% of psoriasis patients in their lifetime[@CIT0003]--[@CIT0005] with manifestations that include pain, swelling, stiffness, and entheses of the peripheral joints and/or axial skeleton. Dactylitis, fusiform full thickness swelling of fingers or toes, can develop in about half of individuals with PsA. In addition, it is also associated with significant comorbidities such as diabetes mellitus, depression, hypertension, cardiovascular disease, and decreased quality of life.[@CIT0001],[@CIT0005] Although the pathogenesis of PsA has not been fully elucidated, it likely results from a combination of genetic, environmental, and immunologic factors.[@CIT0006]
PsA is a heterogeneous disease with diverse clinical presentations, making early diagnosis and management challenging.[@CIT0007] There are five distinct clinical subtypes: the oligoarticular subtype affects four or fewer joints in an asymmetric distribution; the polyarticular subtype affects at least five joints, is symmetrical, and most similar to rheumatoid arthritis; the distal subtype involves the distal interphalangeal joints of the hands and feet; the arthritis mutilans subtype is the most severe and characterized by bone resorption or osteolysis that results in significant deformities; and the spondyloarthritis subtype, which primarily affects the spine and sacroiliac joints.[@CIT0008] Effective treatment of this rheumatologic disease is critical to minimize damage,[@CIT0009] negative impact on quality of life and to mitigate the risk of comorbid disease.
Current PsA treatment guidelines include nonsteroidal anti-inflammatory drugs, glucocorticoids, oral small molecules (methotrexate, sulfasalazine, cyclosporine, leflunomide, apremilast), tumor necrosis factor alpha (TNF-α) inhibitors (etanercept, adalimumab, infliximab, golimumab, certolizumab), interleukin (IL) 12/23 inhibitors (ustekinumab), IL-17 inhibitors (secukinumab, ixekizumab), CTLA4-Ig (abatacept), and the janus kinase inhibitor (tofacitinib).[@CIT0010]--[@CIT0012] Each therapy has a distinct safety and efficacy profile, and treatment plans should be individualized and optimized by taking into account clinical presentation and medical history. While there are many treatment options for the management of PsA, it can be challenging for practitioners to select among therapies. This review addresses key considerations in patient selection for the treatment of psoriatic arthritis with tofacitinib.
Practical considerations {#S0002}
========================
Tofacitinib (Xeljanz, Pfizer Inc.) is an orally administered inhibitor of predominantly JAK1 and JAK3, with functional selectivity over JAK2.[@CIT0004],[@CIT0013],[@CIT0014] Blockade of the JAK receptor downregulates the production of cytokines important in the pathogenesis of PsA, including TNF-α, IL-17, IL-6, IL-23.[@CIT0014],[@CIT0015] Tofacitinib is currently approved for the treatment of PsA, rheumatoid arthritis, and ulcerative colitis.[@CIT0016]
Tofacitinib is indicated for patients with active PsA who have had a poor response or intolerance to methotrexate or other disease modifying antirheumatic drugs (DMARDs). Since tofacitinib has not been evaluated as a monotherapy in PsA, it is recommended that it be used in combination with a conventional synthetic DMARDs (csDMARD), such as methotrexate, sulfasalazine, and leflunomide. Tofacitinib should not be used in combination with biologic DMARDs or potent immunosuppressants, such as cyclosporine and azathioprine due to risk of added immunosuppression.
Tofacitinib is available as a 5 milligram (mg) or 10 mg immediate release tablet and as an 11 mg extended release tablet. The recommended dosing for both PsA and rheumatoid arthritis is immediate release 5 mg twice daily or extended release 11 mg once daily in combination with non-biologic DMARDs. For ulcerative colitis, the recommended dose is 10 mg twice daily for 8 weeks; followed by 5 or 10 mg twice daily, depending on the patient's response. Dose adjustments are recommend for patients concomitantly taking medications that undergo CYP3A4 inhibition as tofacitinib is primarily metabolized by this pathway.[@CIT0016]
Tofacitinib has no contraindications, but there are FDA-required black box warnings for an increased risk of serious infections and malignancies.[@CIT0016] Patients should be evaluated for latent or active tuberculosis infection prior to treatment with tofacitinib, and anti-tuberculosis therapy should occur prior to initiating this medication. Due to the increased risk of serious infections, the risks and benefits of therapy should be weighed in patients with a history of chronic or recurrent infections. Additionally, lymphomas and other malignancies have been observed in patients treated with tofacitinib and an increased risk of EBV-associated post-transplant lymphoproliferative disease was observed in renal transplant patients.
In February 2019, the FDA reported a safety alert regarding the increased risk of pulmonary embolism and death for the 10 mg twice daily dose of tofacitinib, which was noted during a safety trial in patients with rheumatoid arthritis.[@CIT0017] For this reason, patients taking the 10 mg twice daily dose were decreased to a dose of 5 mg twice daily. Further data is needed to elucidate this risk, and the safety trial is anticipated to be completed by the end of 2019. Tofacitinib 10 mg twice daily is not an approved dose for the treatment of PsA. No data has been provided on the risk of tofacitinib 11 mg extended release.
Laboratory monitoring is required due to potential changes in lymphocytes, neutrophils, hemoglobin, liver enzymes, and lipids. Absolute lymphocyte counts (ALC) should be monitored at baseline and every three months thereafter. Absolute neutrophil counts (ANC) should be measured at baseline, after four to eight weeks of treatment, and every three months thereafter. Hemoglobin is monitored at baseline, after four to eight weeks of treatment, and every three months thereafter. Liver function tests (LFTs) should be routinely monitored, and lipids should be assessed four to eight weeks following initiation of therapy.
Psoriatic arthritis clinical trials {#S0003}
===================================
Two phase III randomized controlled trials (RCTs) evaluated the efficacy and safety of tofacitinib in PsA: OPAL Broaden[@CIT0011] and OPAL Beyond.[@CIT0012] In both studies, the co-primary endpoints were the proportion of patients with a 20% improvement of the American College of Rheumatology criteria (ACR20) at three months and the change from baseline in the Health Assessment Questionnaire-Disability Index (HAQ-DI) at three months. The co-primary efficacy endpoints were achieved in both RCTs ([Table 1](#T0001){ref-type="table"}).Table 1Summary of key Phase III clinical trial results of tofacitinib for the treatment of psoriatic arthritis at month threeEndpointsOPAL BroadenOpal BeyondPlacebo (n=105)Tofacitinib 5 mg (n-107)Tofacitinib 10 mg (n=104)Adalimumab 40 mg Q2W (n=106)Placebo (n=131)Tofacitinib 5 mg (n=131)Tofacitinib 10 mg (n=132)Primary Endpoints ACR2033%50%61%52%24%50%47% ΔHAQ-DI−0.18−0.35−0.40−0.38−0.14−0.39−0.35Secondary Endpoints PASI7515%43%44%39%14%21%43% ACR5010%28%40%33%15%30%28% ACR705%17%14%19%10%17%14% ΔLEI−0.4−0.8−1.5−1.1−0.5−1.3−1.3 ΔDSS−2.0−3.5−5.5−4.0−1.9−5.2−5.4 ΔSF-362.15.25.25.21.75.04.1 ΔFACIT-F3.37.06.06.03.07.05.8[^1][^2]
OPAL Broaden was a 12-month double-blind, active-controlled and placebo-controlled, phase 3 trial investigating the efficacy of tofacitinib compared to placebo or adalimumab in patients with active PsA with an inadequate response to csDMARDs. Inclusion criteria required TNF-α inhibitor naivety. Patients were randomly assigned to receive either tofacitinib 5 mg or 10 mg twice daily, adalimumab 40 mg subcutaneously every two weeks, or placebo with a blinded switch to either 5 mg or 10 mg at three months. Patients were also required to receive a stable dose of csDMARD throughout the trial. This study was completed by 373 patients. With regards to the co-primary endpoints, a significantly greater proportion of patients at month three achieved ACR20 on tofacitinib (5 mg: 50%, *p*=0.01; 10 mg: 61%, *p*\<0.001) compared to placebo (33%); for reference, ACR20 in the adalimumab arm was 52%. Tofacitinib was superior to placebo with respect to the mean change from baseline HAQ-DI score (5 mg: −0.35, *p*=0.006; 10 mg: −0.40, *p*\<0.001) versus placebo (−0.18); for reference mean HAQ-DI change with adalimumab was −0.38.
Secondary endpoints of ACR50 (5 mg: 28%, *p*\<0.001; 10 mg: 40, *p*\<0.001; placebo: 10%), ACR70 (5 mg: 17%, *p*=0.004; 10 mg: 14%, *p*=0.02; placebo: 5%), and an improvement in the Psoriasis Area and Severity Index (PASI) of at least 75% (PASI75; 5 mg: 43%, *p*\<0.001; 10 mg: 44%, *p*\<0.001; placebo: 15%) were assessed and both doses of tofacitinib were found to be superior to placebo. The change in the Leeds Enthesitis Index (LEI) was found to be superior to placebo with tofacitinib 10 mg (−1.5 versus −0.4, *p*\<0.001) but not with the tofacitinib 5 mg dose (−0.8 versus −0.4). Additional pre-specified secondary endpoints of Dactylitis Severity Score (DSS), Medical Outcomes Study 36-Item Short Form Health Survey (SF-36), and the Functional Assessment of Chronic Illness Therapy--Fatigue (FACIT-F) were in the same direction as the primary endpoints but could not be assessed for significance due to hierarchical statistical testing of secondary endpoints.[@CIT0011],[@CIT0018] Radiographs of the hands and feet were assessed at baseline and 12 months using the van der Heijde-modified Total Sharp Score (mTSS). At 12 months, 91--98% of patients across all treatment groups met the radiographic criteria for nonprogression of joint damage, defined as a change in mTSS of ≤0.5, ≤0, or ≤0.66 from baseline.[@CIT0011],[@CIT0019]
In OPAL Broaden, adverse events (AEs) occurred more frequently with tofacitinib and adalimumab than placebo at three months (5 mg: 39%, 10 mg: 45%, adalimumab: 46%, placebo: 35%). The most common adverse events were nasopharyngitis, upper respiratory tract infection, and headache. Serious adverse events (SAEs) occurred more frequently with tofacitinib 5 mg at three months (5 mg: 3%, 10 mg: 1%, adalimumab: 1%, placebo: 1%). At 12 months, three tofacitinib patients reported serious infections (influenza, appendicitis, and pneumonia) compared to one adalimumab patient (herpes simplex and streptococcal pyoderma). Herpes zoster occurred in four patients, all whom were treated with tofacitinib. Four cancers (bladder transitional cell carcinoma, squamous cell carcinoma of the vulva, invasive ductal breast carcinoma, and basal cell cancer) were reported, all in patients who received tofacitinib. One death due to cardiac arrest was reported in a patient treated with tofacitinib 5 mg.
At three months, neutrophil counts were more reduced in the active treatment groups than placebo. Elevations in alanine aminotransferase (ALT) and aspartate aminotransferase (AST) three or more times the upper limit of normal was observed in tofacitinib and adalimumab groups.
OPAL Beyond was a six-month RCT evaluating tofacitinib compared to placebo in patients with active PsA who had an inadequate response to TNF-α inhibitors. This study was completed by 345 patients. Similar to the results found in OPAL Broad, a significantly greater proportion of patients at month three achieved ACR20 on tofacitinib (5 mg: 50%, *p*\<0.001; 10 mg: 47%, *p*\<0.001) compared to placebo (24%). Tofacitinib was also found to be superior with respect to the mean change from baseline HAQ-DI score (5 mg: −0.39, *p*\<0.001; 10 mg: −0.35, *p*\<0.001) compared to placebo (−0.14).
Additionally, both doses of tofacitinib were superior to placebo at three months with respect to ACR50 (5 mg: 30%, *p*=0.003; 10 mg: 28%, *p*=0.007; placebo: 15%), but not ACR70 (5 mg: 17%; 10 mg: 14%; placebo: 10%). With regard to PASI75, the 10 mg dose of tofacitinib was superior to placebo (5 mg: 21%; 10 mg: 43%, *p*\<0.001; placebo: 14%), but not the 5 mg dose. Pre-specified endpoints of resolution of LEI, and improvements in DSS, SF-36, FACIT-F were in the same direction as the primary endpoints but could not be analyzed for statistical significance due to hierarchical statistical testing of secondary endpoints.[@CIT0012],[@CIT0020]
In OPAL Beyond, the rate of AEs was higher with tofacitinib than placebo at three months (5 mg: 55%, 10 mg: 53%, placebo: 44%). The rates of SAEs were 1% for tofacitinib 5 mg, 2% for 10 mg, and 2% for placebo. The most common AEs were upper respiratory infection, nasopharyngitis, and headache. Four tofacitinib patients reported serious infections (bilateral pyelonephritis, parotitis, pneumonia, and oral candidiasis). Herpes zoster infection occurred in three patients, all of whom were treated with tofacitinib. Two major adverse cardiovascular events (MACE) occurred in patients treated with tofacitinib (myocardial infarction in 5 mg and ischemic stroke in 10 mg). No deaths or cancers were reported.
Changes from baseline were similar across all groups for hemoglobin and lymphocyte counts at three months. Changes in neutrophil, platelet, creatinine, creatine kinase, and lipid values were observed to be dose dependent. At six months, elevations of AST and ALT greater than the upper limit of normal were observed in patients treated with tofacitinib (AST: 30%, ALT: 32%), and elevations three more times the upper limit of normal was also observed in six patients.
OPAL Balance is an ongoing, open-label, long-term extension (LTE) study evaluating the safety and efficacy of tofacitinib in patients with PsA.[@CIT0005],[@CIT0021],[@CIT0022] Patients from OPAL Broaden and OPAL Beyond were invited to participate in a three year LTE study within three months of completing or discontinuing the parent study. This study treated 686 patients, who were treated with tofacitinib 5 mg BID or 10 mg BID, depending on treatment response. Primary endpoints were incidents and severity of adverse events and changes in baseline laboratory values. The secondary endpoint was long-term efficacy measured up to month 30. By month 36, 73.2% of patients reported AEs (n=502), 10.5% of patients reported SAEs (n=72), and 2.8% reported a herpes zoster event (n=19). MACE was observed in 0.3% of patients (n=2) and malignancies were reported in 1.9% of patients (n=13). Four deaths were reported that were determined to be unrelated to treatment (pancreatic carcinoma, acute cardiac failure, chronic obstructive pulmonary disease, and pulmonary embolism). Four latent tuberculosis adverse events were reported in patients with a previously negative QuantiFERON. Over 36 months of data demonstrated that the safety profile was similar to those reported in previous trials. No new safety risks were reported and efficacy was maintained.[@CIT0005],[@CIT0022]
A 52 week, randomized, double-blind, phase III clinical trial evaluated the safety and efficacy of tofacitinib in Japanese patients with moderate to severe plaque psoriasis (n=87), PsA (n=12), or both (n=5). Patients were randomized 1:1 to tofacitinib 5 mg or 10 mg twice daily for 16 weeks, open-label 10 mg twice daily for 4 weeks, then the dose was adjust to tofacitinib 5 mg or 10 mg at the investigator's discretion to 52 weeks. The primary end points were PASI75, Physician's Global Assessment (PGA) of "clear" or "almost clear" for psoriasis, and ACR20 for PsA. At week 16, 62.8% of patients treated with tofacitinib 5 mg and 72.7% of patients treated with tofacitinib 10 mg achieved PASI75. The PGA primary end point was achieved by 67.4% and 68.2% of patients, respectively. ACR20 was achieved by all patients with PsA. The most common AEs were nasopharyngitis and herpes zoster, with 17% herpes zoster incidence rate. There were no cardiovascular events, deaths, or malignancies reported.[@CIT0023]
Psoriasis clinical trials {#S0004}
=========================
Multiple, randomized clinical trials have been conducted to evaluate the efficacy, safety, and tolerability of tofacitinib in patients with plaque psoriasis ([Table 2](#T0002){ref-type="table"}).Table 2Summary of key clinical trial results of tofacitinib for the treatment of plaque psoriasisStudyPhaseInterventionPASI 75PGAPhase 2b2 mg25.0%24.5%II5 mg40.8%40.8%10 mg66.7%72.9%Placebo2.0%10.0%OPT 15 mg39.9%41.9%III10 mg59.2%59.2%Placebo6.2%9.0%OPT 25 mg46.0%46.0%III10 mg59.6%59.1%Placebo11.4%10.9%OPT Retreatment Withdrawal Period5 mg56.2%49.9%III10 mg62.3%63.9%Placebo 5 mg23.3%22.9%Placebo 10 mg26.1%18.0% Retreatment Period5 mg36.8%44.8%10 mg61.0%57.1%OPT Compare5 mg39.5%47.1%IIII10 mg63.6%68.2%Etanercept58.8%66.3%Placebo5.6%15.0%[^3]
A phase 2b randomized, placebo-controlled, 12-week clinical trial evaluated the efficacy and safety of tofacitinib in patients with moderate-to-severe chronic plaque psoriasis. Patients received 2 mg, 5 mg, 15 mg twice daily or placebo. The primary endpoint was the proportion of patients achieving PASI75 at week 12. PASI75 response rates were significantly higher for all tofacitinib groups compared to placebo (2 mg: 25.0%, 5 mg: 40.8%, 15 mg: 66.7%, placebo: 2.0%). Higher response rates for PASI50, PASI90, and PGA were also seen across all tofacitinib groups compared to placebo. This study demonstrated that short-term treatment with tofacitinib could be an effective treatment in patients with moderate-to-severe plaque psoriasis.[@CIT0024],[@CIT0025]
Two pivotal multi-center, randomized, placebo-controlled, blinded phase III studies, OPT Pivotal 1 and OPT Pivotal 2, were conducted to evaluate the efficacy of 5 mg and 10 mg tofacitinib compared to placebo for patients with moderate-to-severe chronic plaque psoriasis.[@CIT0026]--[@CIT0029] The two primary endpoints were PASI75 and PGA response at 16 weeks. Both studies achieved the coprimary endpoints, with a significantly greater proportion of patients achieving PASI75 (OPT Pivotal 1: 5 mg: 39.9%, 10 mg: 59.2%, placebo: 6.2%; OPT Pivotal 2: 5 mg: 46.0%, 10 mg: 59.6%, placebo: 11.4%) and PGA (OPT Pivotal 1: 5 mg: 41.9%, 10 mg: 59.2%, placebo: 9.0%; OPT Pivotal 2: 5 mg, 46.0%, 10 mg: 59.1%, placebo: 10.9%) compared to placebo. A dose-dependent improvement was observed with a greater proportion of the tofacitinib 10 mg group achieving PASI75 and PGA response, and in a shorter amount of time, than the tofacitinib 5 mg group. Secondary endpoints of percentage change from baseline of body surface area (BSA), Dermatology Life Quality Index (DLQI), Patient Global Assessment (PtGA), Itch Severity Item (ISI), Nail Psoriasis Severity Index (NAPSI) improvement, and PASI90 were also met.
OPT Retreatment was a randomized, blinded, parallel-group, phase III trial evaluating the effects of tofacitinib 5 mg and 10 mg withdrawal and re-treatment in patients with plaque psoriasis. The primary efficacy endpoints were the proportion of patients maintaining a PASI75 and PGA response in the withdrawal period, and the proportion of patients achieving a PASI75 and PGA response in the retreatment period, after relapsing in the withdrawal phase. At the end of the withdrawal phase a greater percentage of tofacitinib patients maintained PASI75 (5 mg: 56.2%, 10 mg: 62.3%, placebo 5 mg: 23.3%, placebo 10 mg: 26.1%,) and PGA response (5 mg: 49.9%, 10 mg: 63.9%, placebo 5 mg: 22.9%, placebo 10 mg: 18.0%). In the retreatment phase, 36.8% and 61.0% of tofacitinib 5 mg and 10 mg, respectively, achieved PASI75; and 44.8% and 57.1% achieved PGA responses. In sum, this study indicated that tofacitinib could be efficacious after withdrawal and retreatment.[@CIT0030],[@CIT0031]
OPT Compare was a randomized, placebo-controlled, 12 week, non-inferiority phase III trial comparing tofacitinib 5 mg and 10 mg to high dose entanercept or placebo in patients with moderate-to-severe plaque psoriasis. The primary endpoints were the proportion of patients with a PASI75 and the proportion of patients with a PGA response at 12 weeks of treatment. A PASI75 at week 12 was achieved by 39.5% of tofacitinib 5 mg, 63.6% of tofacitinib 10 mg, 58.8% of entanercept, and 5.6% of placebo. Both tofacitinib and etanercept achieved significant PGA reductions compared to placebo (5 mg: 47.1%, 10 mg: 68.2%, etanercept: 66.3%, placebo: 15.0%). Tofacitinib 10 mg was found to be non-inferior to etanercept, but tofacitinib 5 mg was not shown to be non-inferior to etanercept.[@CIT0032]
Although study results demonstrated that both the 5 mg and 10 mg doses were effective in treating plaque psoriasis,[@CIT0033]--[@CIT0035] the 10 mg dose was proven to be more efficacious.[@CIT0036] However, this higher dose is associated with increased safety concerns, leading the FDA to decline approval of tofacitinib for moderate-to-severe plaque psoriasis.[@CIT0027],[@CIT0037]
Safety {#S0005}
======
Integrated safety analyses of phase III studies[@CIT0038] and the LTE study[@CIT0022] found tofacitinib to be well tolerated in patients with psoriatic arthritis with a safety profile that is comparable to other biologics.[@CIT0011],[@CIT0039] The adverse events reported in psoriatic arthritis trials were similar to those reported for rheumatoid arthritis and psoriasis.[@CIT0038],[@CIT0039]
The most common adverse events reported in the phase III trials were upper respiratory infections, nasopharyngitis, and headaches.[@CIT0011],[@CIT0012] Tofacitinib has been associated with an increased risk of herpes zoster, which occurred in the active treatment group of both trials. There is also an increased risk of serious infections, including tuberculosis, pneumonia, and influenza.
Due to the immunosuppression induced by tofacitinib, this medication may increase the risk of malignancies.[@CIT0040] No cancers were observed in OPAL Beyond, however in Opal Broaden, 4 cancers were reported in patients who received tofacitinib compared to zero patients in the placebo or adalimumab group. Two of the malignancies were reported within 30 days of the first dose of tofacitinib. Furthermore, the three OPAL trials reported adverse cardiovascular events of myocardial infarction and ischemic stroke ([Table 3](#T0003){ref-type="table"}).Table 3Summary of key phase III clinical trial safety events for tofacitinib for the treatment of psoriatic arthritisStudyTime PeriodnStudy groupPatients reporting AEs, n (%)Patients reporting SAEs, n (%)Patients reporting serious infections, n (%)Herpes Zoster Infections, n (%)Cardiovascular events, n (%)Cancer, excluding nonmelanoma skin cancer, n (%)OPAL BroadenUp to 3 months105Pooled Placebo37 (35%)1 (1%)0000107Tofacitinib 5 mg42 (39%)3 (3%)01 (1%)02 (2%)104Tofacitinib 10 mg47 (45%)1 (1%)0000106Adalimumab49 (46%)1 (1%)0000Up to 12 months52Placebo to tofacitinib 5 mg36 (69)3 (6%)2 (4%)01 (2%)053Placebo to tofacitinib 10 mg34 (64)4 (8%)0000107Tofacitinib 5 mg71 (66)8 (7%)02 (2%)03 (3%)Tofacitinib 10 mg74 (71)4 (4%)1 (1%)2 (2%)00104Adalimumab76 (72%)9 (8%)1 (1%)02 (2%)0OPAL BeyondUp to 3 months131Placebo58 (44%)3 (2%)0000131Tofacitinib 5 mg72 (55%)1 (1%)01 (1%)00132Tofacitinib 10 mg70 (53%)3 (2%)2 (2%)1 (1%)00Up to 6 months66Placebo to tofacitinib40 (61%)2 (3%)000065Placebo to tofacitinib 10 mg38 (58%)1 (2%)0000131Tofacitinib 5 mg93 (71%)5 (4%)2 (2%)1 (1%)1 (1%)0132Tofacitinib 10 mg96 (73%)8 (6%)2 (2%)2 (2%)1 (1%)0[^4]
Tofacitinib has also been associated with a variety of laboratory abnormalities. Due to the blockade of myelopoietic growth factor signals,[@CIT0003] there is an increased risk of cytopenia. Neutropenia and anemia have been observed, but are mild at lower doses. In addition, liver enzyme levels were found to be elevated compared to placebo. Tofacitinib is also associated with a dose-dependent increase in lipid parameters, including total cholesterol, low-density lipoprotein (LDL), high-density lipoprotein (HDL), and triglycerides.[@CIT0041],[@CIT0042] In OPAL Beyond, dose dependent changes were also observed in lymphocyte count, platelet count, creatinine, and creatine kinase. For these reasons, routine laboratory monitoring is recommended, and clinicians should follow up on any abnormalities.
The long-term safety of tofacitinib in psoriatic arthritis was evaluated in OPAL Balance. No new safety risks were identified and tofacitinib's safety profile was similar to phase III trials.
Discussion {#S0006}
==========
Treatment options in PsA have expanded, and choosing among treatments can be a complex decision. Due to the heterogeneous pathogenesis of PsA, patients may not achieve satisfactory disease control with available therapies, which creates a need for alternative treatment options. While TNF-α inhibitors are recommended as first line treatment for active PsA,[@CIT0010],[@CIT0013] tofacitinib can provide an alternative option when injectable biologics are either contraindicated or provide an inadequate response.[@CIT0010] Prior to initiating tofacitinib, a thorough assessment of patient preference, disease presentation, and comorbidities is critical.
As an oral medication, tofacitinib provides patients with an alternative to parenteral therapy. Tofacitinib has a similar efficacy profile to biologics, while also offering patients convenience and independence, potentially improving medication compliance and decreasing healthcare costs. Tofacitinib is an alternative to apremilast, an oral phosphodiesterase-4 inhibitor also indicated for psoriatic arthritis.[@CIT0043] While apremilast has a favorable safety profile, its efficacy is considerably lower than biologics,[@CIT0044] whereas tofacitinib has been demonstrated to have efficacy comparable to biologics.
Tofacitinib can also be considered in patients with difficult to treat disease. The results of OPAL Beyond, which included patients with inadequate response to at least one TNF-α inhibitor, indicate that tofacitinib may be effective in patients with treatment-resistant psoriatic arthritis. In addition, both phase III clinical trials reported improvement in enthesitis and dactylitis, as measured by the LEI and DSS, respectively. These clinical manifestations are associated with an increased burden of disease and decreased quality of life, underscoring the importance of tofacitinib's efficacy in these domains.[@CIT0005],[@CIT0045]
Furthermore, OPAL Broaden evaluated the effect of tofacitinib on radiographic outcomes and reported that greater than 90% of patients achieved radiographic nonprogression.[@CIT0011],[@CIT0019] The overall rate of radiographic progression was low in all patients.
Although tofacitinib has not received an FDA indication for plaque psoriasis, it has demonstrated efficacy in patients with moderate-to-severe chronic plaque psoriasis in multiple clinical trials, as well as in psoriatic arthritis patients with skin manifestations. Tofacitinib can be considered as a treatment option in patients with psoriatic arthritis and comorbid skin disease. Tofacitinib is also being investigated as a treatment for alopecia areata, vitiligo, atopic dermatitis, and dermatomyositis and may be considered in psoriatic arthritis patients with comorbid disease.[@CIT0046],[@CIT0047]
There are a number of considerations with regards to tofacitinib's safety profile. There is a black box warning of an increased risk of serious infections and malignancy. Patients should be tested for latent tuberculosis and treated prior to initiation of therapy, and also monitored for signs of serious infection. Patients with chronic, active, or untreated latent hepatitis infections were excluded from tofacitinib clinical trials and thus the risk of chronic viral hepatitis reactivation is unknown. Due to the immunosuppressive effects of tofacitinib therapy, patients should be screened for viral hepatitis in accordance with clinical guidelines prior to initiating therapy.[@CIT0016]
With regards to the increased risk of malignancy, a pooled analysis from rheumatoid arthritis trials reported that overall rates and types of malignancy remained stable over time with increasing exposure to tofacitinb.[@CIT0048] A meta-analysis of rheumatoid arthritis trials also reported that the risk of malignancy is similar to nonbiologic and biologic DMARDs, concluding that tofacitinib does not increase rate of malignancies in rheumatoid arthritis.[@CIT0040] However, due to these concerns, caution should be exercised in patients with a personal or family history of malignancy. Longer-term studies and registries are needed to evaluate the risk of malignancy in patients taking tofacitinib for psoriatic arthritis.
Furthermore, there was a recent safety alert regarding the risk of pulmonary embolism in rheumatoid arthritis patients treated with tofacitinib 10 mg twice daily. The 36-month OPAL Balance results also reported a death due to pulmonary embolism. Although further data are needed to fully assess this risk, patients should be evaluated for increased risk of a thromboembolic event prior to initiation of therapy.[@CIT0017]
Tofacitinib is also associated with an increased risk of herpes zoster infection.[@CIT0048]--[@CIT0051] Risk factors include increased age, Asian race, tofacitinib dose, concomitant steroid therapy, and prior biologic exposure.[@CIT0052] The nonlive recombinant zoster vaccine (RZV) is the preferred herpes zoster vaccine for psoriasis and psoriatic arthritis patients. Administration is recommended prior to initiation of therapy but may be safely administered during concurrent tofacitinib treatment. The live-attenuated vaccine (Zostavax) can also be administered but should be given prior the start of therapy.[@CIT0053]
Laboratory abnormalities are also associated with tofacitinib use, requiring monitoring of lymphocyte, neutrophil, hemoglobin, lipid levels, and LFTs. Since frequent monitoring can be an inconvenience to patients, this should be discussed prior to initiating therapy. Due to the increased risk of cytopenias, tofacitinib should not be initiated in patients with an ALC of less than 500 cells/mm^3^, ANC of less than 1000 cells/mm^3^, and hemoglobin of less than 9 g/dL. A pooled analysis of tofacitinib in patients with rheumatoid arthritis reported that patients with an ALC less than 500 cells/mm^3^ had an increased risk of infection and identified trends towards an increased rate of herpes zoster and opportunistic infections with lower ALC values.[@CIT0054] Dosing should be discontinued in patients with an ALC and ANC of less than 500 cells/mm^3^. Treatment should be interrupted in patients with a hemoglobin of less than 8 g/dL or a decrease of more than 2 g/dL.[@CIT0016] Compared to placebo, tofacitinib was observed to have a higher incidence of liver enzyme elevations, however, no drug-induced liver injury were reported and incidences of ALT and AST greater than three times the upper limit of normal were rare. Routine monitoring of liver enzymes is recommended and elevations should be investigated.
Of particular concern is the observed increased lipid level, given the association between PsA and cardiovascular disease.[@CIT0055] Pooled analysis from OPAL Broaden, Beyond, and Balance reported an increased in LDL and HDL of 9--14% for tofacitinib 5 mg and 10 mg at three and six months, however no meaningful changes in the ratio of LDL to HDL and total cholesterol to HDL were observed.[@CIT0056] Additionally, five patients experienced MACE, of which two were fatal. A pooled analysis reported the risk of MACE in PsA patients treated with tofacitinib to be similar to the general PsA population.[@CIT0056] Long-term data is needed to fully understand the cardiovascular risk of tofacitinib treatment and patients with a history significant for cardiovascular disease should be closely evaluated.
Conclusion {#S0007}
==========
Tofacitinib is an efficacious treatment option for patients with PsA as demonstrated in two placebo-controlled phase III RCTs using the ACR response definitions and the HAQ-DI. In addition, tofacitinib improved patient reported outcomes and radiographic non-progression in patients with PsA. Tofacitinib therapy, like other biologicals, requires monitoring by the prescribing physician. As an oral therapy, it offers patients an alternative to injectable therapies. The role of tofacitinib in the management of PsA will be elucidated as clinical experience increases and long-term efficacy and safety information become available.
Disclosure {#S0008}
==========
KL, KMB, MPS have no conflicts of interest to disclose. AMO is a consultant for Eli Lilly and Company, Novartis, Pfizer and UCB, and has received grant/research support to Johns Hopkins University from AbbVie, Celgene, Eli Lilly and Company, Horizon, Janssen, and Novartis. WL is funded in part by a grant from the National Institutes of Health (U01AI119125) and has received research grant funding from Abbvie, Amgen, Janssen, Novartis, Regeneron, and Sanofi.
[^1]: **Notes:** All endpoints tested at three months; ACR20/50/70: a minimum of 20, 50, and 70% improvement in the ACR response; PASI75: at least a 75% improvement in the PASI score; ΔHAQ-DI/LEI/DSS/SF-36/FACIT-F: change from baseline score.
[^2]: **Abbreviations:** ACR, American College of Rheumatology; DSS, Dactylitis Severity Score; HAQ-DI, Health Assessment Questionnaire Disability Index; FACIT-F, Functional Assessment of Chronic Illness Therapy--Fatigue; LEI, Leeds Enthesitis Index; PASI, Psoriasis Area-and-Severity Index; SF-36, Study 36-Medical Outcomes Study 36-Item Short Form Health Survey.
[^3]: **Notes:** Phase 2b and OPT Compare endpoints tested at week 12; OPT 1, OPT 2, and OPT Retreatment endpoints tested at week 16; PASI 75, at least 75% reduction in the Psoriasis Area and Severity Index score; PGA, Physician's Global Assessment; Dosing of etanercept: 50 mg SQ twice weekly; tofacitinib dosed twice daily.
[^4]: **Abbreviations:** AE, adverse event; SAE, serious adverse event (any AE resulting in death, is life-threatening, requires inpatient hospitalization, causes prolongation of existing hospitalization, or results in persistent or significant disability or incapacity).
| {
"pile_set_name": "PubMed Central"
} |
1. Introduction
===============
Meckel diverticulum (MD) is the most common congenital anomaly of the gastrointestinal tract.^\[[@R1]\]^ MD originates from an incomplete obliteration of the omphalomesenteric or vitelline duct, which occurs around the fifth week of gestation.^\[[@R2]\]^ It usually appears as a pouch, 3 to 6 cm in length, arising from the antimesenteric border of the ileum at variable lengths from the ileocecal junction.^\[[@R3]\]^ MD has a complication rate of approximately 4%.^\[[@R4]\]^ Bowel obstruction, gastrointestinal bleeding, acute inflammation, and umbilical abnormalities are the most common presentations of MD in children.^\[[@R5]--[@R8]\]^
The mesodiverticular band (MDB) is an embryologic remnant of the vitelline circulation that carries the arterial supply to the MD. In the event of an error of involution, a patent or nonpatent arterial band persists and extends from the mesentery to the apex of the antimesenteric diverticulum. This creates a snare-like opening through which bowel loops may herniate and become obstructed.^\[[@R9]\]^ This kind of internal hernia is a very rare and often overlooked cause of small bowel obstruction.
Here, we report 2 rare cases of internal hernia in children caused by a MDB, and we reviewed the pediatric literature about this rare condition in children.
2. Cases
========
2.1. Presenting concerns
------------------------
Case 1 was a 5-year-old boy who presented to our Emergency Department with colicky abdominal pain. The pain started 1 day before admission and was diffused to all abdominal quadrants. He also had 5 episodes of emesis, the last with bilious vomiting. No past surgical or trauma history was declared.
Case 2, a 12-year-old boy, presented to our Emergency Department complaining of colicky abdominal pain. The pain started 1 day before admission and was located in the mesogastric position. He also had 2 episodes of nonbilious emesis. He had no past surgical or trauma history.
In both patients, family and psychosocial history were negative.
2.2. Clinical findings
----------------------
On physical examination, case 1 showed distension and tenderness of the abdomen. On physical examination, case 2 presented a mild distension of the abdomen and mild tenderness in the mesogastric abdomen. In both cases, no mass was palpable and no peritoneal signs were noted.
2.3. Diagnostic focus and assessment
------------------------------------
In case 1, an abdominal x-ray was performed showing an occlusive picture without an apparent etiology (Fig. [1](#F1){ref-type="fig"}). Abdominal ultrasound was negative for intussusception, but revealed a moderate amount of fluid in the pelvis.
{#F1}
In case 2, an abdominal x-ray was performed showing an occlusive picture without an apparent etiology. Abdominal ultrasound was also conducted to exclude intussusception or other causes of intestinal occlusion, but it revealed only dilated small bowel loops with a small amount of fluid in the pelvis.
2.4. Therapeutic focus and assessment
-------------------------------------
In case 1, an urgent surgical exploration was performed. Laparotomy showed loops of the ileum herniating through an orifice formed by a MDB (Fig. [2](#F2){ref-type="fig"}). The herniated ileal loops were not compromised. The MDB was ligated and cut, freeing the incarcerated small bowel. The MD (Fig. [3](#F3){ref-type="fig"}) was then excised, and an end-to-end ileal anastomosis was performed.
{#F2}
{#F3}
In case 2, owing to the persistence of colicky pain episodes with a new episode of bilious vomiting and the absence of a surgical history, an urgent laparoscopic exploration was performed to obtain the correct diagnosis. A 10-mm trocar was inserted through an umbilical incision. On exploration, the proximally dilated jejunum was encountered (Fig. [4](#F4){ref-type="fig"}). Some loops of the jejunum that had herniated through an orifice formed by a MDB were found (Fig. [5](#F5){ref-type="fig"}). The vascularity of herniated ileal loops was not compromised. After the positioning of 2 operative 5-mm trocars in triangulation, the MDB was coagulated and cut, freeing the incarcerated small bowel. The MD was then exteriorized from the umbilical incision and excised by a 3-cm ileal resection and an end-to-end ileal anastomosis.
{#F4}
{#F5}
2.5. Follow-up and outcomes
---------------------------
The postoperative course was uneventful in both cases, and the boys were discharged on postoperative day 6 without any complications.
2.6. Timeline
-------------
Acute signs, symptoms, and clinical findings upon admission to the Emergency Room raised the suspicion of small bowel occlusion. Abdominal x-ray and abdominal ultrasound were immediately performed and showed an occlusive picture without an apparent etiology. To prevent strangulation and gangrene of the bowel, laparotomy in case 1 and minimally invasive surgery in case 2 were performed and showed that the occlusion was because of an internal hernia caused by a MDB. The MDB was ligated and cut in both cases and the MD was then excised.
3. Discussion
=============
Internal hernia is a rare cause of small bowel obstruction and is defined as a condition where a viscus protrudes through an opening within the abdominal cavity.^\[[@R1]\]^ Symptoms of patients affected by internal hernia include colicky abdominal pain, vomiting, distention, or constipation.^\[[@R5]--[@R8]\]^
The MDB is an error of involution of the vitelline circulation, which carries the arterial supply to the MD.^\[[@R9]\]^ This embryologic remnant may cause different kinds of complications, including hemorrhage and hemoperitoneum owing to traumatic rupture of the patent MDB^\[[@R10]\]^ as well as intermittent small bowel obstruction caused by torsion around the MDB or acute small bowel obstruction caused by an internal hernia,^\[[@R11]\]^ as shown in our cases. The literature reports this last complication as a cause of sudden infant death syndrome for incarceration and the infarct of the herniated loops of small bowel beneath the band.^\[[@R12]\]^ MDB may also be misdiagnosed as acute appendicitis.^\[[@R13]\]^
In addition to our 2 cases, the review of the literature showed 8 other cases of symptomatic MDB in pediatric patients (Table [1](#T1){ref-type="table"}).^\[[@R5],[@R12]--[@R17]\]^ The age of onset ranged from 10 days to 12 years. All cases reported an intestinal occlusion as the clinical picture at onset. Including our cases, internal hernia was the cause of the obstruction in 8 cases, whereas in 2 patients, the occlusion was because of a direct compression of the small bowel by the MDB. Kunitsu et al reported that a relatively long MDB can lead to pathogenesis over a wide age range, from childhood to adulthood, with regard to the internal hernia of an intestinal loop.^\[[@R17]\]^ In contrast, a short MDB can exert direct compression on the digestive tract before or shortly after birth.
######
Review of published pediatric reports with mesodiverticular band MDB, including our presented cases.
All patients were approached with emergent laparotomy except our second case and another one who was treated with an urgent laparoscopy, which confirmed the diagnosis and resolved the clinical condition. Laparoscopy is widely used in pediatric age as an alternative procedure to open surgery for many surgical problems. In the last 2 decades, the use of laparoscopy in the diagnostic confirmation and subsequent laparoscopic excision of MD in children has gained popularity. Laparoscopic intracorporeal or laparoscopic-assisted extracorporeal resection of MD is the possible approach.^\[[@R18]\]^ One of the concerns in total laparoscopic intracorporeal resection of MD is the failure to perform segmental resection of MD because there may be a risk of leaving ectopic gastric mucosa in the adjacent ileum.^\[[@R18]\]^ For this reason, many authors prefer a laparoscopic-assisted technique with an extracorporeal segmental resection of the MD and an open end-to-end intestinal anastomosis of the adjacent ileum.^\[[@R19]\]^ Recently, single incision laparoscopic surgery (SILS) has emerged as a new technique in minimally invasive surgery, but limited numbers of reports on SILS in MD management in children are available.^\[[@R20],[@R21]\]^ As is the usual practice in our Department, the MD in our second case was laparoscopically individuated and excised using a laparoscopic-assisted technique.
For selected children with bowel obstruction, laparoscopy may be an alternative to traditional laparotomy and is associated with reduced morbidity and length of hospital stay. Nevertheless, \>30% of conversions to laparotomy may be required.^\[[@R22]\]^ Our first case was submitted to laparotomy because the abdomen showed distension. However, in our second case, laparoscopy was performed because of the unknown origin of the obstruction in the absence of surgical or trauma history and mild distension of the abdomen, which permitted a sufficient working space during laparoscopy.
In conclusion, although MDBs causing internal hernias are very rare, they should be considered in patients with a clinical picture of small bowel obstruction. In these cases, early surgery is important to prevent strangulation and gangrene of the bowel and to avoid dramatic events. Moreover, laparoscopy is a safe and effective technique in these patients, especially in children with mild abdominal distention without surgical or trauma history, highlighting that further studies on the value of laparoscopy for the treatment of small bowel obstruction in pediatric patients are urgently needed.
4. Patients' parents' perspective
=================================
We are very grateful to the surgeons for their diagnostic and therapeutic approach, which was associated with a favorable clinical evolution.
5. Informed consent
===================
The patients' parents provided their written informed consent for the publication of this study.
Acknowledgments
===============
The authors thank all the surgeons, paediatricians, and nurses involved in the management of these children, as well as their parents.
Abbreviations: MD = Meckel\'s diverticulum, MDB = mesodiverticular band, SILS = single incision laparoscopic surgery.
The authors report no conflicts of interest.
| {
"pile_set_name": "PubMed Central"
} |
Due to restrictions imposed by the Medical Ethics Committee of the Erasmus University Medical Center, requests and inquiries may be sent to the Management Team of the Healthy Ageing and Intellectual Disability Study (<info.gvg@erasmusmc.nl>).
Introduction {#sec005}
============
Intellectual disability is a disability characterised by significant limitations in both intellectual functioning and in adaptive behavior, which covers many everyday social and practical skills. This disability originates before the age of 18, resulting from genetic, neurological, nutritional, social, traumatic or other factors occurring prior to birth, at birth, or during childhood up to the age of brain maturity, that affect intellectual development \[[@pone.0129585.ref001]\]. High health care costs of this population are caused not only by their lifelong dependence on care and support, but also by higher prevalence rates of a large number of health conditions, compared to the general population \[[@pone.0129585.ref002]\]. These health conditions are partly related to the cause of the intellectual disability (such as congenital heart defects in Down syndrome, or cerebral palsy), although lifestyle-related health conditions are also highly prevalent \[[@pone.0129585.ref003]\]. In line with these findings, recent research has shown consistently low levels of physical activity and fitness in people with ID, compared to the general population \[[@pone.0129585.ref004]\], \[[@pone.0129585.ref005]\]. These are likely to cause a cascade of health problems and decline of daily functioning, as is described in the conceptual model of Disability-Associated Low Energy Expenditure Deconditioning Syndrome (DALEEDS) \[[@pone.0129585.ref006]\]. The metabolic effects of physical inactivity, combined with the metabolic effects of antipsychotic drug use, are believed to result in the increased risk of diabetes and other cardiovascular risk factors (metabolic syndrome) in adults with ID \[[@pone.0129585.ref003]\]. Furthermore, nowadays life expectancy of people with ID is approaching that of the general population, resulting in a larger than ever older population with ID. Their ageing process is generally not accompanied by good health, with higher rates of multimorbidity and frailty at younger ages than the general population. \[[@pone.0129585.ref007]\]
*Handgrip strength (HGS*), a measure of maximum voluntary force of the hand, has been described as the simplest method in assessing muscle function \[[@pone.0129585.ref008]\]. It has been shown to be a valid, reliable and feasible measure in multiple populations. It is characterized by overall upper extremity muscle strength \[[@pone.0129585.ref009]\], and correlates with lower extremity strength and power \[[@pone.0129585.ref010]\] \[[@pone.0129585.ref011]\]. Furthermore, this technique has been demonstrated to be a reliable and valid screening tool in the assessment of frailty risk in hospital admission \[[@pone.0129585.ref012]\] as well as a useful indicator of nutritional status in the non-hospitalised population, particularly in identifying individuals with sarcopenia \[[@pone.0129585.ref013]\]. It is an important marker in the assessment of sarcopenia \[[@pone.0129585.ref010]\], nutritional status \[[@pone.0129585.ref014]\], frailty \[[@pone.0129585.ref015]\], and muscular strength as a component of physical fitness \[[@pone.0129585.ref016]\]. Grip strength is a predictor of premature mortality, earlier onset of disability, postoperative complications, increased length of hospital stay \[[@pone.0129585.ref009]\], fractures, and cognitive decline in older adults \[[@pone.0129585.ref017]\] \[[@pone.0129585.ref018]\]. Moreover, data from literature tend to support the fact that HGS may be a good predictor of body cell mass depletion, functional decrease \[[@pone.0129585.ref008]\] during hospitalization, post-surgery complications, and mortality \[[@pone.0129585.ref019]\]. Therefore, it might be valuable to introduce this measurement into the population of people with intellectual disability as a marker for sarcopenia, nutritional status, frailty, and physical fitness.
HGS is a non-invasive and quick measurement of grip strength through use of a hand dynamometer, and is increasingly being used in clinical settings, such as in geriatric practice \[[@pone.0129585.ref010]\]. Measuring grip strength with a hand dynamometer was found to be feasible and reliable in older adults with intellectual disabilities (ID) \[[@pone.0129585.ref020]\]. More information is required about the reference values on grip strength before introducing grip strength measurement into routine diagnostic work-up of adults with ID. The aims of this study were to investigate grip strength in a large sample of people with intellectual disabilities, to establish reference values for adults with intellectual disabilities (ID) and to compare to adults without intellectual disability across age and gender.
Methods {#sec006}
=======
Study design and participants {#sec007}
-----------------------------
This study analysed pooled baseline data from two independent studies of people with ID: Special Olympics Funfitness Spain (n = 801) and the Dutch cross-sectional study 'Healthy aging and intellectual disabilities' (HA-ID; n = 725). For all 1536 adults with ID, details about design, recruitment and representativeness of the sample have been presented elsewhere \[[@pone.0129585.ref021]\] \[[@pone.0129585.ref022]\].
The hours per week of physical activity or sport in Special Olympics Funfitness Spain and physical activity based on pedometer step counts in a subsample of HA-ID were recorded to classify the participants by activity levels.
The waist circumference was measured with a tape measure. Data collection took place between February 2009 and April 2013. Ethical approval was provided by the Medical Ethical Committee of the Erasmus Medical Center (MEC 2008--234) and Ethical Committee of the Faculty of Health Sciences at University of Malaga (FCCSS 314) and by the ethical committees of the participating ID care provider services. Written informed consent was obtained from all participants. This study followed the guidelines of the Declaration of Helsinki (Helsinki, 2008).
Spanish Special Olympics Funfitness {#sec008}
-----------------------------------
This study was focused on describing the level of physical fitness of adults with ID and results were compared across gender and level of sports participation.
All of the participants had been diagnosed with mild ID by a specialised doctor and their parent and/or guardian confirmed the diagnosis. All of the individuals appeared to be healthy, which was determined by their health history obtained from the participants and their parent and/or guardian. The exclusion criteria were: 1) any contraindications to exercise as assessed by a medical history questionnaire; 2) documented atherosclerotic heart disease; 3) documented atlantoaxial instability; 4) uncorrected congenital heart disease; and 5) an implanted pacemaker.
Before starting the investigation we guaranteed the protection of confidential information of all participants \[Act 15/1999 on Protection of Personal Data\]. In all cases we ensured the anonymity of participants. It was also stressed at all times that participation in the study is voluntary and they gave informed consent. All of the participants received counseling and education from the physical therapist and after the screening they were provided with musculoskeletal-specific patient education materials tailored for persons with lower reading levels.
'Healthy aging and intellectual disabilities' (HA-ID) {#sec009}
-----------------------------------------------------
This study was conducted to measure a wide range of health aspects of older adults with ID, including level of physical fitness, and results were compared with reference values from the general population.
All of the participants of the HA-ID were 50 years and over, and had been diagnosed with ID (varying from mild to profound). The Revised Physical Activity Readiness Questionnaire (rPARQ) was administered by the professional caregivers in advance of participation in the physical fitness tests, to determine if the participant could participate safely in these tests \[[@pone.0129585.ref023]\] \[[@pone.0129585.ref024]\]. Statements about the protection of confidential information and anonymity of participants were included in the informed consent form. It was stressed that participation in this study was voluntary, and that participants could withdraw at any time.
Procedure {#sec010}
---------
Data were collected as part of an extensive physical fitness assessment. Assessments were guided by test instructors, who all were physiotherapists, occupational therapists or physical activity instructors with experience with individuals with ID. They all received an instruction manual and followed two days of training for the execution of all assessments.
Handgrip test {#sec011}
-------------
Grip strength \[[@pone.0129585.ref025]\] was measured with the Jamar Hand Dynamometer \[[@pone.0129585.ref026]\]. Reliability and validity in the general population was good \[[@pone.0129585.ref027]\] \[[@pone.0129585.ref028]\]. Test-retest reliability in adults with ID was good (ICC 0.94 \[same day interval\] and 0.90 \[two-week interval\]) \[[@pone.0129585.ref020]\]. In a previous report from the HA-ID study, selective loss of participation was reported. The handle of the dynamometer was placed in the second smallest position according to the instrument's instructions. The middle phalanges had to rest on the handle, otherwise the position was adjusted. An example of the test was provided by the test instructor squeezing a rubber ball. Subsequently, the participant was allowed to squeeze the ball, to assure understanding of the task. The participant squeezed the dynamometer to his or her maximum ability in seated position, according to the recommendations of the American Society of Hand Therapy ASHT \[[@pone.0129585.ref029]\]. The best result of three attempts for both the left and right hand (with a one-minute pause between attempts) was recorded, in kilograms (kg). The test instructor had to be convinced that the participant squeezed with maximal effort; otherwise the result was not recorded. In order to check this, test instructors looked at facial expressions, contracting muscles of the arm and hand, the phalanges turning white, and the consistency of the three attempts.
Comparative reference values from handgrip test for the general population {#sec012}
--------------------------------------------------------------------------
Two meta-analyses were available for grip strength in the general population, the first provided adult normative values (12 studies, 3317 subjects) \[[@pone.0129585.ref030]\], the second provided normative values for adults aged 75 years or over (7 studies, 739 subjects) \[[@pone.0129585.ref031]\]. Means with 95% confidence intervals are presented separately for men and women, for the left hand side and the right hand side, and for 5-year age categories. In these normative values, the distinction between the dominant and the nondominant side is lacking, while the result of the dominant side is widely used to indicate maximum grip strength. To enable comparison with maximum grip strength in our current study, the right hand values of the general population meta-analyses are used as normative values, since this is the dominant side for most people. According to the authors, individual patients whose score is less than the lower limit of the 95% confidence interval of a specific stratum can be considered to be impaired \[[@pone.0129585.ref030]\], or at least below average \[[@pone.0129585.ref031]\].
Statistical analysis {#sec013}
--------------------
Descriptive statistics including measures of central tendency and dispersion were calculated for the hand grip strength across subgroups. Normal distribution was evaluated with the Kolmogorov--Smirnov test. Analysis was performed with SPSS version 21 for Mac.
Results {#sec014}
=======
Descriptives {#sec015}
------------
Participants in this study were 1526 individuals with ID who took part in the HA-ID study and Spanish Special Olympics Games. Characteristics of the study sample are shown in [Table 1](#pone.0129585.t001){ref-type="table"}. The grip strength result for people with ID across gender and age subgroups is presented with CI95% values, and ranges from 25.5--31.0 kg in younger males to 4.3--21.6 kg in older females.
10.1371/journal.pone.0129585.t001
###### Descriptive characteristics of study participants (n = 1526).
{#pone.0129585.t001g}
Descriptive variables Total sample (n = 1526) FunFitness sample (n = 801) HA-ID sample (n = 725)
------------------------------------------------ ------------------------- ----------------------------- ------------------------
Gender (n° males, %) 914 (59.5%) 508 (63.5%) 370 (51%)
Age, years (Mean (SD) 46.9 (16.54) 34.3 (9.5) 61.66 (8.01)
Height, meters (Mean (SD) 1.62 (0.12) 1.63 (0.11)
Weight, kilograms) (Mean (SD) 73.1 (16.4) 73.36 (15.7)
Body Mass Index (Mean (SD) 29.2 (5.1) 27.7 (5.3)
Waist Circumference, centimeters (Mean (SD) 95.1 (13.0) 93.5 (15.3)
Physical Activity /hours per week, n° (%)
Lower 2 hours per week 424 (53.0%)
Higher 2 hour per week 377 (46.8%)
HA-ID: Activity based on pedometer step counts
HA-ID: less than 7500 steps/day 161 (22.2%)
HA-ID: 7500 steps/day or more 90 (12.4%)
HA-ID: unknown 474 (65.4%)
Reference values {#sec016}
----------------
The grip strength results for people with intellectual disabilities across gender and age categories is presented in [Table 2](#pone.0129585.t002){ref-type="table"} (males) and [Table 3](#pone.0129585.t003){ref-type="table"} (females). These results are graphically presented in [Fig 1](#pone.0129585.g001){ref-type="fig"} (males) and [Fig 2](#pone.0129585.g002){ref-type="fig"} (females).
{#pone.0129585.g001}
{#pone.0129585.g002}
10.1371/journal.pone.0129585.t002
###### Mean and CI95% values in kg for females age subgroups in General population (from Bohannon´s studies) and adults with ID.
{#pone.0129585.t002g}
General Population (GP) ID population
-------- ------------------------- --------------- -------------- ------ -------------- ----- ------- ------------ ------- ------------
20--24 133 27.9 (23.1--32.6) 30.6 (26.7--34.4) 48 19.11 16.8--21.3 18.79 16.4--21.1
25--29 142 30.8 (27.2--34.5) 33.8 (29.5--38.1) 57 15.23 13.0--17.3 15.77 13.8--17.6
30--34 141 31.8 (29.0--34.4) 33.8 (28.9--38.6) 52 19.15 16.9--21.3 19.43 17.4--21.4
35--39 142 30.2 (25.8--34.5) 33.2 (28.6--37.8) 48 16.59 14.0--19.1 18.26 15.2--21.2
40--44 133 29.3 (24.5--34.0) 32.8 (28.0--37.6) 32 16.05 13.1--18.9 17.03 14.0--19.9
45--49 133 30.8 (25.8--35.7) 33.9 (28.9--39.0) 38 15.59 12.6--18.5 15.95 13.4--18.3
50--54 116 28.8 (24.0--33.5) 30.9 (26.7--35.2) 96 18.72 17.0--20.4 19.7 17.9--21.4
55--59 123 27.2 (24.6--29.5) 29.9 (26.4--33.6) 103 18.2 16.7--19.6 18.58 17.0--20.1
60--64 132 23 (18.6--27.3) 25.9 (22.2--29.6) 69 18.07 16.3--19.7 18.28 16.4--20.1
65--69 118 22.9 (19.6--26.2) 25.6 (22.5--28.8) 50 19.19 17.3--20.9 19.52 17.7--21.3
70--74 166 22.5 (19.1--25.8) 24.2 (20.7--27.8) 41 16.22 14.3--18.1 16.41 14.1--18.6
75--79 207 18.8 (14.1--23.5) 21.6 (18.6--24.6) 23 19 16.2--21.7 16.57 12.9--20.2
80--84 166 17.1 (14.5--19.6) 17.3 (14.8--19.9) 5 16 8.4--23.5 17.2 9.3--25.0
85--89 96 15.7 (12.2--19.2) 17.1 (12.8--21.4) 4 13 4.3--21.6 11 0.4--21.5
10.1371/journal.pone.0129585.t003
###### Mean and CI95% values in kg for males age subgroups in General population (from Bohannon´s studies) and adults with ID.
{#pone.0129585.t003g}
General Population (GP) ID population
-------- ------------------------- --------------- -------------- ------ -------------- ----- ------- ------------- ------- ------------
20--24 134 47.4 (38.8--56.1) 53.3 (45.2--61.5) 48 28.89 25.5--31.0 28.3 26.2-31-5
25--29 149 50 (41.1--58.9) 53.9 (44.3--63.6) 57 30.36 28.7--34.3 31.57 27.7--32.9
30--34 120 49.2 (40.4--57.9) 52.8 (44.1--61.5) 52 28.59 25.0--30.2 27.65 26.0--31.1
35--39 117 51.6 (44.0--59.3) 53.3 (44.0--62.6) 48 29.5 27.4--32.4 29.97 27.2--31.7
40--44 111 49.8 (42.5--57.1) 54.1 (47.1--61.2) 32 27.24 25.52--30.8 28.21 24.7--29.6
45--49 110 48.7 (40.3--57.2) 50.4 (42.5--58.3) 38 27.86 23.4--30.7 27.09 24.0--31.6
50--54 100 45.2 (39.4--51.1) 50.6 (44.2--56.9) 96 27.04 25.5--29.0 28.07 24.6--28.9
55--59 100 41 (33.7--48.4) 44.1 (36.7--51.4) 103 26.78 24.9--30.2 27 24.6--28.9
60--64 82 38.7 (33.4--44.0) 41.7 (36.8--46.7) 69 26.78 25.8--29.8 28.01 24.6--29.6
65--69 120 38.2 (32.0--44.4) 41.7 (35.4--47.9) 50 27.17 26.0--27.0 28.28 24.7--29.4
70--74 217 36.2 (30.3--42.1) 38.2 (32.0--44.5) 41 25.43 23.0--29.8 26.44 22.0--28.8
75--79 114 31.1 (25.6--36.6) 33 (27.1--38.9) 23 23.41 20.3--27.0 23.71 20.9--25.0
80--84 107 27 (22.2--31.8) 30.1 (14.8--19.9) 5 25.5 21.3--29.0 24 20.3--28.4
85--89 49 25.1 (20.5--29.7) 25.8 (12.8--21.4) 4 16.33 15.3--29.1 17 15.6--22.1
Discussion {#sec017}
==========
This study is the first to present grip strength results of a large sample of people with ID from 20--90 years of age. Although this study provides no information on the validity of measuring grip strength for adverse health outcomes in people with ID, it does provide reference values for people with ID for use in clinical practice. In line with the suggestion provided by Bohannon for the general population, scoring below the 95% confidence interval of the appropriate gender and age category reflects a below average result for that individual.
In the comparison with the data for the general population, this study demonstrates that people with intellectual disabilites have very low levels of grip strength during their entire life. Even at an age of 20--30 years, their grip strength is as low as for 75 year-olds of the general population, which is the age of a nursing home population. There is only a slight decline across the ages, not nearly as much as in the general population. This low level of grip strength probably represents the bare minimum of grip strength necessary to perform basic daily activities. It raises the question whether sarcopenia is already happening at 25, or whether people with ID have never built up any muscle mass to begin with due to physical inactivity.As mentioned in the introduction, grip strength is a strong predictor for a number of negative future health outcomes in the general population \[[@pone.0129585.ref010],[@pone.0129585.ref013]\]. Research on this topic in people with ID is scarce, and somewhat contradictory, which might be explained by the consistently low levels across the entire life span. Oppewal et al. did not find grip strength to be predictive of a decline in basic activities of daily living \[[@pone.0129585.ref032]\] or falls \[[@pone.0129585.ref033]\], but it did prove to be predictive of instrumental activities of daily living \[[@pone.0129585.ref032]\].
One of the strengths of this study is the large dataset used, resulting from combining two highly comparable samples with regards to study procedures, physical activity level, and data collection.
Limitations of this study are the lack of information on the presence of Down syndrome, which has been demonstrated to negatively influence grip strength \[[@pone.0129585.ref034]\]. Since the prevalence of Down syndrome is only around 15% of the total population of people with ID, this influence is considered to be minor. The level of intellectual disability does seem to influence grip strength results \[[@pone.0129585.ref034]\], but the level of ID was not available for full sample in the Fun Fitness ID sample. The second limitation is the comparison with published reference data of the general population. Not being able to work with raw data hampers statistical comparisons. In line with this, the lack of information about physical activity levels of the general population is also a problem, since differences in physical activity levels between the two samples could have confounded the difference in grip strength. A third limitation is reporting the grip strength for the left and the right hand, not taking differences in handedness into account. In the general population, most people are right-handed, and often this is also the strongest hand. In people with intellectual disabilities, not only is a larger percentage left-handed or has no preference for either the right or left hand, but also this does not necessarily result in this hand being the strongest hand \[[@pone.0129585.ref035]\]. We recommend measuring grip strength in both hands for people with ID, and comparing the results with the reference values of both left and right hand.
Overall, these results provide more insight into the development of grip strength across age in people with ID, compared to the development of grip strength in the general population. The implications of these findings for policy and practice are significant, and underline the importance of constant focus on promoting physical activity and exercise across the entire population with ID.
[^1]: **Competing Interests:**The authors have declared that no competing interests exist.
[^2]: Conceived and designed the experiments: ACV TH. Performed the experiments: ACV TH. Analyzed the data: ACV. Contributed reagents/materials/analysis tools: ACV. Wrote the paper: ACV TH.
| {
"pile_set_name": "PubMed Central"
} |
INTRODUCTION
============
Ifosfamide is a commonly used alkylating chemotherapeutic agent of the oxaphosphorines family that is often combined with other drugs, such as cisplatin, for the treatment of sarcomas \[[@sfz183-B1], [@sfz183-B2]\], testicular tumours \[[@sfz183-B3]\] and some refractory lymphomas \[[@sfz183-B4]\], both in children and in adult patients.
Even though the first reported patients with ifosfamide nephrotoxicity were adults \[[@sfz183-B5]\], most recent studies concern paediatric populations, and comprehensive data on adult patients are lacking. According to these studies, ∼30% of the children treated by ifosfamide will consequently develop a chronic kidney disease (CKD) \[[@sfz183-B6]\]. Nevertheless, the reported prevalence of nephrotoxicity ranges from 15% to 60% \[[@sfz183-B7]\], according to the definition of kidney injury, the duration of follow-up and the therapeutic protocols that have been used in each series. The spectrum of nephrotoxic phenotypes depends on the type, the severity and the reversibility of the kidney injury, as well as the time to onset, after administration of the treatment. The most frequent renal symptom is proximal tubular dysfunction (PTD) \[[@sfz183-B8], [@sfz183-B9]\], but acute kidney injury (AKI) often occurs, possibly as a consequence of renal thioredoxin reductase activity inhibition \[[@sfz183-B10]\], and can progress to CKD. In some rare cases, ifosfamide may be responsible for hypokalaemic distal tubular acidosis or nephrogenic diabetes insipidus \[[@sfz183-B11]\]. The aim of this study was to describe the clinical and histological features of the ifosfamide-associated nephrotoxicity in an adult population.
MATERIALS AND METHODS
=====================
A multicentric observational retrospective study was conducted in collaboration with six French nephrology departments. Adult patients who received ifosfamide were included in this study if they had renal failure (AKI and/or CKD) and/or tubular dysfunctions defined by at least two concurrent anomalies among: a hypokalaemic proximal tubular acidosis, a uric acid fractional excretion (UAFE) \>10%, a ratio of tubular maximum reabsorption of phosphate (TmPO4^2-^) to estimated glomerular filtration rate (eGFR), TmPO4^2-^/eGFR \<0.80, a low molecular weight proteinuria, an aminoaciduria and/or a normoglycaemic glycosuria, all occurring after the prescription of an ifosfamide-based chemotherapy regimen.
AKI was defined according to the 2012 Kidney Disease Improving Global Outcomes (KDIGO) guidelines (peak serum creatinine ×1.5 baseline value or delta creatinine \>26.5 μmol/L), and CKD was defined by a chronic renal dysfunction, confirmed by an eGFR \<60 mL/min/1.73 m^2^, by using the Modification of Diet in Renal Disease (MDRD) formula. Proteinuria was defined by a urine protein-to-creatinine ratio \>300 mg/g.
UAFE was calculated using the formula: (urinary uric acid×serum creatinine)/(plasmatic uric acid×urinary creatinine) and TmPO4^2−^/eGFR was calculated using the Bijvoet nomogram, taking into account phosphataemia and phosphate reabsorption rate (PRR): 1−\[(plasmatic creatinine×urine phosphate)/(urine creatinine\*plasmatic phosphate)\].
Clinical, biological and histological data were collected on Day 0, corresponding to the first day of the ifosfamide administration, and after 1, 6 and 12 months following ifosfamide initiation, as well as at the last follow-up.
JMP^®^ version 13.2.1 software was used to calculate means and standard deviation or standard error of mean (SEM), percentages, medians and interquartile range (IQR) and to apply statistical analyses and association analyses (Chi-square).
RESULTS
=======
A total of 34 patients have been explored between 1995 and 2016 in six French nephrology departments. The demographic characteristics of the patients are presented in [Table 1](#sfz183-T1){ref-type="table"}. The male to female ratio was 21:13 and the median age at treatment initiation was 46 years (range 18--75). Sarcoma was the most common indication for ifosfamide treatment (91.2%). The median cumulative dose of ifosfamide was 32.5 g/m^2^ (range 6--102). Fifteen patients (44.1%) received cisplatin treatment in addition to ifosfamide chemotherapy. There was no other nephrotoxic drug found. Acrolein urotoxicity was prevented by hydration and administration of 2-mercaptoethanesulfonic acid (Uromitexan) for all patients.
######
Demographic and oncologic characteristics
Patient characteristics Overall Overall Ifosfamide Overall Ifosfamide + cisplatin P-value
---------------------------------------------------- --------------- --------- ---------------------- --------- ------------------------ ---------
General characteristics
Age at diagnosis \[mean (range)\], years 44.9 (18--75) 19 39.5 (17.75-- 62.25) 15 47 (28--75) NS
Male, *n* (%) 21 (61.7) 19 13 (68.4) 15 8 (53.3) NS
Comorbidities
Hypertension, *n* (%) 8 (23.5) 18 6 (33.3) 15 2 (13.3) NS
Diabetes mellitus, *n* (%) 1 (2.9) 18 1 (5.5) 15 0 (0) NS
Cancer history
Type of cancer
Sarcoma, *n* (%) 31 (91.2) 18 16 (88.8) 15 13 (86.7) NS
Carcinoma, *n* (%) 2 (5.9) 18 1 (5.5) 15 2 (13.3) NS
Germinal tumour, *n* (%) 1 (2.9) 18 1 (5.5) 15 0 (0) NS
Treatment
Cisplatin, *n* (%) 15 (44.1) 19 0 (0) 15 15 (100) --
Ifosfamide cumulative dose (mean ± SEM), mg/m^2^ -- 6 59 800 ± 17 700 10 30 780 ± 8395 --
Use of uromitexan, *n* (%) -- 10 10 (52.6) 12 12 (80) --
NS, not significant.
Renal diseases associated with ifosfamide
-----------------------------------------
Ifosfamide nephrotoxicity was characterized by an isolated PTD in 6 patients (17.7%), among which 4 (11.8%) received ifosfamide alone and 2 (5.9%) ifosfamide and cisplatin; an isolated AKI in 5 patients (14.7%), among which 1 (2.9%) received ifosfamide and 4 (11.8%) ifosfamide in association with cisplatin; or a PTD complicated by AKI in 17 patients (50%), among which 10 (29.4%) received ifosfamide and 7 (20.6%) ifosfamide and cisplatin. In six cases (17.7.%), among which four (11.8%) were only treated by ifosfamide and two (5.9%) by ifosfamide and cisplatin ([Table 2](#sfz183-T2){ref-type="table"}), ifosfamide toxicity was suspected by the diagnosis of delayed-onset CKD, identified after a mean duration of 16.4 months following chemotherapy initiation. These results indicate that the spectrum of nephrotoxicity patterns associated with ifosfamide is wide, with a high prevalence of PTD and CKD, whereas AKI associated with ifosfamide occurs less frequently.
######
Characteristics of nephrotoxicity after ifosfamide therapy
Renal presentation Overall, *n* = 34 Ifosfamide, *n* = 19 Ifosfamide + cisplatin, *n* = 15
-------------------- ------------------- ---------------------- ----------------------------------
Pure PTD, *n* (%) 6 (17.7) 4 (11.8) 2 (5.9)
Pure AKI, *n* (%) 5 (14.7) 1 (2.9) 4 (11.8)
PTD + AKI, *n* (%) 17 (50.0) 10 (29.4) 7 (20.6)
CKD, *n* (%) 6 (17.7) 4 (11.8) 2 (5.9)
Among patients with a PTD ([Table 3](#sfz183-T3){ref-type="table"}), the main anomalies were, in order of frequency, hypokalaemia (20/23 patients), metabolic acidosis (12/23), hypophosphataemia (14/23), low molecular weight proteinuria (10/23), TmPO4^2−^/eGFR \<0.80 (9/23), hypouricaemia (9/23) and UAFE \>10% (5/23). These results indicate that, among individuals with PTD associated with ifosfamide, the addition of cisplatin seems to increase the metabolic alterations reflecting the PTD.
######
Spectrum of PTD after ifosfamide therapy
Proximal dysfunction features Overall, *n* = 23 Overall, *n* Ifosfamide Overall, *n* Ifosfamide + cisplatin P-value
--------------------------------- ------------------- -------------- ------------ -------------- ------------------------ ---------
Hypokalaemia, *n* (%) 20 (87.0) 14 11 (78.6) 9 9 (100) NS
Metabolic acidosis, *n* (%) 12 (52.2) 12 6 (50) 8 6 (75) NS
Hypophosphataemia, *n* (%) 14 (61.0) 14 8 (57.1) 8 6 (75) NS
TmPO4²^−^/eGFR \< 0.80, *n* (%) 9 (39.1) 6 6 (100) 3 3 (100) NS
Hypouricaemia, *n* (%) 9 (39.1) 11 5 (45.4) 5 4 (80) NS
UAFE \> 10%, *n* (%) 5 (21.7) 3 3 (100) 2 2 (100) NS
Low weight proteinuria, *n* (%) 10 (43.5) 10 7 (70) 6 3 (50) NS
NS, not significant.
Pathological findings
---------------------
A kidney biopsy was performed in 14 cases (41.2%) ([Table 4](#sfz183-T4){ref-type="table"}). Although seven of these patients (53.8%) did not receive cisplatin, we found no difference in type and extent of histology findings between the two groups of patients. Under light microscopy ([Figure 1](#sfz183-F1){ref-type="fig"}), acute tubular necrosis with denudation of proximal tubular epithelium was present in 12 biopsies (85.7%), vacuolation of epithelial cells in 11 biopsies (78.6%), nuclear atypia in 10 biopsies (71.4%), interstitial inflammation in 10 biopsies (71.4%) and interstitial fibrosis (IF) (\>10%) in 8 biopsies (57.1%). Focal arteriolar hyalinosis was found in four biopsies (28.6%). Electron microscopy analysis was available in three cases and showed proximal tubular injury with thinning of proximal tubular epithelium, loss of brush bordure and vacuolation of epithelial cells. Moreover, severe mitochondrial abnormalities with irregular mitochondria and disappearance of some mitochondrial ridges were identified ([Figure 2](#sfz183-F2){ref-type="fig"}).
{#sfz183-F1}
{#sfz183-F2}
######
Light microscopy findings on renal biopsy
Light microscopy features
---------------------------------------------------- -----------
Denudation of proximal tubular epithelium, *n* (%) 12 (85.7)
Vacuolation of epithelial cells, *n* (%) 11 (78.6)
Nuclear atypia, *n* (%) 10 (71.4)
Vascular hyalinosis, *n* (%) 4 (28.6)
Tubulo-interstitial inflammation, *n* (%) 10 (71.4)
TA/IF, *n* (%) 8 (57.1)
TA/IF, Tubular Atrophy/Interstitial Fibrosis.
Evolution
---------
In all cases, ifosfamide therapy was discontinued after the diagnosis of the renal disease. The median follow-up duration between drug cessation and last follow-up was 31 months (range 2--245). Mean progression of renal dysfunction during follow-up is described in [Figure 3A](#sfz183-F3){ref-type="fig"}. Baseline median eGFR before initiation of chemotherapy was 67 mL/min/1.73 m^2^ (51--89 mL/min/1.73 m^2^). Median eGFR was 31 mL/min/1.7 m^2^ (7--89.5 mL/min/1.73 m^2^) 1 month after ifosfamide treatment initation, 38 mL/min/1.73 m^2^ (21--65.5 mL/min/1.73 m^2^) at 6 months and 35 mL/min/1.73 m^2^ (11--57 mL/min/1.73 m^2^) at the last follow-up. In 16 patients, renal dysfunction continued to deteriorate \[delta eGFR −26.5 mL/min/1.73 m^2^ range (−4 to −109 mL/min/1.73 m^2^)\] between the cessation of ifosfamide and the last follow-up ([Figure 3B](#sfz183-F3){ref-type="fig"}).
{#sfz183-F3}
Five patients (14.7%) were treated with corticosteroids for inflammatory interstitial nephritis, with no significant improvement of renal function. At the end of follow-up, 15 patients (44.1%) had progressed to Stage 4 CKD and 10 patients (29.4%) to Stage 5 CKD. Overall, six patients required transient (*n* = 1) or definitive (*n* = 5) renal replacement therapy, after a median delay of 4.75 months (0.2--111 months) following the diagnosis of kidney disease. Six patients (17.6%) died from progression of the neoplastic disease during the follow-up period.
On univariate analysis, renal function (eGFR) at last follow-up was associated with age (P = 0.01) and cisplatin co-administration (P = 0.04). Interestingly, renal outcomes were not associated with the cumulative dose of ifosfamide (P = 0.28), the presence of proximal tubulopathy (P = 0.41) or eGFR at baseline (P = 0.38).
DISCUSSION
==========
Our results indicate that ifosfamide promotes tubulo-interstitial injury that can be revealed mostly by an AKI or sometimes a progressive CKD diagnosed several months or years after ifosfamide administration. Ifosfamide also frequently promotes PTD along with a complete or incomplete Fanconi syndrome. Mitochondria could be frequently injured by ifosfamide, and mitochondrial dysfunction may play an important role in proximal tubular injury. However, our retrospective study has some limitations, such as the absence of comprehensive biological tubular data or the lack of follow-up of proximal tubulopathy outcomes. Moreover, the absence of a control group did not allow identification of risk factors of nephrotoxicity.
In our study, the association of ifosfamide and cisplatin appears more toxic than treatment with ifosfamide alone; 44% of patients in the cohort received ifosfamide in combination with cisplatin, which is usually associated with a tubulo-interstitial AKI but not with PTD. Cisplatin frequently affects magnesium reabsorption \[[@sfz183-B12]\] and impairs ability to concentrate urine \[[@sfz183-B13]\], suggesting a tubular injury affecting preferentially the Henle loop and the distal tubule, in contrast to ifosfamide. However, \>50% of our patients did not receive cisplatin and had a PTD, suggesting ifosfamide may promote a specific pattern of nephrotoxicity. These results are consistent with those observed among children treated with ifosfamide \[[@sfz183-B14]\], in whom PTD is a common complication that can sometimes persist for several years after the initial treatment. Moreover, it has been shown that adult survivors of childhood cancer have a residual tubular dysfunction, characterized by an increase of urine β2 microglobulin/creatinine ratio, when ifosfamide cumulative doses are \>16 g/m^2^ \[[@sfz183-B15]\].
Unlike other publications, a nephrotoxicity threshold was not found in our study. It is possible that the threshold of cumulative dose exposing patients to ifosfamide nephrotoxicity is \<60 g/m^2^. Moreover, this threshold may depend upon the age of the patients or their genetic background, such as polymorphisms in genes coding for tubular transporter.
Ifosfamide penetrates the proximal tubular cells through the influx carrier human organic cation transporter 2 \[[@sfz183-B16]\]. After internalization, ifosfamide is processed by two different metabolic pathways \[[@sfz183-B17]\]: a 4-hydroxylation by P450 CYP3A4 and 2B6 \[[@sfz183-B18]\] into an active nitrogen mustard that alkylates and damages DNA, and the urotoxic by-product acrolein \[[@sfz183-B19]\], and a dechloroethylation step that produces chloroacetaldehyde (CAA), which is nephrotoxic \[[@sfz183-B20], [@sfz183-B21]\]. The mechanisms of CAA-induced nephrotoxicity are partially understood. CAA alters cellular energy metabolism by inhibiting oxidative phosphorylation at the level of mitochondrial respiratory chain complex I \[[@sfz183-B8], [@sfz183-B22]\], and increases cellular vulnerability to oxidative stress \[[@sfz183-B21], [@sfz183-B23], [@sfz183-B24]\], and has a profibrotic effect \[[@sfz183-B25]\]. Notably, coadministration of 2-mercaptoethanesulfonic acid prevents haemorrhagic cystitis and allows an increase of the administrated ifosfamide dose, but may facilitate ifosfamide nephrotoxicity. Histological descriptions of ifosfamide nephrotoxicity in adults are uncommon, and light microscopy analysis usually shows non-specific proximal tubular atrophy (TA), proximal tubular basement membrane denudation, proximal tubular cell vacuolation, nuclear atypia and a moderate monomorphic interstitial infiltrate \[[@sfz183-B26], [@sfz183-B27]\], which is consistent with our findings. Our ultrastructural analysis performed in only three patients showed mitochondrial changes may suggest specific toxicity, including irregular mitochondria, mitochondrial enlargement and disappearance of some mitochondrial ridges. This observation is compatible with the CAA nephrotoxicity mechanism. These findings are reminiscent of the nephrotoxicity of tenofovir, which targets proximal tubules and can induce mitochondrial alteration \[[@sfz183-B28]\]. However, tenofovir nephrotoxicity is usually irreversible, which is not always the case for ifosfamide. Ifosfamide also induces nuclear atypia, which can result from the ifosfamide mustard alkylating effect. These atypia are similar to those produced upon viral infection, or during pseudo-caryomegalic nephropathy, which is a rare familial chronic tubule-interstitial nephritis \[[@sfz183-B29]\]. Recently, Zhou *et al.* \[[@sfz183-B30]\] have described the association of caryomegalic nephropathy with mutation in the *Fanconi anaemia-associated nuclease 1* gene, which encodes a DNA repair enzyme.
In our series, 10 patients developed Stage 5 CKD, 6 required haemodialysis and 6 died. Only five still survived without CKD at the end of a mean follow-up of 41.8 months \[(min--max) = (2--71)\]. In line with this, in one study on adults treated by ifosfamide \[[@sfz183-B31]\], the mean eGFR decreased from 81.5 to 57.9 mL/min/1.73 m^2^, with 53% of patients having eGFR \<60 mL/min/1.73 m^2^ 5 years after chemotherapy, although none of them progressed to Stage 5 CKD. In our study, all patients treated with corticosteroids developed CKD. This treatment, sometimes used in cases with significant interstitial inflammation, seems not to have any effect on ifosfamide nephrotoxicity and especially on tubular injury. There is no current strategy for preventing ifosfamide nephrotoxicity. *N*-acetylcysteine (NAC) prophylaxis has showed some beneficial effects, both *in vitro* and *in vivo*, in animal models of ifosfamide nephrotoxicity \[[@sfz183-B32]\], but not in humans. Some reported cases of AKI occurring in children have suggested some benefits obtained by the use of intravenous NAC \[[@sfz183-B33]\].
In conclusion, we have described the pattern of tubulo-interstitial injuries induced by ifosfamide in a cohort of 34 adult patients. Ifosfamide nephrotoxicity can occur in most cases by AKI immediately following administration, sometimes by CKD diagnosed several months or years after the chemotherapy, frequently associated with proximal tubulopathy. Severe CKD is frequent, especially in patients receiving ifosfamide--cisplatin combination. The histological and ultrastructural findings observed in our study suggest that the proximal tubular injury can be mediated by mitochondrial damage. Future studies are necessary to determine adult risk factors prior to the treatment and prophylactic therapies to decrease ifosfamide nephrotoxicity. Careful monitoring of tubular functions is important to ensure timely drug withdrawal or initiation of nephroprotective measures.
CONFLICT OF INTEREST STATEMENT
==============================
None declared.
| {
"pile_set_name": "PubMed Central"
} |
Background
==========
Hypocalcemia is a disorder of mineral metabolism that occurs when PTH is insufficient to act on end organs (kidney, bone, intestine) -hypoparathyroidism, or end organs irresponsiveness in setting of sufficient PTH to normalize serum calcium -- PTH resistance. In the setting of adequately functioning parathyroid glands, and responsive end organs, other causes of hypocalcemia, such as vitamin D deficiency, are characterized by high PTH (secondary hyperparathyroidism). Thus, it is may be useful to broadly characterize hypocalcemia as associated with low PTH or high PTH \[[@b1-amjcaserep-14-113]\]. One of the causes of high PTH associated hypocalcemia is PTH resistance which is often associated with hypocalcemia, hyperphosphatemia.
Case Report
===========
A 40 year old gentleman with no prior medical illness presented to the emergency department with 1 day history of generalized malaise and subjective fever. He denies any history of prior similar symptoms. His past medical and surgical history were unremarkable. His social history is significant for being a carpenter, marriage and two children. His family history was unremarkable for chronic illnesses or diseases. The review of systems was otherwise negative. On examination, the patient looks well and not in distress. BP: 150/82, Temperature: 37.4°C, Pulse: 100 bpm, Respiratory rate: 18/min, O2 sat: 97% on room air, weight: 62 kg, height: 168 cm, BMI: 22, General physical examination was unremarkable. Sample of laboratory investigations were as follows: Corrected Calcium 1.37 mmol/L (2.1--2.6), Ionized Calcium 1.05 (1.18--1.32 mmol/L), Phosphorous 1.72 mmol/L (0.87--1.45), Magnesium 0.72 mmol/L (0.65--1.05), BUN 3.4 mmol/L (1.7--8.3), Creatinine 86 umol/L (62--124), Bicarbonate 26 mmol/L (24--30), Albumin 44 g/L (35--50), Alkaline Phosphatase 114 U/L (40--129), Vitamin D 30 ng/ml (30--80), Parathyroid hormone (immunoreactive) 114 pg/ml (15--65), Urine Calcium 0.2, Urinary Calcium Creatinine ratio 0.02 (\>0.4). Other biochemical analysis including complete blood count, thyroid stimulating hormone, free thyroxine and cortisol levels were all normal. EKG showed normal sinus rhythm with normal Q-T interval. CXR was within normal.
The patient was managed with both parenteral and oral Calcium supplements. The corrected serum calcium reached a plateau of 2.07 mmol/L and phosphorus reached a nadir of 1.42 mmol/L. iPTH was to be repeated following institution of therapy at follow up in out-patient clinic. At the time of this write up, he was discharged to home on oral calcium carbonate tablets with an appointment for follow up in the clinic.
Discussion
==========
Approximately, 500 mg of calcium is removed from the bones daily and replaced by an equal amount. Normally, the amount of calcium absorbed by the intestines is matched by urinary calcium excretion in order to maintain homeostasis. Despite these enormous fluxes of calcium, the levels of ionized calcium remain stable because of the tight control maintained by parathyroid hormone (PTH), vitamin D, and calcitonin through complex feedback loops. These hormones act primarily at the level of bone, kidney, and gastrointestinal tract. Calcium levels are also affected by magnesium and phosphorus \[[@b2-amjcaserep-14-113]\]. The parathyroid gland has a remarkable sensitivity to ionized serum calcium changes. These changes are recognized by the calcium-sensing receptor (CaSR) and a slight decrease in serum ionized calcium stimulates the chief cells of the parathyroid gland to secrete PTH \[[@b1-amjcaserep-14-113]\].
The most common causes of hypocalcemia with accompanied high PTH are hypoalbuminemia, hypomagnesemia, hyperphosphatemia, and PTH resistance. Calcium is transported in blood partly bound to protein (roughly 45--55%), mostly serum albumin. The total serum calcium may be reduced as a result of hypoalbuminemia, but concentration of hormonally active, free ionized calcium remains unchanged \[[@b3-amjcaserep-14-113]\]. Hypomagnesemia may also cause hypocalcemia due to PTH resistance, hence the high PTH. Hyperphosphatemia associated hpocalcemia is primarily due to bone deposition and extraskeletal tissue calcification. Other etiologies include calcium chelation (with citrate, foscanet, lactate for instance) which reduces ionized calcium, but not total serum calcium \[[@b4-amjcaserep-14-113]\], medication effects such as reduced oteoclastic bone resorption in the setting of bisphosphonates and acute inhibition of PTH release with calcimimetic agent, notably cinacalcet \[[@b5-amjcaserep-14-113],[@b6-amjcaserep-14-113]\].
On the other hand, hypocalcemia with low PTH is often encountered following parathyroid gland destruction post surgery and in the rare cases of acquired and/or familial autoimmune disorders (such as in polyglandular autoimmune disorder type 1.) Infiltrative diseases such as hemochromatosis, granulomas can also present with hypocalcemia and low PTH \[[@b7-amjcaserep-14-113]\]. Isolated or familial cases of activating mutation of CaSR may also present in similar manner where PTH is not released at serum calcium concentration that normally stimulate PTH production from the glands.
Our patient had a normal level of serum albumin, magnesium, vitamin D, kidney function, and normal renal calcium excretion (ruling out the possibility of renal tubular loss). Along with the constellation of hypocalcemia, hyperphosphotemia, and high parathyroid hormone levels, he was diagnosed with parathyroid hormone resistance -- pseudohypoparathyroidism.
Pseudohypoparathyroidism is characterized by end-organ resistance to the effects of PTH. PTH binds to the PTH receptor, which, activates cAMP through guanine nucleotide regulatory G-protein for an effective signal transduction. PTH resistance is encountered as a loss-of-function mutation in the gene that encode the alpha subunit of the G-protein that is coupled to the PTH receptor to activate the adenyl cyclase required for signal transduction that produces end-organ response to PTH. There is a failure of signal transduction, hence the PTH resistance \[[@b8-amjcaserep-14-113]\]. Pseudohypoparathyroidism is classified into types I and II. Type I is further subdivided into Ia, Ib, and Ic. *Type Ia* comprises the biochemical features of pseudohypoparathyroidism, blunted urinary cAMP response to exogenous PTH administration along with the somatic features of Albright Hereditary Osteodystrophy (AHO), such as short stature, obesity, developmental delay, subcutaneous calcification, short fourth metacarpal bone, and round face. The mode of inheritance is autosomal dominant pattern with gene imprinting in which only female can transmit the full disease. *Type Ia* is unlikely in our patient who presented in adulthood without AHO's characteristic features \[[@b9-amjcaserep-14-113]\].
*Type Ib* pseudohypoparathyroidism harbors autosomal dominant inherent pattern with familial distribution but without somatic features of AHO. In most patients with PHP-Ib, global G-proteins mutations have not been found, hence the absence of AHO features. The hormonal resistance is usually limited to the PTH-dependent actions in proximal renal tubules in which G-protein is paternally imprinting and consequently the impaired urinary cAMP response to PTH and other biochemistries abnormalities \[[@b8-amjcaserep-14-113]\].
Patient with *type Ic* pseudohypoparathyroidism presents with resistance to multiple hormonal receptors. The patients are phenotypically similar to *type Ia* but with normal functioning G-protein.
In *type II* pseudohypoparathyroidism, PTH raises cAMP normally but fails to increase levels of serum calcium or urinary phosphate excretion, suggesting that the defect is located downstream of the generation of cAMP \[[@b9-amjcaserep-14-113]\]. There is usually no familial pattern (and sporadic if present) as seen with patients in all *Type I* classification. These patients present with hypocalcemia, hypophosphaturia, and elevated immunoreactive PTH (iPTH) levels which are also seen in vitamin D deficiency. These abnormal lab values often return to normal upon adequate vitamin D supplementation.
Conclusions
===========
Our patient was classified as *Type II*. He presented noncontributory familial history and absence of phonotypical features of AHO. Type II is considered the rarest of the classification likely due to its late presentation in adulthood, lack or inconsistent familial distribution, and absence of AHO features. These patients are usually diagnosed incidentally as in our patient and often have benign course.
[^1]: Authors' Contribution:
[^2]: Study Design
[^3]: Data Collection
[^4]: Statistical Analysis
[^5]: Data Interpretation
[^6]: Manuscript Preparation
[^7]: Literature Search
[^8]: Funds Collection
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#cesec10}
============
Incomplete gallbladder surgery, involving leaving a long cystic duct or gallbladder remnant, for example, can occur in both open and laparoscopic procedures ([@bib1]). Incomplete surgery can lead to postcholecystectomy syndrome, which is a recurrence of symptoms similar to those that led to the initial cholecystectomy. The syndrome has been reported in 10--40% of postoperative patients and is seen more commonly in women ([@bib1]). In this case, the patient suffered recurrent cholecystitis with cholelithiasis secondary to a retained gallbladder remnant, 14 years after a cholecystectomy.
Case report {#cesec20}
===========
A 52-year-old male presented to the emergency department with a one-day history of epigastric pain. The patient denied any associated nausea, vomiting, diarrhea, constipation, chest pain, or shortness of breath. He reported having a "difficult" cholecystectomy that was converted to an open procedure in 1994. Further details of the operation were not available. Additional surgical history included a prostatectomy for prostate cancer. Medical history included hypertension, cardiomyopathy, and hyperlipidemia.
Figure 152-year-old man with recurrent cholecystitis and cholelithiasis. Transverse sonographic image at the level of the gallbladder fossa reveals a fluid-filled, saclike structure containing shadowing calculi, consistent with a remnant gallbladder.
Figure 2A-D52-year-old man with recurrent cholecystitis and cholelithiasis. Contrast-enhanced axial CT slices through the level of the gallbladder fossa demonstrate a saclike structure in continuity with the biliary system containing multiple calculi (arrows).
Physical exam revealed a mildly obese, soft abdomen with minimal tenderness, particularly in the right upper quadrant. The remainder of the physical exam was unremarkable. Laboratory results were significant for an elevated bilirubin of 3.9, ALT of 1057, AST of 395, and LDH of 332.
Abdominal ultrasound obtained on the day of admission revealed a fluid-filled saclike structure in the region of the gallbladder fossa resembling a gallbladder and containing shadowing calculi. Review of a recent radiograph before admission revealed right-upper-quadrant surgical clips, confirming the history of prior cholecystectomy. Additionally, there was mild biliary dilatation and an echogenic liver. The possibility of retained gallstones within a dilated cystic duct remnant was raised. Abdominal CT the same day demonstrated a saclike structure in the gallbladder fossa, containing two calculi and appearing to have a connection with the biliary tree. Mild common bile-duct dilatation was again noted. The following day, an MRCP demonstrated at least two small calculi in the distal common bile duct, with associated mild intrahepatic and extrahepatic biliary dilatation. There was a 2.8-cm saccular, fluid-filled structure contiguous with the cystic duct containing two calculi. The patient was diagnosed with cholelithiasis in a gallbladder remnant and choledocholithiasis as the cause of his postcholecystectomy syndrome.
Figure 352-year-old man with recurrent cholecystitis and cholelithiasis. MRCP maximum-intensity projection image displays a fluid-filled sac in continuity with the cystic duct. Small filling defects are seen in the distal common bile duct, consistent with calculi (arrows).
The patient subsequently underwent an ERCP, sphincterotomy, and balloon sweep to clear the common bile duct stones.
A few days later, an open completion cholecystectomy and intraoperative cholangiogram was performed. Intraoperatively, a very small gallbladder remnant was noted. Palpation of the remnant gallbladder revealed two gallstones. The gallbladder was dissected to the level of the hilum, and the cystic duct was noted to be somewhat fused to the common bile duct. The cystic duct was dissected free to its base. An intraoperative cholangiogram revealed excellent flow and no filling defects. The cystic duct was then ligated, and the proximal cystic duct and gallbladder remnant were sent as specimens. A liver biopsy was also obtained.
Surgical pathology revealed a gallbladder remnant with chronic cholecystitis and cholelithiasis as well as a portion of the cystic duct. Liver biopsy revealed fatty change with no evidence of hepatitis, fibrosis, or malignancy.
Discussion {#cesec30}
==========
When gallbladder surgery is incomplete, long cystic duct remnants are more frequently seen in the laparoscopic approach, where the cystic duct is usually divided closer to the gallbladder to avoid iatrogenic common bile duct damage ([@bib2], [@bib3]). A common cause for leaving a gallbladder remnant is failure to properly identify the gallbladder-cystic junction, which can occur with incomplete mobilization of the cystic duct ([@bib3], [@bib4]). This complication is thought to be more common in the setting of acute cholecystitis ([@bib4]). Incomplete surgery can lead to postcholecystectomy syndrome, described as a recurrence of symptoms similar to those that led to the initial cholecystectomy ([@bib1], [@bib2], [@bib4], [@bib5]). This usually presents as right-upper-quadrant pain and dyspepsia without jaundice ([@bib2]).
Although it is extremely rare that incomplete surgery causes significant symptoms, one should always suspect a gallbladder remnant or long cystic duct in a patient with postcholecystectomy right-upper-quadrant symptoms ([@bib6]). In one study of 500 cholecystectomized patients, 82% of patients with severe postoperative biliary distress had either a long cystic duct (\>1cm) or a gallbladder remnant, where only 40% of asymptomatic or mildly symptomatic cholecystectomized patients had a long duct or remnant ([@bib7]).
Other causes of postcholecystectomy syndrome include neuromas in the surgical bed, scar tissue around the cystic duct stump, biliary strictures, Sphincter of Oddi dyskinesia, and psychological conditions such as affective disorders ([@bib1], [@bib2], [@bib3], [@bib4], [@bib5], [@bib8]). Symptom onset can arise two days to 25 years after surgery ([@bib2]).
Due to the numerous causes of postcholecystectomy syndrome, the etiology can be difficult to determine ([@bib2]). Radiographic studies, such as ultrasound, CT, MRCP, and ERCP are helpful in evaluating the cause of postcholecystectomy syndrome. Abdominal ultrasound, often the initial test for evaluating right-upper-quadrant pain, can miss nearly half of biliary tract abnormalities compared to ERCP and MRCP. Both ERCP and MRCP have a sensitivity that has been reported anywhere between 85%--100% for choledocolithiasis; ERCP is more specific and should be used when intervention is clearly indicated ([@bib2]). In a patient presenting with postcholecystectomy symptoms, with a gallbladder-like structure seen on radiographic studies, retained gallbladder remnant and/or retained cystic duct remnant should be the primary consideration.
Figure 4A-D52-year-old man with recurrent cholecystitis and cholelithiasis. T2 coronal images demonstrate a saclike structure contiguous with the cystic duct containing two large calculi (arrow). Multiple smaller calculi appear in the distal common bile duct (arrowhead). Mild extrahepatic and minimal intrahepatic biliary dilatation appears.
Figure 552-year-old man with recurrent cholecystitis and cholelithiasis. Intraoperative cholangiogram shows no residual filling defects in the common bile duct.
Published: March 30, 2010
[^1]: Dr. Calhoun is in Interventional Radiology and is Assistant Program Director of the Radiology Residency Program, and Dr. Piechowiak is a PGY-3 radiology resident, both at Morristown Memorial Hospital, Morristown NJ.
| {
"pile_set_name": "PubMed Central"
} |
Excessive plasma triglyceride and cholesterol levels contribute to the development of several prevalent cardiovascular risk factors, such as hypertriglyceridemia, hypercholesterolemia, obesity and diabetes. Plasma triglyceride concentrations are maintained within a narrow range and exhibit circadian rhythmicity in humans and rodents[@b1][@b2]. Lipoprotein production is highly regulated to maintain plasma lipids. Lipoprotein production is dependent on a structural protein, apolipoprotein B (apoB), and a chaperone, microsomal triglyceride transfer protein (MTP)[@b3]. We have shown that plasma lipids and MTP expression exhibit in sync circadian changes and have suggested that changes in MTP expression contribute to daily variations in plasma lipids[@b4]. Cholesterol transported via lipoproteins is either delivered to peripheral tissues or to the liver. In the liver, cholesterol is secreted into circulation as lipoproteins or into the bile. Secretion to blood is dependent on MTP, whereas a heterodimeric complex of ATP binding cassette family G protein 5 and protein 8 (Abcg5 and Abcg8) transporters assist in the secretion of cholesterol to the bile[@b5][@b6]. Expression of Abcg5/Abcg8 is regulated by Lxr, Hnf4α and Gata4 (ref. [@b6]). It is unknown whether Abcg5/Abcg8 expression changes within a day.
Daily variations in various biological, behavioral and physiological processes are controlled by several transcription factors, known as 'clock genes\', expressed in the suprachiasmatic nuclei of the brain[@b7][@b8][@b9][@b10]. These central clock genes form a hierarchical system to control most of the physiologic systems. Besides the suprachiasmatic nuclei, all peripheral tissues also express these clock genes[@b11]. This raises the question whether peripheral clock genes have autonomous function or they are subservient to the central regulatory system. Understanding the roles of peripheral clock genes in the diurnal regulation of peripheral tissues is an area of active research. The Clock and Bmal1 are two key transcription factors that increase the expression of other transcription factors to control rhythmicity of different biological functions[@b1][@b7][@b8][@b12][@b13][@b14]. We have examined the role of Clock by studying mice that express a dominant negative Clock mutant (Clock^Δ19/Δ19^) protein[@b15]. We showed that plasma triglyceride in *Clock*^*Δ19/Δ19*^ mice do not exhibit circadian rhythms, instead plasma triglyceride levels are high at all times[@b16]. Molecular studies showed that the Clock^Δ19/Δ19^ protein disrupts plasma triglyceride homoeostasis by de-regulating diurnal transcriptional regulation of Shp and Mtp[@b16]. Furthermore, we have shown that the presence of Clock^Δ19/Δ19^ protein in mice enhances atherosclerosis by increasing hepatic lipoprotein production and reducing cholesterol efflux from macrophages[@b17]. Clock^Δ19/Δ19^ affected the expression of MTP and ABCA1 by modulating the expression of Shp and Usf2 in hepatocytes and macrophages, respectively[@b17].
Besides the Clock protein, Bmal1 is another key transcription factor acts in concert with Clock to regulate circadian mechanisms. The role of Bmal1 in circadian regulations has been gleaned from studies in global and tissue-specific Bmal1-deficient mice. Global Bmal1 deficiency results in arrhythmias, hyperglycemia and hypoinsulinemia[@b18]. Arrhythmias are more likely due to its critical role in the central regulation of circadian rhythms[@b19]. Bmal1 affects plasma glucose and insulin levels by regulating the secretion of insulin from pancreatic cells[@b20]. Tissue-specific deletion studies have shown that endothelial[@b21], smooth muscle[@b22] and bronchiolar[@b23] Bmal1 affects endothelial function, blood pressure and pulmonary inflammation. Bmal1 deficiency in adipose tissues results in obesity[@b24]. Bmal1-deficient mice exhibit dyslipidemia[@b25], but it is unknown how Bmal1 regulates plasma lipids. Transplantation of Bmal1-deficient aortic grafts into wild-type mice results in robust lesion development[@b26], but it is unknown whether Bmal1 deficiency affects atherosclerosis. On the basis of these studies, we hypothesized that Bmal1 may play an important regulatory role in plasma lipid metabolism and atherosclerosis.
Here we show that Bmal1 deficiency increases hepatic lipoprotein production, cholesterol excretion to bile and atherosclerosis. Mechanistic studies show that Bmal1 regulates Shp and MTP to regulate hepatic lipoprotein production. Further, it regulates the expression of Abcg5/Abcg8 and biliary cholesterol excretion by modulating the expression of Gata4. Thus, Bmal1 is an anti-atherogenic transcription factor that controls hepatic lipoprotein production and biliary cholesterol excretion.
Results
=======
Global Bmal1 deficiency increases atherosclerosis in mice
---------------------------------------------------------
We investigated the effects of Bmal1 deficiency on atherosclerosis and observed that 3--12 months old, male *Bmal1*^*−/−*^*Apoe*^*−/−*^ mice had more aortic lesions compared with *Bmal1*^*+/+*^*Apoe*^*−/−*^ mice ([Fig. 1a](#f1){ref-type="fig"}). Further, lipid deposition in the abdominal aorta was higher in *Bmal1*^*−/−*^*Apoe*^*−/−*^ mice of all ages ([Fig. 1b](#f1){ref-type="fig"}). Brachiocephalic arteries (BCA) and cardiac/aortic junctions had more lipids in *Bmal1*^*−/−*^*Apoe*^*−/−*^ mice compared with *Bmal1*^*+/+*^*Apoe*^*−/−*^ mice ([Fig. 1c,d](#f1){ref-type="fig"}). Further, atherosclerotic plaques at the cardiac/aortic junctions were enriched in necrotic core, collagen and macrophage content ([Fig. 1e--h](#f1){ref-type="fig"}). The development of atherosclerosis was more pronounced when these mice were fed a Western diet ([Supplementary Fig. 1A](#S1){ref-type="supplementary-material"}). Further, higher aortic lesions were also seen in *Bmal1*^*−/−*^*Ldlr*^*−/−*^ mice fed a chow and western diets ([Supplementary Fig. 1B](#S1){ref-type="supplementary-material"}) compared with *Bmal1*^*+/+*^*Ldlr*^*−/−*^ mice. Thus, Bmal1 deficiency increases atherosclerosis in different mouse models.
Increased hyperlipidaemia in *Bmal1* ^−/−^ *Apoe* ^−/−^ mice
------------------------------------------------------------
To understand why Bmal1 deficiency augments atherosclerosis, we examined multiple metabolic parameters in *Bmal1*^*+/+*^*Apoe*^*−/−*^ and *Bmal1*^*−/−*^*Apoe*^*−/−*^ mice fed a chow diet. The Bmal1 deficiency significantly reduced hepatic expression of different clock genes with no effect on RORα expression in the liver ([Supplementary Fig. 2A](#S1){ref-type="supplementary-material"}). Livers from *Bmal1*^*−/−*^*Apoe*^*−/−*^ mice had higher triglyceride and cholesterol but normal phospholipid levels compared with *Bmal1*^*+/+*^*Apoe*^*−/−*^ mice ([Supplementary Fig. 2B](#S1){ref-type="supplementary-material"}). Plasma cholesterol ester, non-esterified fatty acids, phospholipids, glucose and insulin were higher in *Bmal1*^*−/−*^*Apoe*^*−/−*^ mice compared with *Bmal1*^*+/+*^*Apoe*^*−/−*^ mice ([Supplementary Fig. 2C](#S1){ref-type="supplementary-material"}), but plasma leptin, AST and ALT were not different in these two groups ([Supplementary Fig. 2D](#S1){ref-type="supplementary-material"}). *Bmal1*^*−/−*^*Apoe*^*−/−*^ mice had similar liver, intestine and kidney weights, but had significantly lower body weight and adipose tissue compared with controls ([Supplementary Fig. 2E](#S1){ref-type="supplementary-material"}). Total plasma triglyceride, cholesterol were higher in *Bmal1*^*−/−*^*Apoe*^*−/−*^ mice of different ages compared with *Bmal1*^*+/+*^*Apoe*^*−/−*^ mice mainly due to increases in non-HDL lipoproteins ([Supplementary Fig. 2F](#S1){ref-type="supplementary-material"}). Thus, Bmal1 deficiency causes hyperlipidaemia and hepatosteatosis. However, hepatosteatosis was not associated with increases in plasma transaminases.
*Bmal1*^*−/−*^*Apoe*^*−/−*^ mice had 1.5- and 4-fold more plasma apoB48 and apoB100 levels compared with *Bmal1*^*+/+*^*Apoe*^*−/−*^ mice, respectively; however, Bmal1 ablation had no effect on plasma apoAI levels resulting in higher apoB/apoAI ratio ([Fig. 2a](#f2){ref-type="fig"}). *Bmal1*^*−/−*^*Apoe*^*−/−*^ mice had higher triglyceride and cholesterol in VLDL/LDL with no change in HDL lipids ([Fig. 2b](#f2){ref-type="fig"}). Density gradient ultracentrifugation also showed that VLDL triglyceride, cholesterol, free cholesterol and cholesterol esters were increased in *Bmal1*^*−/−*^*Apoe*^*−/−*^ mice with no significant change in their HDL concentrations ([Supplementary Fig. 3](#S1){ref-type="supplementary-material"}). Negative staining and electron microscopy demonstrated that *Bmal1*^*−/−*^*Apoe*^*−/−*^ mice had larger VLDL particles with diameters ranging between 80 and 140 nm, while there were no significant differences in the size of LDL and HDL particles compared with *Bmal1*^*+/+*^*Apoe*^*−/−*^ mice ([Fig. 2c](#f2){ref-type="fig"}). These studies show that Bmal1 deficiency in *Apoe*^*−/−*^ mice increases plasma concentrations of larger lipid-rich apoB-containing lipoproteins.
Lipoprotein production in *Bmal1* ^*−/−*^ *Apoe* ^*−/−*^ mice
-------------------------------------------------------------
To explain reasons for higher VLDL plasma lipids, we studied hepatic lipoprotein production after inhibiting lipases by the injection of poloxamer 407 (P407). Plasma triglyceride levels increased faster and remained higher at all times in lipoprotein lipase inhibited *Bmal1*^*−/−*^*Apoe*^*−/−*^ mice indicating higher triglyceride production rates ([Fig. 2d](#f2){ref-type="fig"}). Due to the uncertainty about the effects of P407, we did not measure hepatic lipids in these mice. The effect of Bmal1 deficiency was confirmed further by studying lipoprotein production in isolated primary hepatocyte ([Supplementary Fig. 4](#S1){ref-type="supplementary-material"}). *Bmal1*^*−/−*^*Apoe*^*−/−*^ hepatocytes secreted significantly higher amounts of triglyceride and phospholipid and had decreased cellular triglyceride with no significant differences in cellular phospholipid and protein ([Supplementary Fig. 4](#S1){ref-type="supplementary-material"}); cholesterol mass was too low to measure. These studies indicated that *Bmal1*^*−/−*^*Apoe*^*−/−*^ hepatocytes secrete higher amounts of triglyceride-rich larger lipoproteins.
Similar studies in *Bmal1*^*−/−*^*Ldlr*^*−/−*^ mice revealed that these mice also have higher plasma triglyceride and cholesterol when fed chow or western diets compared with *Bmal1*^*+/+*^*Ldlr*^*−/−*^ mice ([Supplementary Fig. 5A](#S1){ref-type="supplementary-material"}). Further, hepatic triglyceride production was higher in P407 injected *Bmal1*^*−/−*^*Ldlr*^*−/−*^ mice compared with *Bmal1*^*+/+*^*Ldlr*^*−/−*^ mice ([Supplementary Fig. 5B](#S1){ref-type="supplementary-material"}). Thus, Bmal1 deficiency augments hepatic lipoprotein production and plasma lipids in both *Apoe*^*−/−*^and *Ldlr*^*−/−*^ mice.
Hepatic triglyceride-rich lipoprotein production is dependent on MTP. Livers of *Bmal1*^*−/−*^*Apoe*^*−/−*^ mice had higher MTP activity, protein and mRNA ([Fig. 2e](#f2){ref-type="fig"}). Expression analysis of transcription factors that regulate MTP[@b27] revealed that increases in MTP might be secondary to reduced expression of Shp, a repressor ([Fig. 2f,g](#f2){ref-type="fig"}). Thus, Bmal1 deficiency might augment MTP activity, enhance VLDL production and increase plasma lipids.
Reduced cholesterol excretion to bile in Bmal1-deficient mice
-------------------------------------------------------------
Besides VLDL production, hepatocytes efflux cholesterol by two additional mechanisms: one to HDL towards the basolateral side via Abca1 and Abcg1, and second towards the apical side to bile acids via Abcg5 and Abcg8. We found that hepatic mRNA and protein levels of Sr-b1, Abca1 and Abcg1 were unaffected; Npc1L1 were increased; and Abcg5 and Abcg8 were significantly reduced in Bmal1-deficient mice ([Fig. 2f,g](#f2){ref-type="fig"}). Since Abcg5/Abcg8 is involved in cholesterol secretion to bile, we determined whether Bmal1 deficiency affects cholesterol efflux to bile. For this purpose, we intravenously injected \[^3^H\]cholesterol as lipid emulsions and studied its appearance in the bile[@b28][@b29]. *Bmal1*^*−/−*^*Apoe*^*−/−*^ mice had lower amounts of cholesterol in their bile compared with *Bmal1*^*+/+*^*Apoe*^*−/−*^ mice ([Fig. 2h](#f2){ref-type="fig"}). Further, *Bmal1*^*−/−*^*Apoe*^*−/−*^ primary hepatocytes effluxed significantly lower amounts of cholesterol to bile acid acceptors ([Supplementary Fig. 6](#S1){ref-type="supplementary-material"}). In addition, *Bmal1*^*−/−*^*Apoe*^*−/−*^ mice also had lower amounts of cholesterol in their bile and feces compared with *Bmal1*^*+/+*^*Apoe*^*−/−*^ mice ([Supplementary Fig. 7](#S1){ref-type="supplementary-material"}). These studies showed that Bmal1 deficiency reduces the expression of Abcg5/Abcg8 and cholesterol excretion to the bile.
Delivery of cholesterol to bile is an important step in reverse cholesterol transport (RCT)[@b29][@b30]. Therefore, we also performed RCT by intraperitoneal injections of equal amounts of ^3^H-cholesterol loaded J774 macrophages in *Bmal1*^*−/−*^*Apoe*^*−/−*^ and *Bmal1*^*+/+*^*Apoe*^*−/−*^ mice. The amounts of cholesterol in the bile and feces were significantly less in *Bmal1*^*−/−*^*Apoe*^*−/−*^ mice compared with controls ([Fig. 2i](#f2){ref-type="fig"}). This was not secondary to reduced delivery of cholesterol to the liver as the amounts of cholesterol delivered to the liver were similar in these mice. These studies indicated that cholesterol delivered to the liver via RCT is not excreted efficiently to the bile and feces in Bmal1-deficient mice.
Lxrα, Hnf4α and Gata4 are the major transcription factors involved in the regulation of ABCG5/G8 expression[@b6]. Lxrα and Hnf4α levels did not differ between the livers of *Bmal1*^*−/−*^*Apoe*^*−/−*^ and *Bmal1*^*+/+*^*Apoe*^*−/−*^mice, but *Bmal1*^*−/−*^*Apoe*^*−/−*^ livers had significantly reduced the levels of Abcg5, Abcg8 and Gata4 mRNA and protein levels ([Fig. 2f,g](#f2){ref-type="fig"}) suggesting that Bmal1 might modulate Gata4 expression to regulate Abcg5/Abcg8 expression and biliary cholesterol secretion.
Hepatic Bmal1 deficiency increases atherosclerosis
--------------------------------------------------
The above studies showed that global Bmal1 deficiency increases hepatic lipoprotein production, plasma lipids and atherosclerosis while reducing cholesterol secretion to bile. To address whether this is a consequence of global Bmal1 deficiency or liver-specific function of Bmal1, we generated liver-specific Bmal1-deficient Apoe^*−/−*^mice (*L-Bmal1*^*−/−*^*Apoe*^*−/−*^). Visualization of aortic arches revealed increased lesions in chow fed *L-Bmal1*^*−/−*^*Apoe*^*−/−*^ mice compared with *Bmal1*^*fl/fl*^*Apoe*^*−/−*^ *mice* ([Fig. 3a](#f3){ref-type="fig"}). Further, these mice had higher lipids in their aortas ([Fig. 3b](#f3){ref-type="fig"}). Amounts of hepatic triglyceride and cholesterol, but not phospholipids, were significantly increased in *L-Bmal1*^*−/−*^*Apoe*^*−/−*^ mice compared with *Bmal1*^*fl/fl*^*Apoe*^*−/−*^ controls ([Fig. 3c](#f3){ref-type="fig"}). *L-Bmal1*^*−/−*^*Apoe*^*−/−*^ mice had significantly higher plasma triglyceride and cholesterol in non-HDL ([Fig. 3d](#f3){ref-type="fig"}), VLDL/LDL ([Supplementary Fig. 8A](#S1){ref-type="supplementary-material"}) fractions. Plasma of *L-Bmal1*^*−/−*^*Apoe*^*−/−*^ mice contained higher amounts of apoB100, while apoB48 and apoAI levels were similar ([Fig. 3d](#f3){ref-type="fig"} and [Supplementary Fig. 8B](#S1){ref-type="supplementary-material"}). Thus, liver-specific Bmal1 deficiency causes hyperlipidaemia due to increases in apoB100-containing triglyceride and cholesterol enriched lipoproteins.
Hepatic Bmal1 deficiency increases lipoprotein production
---------------------------------------------------------
Physiologic studies revealed that triglyceride production rates were higher in *L-Bmal1*^*−/−*^*Apoe*^*−/−*^ mice compared with *Bmal1*^*fl/fl*^*Apoe*^*−/−*^ mice ([Fig. 3e](#f3){ref-type="fig"}). Further, MTP activity, mRNA and protein levels were significantly increased in the livers of *L-Bmal1*^*−/−*^*Apoe*^*−/−*^ mice ([Fig. 3f](#f3){ref-type="fig"}). We previously showed that MTP expression was increased in *Clk*^*Δ19/Δ19*^ mice due to reduced Shp expression[@b16]. Shp mRNA and protein levels were also reduced in the livers of *L-Bmal1*^*−/−*^*Apoe*^*−/−*^ mice compared with *Bmal1*^*fl/fl*^*Apoe*^*−/−*^ mice ([Fig. 3f](#f3){ref-type="fig"}). These studies suggest that hepatic Bmal1 deficiency increases lipoprotein production most likely by reducing Shp and increasing MTP expression.
Hepatic Bmal1 deficiency reduces cholesterol efflux to bile
-----------------------------------------------------------
Next, we looked at the secretion of cholesterol to the bile in *L-Bmal1*^*−/−*^*Apoe*^*−/−*^ and *Bmal1*^*fl/fl*^*Apoe*^*−/−*^ mice. Bile flow rates were similar in these mice. Fecal cholesterol levels were significantly lower in *L-Bmal1*^*−/−*^*Apoe*^*−/−*^ compared to *Bmal1*^*fl/fl*^*Apoe*^*−/−*^ mice ([Supplementary Fig. 9A,B](#S1){ref-type="supplementary-material"}). *L-Bmal1*^*−/−*^*Apoe*^*−/−*^ mice secreted less cholesterol to bile when injected intravenously ([Fig. 3g](#f3){ref-type="fig"}). And, *L-Bmal1*^*−/−*^*Apoe*^*−/−*^ hepatocytes effluxed less cholesterol to bile acid acceptors ([Supplementary Fig. 9C](#S1){ref-type="supplementary-material"}). Further, mRNA levels of Gata4 and Gata6 were significantly reduced ([Supplementary Fig. 9D](#S1){ref-type="supplementary-material"}). The amounts of cholesterol delivered from J774 macrophage to the liver were not affected but those excreted to the bile and feces during RCT were significantly reduced in these mice ([Fig. 3h](#f3){ref-type="fig"}). Moreover, cholesterol, but not total phospholipids and bile acid, mass in the bile of *L-Bmal1*^*−/−*^*Apoe*^*−/−*^ mice was significantly less compared with controls ([Fig. 3i](#f3){ref-type="fig"}). The mRNA and protein ([Fig. 3j](#f3){ref-type="fig"}) levels of Abcg5 and Abcg8 were significantly decreased in the livers of *L-Bmal1*^*−/−*^*Apoe*^*−/−*^ mice compared with *Bmal1*^*fl/fl*^*Apoe*^*−/−*^ mice. These studies suggested that liver-specific Bmal1 deficiency reduces Abcg5 and Abcg8 expression to lower excretion of cholesterol to bile.
Western diet enhanced atherosclerosis in *L-Bmal1*^*−/−*^*Apoe*^*−/−*^ mice ([Supplementary Fig. 10A,B](#S1){ref-type="supplementary-material"}). It increased hepatic triglyceride, cholesterol and cholesterol esters but had no effect on free cholesterol ([Supplementary Fig. 10C](#S1){ref-type="supplementary-material"}). Plasma fractionation studies showed higher triglyceride and cholesterol in VLDL ([Supplementary Fig. 10D](#S1){ref-type="supplementary-material"}). Thus, western diet augments atherosclerosis in liver-specific Bmal1-deficient *Apoe*^*−/−*^ mice.
Effects of hepatic Bmal1 overexpression
---------------------------------------
To determine further the role of hepatic Bmal1, we expressed human BMAL1 using adenoviruses (Adv-BMAL1) in western diet fed *L-Bmal1*^*−/−*^*Apoe*^*−/−*^ mice. After 4 weeks of transduction, hepatic BMAL1 mRNA and protein levels were ∼9-fold higher compared with controls ([Fig. 4a](#f4){ref-type="fig"}). Overexpression of BMAL1 had no effect on Gata6 mRNA; increased Gata4, Abcg5, Abcg8 and Shp; and decreased MTP mRNA and protein ([Fig. 4a,b](#f4){ref-type="fig"}) as well as activity ([Fig. 4c](#f4){ref-type="fig"}). Further, it significantly reduced hepatic cholesterol and triglyceride compared with mice transduced with Ad-GFP control virus ([Fig. 4d](#f4){ref-type="fig"}). BMAL1 expressing mice had lower plasma triglyceride and cholesterol due to reductions in non-HDL ([Fig. 4e](#f4){ref-type="fig"}). Moreover, these mice had ∼75% less atherosclerosis ([Fig. 4f,g](#f4){ref-type="fig"}). Overexpression of Bmal1 increased cholesterol and bile acids in the bile but had no effect on phospholipids. Further, it increased fecal cholesterol but had no effect on bile acids ([Supplementary Fig. 11](#S1){ref-type="supplementary-material"}). Cholesterol efflux to bile after intravenous injection was increased in BMAL1 expressing mice compared with GFP expressing mice ([Fig. 4h](#f4){ref-type="fig"}). Further, amounts of cholesterol delivered to the bile and feces from J774 cells during RCT were increased in Bmal1 expressing mice ([Fig. 4i](#f4){ref-type="fig"}). Similar observations were made in *Bmal1*^*fl/fl*^*Apoe*^*−/−*^ mice after overexpressing Bmal1 ([Supplementary Fig. 12](#S1){ref-type="supplementary-material"}). Thus, hepatic over expression of Bmal1 lowers plasma lipids, mitigates atherosclerosis and enhances biliary cholesterol secretion in *L-Bmal1*^*−/−*^*Apoe*^*−/−*^ and in *Bmal1*^*fl/fl*^*Apoe*^*−/−*^ mice. In short, these studies involving both hepatic Bmal1 ablation and over expression indicate that hepatic Bmal1 is a major regulator of plasma and tissue lipid levels.
Effects of forced hepatic SHP expression
----------------------------------------
Bmal1 ablation studies indicated that Bmal1 might regulate Shp to affect MTP expression and hepatic lipoprotein production. To determine whether Shp is an intermediary transcription factor used by Bmal1 to control plasma lipids, we injected *L-Bmal1*^*−/−*^*Apoe*^*−/−*^ mice with adenoviruses expressing human SHP (Adv-SHP) or green fluorescence protein (Adv-GFP) and started on a Western diet. Four weeks later, *L-Bmal1*^*−/−*^*Apoe*^*−/−*^ mice transduced with Adv-SHP had six-fold higher hepatic SHP expression ([Fig. 5a](#f5){ref-type="fig"}) and ∼50% lower MTP mRNA, activity, and protein ([Fig. 5a--c](#f5){ref-type="fig"}). SHP expressing mice had significantly lower plasma triglyceride and cholesterol mainly in non-HDL particles ([Fig. 5d](#f5){ref-type="fig"}) most likely due to lower hepatic lipoprotein production ([Supplementary Fig. 13](#S1){ref-type="supplementary-material"}). In contrast, hepatic triglyceride and cholesterol were increased ([Fig. 5e](#f5){ref-type="fig"}). Further, SHP expressing mice had ∼50% less atherosclerotic lesions compared with controls ([Fig. 5f,g](#f5){ref-type="fig"}). Gene expression analysis showed that SHP overexpression had no effect on Abcg5 and Abcg8 mRNA ([Fig. 5h](#f5){ref-type="fig"}) and biliary cholesterol secretion ([Fig. 5i](#f5){ref-type="fig"}). Similar results were obtained in *Bmal1*^*fl/fl*^*Apoe*^*−/−*^ mice transduced with Adv-SHP ([Supplementary Fig. 14](#S1){ref-type="supplementary-material"}). Thus, Shp is involved in the regulation of MTP and VLDL production but not in the regulation of biliary cholesterol secretion in these mice.
Bmal1 regulates Gata4 to modulate Abcg5
---------------------------------------
Since Shp expression had no effect on cholesterol efflux to bile in *L-Bmal1*^*−/−*^*Apoe*^*−/−*^ mice, we hypothesized that Bmal1 might regulate this pathway involving another transcription factor. To test this, wild-type primary hepatocytes were transfected with siBmal1 and changes in several candidate genes known to regulate Abcg5/Abcg8 were quantified. SiBmal1 significantly reduced Bmal1, Abcg5, Abcg8, Gata4, Fog1 (a Gata4 response gene); but had no effect on Gata6 and Gapdh mRNA and protein levels ([Fig. 6a](#f6){ref-type="fig"}). Further, it reduced cholesterol efflux from hepatocytes to bile acid acceptors ([Fig. 6b](#f6){ref-type="fig"}, left). Similar reductions in gene expression and cholesterol excretion to bile acid acceptors were observed in siBmal1-treated human hepatoma Huh-7 cells ([Supplementary Fig. 15](#S1){ref-type="supplementary-material"}). These studies indicated that Bmal1 deficiency reduces Gata4, Abcg5, Abcg8 expression and cholesterol efflux to bile acid acceptors in liver cells.
To evaluate the role of Gata4 in the regulation of Abcg5 and Abcg8, wild-type hepatocytes were treated with siGata4. SiGata4 decreased cholesterol efflux to bile acid acceptors ([Fig. 6b](#f6){ref-type="fig"} right); reduced Gata4, Abcg5, Abcg8, and Fog1 expression; and had no effect on Gata6 and Gapdh mRNA and protein levels ([Fig. 6c](#f6){ref-type="fig"}). These studies suggest that Gata4 regulates Abcg5 and Abcg8 expression and cholesterol efflux to bile acids.
To determine whether Bmal1 acts via Gata4 to regulate Abcg5 and Abcg8, hepatocytes were transfected with siBmal1, siGata4 or siBmal1+siGata4. SiBmal1 significantly reduced Bmal1, Gata4, Abcg5 and Abcg8 without affecting Gapdh, Abca1 and Gata6 mRNA and protein levels ([Fig. 6d](#f6){ref-type="fig"}). As expected, siGata4 reduced its own expression as well as that of Abcag5 and Abcg8, but had no effect on Gapdh, Abca1 and Gata6. A combination of siBmal1+siGata4 reduced the levels of Abcg5 and Abcg8 to the same extent as individual siRNAs indicating that both Bmal1 and Gata4 are in the same pathway.
Bmal1 modulates cyclic expression of Abcg5 and Gata4
----------------------------------------------------
Next, we asked whether Abcg5 shows cyclic expression and whether Gata4 is involved in the rhythmic regulation of Abcg5 by Bmal1. wild-type hepatocytes were transfected with siControl or siBmal1 and then treated with 50% serum for 2 h. Cyclic expression of candidate genes was followed over time in normal media. Bmal1 expression showed cyclic expression with first peak at 12--16 h followed by a second peak at 36--40 h. SiBmal1 significantly reduced Bmal1 expression and residual levels did not show cyclic changes ([Fig. 6e](#f6){ref-type="fig"}). SiControl-treated cells showed peak expressions of Abcg5 and Gata4 expression at ∼12 h after serum supplementation and a second peak at ∼36 h ([Fig. 6e](#f6){ref-type="fig"}). Gata6 mRNA exhibited peak expressions at 8 and 32 h. In siBmal1-treated cells, peak expression levels of Abcg5 and Gata4 at 12 h were significantly reduced. However, there were more pronounced reductions in their expressions at the second 36 h peak. SiBmal1 had no significant effect on the peak expression of Gata6 at 8 h and appears to dampen second peak expression at 32--36 h. These studies indicate that Abcg5, Abcg8 and Gata4 exhibit cyclic changes in hepatocytes after serum synchronization and these changes are diminished in the absence of Bmal1.
Further, we studied the cyclic expression of Bmal1, Abcg5, Gata4 and Gata6 in primary hepatocytes isolated from *Bmal1*^*+/+*^*Apoe*^*−/−*^ and *Bmal1*^*−/−*^*Apoe*^*−/−*^ mice ([Supplementary Fig. 16](#S1){ref-type="supplementary-material"}), and in *L-Bmal1*^*−/−*^*Apoe*^*−/−*^ and *Bmal1*^*fl/fl*^*Apoe*^*−/−*^ mice ([Fig. 6f](#f6){ref-type="fig"}). As expected, Bmal1 showed cyclic expression in *Bmal1*^*fl/fl*^*Apoe*^*−/−*^ and in *Bmal1*^*+/+*^*Apoe*^*−/−*^ but not in *L-Bmal1*^*−/−*^*Apoe*^*−/−*^ and *Bmal1*^*−/−*^*Apoe*^*−/−*^ mice. Abcg5 expression showed two peaks in the *Bmal1*^*fl/fl*^*Apoe*^*−/−*^ and *Bmal1*^*+/+*^*Apoe*^*−/−*^ hepatocytes. In *L-Bmal1*^*−/−*^*Apoe*^*−/−*^ and *Bmal1*^*−/−*^*Apoe*^*−/−*^ hepatocytes, the first peak was lower and the second peak was absent. In *L-Bmal1*^*−/−*^*Apoe*^*−/−*^ and *Bmal1*^*−/−*^*Apoe*^*−/−*^hepatocytes, peak expressions levels of Gata4, but not Gata6, were significantly lower compared with control *Bmal1*^*fl/fl*^*Apoe*^*−/−*^ and *Bmal1*^*+/+*^*Apoe*^*−/−*^ hepatocytes. Thus, Bmal1 deficiency appears to dampen the peak expression of Gata4; in contrast it abolishes the second peak expression of Abcg5 seen in control hepatocytes.
Next, we asked whether expressions of Abcg5, Abcg8 and Gata4 change within a day and whether these changes are controlled by Bmal1 in mice. In *Bmal1*^*+/+*^*Apoe*^*−/−*^ livers, high and low expressions of Abcg5, Abcg8, Gata4 and Gata6 were seen at 16:00 and 4:00 h, respectively ([Fig. 7a](#f7){ref-type="fig"}). In *Bmal1*^*−/−*^*Apoe*^*−/−*^ livers, the peak expressions of these genes at 16:00 h were significantly reduced but not the nadirs seen at 4:00 h. Similar to global deficiency of Bmal1, liver-specific deficiency also significantly reduced mRNA and protein levels of Gata4 and Gata6 at 16 h ([Fig. 7b](#f7){ref-type="fig"}). These studies indicated that Bmal1 augments expression of these genes just before the nighttime suggesting that Bmal1 contributes to their peak expressions.
Bmal1 binds to the Gata4 and Gata4 interacts with the Abcg5 promoter
--------------------------------------------------------------------
The above studies suggested that Bmal1 might regulate Abcg5 and Abcg8 expression via Gata4 to modulate cholesterol excretion to bile. We then asked whether Bmal1 can interact with the *Gata4* promoter. Bioinformatics analysis revealed several E-boxes in the promoter of Gata4. We concentrated on the one E-box that was proximal to the transcription start site in the promoter of *Gata4* ([Supplementary Fig. 17A](#S1){ref-type="supplementary-material"}); therefore, we performed semi-quantitative ([Fig. 7c,d](#f7){ref-type="fig"}) and quantitative ([Fig. 7e,f](#f7){ref-type="fig"}) chromatin immunoprecipitation (ChIP) in the livers at different times to determine the binding of Bmal1. In *Bmal1*^*+/+*^*Apoe*^*−/−*^ livers, the binding of Bmal1 to *Gata4* E-box was high at 16:00 h than at 4:00 h ([Fig. 7c,e](#f7){ref-type="fig"}). In *Bmal1*^*−/−*^*Apoe*^*−/−*^ liver, as expected, Bmal1 was not associated with the promoter. Bmal1 did not appear to bind the *Gata6* promoter ([Fig. 7c](#f7){ref-type="fig"}). We then studied the binding of Gata4 and Gata6 to the *Abcg5* promoter that contains a GATA-box ([Supplementary Fig. 17B](#S1){ref-type="supplementary-material"}). In *Bmal1*^*+/+*^*Apoe*^*−/−*^ control mice, Gata4 binding to the *Abcg5* promoter was high at 16:00 h and low at 4:00 h ([Fig. 7c,e](#f7){ref-type="fig"}). In *Bmal1*^*−/−*^*Apoe*^*−/−*^ mice, the binding of Gata4 to the *Abcg5* promoter at 16:00 h was significantly reduced but this binding was not reduced at 4:00 h. Gata6 did not bind to the *Abcg5* promoter. Similar results were obtained in liver-specific Bmal1 ablated mice ([Fig. 7d,f](#f7){ref-type="fig"}). Bmal1 interacted with the *Gata4*, but not with *Gata6*, promoter and this binding was not seen in *L-Bmal1*^*−/−*^*Apoe*^*−/−*^ livers. Binding of Bmal1 to Gata4 promoter was higher at 16:00 h compared with 4:00 h. Gata4 binding to the *Abcg5* promoter was high in *Bmal1*^*fl/fl*^*Apoe*^*−/−*^ mice at 16:00 h but was significantly attenuated in *L-Bmal1*^*−/−*^*Apoe*^*−/−*^ mice. These data suggest that Bmal1 interacts with *Gata4* promoter at the onset of nighttime to increase expression resulting in enhanced binding of Gata4 to the *Abcg5* promoter.
GATA4 affects atherosclerosis and biliary cholesterol excretion
---------------------------------------------------------------
Studies described above indicated that Gata4 might be an intermediary transcription factor regulating *Abcg5* and *Abcg8* gene expression by Bmal1. To test this further, *L-Bmal1*^*−/−*^*Apoe*^*−/−*^ mice were transduced with Adv-GATA4 expressing human GATA4. This transduction increased GATA4 expression by 10-fold ([Fig. 8a](#f8){ref-type="fig"}). GATA4 overexpression had no effect on Gata6 and Gapdh mRNA and protein levels, but increased Abcg5, Abcg8, Fxr and Shp mRNA and protein while decreasing MTP expression when compared with mice transduced with Ad-GFP ([Fig. 8a](#f8){ref-type="fig"}). Further, Gata4 overexpression decreased MTP activity ([Fig. 8b](#f8){ref-type="fig"}), and plasma triglyceride and cholesterol in non-HDL lipoproteins ([Fig. 8c](#f8){ref-type="fig"}). Hepatic cholesterol was unaffected but triglyceride was increased ([Fig. 8d](#f8){ref-type="fig"}). In addition, bile and fecal cholesterol levels were increased after the overexpression of GATA4 ([Supplementary Fig. 18](#S1){ref-type="supplementary-material"}). Mice overexpressing GATA4 accumulated more cholesterol in the bile compared with mice overexpressing GFP after intravenous injection of ^3^H-cholesterol ([Fig. 8e](#f8){ref-type="fig"}). Further, cholesterol efflux from J774 macrophages during RCT to the liver, feces and bile was higher in GATA4 expressing mice compared with GFP expressing mice ([Fig. 8f](#f8){ref-type="fig"}). GATA4 expression reduced lesion areas in the aortic arches and lipid staining in the abdominal aortas by ∼60--75% ([Fig. 8g,h](#f8){ref-type="fig"}). Similar effects were observed in the *Bmal1*^*fl/fl*^*Apoe*^*−/−*^ mice after injecting Adv-GATA4 mice ([Supplementary Fig. 19](#S1){ref-type="supplementary-material"}). Thus, overexpression of GATA4 enhances hepatic expression of Abcg5 and Abcg8, and cholesterol excretion to the bile in *L-Bmal1*^*−/−*^*Apoe*^*−/−*^ and *Bmal1*^*fl/fl*^*Apoe*^*−/−*^ mice, while reducing MTP expression and atherosclerosis.
Bmal1 regulates cholesterol excretion to bile in wild type mice
---------------------------------------------------------------
The above studies were performed in *Bmal1*^*+/+*^*Apoe*^*−/−*^ mice. To determine whether this regulation is specific to these mice, we studied the effect of global (*Bmal1*^*−/−*^) and liver-specific (*L-Bmal1*^*−/−*^) Bmal1 deficiency in C57Bl6J mice on the hepatic expression of Abcg5 and Gata4 ([Fig. 9](#f9){ref-type="fig"}). Livers from *Bmal1*^*−/−*^ and *L-Bmal1*^*−/−*^ mice had significantly lower levels of Gata4, Gata6 and Fog1 compared with their respective controls ([Fig. 9a,b](#f9){ref-type="fig"}). Overexpression of Bmal1+Clock in wild-type hepatocytes significantly increased the expression of Bmal1, Gatat4, Fog1, Abcg5 and Abcg8 ([Fig. 9c](#f9){ref-type="fig"}), and increased cholesterol efflux to bile acid acceptors ([Fig. 9d](#f9){ref-type="fig"}). These studies indicate that Bmal1 regulates expression of Gata4 and Abcg5.
In addition, we studied cyclic expression of these genes in *Bmal1*^*−/−*^ deficient hepatocytes subjected to serum shock ([Fig. 9e](#f9){ref-type="fig"}). Bmal1 showed cyclic expression with two peaks at 16 and 36--40 h. Bmal1 expression was not seen in *Bmal1*^*−/−*^ hepatocytes. Abcg5 and Gata4 mRNA showed peak expressions at 12 and 36 h consistent with ([Fig. 6e,f](#f6){ref-type="fig"}). Gata6 showed peak expressions at 8 and 32 h, and these peaks were significantly reduced in Bmal1-deficient hepatocytes. Thus, Bmal1 deficiency significantly reduced the expression of Abcg5, Gata4 and Gata6 at peak hours.
We also studied temporal changes within a day in the hepatic expression of these genes in *Bmal1*^*−/−*^ and *L-Bmal1*^*−/−*^ mice ([Fig. 9f,g](#f9){ref-type="fig"}). Expression of Abcg5, Abcg8 and Gata4 showed maximum expression in the daytime with a peak at 16:00 h. In Bmal1-deficient animals, mRNA levels of Abcg5, Abcg8 and Gata4 were very low and they did not show significant changes within a day. Gata6 expression was also higher in the daytime but the peak was at 20:00 h. These data are consistent with CircaDB[@b31]. In Bmal1-deficient mice expression of Gata6 was significantly reduced and these low levels showed peak expression at 20:00 h.
Changes in the expression of these genes within a day were coincident with the binding of Bmal1 to the *Gata4* promoter and the binding of Gata4 to the *Abcg5* promoter. The binding of Bmal1 to the *Gata4* promoter was the highest at 16:00 h and low at 24:00 h ([Fig. 9h](#f9){ref-type="fig"}). This binding was not seen in *Bmal1*^*−/−*^ livers. The binding of Gata4 to the *Abcg5* promoter was high at 16--24 h. These studies indicate that Bmal1 regulates diurnal expression of Abcg5 by upregulating Gata4 in wild-type C57Bl6J mice.
Discussion
==========
We used global and liver-specific Bmal1-deficient *Apoe*^*−/−*^ mice to examine the role of Bmal1 in the regulation of plasma and hepatic lipids as well as development of atherosclerosis. Both *Bmal1*^*−/−*^*Apoe*^*−/−*^ and *L-Bmal1*^*−/−*^*Apoe*^*−/−*^ mice showed increased hyperlipidaemia and atherosclerosis compared with apoE-deficient mice. Adenovirus mediated hepatic overexpression of Bmal1 in *L-Bmal1*^*−/−*^*Apoe*^*−/−*^ mice reduced hyperlipidaemia and atherosclerosis. On the basis of these knockout and overexpression studies, we conclude that hepatic Bmal1 plays a significant role in the regulation of plasma lipids and atherosclerosis. Mechanistic studies showed that hepatic Bmal1 regulates plasma and hepatic lipids by regulating Shp and Gata4 ([Fig. 10a](#f10){ref-type="fig"}). Bmal1 deficiency reduces Shp and increases Mtp expression and VLDL production. Further, it reduces Gata4 expression leading to diminished Abcg5/Abcg8 expression and cholesterol efflux to bile ([Fig. 10b](#f10){ref-type="fig"}). Thus, hepatic Bmal1 regulates at least two pathways (VLDL production and cholesterol efflux to bile) and disruptions in these pathways increase plasma lipids and atherosclerosis.
These studies for the first time showed that Bmal1 deficiency affects cholesterol secretion to the bile. Several lines of evidence suggest that Bmal1 regulates this pathway by modulating the expression of Abcg5/Abcg8 via Gata4. First, simultaneous knockdown of Bmal1 and Gata4 reduce Abcg5/Abcg8 to the same extent as their individual knockdowns ([Fig. 6d](#f6){ref-type="fig"}). Second, Bmal1 deficiency reduces mRNA and protein levels of Abcg5, Abcg8 and Gata4 in the liver ([Fig. 2f,g](#f2){ref-type="fig"}), and BMAL1 overexpression increases their levels ([Fig. 4a,b](#f4){ref-type="fig"}). Third, temporal variations in the expression of Gata4, Abcg5 and Abcg8 were correlated with changes in Bmal1 levels within 24 h and these changes were dampened in Bmal1 deficiency ([Figs 6e,f](#f6){ref-type="fig"} and [9e--g](#f9){ref-type="fig"}). Fourth, overexpression of GATA4 in Bmal1-deficient mice increased Abcg5/Abcg8 expression and cholesterol secretion to bile in *Apoe*^*−/−*^ mice ([Fig. 8a,b,e,f](#f8){ref-type="fig"} and [Supplementary Fig. 18](#S1){ref-type="supplementary-material"}). Fifth, Bmal1 showed time-dependent interactions with the Gata4 promoter ([Fig. 7c-f](#f7){ref-type="fig"}). Sixth, temporal binding of Gata4 to the *Abcg5* promoter was dampened in the absence of Bmal1 ([Figs 7c--f](#f7){ref-type="fig"} and [9h](#f9){ref-type="fig"}).
Excretion of cholesterol to bile is a critical step in RCT, an important physiologic process that extracts cholesterol from peripheral tissues, mainly macrophages, delivers it to the liver for excretion out of the body[@b32][@b33][@b34]. We did not see significant differences in the delivery of cholesterol from macrophages to the liver in *L-Bmal1*^*−/−*^*Apoe*^*−/−*^ mice. Instead, we observed significant reduction in the excretion of macrophage-derived radiolabelled cholesterol from the liver to the bile and feces. We infer this to suggest that Bmal1 deficiency might not affect cholesterol efflux from macrophages and subsequent delivery to the liver. This is supported by the observation that hepatic Abca1 and Sr-b1 were not affected in Bmal1-deficient mice ([Fig. 2g](#f2){ref-type="fig"}). Instead, Bmal1 deficiency selectively regulates cholesterol secretion to the bile and plasma.
Due to strong epidemiological evidence for an inverse relationship between HDL and cardiovascular disease, attempts are underway to increase HDL functionality. This is usually assayed by studying cholesterol efflux from cultured macrophages using plasma or isolated HDL. Our studies suggest that monitoring of this first step might not be sufficient. There is a possibility of defects in subsequent steps of RCT, such as we have observed for cholesterol excretion into bile in Bmal1-deficient mice. Therefore, measurement of cholesterol excretion to the feces might be needed to evaluate efficacy of new therapeutic drugs.
Bmal1 deficiency increased triglyceride concentration in the plasma. Physiological studies suggested that increases in plasma triglyceride might be secondary to overproduction of hepatic lipoproteins. Molecular studies indicated that this might be related to enhanced expression of MTP. Mechanistic studies showed that this increase might be due to reduced expression of Shp, a repressor of MTP. The critical role of Shp in the regulation of MTP and lipoprotein production was evaluated by over expressing Shp in *L-Bmal1*^*−/−*^*Apoe*^*−/−*^ mice ([Fig. 5](#f5){ref-type="fig"} and [Supplementary Fig 13](#S1){ref-type="supplementary-material"}). Expression of SHP reduced MTP expression, plasma lipids and atherosclerosis supporting the hypothesis that Bmal1 regulates hepatic lipoprotein production by regulating Shp.
In global Bmal1-deficient mice, both enhanced hepatic lipoprotein production and reduced cholesterol secretion to the bile might play an important role in the development of atherosclerosis. This is based on the observations that reductions in MTP activity reduces plasma lipids and atherosclerosis[@b35], and global ablation of Abcg5 and Abcg8 increases atherosclerosis[@b36][@b37][@b38]. However, it is unclear if reductions in cholesterol secretion to the bile could contribute to atherosclerosis in *L-Bmal1*^*−/−*^*Apoe*^*−/−*^ mice. It has been shown that the liver-specific overexpression of Abcg5 in *Abcg5*^*−/−*^ mice does not reduce plasma cholesterol and atherosclerosis most likely due to increased absorption of cholesterol in the intestine[@b38]. However, our studies involving overexpression of BMAL1, SHP or GATA4 in *L-Bmal1*^*−/−*^*Apoe*^*−/−*^ mice indicate that a combination of hepatic lipoprotein over production and reduced biliary cholesterol excretion might contribute additively to atherosclerosis. This is based on the observation that expression of BMAL1 or GATA4 reduced atherosclerosis to similar extents (∼75%), whereas Shp expression reduced it by ∼50%. Since, Bmal1 and Gata4 affect both lipoprotein production and cholesterol secretion to bile, it is likely that their higher efficacy in reducing atherosclerosis compared with Shp is secondary to their effects on both these processes.
It is unclear why Bmal1 coordinately regulates hepatic lipoprotein production and cholesterol excretion. It is possible that increased VLDL production might protect liver from excess cholesterol accumulation that might ensue when excretion to the bile is curtailed. This knowledge about the modulation of two functions probably indicates that Bmal1 might regulate hepatic lipid metabolism by affecting several different pathways.
By studying atherosclerosis in *Bmal1*^*−/−*^*Apoe*^*−/−*^ and *L-Bmal1*^*−/−*^*Apoe*^*−/−*^ mice we tried to learn about the contribution of hepatic Bmal1 to atherogenesis. The atherosclerotic lesion areas were smaller in *L-Bmal1*^*−/−*^*Apoe*^*−/−*^ mice indicating that Bmal1 deficiency in other tissues also contributes to atherosclerosis. In this regards, it would be interesting to explore in the future the role of intestine and macrophage-specific Bmal1 on atherosclerosis. Aortic transplantation studies have shown that aortic Bmal1 contributes to atherosclerosis. Thus, further studies can also be conducted to address the role of endothelial and smooth muscle cell Bmal1 on atherogenesis.
Our studies showed that Bmal1 deficiency does not affect hepatic Abca1 and Abcg1 levels in the liver. This was surprising as we had previously noted that Clock^Δ19/Δ19^ regulates cholesterol efflux from macrophages by modulating Abca1 expression[@b17]. More studies are needed to fully explain differential, tissue-specific mechanisms involved in the regulation of Abca1 by clock genes.
Bmal1 is an integrated component of the clock regulatory transcription factors that form an auto-regulatory loop. Therefore, it is possible that the effect of Bmal1 deficiency might be a reflection of disruptions in the whole circadian regulatory mechanisms. More studies are needed to find out whether the observed effects are due to Bmal1 deficiency or they are related to altered expression of downstream targets such as Pers, Crys and other transcription factors. In fact, it has been shown that Rev-erba regulates lipid metabolism by both circadian and epigenetic mechanisms[@b39]. Studies in mice deficient in individual or several of these transcription factors may provide information about circadian and transcription factor specific effects of clock genes in the regulation of plasma lipids and atherosclerosis.
In short, these studies show that Bmal1 deficiency affects two different hepatic pathways of cholesterol secretion. Bmal1 deficiency increases cholesterol secretion via VLDL and reduces cholesterol secretion to bile. Increased VLDL assembly and secretion is due to suppression of Shp and induction of MTP expression. The reductions in cholesterol secretion to the bile are due to lower expressions of Gata4, Abcg5 and Abcg8. Thus, Bmal1 regulates a repressor and an activator to control hepatic lipid homeostasis. It is likely that Bmal1 might regulate additional pathways and acts as a master regulator of hepatic lipid metabolism.
Methods
=======
Materials
---------
\[^3^H\]Cholesterol and \[^3^H\]glycerol were from NEN LifeScience Products. AcLDL was from Biomedical Technologies, Inc.
Animals
-------
Bmal1^+/−^, Bmal1^+/−^Ldlr^−/−^ and Bmal1^+/−^Apoe^−/−^ heterozygous mice were bred to obtained Bmal1^+/−^, Bmal1^−/−^, Bmal1^−/−^Ldlr^−/−^, Ldlr^−/−^, Bmal1^−/−^Apoe^−/−^ and Apoe^−/−^ siblings. To obtain liver-specific Bmal1 gene deleted (Alb-cre-Bmal1^fl/fl^, L-Bmal1^−/−^) mice, Bmal1^fl/fl^ mice were crossed with Alb-Cre (Jackson Laboratories) mice. They were then bred with Apoe^−/−^ mice to get Bmal1^fl/fl^Apoe^−/−^ and L-Bmal1^−/−^Apoe^−/−^ mice. All mice were on C57Bl/6J background. Knockout and wild-type siblings from heterozygous parents were used for studies. Global Bmal1 deficient and their controls were provided food in the bottom of their cages as these mice are unable to reach to the top for food[@b40][@b41]. Provision of food in the bottom of the cage prevents early death reported in these mice[@b42]. They were fed a chow diet (LabDiet, \#5001). Unless indicated otherwise, male, 3--4 months old mice were fed a Western diet (Harland Tekland, TD88137) for atherosclerosis studies. Mice were housed with a 12-h lighting schedule (7:00--19:00 h). Animal experiments were approved by the Animal Care and Use Committee of the SUNY Downstate Medical Center.
Other analyses
--------------
After 4 h fast, plasma was obtained to measure lipids using kits. Mice were not fasted when daily changes in plasma and tissue lipids were studied. Triglyceride, cholesterol, free cholesterol and phospholipid in tissues, bile and feces were determined using kits[@b4][@b16]. All the feces collected during 48 h were used for lipid extraction.
Characterization of lipoproteins
--------------------------------
Plasma was subjected to FPLC and different fractions were collected. In addition, plasma was subjected to sequential density gradient centrifugation to isolate VLDL, LDL and HDL. Plasma was also subjected to precipitation with tungstate/MgCl~2~ to obtain values in HDL and non-HDL fractions[@b43]. Lipid synthesis and secretion were also studied[@b43].
Hepatic lipoprotein production
------------------------------
Mice were fasted for 4 h and injected intraperitoneally with 0.5 ml of Poloxamer P407 in PBS (1:6, v/v). Plasma was collected at indicated times to measure lipids and determine triglyceride production rates[@b16]. MTP activity was determined using a kit[@b44].
Western blot analyses
---------------------
Proteins were separated on polyacrylamide gels, transferred to nitrocellulose membranes, blocked for 2 h in 20 mM Tris, 137 mM NaCl, pH 7.5, containing 0.1% Tween 20 and 5% nonfat dry milk at room temperature. The blots were washed three times and incubated overnight at 4 °C in the same buffer containing 0.5% dry milk with different antibodies ([Supplementary Table 2](#S1){ref-type="supplementary-material"}), washed, and then incubated with mouse horseradish peroxidase-conjugated secondary antibody (1:1,000--1:4,000) in 1% skim milk for 1 h. Immune reactivity was detected by chemiluminescence[@b4][@b16][@b45] . Full scans of western blots and gels are presented in [Supplementary Fig. 20](#S1){ref-type="supplementary-material"}.
Electron microscopy of lipoproteins
-----------------------------------
VLDL, LDL and HDL fractions were isolated from the plasma by ultracentrifugation. Fractions were mixed 1:1 (v/v) with 2% phosphotungstate for 1 min and applied to a carbon-formvar grid[@b46]. Excess sample was removed by blotting, air-dried and viewed under an electron microscope.
Cholesterol efflux from hepatocytes to bile acid acceptors
----------------------------------------------------------
Primary hepatocytes (10 × 10^6^ cells in each well of a six-well plate) from different mice were incubated in triplicate with 1 μCi ml^−1^ of \[^3^H\]cholesterol for 16 h. Cells were washed and then incubated in fresh media containing 100 μM tauroursodeoxycholate (TUDC) for different time[@b47]. After the incubation in a humidified 37 °C incubator in the presence of 5% CO~2~, media from each well were collected and centrifuged (12,000 r.p.m., 5 min) to remove cellular debris. Aliquots of the supernatant were taken for counting radioactivity. The remaining cell-associated \[^3^H\]cholesterol was determined after extraction for at least 1 h with β-propanol. The radioactivity released to the medium was expressed as the fraction of the total radioactive cholesterol present in the media and cells.
Evaluation of atherosclerosis
-----------------------------
The proximal aorta was collected after saline perfusion. The aortic root and ascending aorta were sectioned at a thickness of 10 μm. Alternate sections were stained for lipids using Oil Red O (ref. [@b17]). Further sections were treated with Trichrome to measure collagen deposition and with antibodies AIA31240 (Accurate Chemical and Scientific Corp) to assess macrophage infiltration[@b17].
Synchronization of cells to determine cyclic changes
----------------------------------------------------
Primary hepatocytes (10 × 10^6^) were plated in 12-well plates. Hepatocytes were starved in the same medium with no FBS for 18 h. They were then incubated in media containing 50% horse serum for 2 h. Media was changed to serum-free media and cells were harvested at 4 h intervals for analyses[@b16][@b17].
*In vivo* liver cholesterol biliary excretion
---------------------------------------------
Separate groups of mice (*n*=6) received an intravenous (tail vein) injection of \[^3^H\]cholesterol (1 μCi) with unlabelled cholesterol as lipid emulsions[@b28][@b29]. At 4 h, mice were anaesthetized by intraperitoneal injection with ketamine and xylazine. Bile was collected by cannulation of the gallbladder for 30 min, during which body temperature was stabilized using a heating humidified incubator.
*In vivo* reverse cholesterol transport
---------------------------------------
J774A.1 cells were radiolabelled with 5 μCi ml^−1^ of \[^3^H\]cholesterol and 50 μg ml^−1^ acetylated LDL for 48 h. These labelled foam cells were washed twice, equilibrated in medium with 0.2% bovine serum albumin for 8 h, centrifuged and resuspended in RPMI medium immediately before use. The washed cells (2 × 10^6^) were injected intraperitoneally into different mice housed individually with unlimited access to food and water. After 48 h of injection, mice were exsanguinated and perfused with cold PBS. Bile, feces and liver were collected and the radioactivity was measured by liquid scintillation counting (Packard Tri-Carb model 1,500 liquid scintillation counter)[@b4][@b17].
Quantitative RT--PCR
--------------------
Total RNA from tissues and cells were isolated using TriZol. The first-strand complementary DNA was then synthesized using Omniscript RT (Qiagen) kit, and then used for quantitative RT--PCR using SYBR Green to quantify changes in mRNA. The data were analysed using the ΔΔC~T~ method and presented as arbitrary units[@b16][@b17].
Chromatin immunoprecipitation
-----------------------------
ChIP[@b16][@b17] was performed to study the binding of different transcription factors to the promoters of different genes using polyclonal antibodies ([Supplementary Table 1](#S1){ref-type="supplementary-material"}). DNA samples recovered after immunoprecipitation were subjected to quantitative PCR using specific primers ([Supplementary Table 2](#S1){ref-type="supplementary-material"}). As negative controls, ChIP was performed in the absence of antibody or in the presence of rabbit IgG. These experiments were repeated three to four times with similar results and a representative experiment is shown.
Statistical analyses
--------------------
Data are presented as mean±s.d., *n*=6--12 animals for each time point. Statistical testing was performed by the unpaired Student\'s *t*-test. For comparison between two groups one-way analysis of variance (ANOVA) was performed followed by Dunnett\'s correction. Temporal comparisons between two groups were performed using two-way ANOVA as indicated in figure legends. Differences were considered statistically significant when *P*\<0.05. GraphPad Prism was used for graphing and statistical evaluations.
Data availability
-----------------
All relevant data are available from the authors upon request.
Additional information
======================
**How to cite this article**: Pan, X. *et al*. Global and hepatocyte-specific ablation of Bmal1 induces hyperlipidaemia and enhances atherosclerosis. *Nat. Commun.* **7**, 13011 doi: 10.1038/ncomms13011 (2016).
Supplementary Material {#S1}
======================
###### Supplementary Information
Supplementary Figures 1-20 and Supplementary Tables 1-2
We are thankful to Drs Roman Kondratov and Antoch Marino for providing plasmids expressing Clock. Technical assistance of Joyce Queiroz in the early analysis of atherosclerotic plaques is highly appreciated. We thank Wei Quan for technical assistance in electron microscopy. This work was supported in part by National Institutes of Health grant DK-81879 and VA Merit Award (BX001728) to M.M.H., R37 ES005703 to C.A.B. and AHA Scientist Development Grant (2300158) to X.P.
**Author contributions** X.P. conceived, designed and performed experiments, analysed data, interpreted results and wrote a draft of the paper. C.A.B. provided global Bmal1-deficient mice and commented on the manuscript. M.M.H. funded the project, discussed experiments, supervised the progress and thoroughly edited the manuscript.
{#f1}
{#f2}
{#f3}
{#f4}
{#f5}
{#f6}
{ref-type="supplementary-material"}), separated on agarose gels and photographed. Representative of three experiments. (**d**) Liver from *Bmal1*^*fl/fl*^*Apoe*^*−/−*^ and *L-Bmal1*^*−/−*^*Apoe*^*−/−*^ mice were collected at 16:00 h and at 4:00 h to study the binding of Bmal1, Gata4 or Gata6 to the promoters of *Gata4*, *Gata6* and *Abcg5* genes by ChIP (representative of three different experiments). (**e**,**f**) Occupancy of Bmal1 on the Gata4 promoter and the occupancy of Gata4 on the Abcg5 promoter was studied by ChIP and then quantified by qRT--PCR in the livers of *Bmal1*^*−/−*^*Apoe*^*−/−*^ and *Bmal1*^*+/+*^*Apoe*^*−/−*^ (**e**) and *Bmal1*^*fl/fl*^*Apoe*^*−/−*^ and *L-Bmal1*^*−/−*^*Apoe*^*−/−*^ (**f**) mice. Data in **a**,**b**,**e** and **f** were examined using the unpaired Student\'s *t*-test. Values are mean±s.d., *n*=6 per group. \**P*\<0.05, \*\**P*\<0.01 and \*\*\**P*\<0.001. Error bars represent s.d.](ncomms13011-f7){#f7}
{#f8}
{#f9}
{#f10}
| {
"pile_set_name": "PubMed Central"
} |
Related literature {#sec1}
==================
For the isostructural chloro analog, see: Sattarzadeh *et al.* (2009[@bb3]).
Experimental {#sec2}
============
{#sec2.1}
### Crystal data {#sec2.1.1}
(C~10~H~10~NO)\[ZnBr~2~(C~10~H~8~NO)\]·CH~4~O*M* *~r~* = 575.60Monoclinic,*a* = 9.9704 (8) Å*b* = 13.9954 (11) Å*c* = 15.8815 (12) Åβ = 105.815 (1)°*V* = 2132.2 (3) Å^3^*Z* = 4Mo *K*α radiationμ = 4.93 mm^−1^*T* = 100 K0.35 × 0.30 × 0.25 mm
### Data collection {#sec2.1.2}
Bruker SMART APEX diffractometerAbsorption correction: multi-scan (*SADABS*; Sheldrick, 1996[@bb4]) *T* ~min~ = 0.278, *T* ~max~ = 0.37219908 measured reflections4889 independent reflections4044 reflections with *I* \> 2σ(*I*)*R* ~int~ = 0.040
### Refinement {#sec2.1.3}
*R*\[*F* ^2^ \> 2σ(*F* ^2^)\] = 0.025*wR*(*F* ^2^) = 0.057*S* = 1.034889 reflections267 parametersH-atom parameters constrainedΔρ~max~ = 0.49 e Å^−3^Δρ~min~ = −0.37 e Å^−3^
{#d5e453}
Data collection: *APEX2* (Bruker, 2009[@bb2]); cell refinement: *SAINT* (Bruker, 2009[@bb2]); data reduction: *SAINT*; program(s) used to solve structure: *SHELXS97* (Sheldrick, 2008[@bb5]); program(s) used to refine structure: *SHELXL97* (Sheldrick, 2008[@bb5]); molecular graphics: *X-SEED* (Barbour, 2001[@bb1]); software used to prepare material for publication: *publCIF* (Westrip, 2010[@bb6]).
Supplementary Material
======================
Crystal structure: contains datablocks global, I. DOI: [10.1107/S1600536810036706/bt5354sup1.cif](http://dx.doi.org/10.1107/S1600536810036706/bt5354sup1.cif)
Structure factors: contains datablocks I. DOI: [10.1107/S1600536810036706/bt5354Isup2.hkl](http://dx.doi.org/10.1107/S1600536810036706/bt5354Isup2.hkl)
Additional supplementary materials: [crystallographic information](http://scripts.iucr.org/cgi-bin/sendsupfiles?bt5354&file=bt5354sup0.html&mime=text/html); [3D view](http://scripts.iucr.org/cgi-bin/sendcif?bt5354sup1&Qmime=cif); [checkCIF report](http://scripts.iucr.org/cgi-bin/paper?bt5354&checkcif=yes)
Supplementary data and figures for this paper are available from the IUCr electronic archives (Reference: [BT5354](http://scripts.iucr.org/cgi-bin/sendsup?bt5354)).
We thank Shahid Beheshti University and the University of Malaya for supporting this study.
Comment
=======
An earlier study reported C~10~H~10~NO^+.^ZnCI~2~(C~10~H~8~NO)^.^CH~3~OH, which feature cations, tetrahedral anions and solvent molecules linked by N···O, O···O and O···Cl hydrogen bonds into a linear chain (Sattarzadeh *et al.*, 2009). The present bromide analog (Scheme I, Fig. 1) is isostructural, the two compounds displaying matching cell dimensions. The hydroxy unit of the cation is hydrogen-bond donor to the alkoxide O atom of the tetrahedrally coordinated anion whereas the ammonium unit is hydrogen-bond donor to the methanolic O atom. Adjacent ion-pairs and solvent molecules are linked by a *O*--H~methanol~···Br hydrogen bond to generate a linear chain running along the *a*-axis of the monoclinic unit cell.
Experimental {#experimental}
============
Zinc bromide (0.10 g, 0.75 mmol) and 2-methyl-8-hydroxyquinoline (0.24 g, 1.5 mmol) were loaded into a convection tube; the tube was filled with dry methanol and kept at 333 K. Crystals were collected from the side arm after several days.
Refinement {#refinement}
==========
Hydrogen atoms were placed in calculated positions (C--H 0.95--0.98, N--H 0.86 and O--H 0.84 Å) and were included in the refinement in the riding model approximation, with *U*(H) set to 1.2--1.5*U*~eq~(C,*N*,*O*).
Figures
=======
{#Fap1}
Crystal data {#tablewrapcrystaldatalong}
============
----------------------------------------------- ---------------------------------------
(C~10~H~10~NO)\[ZnBr~2~(C~10~H~8~NO)\]·CH~4~O *F*(000) = 1144
*M~r~* = 575.60 *D*~x~ = 1.793 Mg m^−3^
Monoclinic, *P*2~1~/*n* Mo *K*α radiation, λ = 0.71073 Å
Hall symbol: -P 2yn Cell parameters from 6099 reflections
*a* = 9.9704 (8) Å θ = 2.6--28.2°
*b* = 13.9954 (11) Å µ = 4.93 mm^−1^
*c* = 15.8815 (12) Å *T* = 100 K
β = 105.815 (1)° Prism, yellow
*V* = 2132.2 (3) Å^3^ 0.35 × 0.30 × 0.25 mm
*Z* = 4
----------------------------------------------- ---------------------------------------
Data collection {#tablewrapdatacollectionlong}
===============
--------------------------------------------------------------- --------------------------------------
Bruker SMART APEX diffractometer 4889 independent reflections
Radiation source: fine-focus sealed tube 4044 reflections with *I* \> 2σ(*I*)
graphite *R*~int~ = 0.040
ω scans θ~max~ = 27.5°, θ~min~ = 2.0°
Absorption correction: multi-scan (*SADABS*; Sheldrick, 1996) *h* = −12→12
*T*~min~ = 0.278, *T*~max~ = 0.372 *k* = −18→18
19908 measured reflections *l* = −19→20
--------------------------------------------------------------- --------------------------------------
Refinement {#tablewraprefinementdatalong}
==========
------------------------------------- -------------------------------------------------------------------------------------------------
Refinement on *F*^2^ Primary atom site location: structure-invariant direct methods
Least-squares matrix: full Secondary atom site location: difference Fourier map
*R*\[*F*^2^ \> 2σ(*F*^2^)\] = 0.025 Hydrogen site location: inferred from neighbouring sites
*wR*(*F*^2^) = 0.057 H-atom parameters constrained
*S* = 1.03 *w* = 1/\[σ^2^(*F*~o~^2^) + (0.0224*P*)^2^ + 0.8543*P*\] where *P* = (*F*~o~^2^ + 2*F*~c~^2^)/3
4889 reflections (Δ/σ)~max~ = 0.001
267 parameters Δρ~max~ = 0.49 e Å^−3^
0 restraints Δρ~min~ = −0.37 e Å^−3^
------------------------------------- -------------------------------------------------------------------------------------------------
Fractional atomic coordinates and isotropic or equivalent isotropic displacement parameters (Å^2^) {#tablewrapcoords}
==================================================================================================
------ -------------- --------------- --------------- -------------------- --
*x* *y* *z* *U*~iso~\*/*U*~eq~
Br1 1.00294 (2) 0.228905 (17) 0.184698 (15) 0.01708 (6)
Br2 1.12097 (3) 0.497149 (17) 0.227454 (16) 0.02018 (7)
Zn1 0.99449 (3) 0.369708 (19) 0.266658 (17) 0.01403 (7)
O1 0.79813 (16) 0.40171 (12) 0.25997 (10) 0.0166 (4)
O2 0.58392 (16) 0.36051 (12) 0.13590 (10) 0.0155 (3)
H2 0.6590 0.3718 0.1741 0.023\*
O3 0.30839 (18) 0.27619 (14) 0.15184 (11) 0.0255 (4)
H3 0.2240 0.2660 0.1482 0.038\*
N1 1.01652 (19) 0.36488 (13) 0.39831 (12) 0.0120 (4)
N2 0.35285 (19) 0.35026 (13) 0.00209 (12) 0.0129 (4)
H2N 0.3540 0.3348 0.0547 0.015\*
C1 0.8930 (2) 0.38963 (15) 0.41453 (15) 0.0124 (5)
C2 0.7784 (2) 0.40932 (16) 0.33955 (15) 0.0141 (5)
C3 0.6537 (2) 0.43641 (17) 0.35416 (16) 0.0168 (5)
H3a 0.5762 0.4505 0.3058 0.020\*
C4 0.6399 (3) 0.44345 (17) 0.43988 (16) 0.0181 (5)
H4 0.5525 0.4619 0.4480 0.022\*
C5 0.7483 (3) 0.42455 (17) 0.51175 (16) 0.0185 (5)
H5 0.7362 0.4295 0.5689 0.022\*
C6 0.8786 (2) 0.39750 (16) 0.50003 (15) 0.0143 (5)
C7 0.9995 (3) 0.37848 (16) 0.56932 (15) 0.0169 (5)
H7 0.9952 0.3826 0.6283 0.020\*
C8 1.1213 (3) 0.35443 (16) 0.55191 (15) 0.0157 (5)
H8 1.2016 0.3415 0.5987 0.019\*
C9 1.1288 (2) 0.34858 (15) 0.46464 (15) 0.0140 (5)
C10 1.2609 (2) 0.32358 (17) 0.44237 (16) 0.0177 (5)
H10A 1.2913 0.3781 0.4135 0.027\*
H10B 1.2447 0.2683 0.4030 0.027\*
H10C 1.3332 0.3079 0.4961 0.027\*
C11 0.4758 (2) 0.37589 (15) −0.01425 (15) 0.0124 (5)
C12 0.5986 (2) 0.38162 (15) 0.05651 (15) 0.0127 (5)
C13 0.7208 (2) 0.40884 (16) 0.03862 (15) 0.0144 (5)
H13 0.8044 0.4129 0.0847 0.017\*
C14 0.7223 (2) 0.43073 (16) −0.04781 (16) 0.0157 (5)
H14 0.8076 0.4494 −0.0588 0.019\*
C15 0.6052 (2) 0.42585 (16) −0.11591 (15) 0.0157 (5)
H15 0.6090 0.4413 −0.1735 0.019\*
C16 0.4782 (2) 0.39768 (15) −0.10042 (15) 0.0133 (5)
C17 0.3494 (2) 0.39320 (16) −0.16628 (16) 0.0164 (5)
H17 0.3470 0.4067 −0.2253 0.020\*
C18 0.2294 (3) 0.36989 (16) −0.14609 (16) 0.0178 (5)
H18 0.1438 0.3684 −0.1908 0.021\*
C19 0.2312 (2) 0.34802 (16) −0.05948 (15) 0.0152 (5)
C20 0.1026 (2) 0.32331 (18) −0.03312 (17) 0.0212 (5)
H20A 0.1008 0.3598 0.0192 0.032\*
H20B 0.1025 0.2548 −0.0204 0.032\*
H20C 0.0202 0.3392 −0.0809 0.032\*
C21 0.3867 (3) 0.27543 (19) 0.24135 (16) 0.0218 (5)
H21A 0.4864 0.2742 0.2452 0.033\*
H21B 0.3623 0.2186 0.2701 0.033\*
H21C 0.3652 0.3329 0.2704 0.033\*
------ -------------- --------------- --------------- -------------------- --
Atomic displacement parameters (Å^2^) {#tablewrapadps}
=====================================
----- -------------- -------------- -------------- --------------- -------------- --------------
*U*^11^ *U*^22^ *U*^33^ *U*^12^ *U*^13^ *U*^23^
Br1 0.01686 (12) 0.02067 (12) 0.01393 (12) −0.00406 (9) 0.00458 (9) −0.00180 (9)
Br2 0.02141 (13) 0.02021 (12) 0.02170 (14) −0.00056 (10) 0.01060 (11) 0.00474 (10)
Zn1 0.01209 (14) 0.02082 (14) 0.00922 (14) −0.00032 (11) 0.00298 (10) 0.00126 (11)
O1 0.0102 (8) 0.0304 (9) 0.0080 (8) 0.0014 (7) 0.0003 (6) 0.0007 (7)
O2 0.0124 (8) 0.0244 (9) 0.0078 (8) −0.0009 (7) −0.0005 (6) 0.0012 (7)
O3 0.0159 (9) 0.0454 (12) 0.0159 (9) −0.0032 (9) 0.0056 (7) 0.0036 (8)
N1 0.0126 (9) 0.0126 (9) 0.0103 (9) −0.0021 (8) 0.0024 (8) 0.0006 (7)
N2 0.0143 (10) 0.0131 (9) 0.0103 (10) 0.0007 (8) 0.0019 (8) 0.0002 (7)
C1 0.0154 (12) 0.0108 (10) 0.0119 (11) −0.0042 (9) 0.0051 (9) −0.0005 (8)
C2 0.0157 (12) 0.0144 (11) 0.0119 (12) −0.0035 (9) 0.0032 (9) 0.0000 (9)
C3 0.0130 (11) 0.0203 (12) 0.0154 (12) −0.0033 (9) 0.0011 (10) 0.0006 (9)
C4 0.0150 (12) 0.0197 (12) 0.0217 (13) −0.0031 (10) 0.0084 (10) −0.0042 (10)
C5 0.0270 (14) 0.0165 (12) 0.0152 (12) −0.0053 (10) 0.0113 (11) −0.0031 (10)
C6 0.0189 (12) 0.0124 (11) 0.0116 (11) −0.0042 (9) 0.0043 (10) 0.0006 (9)
C7 0.0262 (13) 0.0152 (12) 0.0090 (12) −0.0049 (10) 0.0045 (10) −0.0006 (9)
C8 0.0197 (12) 0.0131 (11) 0.0120 (12) −0.0016 (9) 0.0003 (10) 0.0007 (9)
C9 0.0177 (12) 0.0097 (11) 0.0144 (12) −0.0032 (9) 0.0039 (10) 0.0010 (9)
C10 0.0158 (12) 0.0202 (12) 0.0153 (13) 0.0008 (10) 0.0013 (10) 0.0005 (10)
C11 0.0160 (12) 0.0095 (10) 0.0118 (11) 0.0033 (9) 0.0041 (9) −0.0007 (8)
C12 0.0153 (11) 0.0120 (11) 0.0100 (11) 0.0022 (9) 0.0018 (9) −0.0016 (8)
C13 0.0114 (11) 0.0165 (11) 0.0140 (12) 0.0022 (9) 0.0012 (9) −0.0018 (9)
C14 0.0142 (12) 0.0155 (11) 0.0199 (13) 0.0023 (9) 0.0090 (10) −0.0004 (9)
C15 0.0212 (13) 0.0161 (11) 0.0112 (12) 0.0023 (10) 0.0065 (10) −0.0001 (9)
C16 0.0168 (12) 0.0108 (10) 0.0115 (11) 0.0049 (9) 0.0021 (9) −0.0008 (9)
C17 0.0217 (13) 0.0139 (11) 0.0134 (12) 0.0052 (9) 0.0042 (10) −0.0012 (9)
C18 0.0175 (12) 0.0186 (12) 0.0144 (12) 0.0021 (10) −0.0009 (10) −0.0022 (9)
C19 0.0152 (12) 0.0114 (11) 0.0175 (13) 0.0007 (9) 0.0017 (10) −0.0005 (9)
C20 0.0145 (12) 0.0254 (13) 0.0220 (14) −0.0006 (10) 0.0023 (11) 0.0035 (11)
C21 0.0207 (13) 0.0275 (14) 0.0160 (13) −0.0011 (11) 0.0032 (10) 0.0022 (10)
----- -------------- -------------- -------------- --------------- -------------- --------------
Geometric parameters (Å, °) {#tablewrapgeomlong}
===========================
--------------------- -------------- ----------------------- --------------
Br1---Zn1 2.3758 (4) C8---H8 0.9500
Br2---Zn1 2.3637 (4) C9---C10 1.496 (3)
Zn1---O1 1.9828 (16) C10---H10A 0.9800
Zn1---N1 2.0431 (19) C10---H10B 0.9800
O1---C2 1.335 (3) C10---H10C 0.9800
O2---C12 1.341 (3) C11---C16 1.409 (3)
O2---H2 0.8400 C11---C12 1.421 (3)
O3---C21 1.423 (3) C12---C13 1.378 (3)
O3---H3 0.8400 C13---C14 1.411 (3)
N1---C9 1.331 (3) C13---H13 0.9500
N1---C1 1.370 (3) C14---C15 1.359 (3)
N2---C19 1.335 (3) C14---H14 0.9500
N2---C11 1.368 (3) C15---C16 1.411 (3)
N2---H2N 0.8600 C15---H15 0.9500
C1---C6 1.408 (3) C16---C17 1.419 (3)
C1---C2 1.434 (3) C17---C18 1.360 (3)
C2---C3 1.379 (3) C17---H17 0.9500
C3---C4 1.409 (3) C18---C19 1.404 (3)
C3---H3a 0.9500 C18---H18 0.9500
C4---C5 1.367 (3) C19---C20 1.494 (3)
C4---H4 0.9500 C20---H20A 0.9800
C5---C6 1.414 (3) C20---H20B 0.9800
C5---H5 0.9500 C20---H20C 0.9800
C6---C7 1.418 (3) C21---H21A 0.9800
C7---C8 1.359 (3) C21---H21B 0.9800
C7---H7 0.9500 C21---H21C 0.9800
C8---C9 1.411 (3)
O1---Zn1---N1 83.77 (7) H10A---C10---H10B 109.5
O1---Zn1---Br2 113.91 (5) C9---C10---H10C 109.5
N1---Zn1---Br2 112.22 (5) H10A---C10---H10C 109.5
O1---Zn1---Br1 109.87 (5) H10B---C10---H10C 109.5
N1---Zn1---Br1 121.55 (5) N2---C11---C16 119.6 (2)
Br2---Zn1---Br1 112.362 (14) N2---C11---C12 119.2 (2)
C2---O1---Zn1 111.41 (13) C16---C11---C12 121.1 (2)
C12---O2---H2 109.5 O2---C12---C13 125.7 (2)
C21---O3---H3 109.5 O2---C12---C11 116.1 (2)
C9---N1---C1 119.98 (19) C13---C12---C11 118.2 (2)
C9---N1---Zn1 130.41 (16) C12---C13---C14 120.4 (2)
C1---N1---Zn1 109.43 (14) C12---C13---H13 119.8
C19---N2---C11 123.3 (2) C14---C13---H13 119.8
C19---N2---H2N 118.3 C15---C14---C13 121.9 (2)
C11---N2---H2N 118.3 C15---C14---H14 119.0
N1---C1---C6 122.3 (2) C13---C14---H14 119.0
N1---C1---C2 116.5 (2) C14---C15---C16 119.5 (2)
C6---C1---C2 121.2 (2) C14---C15---H15 120.3
O1---C2---C3 123.6 (2) C16---C15---H15 120.3
O1---C2---C1 118.8 (2) C11---C16---C15 118.9 (2)
C3---C2---C1 117.6 (2) C11---C16---C17 117.1 (2)
C2---C3---C4 120.8 (2) C15---C16---C17 124.0 (2)
C2---C3---H3a 119.6 C18---C17---C16 120.9 (2)
C4---C3---H3a 119.6 C18---C17---H17 119.5
C5---C4---C3 122.0 (2) C16---C17---H17 119.5
C5---C4---H4 119.0 C17---C18---C19 120.4 (2)
C3---C4---H4 119.0 C17---C18---H18 119.8
C4---C5---C6 119.2 (2) C19---C18---H18 119.8
C4---C5---H5 120.4 N2---C19---C18 118.6 (2)
C6---C5---H5 120.4 N2---C19---C20 118.6 (2)
C1---C6---C5 119.1 (2) C18---C19---C20 122.8 (2)
C1---C6---C7 116.5 (2) C19---C20---H20A 109.5
C5---C6---C7 124.4 (2) C19---C20---H20B 109.5
C8---C7---C6 120.4 (2) H20A---C20---H20B 109.5
C8---C7---H7 119.8 C19---C20---H20C 109.5
C6---C7---H7 119.8 H20A---C20---H20C 109.5
C7---C8---C9 120.3 (2) H20B---C20---H20C 109.5
C7---C8---H8 119.9 O3---C21---H21A 109.5
C9---C8---H8 119.9 O3---C21---H21B 109.5
N1---C9---C8 120.6 (2) H21A---C21---H21B 109.5
N1---C9---C10 117.2 (2) O3---C21---H21C 109.5
C8---C9---C10 122.1 (2) H21A---C21---H21C 109.5
C9---C10---H10A 109.5 H21B---C21---H21C 109.5
C9---C10---H10B 109.5
N1---Zn1---O1---C2 3.23 (14) C6---C7---C8---C9 0.3 (3)
Br2---Zn1---O1---C2 −108.35 (14) C1---N1---C9---C8 1.4 (3)
Br1---Zn1---O1---C2 124.58 (13) Zn1---N1---C9---C8 176.10 (15)
O1---Zn1---N1---C9 −178.0 (2) C1---N1---C9---C10 −178.93 (19)
Br2---Zn1---N1---C9 −64.7 (2) Zn1---N1---C9---C10 −4.3 (3)
Br1---Zn1---N1---C9 72.5 (2) C7---C8---C9---N1 −1.1 (3)
O1---Zn1---N1---C1 −2.90 (14) C7---C8---C9---C10 179.2 (2)
Br2---Zn1---N1---C1 110.42 (13) C19---N2---C11---C16 −2.7 (3)
Br1---Zn1---N1---C1 −112.43 (13) C19---N2---C11---C12 176.3 (2)
C9---N1---C1---C6 −0.9 (3) N2---C11---C12---O2 0.2 (3)
Zn1---N1---C1---C6 −176.60 (17) C16---C11---C12---O2 179.15 (19)
C9---N1---C1---C2 177.82 (19) N2---C11---C12---C13 −179.13 (19)
Zn1---N1---C1---C2 2.1 (2) C16---C11---C12---C13 −0.2 (3)
Zn1---O1---C2---C3 176.09 (18) O2---C12---C13---C14 −179.0 (2)
Zn1---O1---C2---C1 −3.1 (2) C11---C12---C13---C14 0.3 (3)
N1---C1---C2---O1 0.6 (3) C12---C13---C14---C15 −0.1 (3)
C6---C1---C2---O1 179.3 (2) C13---C14---C15---C16 −0.3 (3)
N1---C1---C2---C3 −178.6 (2) N2---C11---C16---C15 178.7 (2)
C6---C1---C2---C3 0.1 (3) C12---C11---C16---C15 −0.2 (3)
O1---C2---C3---C4 −179.7 (2) N2---C11---C16---C17 1.1 (3)
C1---C2---C3---C4 −0.6 (3) C12---C11---C16---C17 −177.8 (2)
C2---C3---C4---C5 0.4 (4) C14---C15---C16---C11 0.4 (3)
C3---C4---C5---C6 0.3 (4) C14---C15---C16---C17 177.9 (2)
N1---C1---C6---C5 179.2 (2) C11---C16---C17---C18 0.8 (3)
C2---C1---C6---C5 0.6 (3) C15---C16---C17---C18 −176.8 (2)
N1---C1---C6---C7 0.1 (3) C16---C17---C18---C19 −1.2 (3)
C2---C1---C6---C7 −178.6 (2) C11---N2---C19---C18 2.2 (3)
C4---C5---C6---C1 −0.7 (3) C11---N2---C19---C20 −177.2 (2)
C4---C5---C6---C7 178.4 (2) C17---C18---C19---N2 −0.3 (3)
C1---C6---C7---C8 0.2 (3) C17---C18---C19---C20 179.1 (2)
C5---C6---C7---C8 −178.9 (2)
--------------------- -------------- ----------------------- --------------
Hydrogen-bond geometry (Å, °) {#tablewraphbondslong}
=============================
------------------ --------- --------- ------------- ---------------
*D*---H···*A* *D*---H H···*A* *D*···*A* *D*---H···*A*
O2---H2···O1 0.84 1.71 2.546 (2) 172
O3---H3···Br1^i^ 0.84 2.48 3.2941 (17) 163
N2---H2n···O3 0.86 1.91 2.739 (3) 162
------------------ --------- --------- ------------- ---------------
Symmetry codes: (i) *x*−1, *y*, *z*.
###### Hydrogen-bond geometry (Å, °)
*D*---H⋯*A* *D*---H H⋯*A* *D*⋯*A* *D*---H⋯*A*
---------------- --------- ------- ------------- -------------
O2---H2⋯O1 0.84 1.71 2.546 (2) 172
O3---H3⋯Br1^i^ 0.84 2.48 3.2941 (17) 163
N2---H2n⋯O3 0.86 1.91 2.739 (3) 162
Symmetry code: (i) .
| {
"pile_set_name": "PubMed Central"
} |
###### Strengths and limitations of this study
- This is the first study to assess the frequency and types of conflict of interest (COI) disclosed by authors of primary studies of health policy and systems research.
- The study used a rigorous methodology that included a search strategy specific to health policy and services journals and duplicate study selection and data abstraction processes.
- We used a comprehensive framework for the classification of COI.
- The study focused on reported COI, thus these statements depend on journals' COI policy requirements, and whether authors' disclosures are accurate or complete remains uncertain.
Background {#s1}
==========
Evidence-informed health policymaking aims to ensure that policymaking is well informed by the best available evidence.[@R1] Evidence from health policy and systems research (HPSR) can inform health system policy decisions including who delivers health services and where, and how these services are financed and organised.[@R2] Furthermore. policymakers are increasingly recognising the importance of the use of research evidence in improving health, reducing health inequities and contributing to economic development.[@R4] However, conflict of interest (COI) of researchers may influence the conduct and reporting of HPSR.
COI is defined as 'a financial or intellectual relationship that may impact an individual's ability to approach a scientific question with an open mind'.[@R6] For instance, one study assessing the frequency and influence of financial COI on economic analyses in oncology found that the studies disclosing financial COI directly or indirectly consistently supported the sponsor's product.[@R7] Additionally, Forsyth *et al* found that opinion articles sceptical of the use of systematic reviews for policymaking were more likely to have industry ties than articles supportive of their use.[@R8]
Reporting of COI in HPSR is important given its potential influence on public policy and decision-making. We previously assessed the reporting of COI in HPSR systematic reviews.[@R9] We found that 20% of those reviews did not include a COI disclosure statement, and only 15% of disclosure statements reported the existence of any COI. Furthermore, the reporting of COI in primary studies is important for both policymakers, relying on their findings for decision-making, as well as for authors of systematic reviews assessing the potential bias associated with the COI of study investigators.[@R10] Therefore, this study aims to assess the types and frequency of COI disclosed by authors of primary studies of HPSR.
Methods {#s2}
=======
Design overview and definitions {#s2-1}
-------------------------------
We conducted a cross-sectional survey using standard systematic review methodology for study selection and data extraction. We defined COI disclosure as the reporting of whether a COI exists or not. We classified the types of disclosed COIs as shown in [figure 1](#F1){ref-type="fig"} and detailed in [online supplementary appendix S1](#SP1){ref-type="supplementary-material"}. Our classification of COIs relies on a framework informed by a literature review, the findings of recent studies assessing COIs reported by authors of clinical systematic reviews, HPSR systematic reviews and randomised controlled trials[@R9] and the International Committee of Medical Journal Editors (ICMJE) COI disclosure form.[@R13] We used the word '*loogly*' *to label* '*any additional statement in the COI disclosure that attempts to downplay a disclosed relationship by suggesting that it is unrelated to COI*' (eg, 'this relationship did not influence the content of the manuscript').[@R11]
10.1136/bmjopen-2019-032425.supp1
{#F1}
Eligibility criteria {#s2-2}
--------------------
We included articles meeting the following eligibility criteria:
- Type of study: primary studies (eg, randomised controlled trials, cohort studies, qualitative studies). We excluded systematic and literature reviews, case studies, technical reports, conference reports, proceedings, editorials and opinion pieces; type of field: HPSR; we used the taxonomy of health systems topics used to code Health Systems Evidence database of McMaster Health Forum to assess eligibility: governance, financial, delivery arrangements and implementation strategies.[@R14] Governance arrangements cover five topics: policy authority, organisational authority, commercial authority, professional authority, and consumer and stakeholder involvement. Financial arrangements include topics on financing systems, funding organisations, remuneration providers, purchasing products and services, and incentivising consumers. Delivery arrangements cover topics related to how care is designed to meet consumers' needs, by whom care is provided, where care is provided and with what supports is care provided. Implementation strategies comprise topics on consumer-targeted strategy, provider-targeted strategy and organisation-targeted strategy.
- Articles published in English in 2016.
Search strategy {#s2-3}
---------------
We searched for papers published in peer-reviewed health policy and services journals. We ran the search in the Web of Science database limiting to 'Health Policy and Services' journal category, 'article' document type, English language and to the year 2016. [Online supplementary appendix S2](#SP2){ref-type="supplementary-material"} presents the detailed search strategy.
10.1136/bmjopen-2019-032425.supp2
Selection process {#s2-4}
-----------------
We drew a random sample of 200 papers from the set of citations retrieved by the search to undergo the selection process using an online random sequence generator ([www.random.org/sequences](www.random.org/sequences)). This sample of 200 primary studies is a subset of our previously published study on the reporting of funding in HPSR.[@R16]
Citations were exported to EndNote X7.5 software (Thomson Reuters, Philadelphia, PA, USA). Reviewers completed calibration exercises before starting the selection process. Two reviewers screened title and abstracts for eligibility in duplicate and independently using EndNote. We ensured that papers retrieved by our search were effectively on HPSR. We retrieved the full text of citations judged as potentially eligible by at least one of the two reviewers. The two reviewers screened the full texts in duplicate and independently. The reviewers resolved their disagreements by discussion, and consulted a third reviewer when consensus could not be reached. We used a standardised and pilot-tested full-text screening form. We recorded reasons for exclusion and summarised the selection process results in a Preferred Reporting Items for Systematic Reviews and Meta-Analyses study flow diagram.[@R17]
Data extraction process {#s2-5}
-----------------------
We developed and pilot-tested a standardised data extraction form with detailed instructions (see [online supplementary appendix S3](#SP3){ref-type="supplementary-material"}). Two teams of eight reviewers completed calibration exercises and extracted data in duplicate and independently. Reviewers extracted study data using Research Electronic Data Capture tool, a secure, web-based application designed to support data capture for research studies.[@R18] The reviewers compared results and resolved disagreements through discussion, or with the help of a third person when consensus could not be reached.
10.1136/bmjopen-2019-032425.supp3
Extracted data {#s2-6}
--------------
We extracted the following general characteristics of each article:
- Number of authors.
- Reported affiliation(s) of first and last authors (private or public academic institution, government, not-for-profit organisation, private for profit, intergovernmental).
- Country of affiliation of the first author and its classification (as per World Bank list of economies issued in September 2016).
- Health systems arrangement of the paper (governance, financial, delivery arrangements and implementation strategies).
We extracted the following characteristics of the reported COI disclosures (as defined above):
- Whether authors reported COI or not.
- Form in which COI disclosures were provided (a narrative statement, an online document, available on request).
- Number of authors per paper who report any type of COI.
- Number of authors per paper who report each specific type of COI, and when applicable, the different subtypes of COI.
- Whether the paper reports relevant characteristics of the COI (source, monetary value, duration).
- Whether individuals other than the authors provided COI disclosures (eg, editors, peer reviewers, external writers, others).
We extracted the following information on the characteristics of the journal:
- Impact factor.
- Existence of a COI disclosure policy.
Data analysis {#s2-7}
-------------
For eligible articles, we conducted descriptive analyses, focusing on the reported COI disclosures. For continuous variables, we present summary data as medians and quartiles since the application of the Kolmogorov-Smirnov test did not demonstrate normality. We presented the results for categorical variables as frequencies and percentages, and analysed them using the χ^2^ test or, if the expected event number proved less than 5, the Fisher's exact test. We considered a p value of \<0.05 as statistically significant. We performed the analysis using SPSS V.21.0 for Windows (SPSS).
Results {#s3}
=======
Out of the 2648 citations identified, we included 200 eligible primary studies that were published in 55 'Health Policy & Services' journals. [Figure 2](#F2){ref-type="fig"} shows the study flow diagram.
{#F2}
General characteristics of the included primary studies {#s3-1}
-------------------------------------------------------
[Table 1](#T1){ref-type="table"} presents the general characteristics of the included primary studies. The median number of authors per study was 4. The majority of studies were conducted by authors affiliated with institutions located in high-income countries (92%) where most articles were conducted in the USA (54%) followed by UK (8%). Most articles addressed the topic of delivery arrangements (72%). Most first authors and last authors were affiliated with public academic institutions (68% and 65%, respectively).
######
General characteristics of the included primary studies (n=200)
Overall
------------------------------------------------------------------------------------------ ----------
Number of authors; median (IQR) 4 (3--6)
Classification of the country of the institution to which the first author is affiliated
High income 183 (92)
USA 107 (54)
UK 16 (8)
Australia 13 (7)
Canada 9 (5)
The Netherlands 7 (4)
Other high-income countries 31 (16)
Upper middle income 10 (5)
China 3 (2)
South Africa 3 (2)
Other upper middle-income countries 4 (2)
Lower middle income 4 (2)
Kenya 1 (0.5)
Philippines 1 (0.5)
Bangladesh 1 (0.5)
Vietnam 1 (0.5)
Low income 3 (2)
Uganda 3 (2)
Affiliation of first author\*
Public academic institution 135 (68)
Private academic institution 46 (23)
Government 18 (9)
Not-for-profit organisation 23 (12)
Private for profit 2 (1)
Intergovernmental 1 (1)
Affiliation of last author\*
Public academic institution 129 (65)
Private academic institution 51 (26)
Government 21 (11)
Not-for-profit organisation 20 (10)
Private for profit 3 (2)
Intergovernmental 0 (0)
Type of health systems arrangement\*
Delivery arrangement 143 (72)
Implementation strategies 25 (13)
Governance arrangement 23 (12)
Financial arrangement 67 (34)
\*Studies may have more than one option that applies.
Characteristics of the reported COI disclosures {#s3-2}
-----------------------------------------------
Of the 200 primary studies, 66% (132/200) included COI disclosure statements of authors. All but one study provided COI disclosures narratively in the main document; the single study provided them in an online form that was not accessible. None of the included studies reported COI by individuals other than the authors (eg, editors or peer reviewers).
[Table 2](#T2){ref-type="table"} presents the reporting of the different types of COI in the 132 studies that included COI disclosure statements. Of these 132 studies that included COI disclosure statements, 19 (14%) had at least one author reporting at least one type of COI while 113 (86%) studies had their authors reporting that they had no COI. The most frequently reported type was individual financial COI (n=15, 11%), with the median percentage of authors reporting this type of COI being 25%. None of the authors reported individual intellectual COIs or personal COIs. Of the 132 primary studies that provided COI disclosure statements, more had at least one author reporting financial COIs compared with non-financial COIs (n=16, 12% vs n=3, 2%; p=0.04). More studies had at least one author reporting individual COIs compared with institutional COIs (n=15, 11% vs n=5, 4%; p=0.01).
######
Reporting by primary study authors of the different types of conflict of interest (COI) (n=132)
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Studies with at least one author reporting a specific type of COI\*\ Distribution of the percentage of authors per study reporting that type of COI†\
n (%) Median (IQR)
------------------------------------------------------------ ---------------------------------------------------------------------- ----------------------------------------------------------------------------------
At least one type 19 (14) 25 (17--50)
Individual financial (direct benefit) 15 (11) 25 (15--50)
Individual financial (benefit through professional status) 0 (0) N/A
Individual intellectual 0 (0) N/A
Individual personal 0 (0) N/A
Institutional financial 2 (2) ‡
Institutional intellectual 3 (2) §
Institutional cultural 0 (0) N/A
'Other types'¶ 4 (3) 30 (18--85)
Provided a 'loogly statement' 3 (2) \*\*
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
\*One study can have authors reporting more than one type of COI.
†Calculated using the number of papers with at least one author reporting the specific type of COI (ie, papers counted in the preceding column) as the denominator.
‡Authors of only two studies reported institutional financial COI, with the percentages being 20% and 100%.
§Authors of only three studies reported institutional intellectual COI, with the percentages being 20%, 25% and 33%.
¶'Other types' of COIs included: 'implementing national clinical audit' (n=1), 'non-compensated affiliations' (n=1), 'attended meetings' (n=1) and relationship with a publishing entity (n=1). We consider these as individual and non-financial types of COI.
\*\*Authors of only three studies provided a 'loogly statement', with the percentages being 10%, 25% and 100%.
N/A, not applicable.
### Individual financial COI {#s3-2-1}
[Table 3](#T3){ref-type="table"} presents the reporting of the different subtypes of individual financial COI in the 15 primary studies with at least one author reporting individual financial COI. The two most frequently reported subtypes were 'personal fees' (n=9, 60%) and 'grant' (n=6, 40%). The median percentages of authors reporting these two subtypes were 20% and 18%, respectively.
######
Reporting of primary study authors of different subtypes of individual financial conflict of interest (COI) (n=15)
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Studies with at least one author reporting the subtype of individual financial COI\*\ Distributions of the percentage of authors per study reporting that subtype of COI†\
n (%) Median (IQR)
------------------------------------------------ --------------------------------------------------------------------------------------- --------------------------------------------------------------------------------------
Grant 6 (40) 18 (9--27)
Employment 2 (13) ‡
Personal fees (other than employment) 9 (60) 20 (12--38)
Non-monetary support 1 (7) §
Study supplies/services 0 (0) N/A
Patent(s) 0 (0) N/A
Stocks, bonds, stock options, other securities 3 (20) ¶
'Other subtypes' 0 (0) N/A
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
\*One study can have authors reporting more than one type of COI.
†Calculated using the number of papers with at least one author reporting the specific type of COI (ie, papers counted in the preceding column) as the denominator.
‡Authors of only two studies reported 'Employment', with the percentages being 50% and 100%.
§Authors of only one study reported 'Non-monetary support', with the percentage being 17%.
¶Authors of only three studies reported 'Stocks, bonds, stock options, other securities', with the percentages being 20%, 25% and 33%.
N/A, not applicable.
Of the 15 studies with at least one author reporting individual financial COI, 14 reported the source of financial COI. Only two of these 14 studies specified the relationship of the source to the field under study; in both cases, the sources produced a product not the subject of the study but under the same field. Only one of the 15 studies reported on the timing of the conflicted relationship relative to the conduct of the study; in that case, the relationship occurred during the conduct of the study. None of the studies reported on the monetary value of the financial COI.
Characteristics of the journals {#s3-3}
-------------------------------
The median impact factor of the 55 journals that published the included primary studies was 1.66 (IQR=1.36--2.41). Ninety-six per cent (53/55) of the journals had a COI disclosure policy requiring authors to report their COI. Of the 68 papers that did not include a COI statement, 90% (61/68) were published in journals that did have a COI disclosure policy. The percentage of papers that included a COI statement was 68.2% in journals with a COI disclosure policy and 12.5% in journals without a COI disclosure policy (p=0.012). We provided the list of the 55 journals that published the included primary studies in [online supplementary appendix S4](#SP4){ref-type="supplementary-material"}.
10.1136/bmjopen-2019-032425.supp4
Discussion {#s4}
==========
Summary of findings {#s4-1}
-------------------
In summary, 66% of 200 HPSR primary studies included COI disclosure statements of authors, with only one using an inaccessible online disclosure form. Of these studies, 14% had at least one author reporting at least one type of COI. Most frequently, the authors reported individual financial COI. Very few studies reported non-financial or institutional COIs. The two most frequently reported subtypes of individual financial COI were 'personal fees' and 'grant'. None of the studies reported on the monetary value of the financial COI, or provided disclosure by individuals other than the authors such as editors or reviewers.
Strengths and limitations {#s4-2}
-------------------------
This is the first study to assess the frequency and types of COI disclosed by authors of primary studies of HPSR. We have used a rigorous methodology that included a search strategy specific to health policy and services journals and duplicate study selection and data abstraction processes. We used a comprehensive framework for the classification of COI used in previous studies.[@R9] Our study focused on reported COI, thus these statements depend on journals' COI policy requirements, and whether authors' disclosures are accurate or complete remains uncertain.
Comparison to other studies {#s4-3}
---------------------------
Our findings, in relation to similar studies, demonstrate that COI disclosure statements are less frequently included in HPSR primary studies (66%) compared with HPSR systematic reviews (80%), clinical randomised controlled trials (94%) and clinical systematic reviews (97%) ([figure 3](#F3){ref-type="fig"}).[@R9] Factors that may be contributing to these differences include the less rigorous COI policies in HPSR journals compared with core clinical journals, and potentially a less strict implementation: 93% of HPSR journals (including the 55 journals that published the primary studies included in this study) have a COI disclosure policy compared with 99% for core clinical journals.[@R19]
{#F3}
The percentage of authors reporting any type of COI in HPSR primary studies (14%) was comparable to that of HPSR systematic reviews (15%). However, that percentage is much lower compared with that of clinical systematic reviews (41%) and clinical trials (57%).[@R9] '*Possible explanations for this low rate of disclosure could be either an actual low prevalence of COI in this field, or an underreporting by HPSR authors of their COIs*'. Indeed, an increasing number of studies are using resources such as the Open Payment database to verify the accuracy of the COI disclosures of health researchers.[@R21] They are consistently showing that researchers tend to under-report their COIs (up to 81% in one study[@R25]).
Reporting of financial COI was higher than non-financial COI in HPSR primary studies. This is consistent with the findings of previous studies that focused on COI reporting in HPSR systematic reviews, clinical systematic reviews and randomised controlled trials.[@R9] Although this might reflect how frequently these types of COI exist, it might also be that authors are less aware of the concept of non-financial COI, or of what exactly qualifies as a non-financial COI. Another explanation could be related to the extent of use of standard COI disclosure forms: we found that only one study used a standardised form to report COI, compared with 12% in clinical trials.[@R12]
Implications for practice and research {#s4-4}
--------------------------------------
As HPSR may be used to inform policy decisions, COI of HPSR authors may bias their research output and subsequently lead to misguided public policies and decisions.[@R26] For example, Bes-Rastrollo *et al* found that financial COI may bias findings of systematic reviews of the effects of sugar-sweetened beverages consumption on weight gain and obesity.[@R28] In turn, such biased conclusions might adversely influence policymaking related to regulation of sugar-sweetened beverages. Consequently, the appropriate disclosure and management of COIs are essential for the credibility and trust in HPSR, and hence might increase its uptake in policymaking. For that reason, HPSR journals strengthen their COI disclosure policies, and the implementation of existing policies. One approach to help authors better recognise and disclose their COIs would be to develop a standardised COI disclosure form similar to that of the ICMJE but more specific to HPSR. Journals publishing HPSR should also consider collecting and publishing the COIs of editors and peer reviewers. Future research should investigate the reasons behind the higher reporting of financial COI compared with non-financial COI in HPSR primary studies. Investigation of the accuracy and completeness of reporting of COI may also provide insight into the low rates of disclosed COI.
Supplementary Material
======================
###### Reviewer comments
###### Author\'s manuscript
**Contributors:** MBH, LBK, FEJ, GG and EAA conceived and designed the study. MBH coordinated the study throughout. EAA had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. MBH and LBK ran the search and study selection processes. MBH, LBK, MAG, AMK, ASR, SB, AA and FA extracted the data. MBH, LBK and EAA analysed and interpreted the data. MBH and LBK wrote the first draft of the manuscript with EAA. All authors critically revised the manuscript and approved the final version. The lead author (EAA) affirms that this manuscript is an honest, accurate and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned have been explained.
**Funding:** This work was supported by the American University of Beirut Faculty of Medicine's Medical Practice Plan (MPP) funds.
**Competing interests:** MBH, GG and EAA have competing interests related to their research in the area of conflicts of interest.
**Patient and public involvement:** Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
**Patient consent for publication:** Not required.
**Provenance and peer review:** Not commissioned; externally peer reviewed.
**Data availability statement:** All data relevant to the study are included in the article or uploaded as supplementary information.
| {
"pile_set_name": "PubMed Central"
} |
Introduction
============
Parkinson's disease (PD) is a progressive neurodegenerative disorder characterized primarily by motor symptoms of bradykinesia, tremor, rigidity, and gait/postural disturbance. However, growing evidence suggests it may commonly lead to many non-motor psychiatric symptoms as well since at the early stage of this disease, such as cognitive impairment (e.g., verbal, visuospatial, executive, and memory functions) and depression ([@B12]), and dementia is a frequently final complication of the late symptoms in PD patients ([@B2]; [@B21]). Moreover, mild cognitive impairment (MCI) has been identified as an important risk factor to dementia at the early PD diagnosis and strongly predicts the progression to PD dementia ([@B5]), since it occurs most frequently with aging and duration of PD ([@B31]). Thus, the notion of MCI in PD is valuable of understanding PD development and associated dementia. However, it has not been particularly well understood of the neural basis underlying the MCI in PD.
Over the past decade, resting-state functional magnetic resonance imaging (R-fMRI) has been widely developed as a non-invasive and effective technology to examine spontaneous brain activity in both normal functions and neuropsychiatric disease pathophysiology ([@B9]; [@B18]). Amplitude of low-frequency fluctuations (ALFF) is a widely used measurement for the local neural activity in low-frequency (0.01-0.1 Hz) fluctuations of the blood oxygen level dependent (BOLD) signal during rest ([@B50]). It has also been effectively used to investigate abnormal spatial patterns of spontaneous brain activity in neurodegenerative diseases, including PD ([@B26]; [@B13]; [@B45]; [@B39]; [@B41]), Alzheimer's disease ([@B42]; [@B32]) and MCI ([@B22]; [@B42]). On the one hand, studies in PD have shown widespread abnormalities of the ALFF of the whole brain spontaneous activity in a range of brain regions, such as the motor cortices, striatum, cerebellum, and brain stem, but also emphasize that these different findings depend on PD subtypes (i.e., tremor-dominant, posture instability gait difficulty, and non-motor symptoms) ([@B13]; [@B39]) and BOLD fluctuation frequency ([@B26]). On the other hand, studies in MCI have shown the whole brain ALFF values are decreased mainly in the default mode network and increased in several widespread areas including lateral temporal regions, superior frontal, occipital and parietal regions ([@B22]; [@B42]). However, underlying neural correlates of cognitive impairment in early PD is unclear.
Recently, several studies have reported a range of abnormal brain areas that are associated with cognitive decline in PD. For example, R-fMRI functional connectivity analysis has suggested patients of PD with MCI could be characterized by dysfunction in the default mode network and dorsal attention network ([@B8]; [@B27]). A longitudinal study using task-fMRI has reported reduction of BOLD signals that are related to working-memory brain activity in the right fusiform gyrus and right vermis in patients of PD with MCI ([@B16]). Furthermore, a quite similar work with the present work applied ALFF to measure abnormalities of spontaneous brain activity in patients of PD with MCI ([@B20]). But the patients used in their study were not early PD diagnosis, without drug-naïve and depression assessment (depression is also another non-motor symptom in PD), so the results could not get rid of the influence of drug-treatment and possible mood disorder.
In the present study, we employed MRI and clinical/neuropsychological data of baseline PD patients without drug-treatment from Parkinson's progression markers initiative (PPMI)^[1](#fn01){ref-type="fn"}^ to detect the abnormalities of brain activity that is widely associated with cognitive decline in PD. To address this issue, we used early diagnosed PD patients who do not suffer from depression and are untreated. Then, we used R-fMRI and ALFF measurement to investigate regional differences in spontaneous brain activity in the patients of PD with MCI (PD-MCI), PD with normal cognition (PD-NC) and healthy controls (HC), and also examined correlations of the ALFF values with scores of clinical and neuropsychological assessments.
Materials and Methods {#s1}
=====================
Participants
------------
All MRI and clinical data used in this study was downloaded from the PPMI^1^, which is a first large-scale multicenter project of PD progression biomarkers ([@B34]). According to the PPMI cohort inclusion criteria, the included baseline PD patients were required to: (i) have at least two criteria of bradykinesia, ridigity, and resting tremor or either asymmetric resting tremor or asymmetric bradykinesia; (ii) satisfy an early clinical disease stage \[diagnosis of PD for less than 2 years and Hoehn and Yahr (H&Y) stage I or II\]; (iii) be untreated for PD; (iv) have a dopamine transporter (DAT) deficit on imaging; (v) not have dementia as determined by the site investigators; and (vi) not suffer from depression with 15-item Geriatric Depression Scale (GDS-15) \< 5 ([@B35]). Healthy controls (HC) were required to have: (i) no significant neurologic dysfunction; (ii) no first-degree family member with PD; and (iii) a Montreal Cognitive Assessment (MoCA) score greater than 26. The patients of PD with mild cognitive impairment (PD-MCI) were required to meet diagnosis of MCI defined by PD-MCI level I of the Movement Disorder Society (MDS) Task Force. The patients of PD with normal cognition (PD-NC) were required to satisfy the criteria of PD (i--v) and a MoCA greater than 26. The PPMI overall study was approved by the Research Subjects Review Board at the University of Rochester, and each site was approved by the institutional review board, and participants provided written informed consent. Considering that all participants have resting-state fMRI data and the same data-centers have both patients of PD with and without cognitive impairment, we totally acquired 19 PD-NC, 10 PD-MCI, and 13 age- and gender-matched HC participants. Data was downloaded on September 22nd, 2015 for PD and on November 25th, 2016 for HC.
These participants were totally enrolled at four PPMI sites that used a standardized protocol for three Tesla machines (all Siemens Healthcare, United States). A 3D magnetization prepared rapid gradient echo (MPRAGE) sequence was used for imaging brain anatomy (176 axial slices, repetition time = 2300 ms, echo time = 2.98 ms, flip angle = 9°, voxel size 1 mm × 1 mm × 1 mm). A gradient-echo echo-planar imaging (GE-EPI) sequence was used for imaging brain functional activity over 210 volumes or time points during resting state (40 axial slices, repetition time = 2400 ms, echo time = 25 ms, flip angle = 80°, voxel size 3.3 mm × 3.3 mm × 3.3 mm).
Clinical and Neuropsychological Assessments
-------------------------------------------
For the clinical characteristics, the disease stage was scored using the H&Y stage score, and the disease severity was captured by the Movement Disorder Society-Unified Parkinson's Disease Rating Scale (MDS-UPDRS). Global cognition was assessed with the MoCA, because it is sensitive to the detection of executive deficits in PD ([@B49]; [@B38]) and consequently is a sensitive tool for the detection of PD dementia in the clinic than the traditional Mini-Mental State Examination (MMSE) ([@B25]; [@B10]). In addition, more details in cognitive domains were also assessed ([@B43]), including (i) memory \[Hopkins Verbal Learning Test-Revised (HLVT-R)\]; (ii) visuospatial function \[Benton Judgment of Line Orientation (BJLO)\] 15-item (split-half) version; (iii) processing speed-attention \[Symbol-Digit Modalities Test (SDMT)\]; and (iv) executive function and working memory \[Letter-Number Sequencing (LNS) and semantic fluency (SF, animals, vegetables and fruits)\]. Single scores were generated for each test; except for the HLVT-R, two scores were generated \[i.e., immediate free recall (IFR) and delayed recognition hits (DRH)\]. According to the diagnosis criteria of PD-MCI level I ([@B31]), cognitive impairment was defined by two ways: (i) a recommended MoCA cut-off for PD of \<26 was used ([@B15]); (ii) using the detailed cognitive battery, cognitive impairment was defined as at least two test scores \>1.5 standard deviations below the standardized mean score to support a PD-MCI diagnosis ([@B31]). Clinical characteristics and neuropsychological assessments were presented in **Table [1](#T1){ref-type="table"}**.
######
Demographics of PD-NC, PD-MCI and HC groups.
PD-MCI (*n* = 10) PD-NC (*n* = 19) HC (*n* = 13) *p*
--------------------------- ------------------------- -------------------------- ----------------------- -------------
Gender (M:F) 5:5 14:5 11:2 0.182^a^
Age (years) 64.7 ± 7.0 (55.4--72.8) 59.1 ± 12.3 (38.2--77.2) 62.9 ± 9.0 (44--78.8) 0.332^a^
Disease duration (months) 5.4 ± 7.9 (1--25) 9.5 ± 10.8 (1--32) -- 0.304^b^
UPDRS total score 29.3 ± 9.6 (12--42) 25.7 ± 11.8 (10--53) -- 0.411^b^
H&Y stage 1.5 ± 0.5 (1--2) 1.3 ± 0.5 (1--2) -- 0.349^b^
MoCA 24.6 ± 2.4 (21--29) 28.4 ± 1.3 (26--30) 27.9 ± 1.1 (27--30) \<0.0001^a^
BJLO 11.7 ± 2.1 (8--14) 13.5 ± 1.8 (8--15) 12.8 ± 1.7 (10--15) 0.0487^a^
HVLT-R-DRH 11.2 ± 1.0 (9--12) 11.7 ± 0.5(11--12) 11.4 ± 1.1 (8--12) 0.328^a^
HVLT-R-IFR 22.6 ± 5.7 (13--32) 26.9 ± 4.5(18--35) 26.3 ± 4.4 (19--33) 0.0687^a^
SF total score 46.7 ± 11.3 (25--66) 47.6 ± 8.2 (33--63) 48.2 ± 9.5 (36--63) 0.928^a^
SDMT 35.8 ± 10.2 (20--49) 43.7 ± 7.6 (32--60) 47.5 ± 10.1 (28--65) 0.0133^a^
a
Comparisons were made among the three groups of PD-MCI, PD-NC and HC.
b
Comparisons were made between PD-MCI and PD-NC groups.PD, Parkinson's disease; PD-NC, Parkinson's disease with normal cognition; PD-MCI, Parkinson's disease with mild cognitive impairment; HC, healthy control. UPDRS, Unified Parkinson's Disease rating scale; H&Y, Hoehn and Yahr; MoCA, Montreal cognitive assessment; BJLO, Benton judgment of line orientation; HVLT-R-DRH, delayed recognition hits in Hopkins verbal learning test-revised; HVLT-R-IFR, immediate free recall in Hopkins verbal learning test-revised; SF, semantic fluency; SDMT, symbol digit modalities test.
Data Preprocessing
------------------
Data preprocessing was carried out using Statistical Parametric Mapping (SPM12)^[2](#fn02){ref-type="fn"}^ and Data Processing Assistant for R-fMRI (DPARSF)^[3](#fn03){ref-type="fn"}^ toolkit ([@B46]). For MRI signal equilibrium and the participants' adaptation to the scanning circumstance, the first 10 volumes of the functional images were discarded. The remaining 200 volumes were then corrected for intra-volume acquisition time delay between slices and for inter-volume geometrical displacement due to head motion. No participant was excluded under a criterion of the head displacement of \>2 mm or angular rotation of \>2° in any direction. Next, the individual T1-weighted images were co-registered to the mean realigned functional images using a linear transformation ([@B14]), then were segmented into gray matter (GM), white matter (WM), and cerebrospinal fluid (CSF) tissue maps using a unified segmentation algorithm ([@B7]), and finally non-linearly normalized into the Montreal Neurological Institute (MNI) space using DARTEL approach ([@B6]). With the transformation parameters, all corrected functional volumes were spatially normalized into the MNI space, resampled to 3-mm isotropic voxels and spatially smoothed with a 4 mm full width at half maximum Gaussian kernel. Then, nuisance signals of 24 head motion parameters ([@B19]), global signal, CSF, and WM time series as well as linear trend were regressed out from each voxel's time course. Finally, temporal band-pass filtering (0.01-0.1 Hz) was performed on the residual time series to reduce the effect of low-frequency drifts and high-frequency noise ([@B9]; [@B33]).
The ALFF calculation was performed using the DPARSF toolkit within a group GM mask that was generated by greater than 0.2 of the mean GM probability map of all 42 participants. As described in previous studies ([@B47]; [@B50]), the ALFF of each voxel is defined by averaged square root of the power spectrum of time series, computed in a frequency domain based on a fast Fourier transformation and averaged across the 0.01--0.1 Hz frequency interval. It was further divided by the global mean value to reduce the global effects of variability across participants.
Statistical Analysis
--------------------
To examine between-group differences in ALFF, a one-way analysis of covariance (ANCOVA) was first performed in a voxel-wise manner (within the group GM mask) under non-parametric permutation tests in a general linear model (GLM) among the three groups, with age, gender and data center treated as additional covariant factors. Here, we used Statistical non-Parametric Mapping software (SnPM13)^[4](#fn04){ref-type="fn"}^ to determine the significant voxels, and it provides an extensible framework for non-parametric permutation tests based on a GLM. The ANCOVA statistical significance level was not corrected by multiple comparisons but to just obtain a small volume mask using a threshold of *p* \< 0.05 and cluster size \>50 voxels, which could as much as possibly contain candidate voxels that may show truly significant differences between groups. The following subsequent *post hoc* analysis was performed within the mask, and the multiple comparisons were used with small volume correction based on non-parametric permutation tests, at a cluster-level threshold of *p* \< 0.05 (*n* = 10000 permutations, FWE-corrected) with a cluster-forming threshold at the voxel level *p* \< 0.01.
To investigate the relationship between regional ALFF values and clinical scores (i.e., UPDRS, MoCA) as well as cognitive domains (i.e., HLVT-R, BJLO, SDMT, and SF), we performed multiple linear regression analysis with age, gender and data center treated as additional covariates. First, we examined correlations of regional ALFF values within the regions that showed significant differences between the PD-MCI and HC with the UPDRS total score and the MoCA score in the PD-MCI group. Second, to examine possible contributions of cognitive domains in the relationships between the regional ALFF values and cognitive impairment, we examined correlations of regional ALFF values within the regions that showed significant differences between the PD-MCI and PD-NC with the test scores of cognitive domains (i.e., HLVT-R, BJLO, SDMT, and SF) in the PD-MCI group.
Results
=======
Clinical and Demographic Testing of Sample
------------------------------------------
Clinical and demographic profiles of the 42 participants are shown in **Table [1](#T1){ref-type="table"}**. There was no significant difference among the three groups in gender, age, DRH and IFR in HVLT-R (HVLT-R-DRH and HVLT-R-IFR) and SF scores. Moreover, no significant difference in disease duration, UPDRS total score and H&Y stage score was found between the PD-MCI and PD-NC subtypes. However, the PD-MCI patients had significantly lower MoCA, BJLO, and SDMT scores than the PD-NC and HC group, in consistent with clinical diagnosis of each subtype.
Comparisons of ALFF Values Between Groups
-----------------------------------------
ANCOVA analysis revealed a range of regions showing uncorrected significant differences in ALFF among the PD-MCI, PD-NC, and HC groups (*p* \< 0.05, cluster size *k* \> 50, uncorrected; see **Figure [1](#F1){ref-type="fig"}**). These areas mainly covered regions of inferior cerebellum, occipital, right fusiform and right inferior frontal lobes. This analysis aimed to result as a small volume mask for the following *post hoc* analysis, which could as much as possibly contain candidate voxels that may show truly significant differences between PD-MCI and NC, between PD-NC and HC, or between PD-MCI and PD-NC.
{#F1}
Compared with the HC, both PD groups exhibited ALFF decreases in the occipital areas (i.e., Calcarine_R/Cuneus_R), and the PD-NC group additionally showed ALFF decreased in the right fusiform area as well (*p* \< 0.05; see **Figures [2A,B](#F2){ref-type="fig"}** and **Table [2](#T2){ref-type="table"}**). Specially, the PD-MCI group additionally exhibited increased ALFF in the opercular part of right inferior frontal lobe (i.e., Frontal_Inf_Oper_R) (*p* \< 0.05; see **Figure [2B](#F2){ref-type="fig"}** and **Table [2](#T2){ref-type="table"}**). Comparing the PD-NC, the PD-MCI group exhibited significantly higher ALFF in the Frontal_Inf_Oper_R and left fusiform gyrus (*p* \< 0.05; see **Figure [2C](#F2){ref-type="fig"}** and **Table [2](#T2){ref-type="table"}**).
{ref-type="table"}**.](fphys-09-01093-g002){#F2}
######
Between-group ALFF differences of regional brain activity.
Brain regions (AAL) Brodmann area MNI coordinates (mm) *T* value Cluster size
---------------------- --------------- ---------------------- ----------- -------------- ------- ----
**PD-MCI vs. HC**
Frontal_Inf_Oper_R BA 44/45 42 12 3 5.16 54
Calcarine_R/Cuneus_R BA 18 12 -84 12 -3.84 30
**PD-NC vs. HC**
Calcarine_R/Cuneus_R BA 18/17 15 -90 12 -6.58 39
Fusiform_R BA 37 27 -42 -15 -3.61 20
**PD-MCI vs. PD-NC**
Frontal_Inf_Oper_R BA 45 57 18 6 4.10 45
Fusiform_L BA 36 -27 -36 -21 4.13 34
PD, Parkinson's disease; PD-NC, Parkinson's disease with normal cognition; PD-MCI, Parkinson's disease with mild cognitive impairment; HC, healthy control; BA, Brodmann area; R, right; L, left. All the coordinates are denoted by MNI space coordinates.
Correlation of ALFF Values With Clinical and Neuropsychological Assessment in the PD-MCI Group
----------------------------------------------------------------------------------------------
First, we examined correlations of regional (i.e., Calcarine_R/Cuneus_R and Frontal_Inf_Oper_R) ALFF values with the UPDRS total score and the MoCA score in the PD-MCI group. We found that the neural activity in the Frontal_Inf_Oper_R area was positively correlated with the UPDRS total score (*p* \< 0.05), but marginally negatively correlated with the MoCA score (see **Figure [3](#F3){ref-type="fig"}**). Second, we also examined correlations of regional (i.e., Frontal_Inf_Oper_R and left fusiform gyrus) ALFF values with the test scores of cognitive domains. Interestingly, the neural activity in the Frontal_Inf_Oper_R showed a significantly negative correlation with the score of SF test (the executive function and working memory) (*p* \< 0.01) and a marginally negative correlation with the score of SDMT (the processing speed-attention) (see **Figure [4](#F4){ref-type="fig"}**).
{#F3}
{#F4}
Discussion
==========
Cognitive dysfunction occurs frequently in patients with PD. They are highly relevant, and limit the patients' quality of life and increase caregiver burden ([@B29]). A range of 18.9--38.2% of PD patients are diagnosed as PD-MCI across studies as a risk factor or a harbinger of subsequent dementia ([@B30]), and about 30% of PD patients develop to PD dementia ([@B3]). Moreover, common features such as visual hallucinations, depression, fluctuating cognition, parasomnias, autonomous disturbance, and apathy in PD are also associated with cognitive impairment. Thus, investigation of abnormality of brain activity related to mild cognitive impairment in PD is very important not only for the PD treatment and management but also the further understanding of neuronal and pathophysiological mechanisms in PD development.
Recent R-fMRI studies have indicated that the ALFF is physiologically meaningful for measuring intrinsic or spontaneous neuronal activity of the brain ([@B9]; [@B50]; [@B42]). By using the ALFF measurement, our study revealed that both PD-MCI and PD-NC groups displayed hypoactivity (i.e., reduced spontaneous brain activity) in the occipital cortex (i.e., Calcarine_R/Cuneus_R) compared with the HC group. This finding was also reported in a previous study using a dataset of 58 patients with early to moderate stage PD and 54 healthy controls ([@B44]) and a meta-analysis ([@B39]). More emerging evidence has indicated crucial involvement of the occipital lobe in the pathophysiology of PD. Morphological analysis indicated reduction in cortical thickness in the occipital lobe ([@B8]). Using graph-theoretical analysis, [@B17] found that in the early disease stage, patients with PD displayed significant lower network efficiency in the bilateral occipital lobe ([@B17]). Furthermore, a recent PET study indicated that both the cognitively normal PD and the PD-MCI groups had reduced FDG metabolism in the occipital lobe, as well as in the inferior parietal and posterior temporal regions in the PD-MCI group. To a certain degree, the reason may be due to a crucial rule in visual-spatial processing in the occipital lobe, which should be related with hallucination ([@B48]) and other common dysfunction in PD. Moreover, a longitudinal study over 2 years using H2(15)O PET has observed decreased learning-related activation in parietal and temporal-occipital association areas at the second session in PD ([@B11]). Therefore, we postulate that reduction in occipital activity is associated with the PD development.
The cognitive profile in PD-MCI is heterogeneous, but differs from that of MCI due to AD by showing relatively more severe visuospatial and executive deficits and relatively less severe memory impairment ([@B1]). In the present study, we found lower level of visuospatial (i.e., measured by BJLO) and processing speed-attention functions (i.e., measured by SDMT) in the PD-MCI, which is consistent with the characteristics of cognitive impairment in PD-MCI. Moreover, hyperactivity in the opercular part of right inferior frontal lobe (i.e., Frontal_Inf_Oper_R) was observed in the PD-MCI, compared to both HC and PD-NC groups. This region has been broadly suggested to play an important role in executive control function in many task-fMRI studies ([@B23]). Hyperactivity in the region could be explained by greatly improving its functional performance in favor of compensating a certain degree of cognitive decline and preserving the global cognition in early PD with MCI. The ALFF-clinical correlation results may support this presumption. The regional ALFF in this region was marginally negatively correlated with general cognition as measured by MoCA in the PD-MCI (see **Figure [3](#F3){ref-type="fig"}**). With regarding to cognitive domains, the present study showed a significantly negative correlation with the score of SF test (the executive function and working memory) and a marginally negative correlation with the score of SDMT (the processing speed-attention) (see **Figure [4](#F4){ref-type="fig"}**). Thus, these findings may indicate that more cognitive deficit induced higher neural activity in the right inferior frontal lobe. In a cognitive task study (planning set-shift task), [@B36] found a similar phenomenon, that was increased activation in the frontal and other related areas during the planning the set-shift task at the second session when the behavior performance had inclined to be decreased in the PD-MCI patients. So, we suggested that hyperactivity in the opercular part of right inferior frontal lobe may reflect an adaptive compensatory effect in response to a modest cognitive decline in early PD with MCI.
Moreover, the presence of MCI at early stages of PD could affect both cognitive and motor corticostriatal loops ([@B37]). The spontaneous neural activity in the right inferior frontal gyrus was additionally positively correlated with disease severity as measured by UPDRS total score, which may indicate the region as an indicator of disease progression of total PD symptoms as well. Nevertheless, whether and how the right inferior frontal gyrus contributes or interacts with cognitive function in PD remain to be studied in the future.
In addition, we also found hyperactivity in the left fusiform in the PD-MCI. Similar findings have been found in AD or MCI patients during both cognitive task and resting state. For example, [@B42] have showed increased activity in ALFF in both AD and MCI patients compared to healthy controls, and so does the functional homogeneity of spontaneous brain activity in AD patients ([@B24]). [@B40] observed that the fusiform increased functional activation in AD patients who were performing a visuospatial processing task. Because the fusiform spatially locates close to the parahippocampal gyrus, it may serve as a role of compensation for abnormality of the parahippocampal gyrus associated memory deficit in PD-MCI. Indeed, the region is mainly involved in the processing of memory ([@B28]) except for facial processing. However, it is unclear about the differences of compensatory mechanism in the fusiform between the MCI due to AD and MCI in early PD.
Limitations
===========
We acknowledge several limitations in our study. First, the sample size in the present study was relatively small. Because the PPMI project mainly focused on structural MRI and diffusion tensor imaging (DTI), very few R-fMRI scanning has been conducted. Second, other MRI modality and gene data were not investigated in this study. The underlying basis of structural (cortical Morphological) and functional connectivity or network and genetics of impaired brain activity in the occipital lobe and hyperactivity in the inferior frontal gyrus needs to be investigated in the future. Recently, several studies have found a range of abnormal functional network that is associated with cognitive decline in PD. For example, a white-matter (WM) study has revealed that PD-MCI patients showed a distributed pattern of WM abnormalities, including anterior and superior corona radiata, genu, and body of the corpus callosum, and anterior inferior fronto-occipital, uncinate, and superior longitudinal fasciculi ([@B4]). R-fMRI functional connectivity analysis suggested a wide range of connectivity dysfunction in the default mode network, dorsal attentional network and frontoparietal networks ([@B8]; [@B27]).
Conclusion
==========
In summary, our study concluded hyperactivity (i.e., reflect a compensatory recruitment) in the opercular part of right inferior frontal gyrus and hypoactivity in the occipital areas in early PD with MCI, which may be associated with cognitive decline and contribute to thought of PD dementia progression and further clinical treatment in PD. Further investigations of longitudinal changes and interaction in brain activity with the right inferior frontal gyrus as well as intervention are needed in future.
Author Contributions
====================
ZW conceived and analyzed the study, and wrote the manuscript. XJ arranged the clinical data and participated in the manuscript revision. HC and TF were involved in the writing and discussion. HW contributed to clinical diagnosis and was involved in writing and discussion. All authors read and approved the final manuscript.
Conflict of Interest Statement
==============================
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The reviewer CL declared a shared affiliation, with no collaboration, with several of the authors, XJ, HC, and TF to thehandling Editor.
**Funding.** This work was supported by National Natural Science Foundation of China (81701777) and National Key Research and Development Program of China (Z161100002616021).
Data used in the preparation of this article were obtained from the Parkinson's Progression Markers Initiative database ([www.ppmi-info.org/data](http://www.ppmi-info.org/data)). For up-to-date information on the study, visit [www.ppmi-info.org](http://www.ppmi-info.org). Parkinson's Progression Markers Initiative -- a public-private partnership -- is sponsored by the Michael J. Fox Foundation for Parkinson's Research (MJFF) and is co-funded by MJFF, Abbvie, Avid Radiopharmaceuticals, Biogen Idec, Bristol-Myers Squibb, Covance, Eli Lilly & Co., F. Hoffman-La Roche, Ltd., GE Healthcare, Genentech, GlaxoSmithKline, Lundbeck, Merck, MesoScale, Piramal, Pfizer and UCB.
<http://www.ppmi-info.org/>
<http://www.fil.ion.ucl.ac.uk/spm>
<http://www.restfmri.net/forum/DPARSF>
<http://warwick.ac.uk/snpm>
[^1]: Edited by: Chunhua Bian, Nanjing University, China
[^2]: Reviewed by: Chunlin Li, Capital Medical University, China; Qixiang Lin, Emory University, United States
[^3]: This article was submitted to Fractal Physiology, a section of the journal Frontiers in Physiology
| {
"pile_set_name": "PubMed Central"
} |
Background
==========
Glass ionomer cements (GICs) are a class of biomaterial in widespread use in modern dentistry \[[@B1]\]. They are used for a multitude of applications including for filling cavities caused by tooth decay or wear, as cavity liners, as fissure sealants, and as cements to form an adhesive bond between the tooth and a prosthetic dental restoration such as a crown or bridge. GICs have several favorable properties which make them suitable for these applications: They are tooth colored and available in a range of shades to allow matching to a patient's natural dentition, they have a good biocompatibility profile, and they have an inherent adhesion to enamel and dentine and thus require minimal preparation of the tooth surface prior to application. In comparison to the other main clinical material used for direct tooth-coloured restorations, methacrylate resin-based silica-filled composites, GICs are less compromised by moisture contamination. This means that the clinician does not have to comply with such stringent requirements to thoroughly dry the tooth and surrounding area, which is a significant benefit over the less forgiving resin-based materials, although of course moisture control and careful preparation of the field is important with GICs and indeed all restorative materials. Aesthetically GICs are generally considered good, although the opacity and thus the \"life-like\" appearance are inferior to that of the resin-based materials. As well as for the applications described above, GICs also find application in Atraumatic Restorative Treatment (ART), whereby a dental filling is placed rapidly and without the use of drills or anesthetics \[[@B2]\]. This is particularly beneficial for pediatric and elderly patients as well as those with dental anxiety or learning difficulties.
GICs are acid--base cements composed of glass filler particles and polyacid molecules. The setting of the material is initiated by mixing with water, whereupon the polyacid molecules dissociate and cause dissolution of the glass, resulting in the release of ions which allow cross-linking of the polyacids. The transition from viscous liquid to rigid solid takes place typically over around 2 minutes, although the reaction does not reach completion until several days have elapsed. Even after setting the GICs can still participate in ion exchange with the oral fluids \[[@B3]\]. This has the outcome that GICs can release and absorb fluoride; the glasses in GICs contain calcium fluoride which leaches soluble fluoride into the mouth during normal function. The fluoride in the glass can be replenished by exposure to fluoride-containing oral care products such as dentifrice and mouth rinse, thus creating a rechargeable fluoride \"reservoir\" and allowing for sustained release of fluoride in the vicinity of a GIC restoration. The intention, when GICs were first developed, was that this fluoride release would protect the surrounding tooth tissue from further decay. While there is no doubt that fluoride in drinking water and oral care products serves to protect the teeth and improve oral health at an individual and community level \[[@B4]\], clinical data is not supportive of an anti-caries effect of GICs \[[@B5]\], and a recent Cochrane Database Systematic Review could find no evidence to support any beneficial result of fluoride-releasing restorative materials \[[@B6]\].
A GIC which offers a genuinely antimicrobial and antibiofilm efficacy would be of considerable clinical benefit \[[@B7]\]. Such a material could reduce recurrent decay in the vicinity of a restoration and could provide an antibacterial seal under other materials, protecting the pulp from bacterial ingress. It could be useful in ART, and as a fissure sealant, providing a protective seal over the occlusal surfaces of caries-vulnerable teeth. The fact that ions can readily travel in and out of the material offers the opportunity to dope the cement with other soluble antimicrobials. New advances in nanotechnology may provide the means to developing such a material \[[@B8]\].
In this manuscript the development of a GIC which contains novel antimicrobial nanoparticles composed of chlorhexidine hexametaphosphate (CHX-HMP nanoparticles) is reported. A recent publication describes surface functionalization of other materials using CHX-HMP nanoparticles prepared using a similar technique, and it was found that they acted as slow release devices for soluble chlorhexidine (CHX) \[[@B9]\] which is a potent antimicrobial agent in widespread use in medicine and dentistry. The CHX-HMP nanoparticles are formed by a precipitation reaction on mixing of aqueous CHX and HMP solutions with HMP in excess. The aim of this study was to establish whether it was possible to create viable GICs containing CHX-HMP nanoparticles and to investigate the properties of those GICs. The ultimate aim of this work is to create a GIC which exhibits a lasting antimicrobial effect *in vivo* without compromising other useful properties of the material.
Results and discussion
======================
Overview
--------
GIC specimens with substitutions of 1, 2, 5, 10 and 20% CHX-HMP nanoparticles for GIC powder were successfully created and compared with unmodified GICs (0% substitution). Those with 30% substitution of CHX-HMP nanoparticles were difficult to handle and the set material was crumbly so these were discarded without further analysis.
Chlorhexidine release
---------------------
CHX release over 791 h (33 days) normalized to surface area and CHX-free controls is shown in Figure [1](#F1){ref-type="fig"}. CHX release persisted for the duration of the study with a rate of release which decreased with time. A dose--response was evident in that specimens with a higher substitution of CHX-NPs exhibited a larger CHX release, although the relationship was not directly proportional.
{#F1}
Cumulative CHX release at 1 and 24 h and 8, 15 and 33 days for the 6 specimen groups and the outcome of the statistical analyses are shown in Table [1](#T1){ref-type="table"}. In this table, values for the 0% (unsubstituted) GICs are also shown; it can be seen that these readings were small compared to the actual CHX concentration readings, but these are included in the statistical analysis nevertheless so as not to manipulate the data unnecessarily and to allow a figure for comparison. The outcome was the same for each time point in that 0% and 1% were not statistically significantly different from one another, and 1% and 2% were not statistically significantly different from one another, but all other pairings were significantly different indicating a clear increase in chlorhexidine release correlated with an increase in nanoparticle substitution at all measured times.
######
Cumulative CHX release for specimens with differing levels of nanoparticle substitution at each of 5 time points
**Cumulative CHX release \[nmol.mm**^**-2**^**\] (standard deviation in parentheses)**
-------- ---------------------------------------------------------------------------------------- -------------------- -------------------- -------------------- --------------------
**0** 0.01 (0.003)^a^ 0.01 (0.02)^a^ 0.11 (0.06)^a^ 0.12 (0.07)^a^ 0.16 (0.08)^a^
**1** 0.16 (0.08)^a,\ b^ 0.20 (0.09)^a,\ b^ 0.48 (0.15)^a,\ b^ 0.56 (0.16)^a,\ b^ 0.65 (0.17)^a,\ b^
**2** 0.66 (0.18)^b^ 0.70 (0.19)^b^ 1.04 (0.21)^b^ 1.17 (0.22)^b^ 1.30 (0.24)^b^
**5** 1.30 (0.24)^c^ 1.39 (0.25)^c^ 2.03 (0.34)^c^ 2.30 (0.36)^c^ 2.51 (0.38)^c^
**10** 2.52 (0.38)^d^ 2.65 (0.40)^d^ 3.40 (0.43)^d^ 3.73 (0.44)^d^ 4.01 (0.45)^d^
**20** 4.02 (0.48)^e^ 4.20 (0.50)^e^ 5.09 (0.55)^e^ 5.46 (0.59)^e^ 5.94 (0.67)^e^
Within each time point, superscript letters indicate statistically homogeneous groups, so figures with different superscript letters are statistically significantly different to a 95% confidence level at that time.
Since CHX is efficacious against a wide range of bacteria and yeasts, this may confer antimicrobial and thus anti-caries properties on these nanofunctionalized dental filling materials. CHX disrupts the bacterial cell membrane \[[@B10]\] and results in the loss of intracellular components; this has the outcome that the evolution of bacterial resistance to CHX is considered unlikely \[[@B11]\].
Since CHX is an appealing option for the development of a dental cement which reduces the incidence of recurrent tooth decay, is it not surprising that there have been other attempts to incorporate CHX into GICs. CHX diacetate was added to a resin-modified GIC and this resulted in CHX release, but this was only sustained at significant levels for one week \[[@B12]\] and thus offered limited scope for lasting anti-caries effects. Incorporating CHX diacetate into a conventional (not resin-modified) GIC also yielded a CHX-releasing material, but again the CHX release was sustained for only around a week with all except the highest substitutions, and these high substitutions resulted in a deterioration of the mechanical properties of the material \[[@B13]\]. Dental composite resins supplemented with pulverized CHX diacetate also showed CHX release which reached a plateau after around 7 days \[[@B14]\]. The CHX release observed in the study reported here was more prolonged, and it is thought that this ie because the nanoparticlces themselves exhibit a gradual release of soluble CHX \[[@B9]\] rather than the already soluble and thus readily lost CHX in the studies described above.
Another report describes a longer-term effect of incorporating CHX -- as ground CHX diacetate powder or as CHX digluconate solution -- into GICs \[[@B15]\]. The CHX release per se was not measured but it was shown that an antimicrobial effect persisted for between 40 and 90 days. The peak of efficacy was the first 24 h for all GIC specimens, suggesting that most CHX may have been released during this initial period, and most specimens showed no antimicrobial behavior after 60--90 days. For some formulations, a limited deterioration in mechanical properties was observed. CHX digluconate solution has also been incorporated into GICs in combination with another antimicrobial agent, cetrimide, and this too had an antimicrobial effect on oral bacteria \[[@B16]\]. The authors indicate that this effect persisted for up to 180 days, but whether this was due to antimicrobial still leaching at the 180 day point, or that which had earlier leached and was still present in the agar plate, is not clear.
Fluoride release
----------------
Fluoride release over 791 h (33 days) normalized to surface area can be seen in Figure [2](#F2){ref-type="fig"}. All of the GIC specimens released fluoride continually over the duration of the experiment. The initial release rate was the most rapid and this gradually slowed over the experimental period, as has been observed for conventional GICs by other researchers \[[@B17]\].
{#F2}
Cumulative fluoride release at 1 and 24 h and 8, 15 and 33 days for the 6 specimen groups and the outcome of the statistical analysis are shown in Table [2](#T2){ref-type="table"}. At 1 h, the unmodified GIC released significantly more fluoride than nanoparticle-substituted cements but there were no statistically significant differences between the different substitutions. At later times, in numerical terms fluoride release displayed a pattern of 20% \> 5, 10% \> 1, 2% with a complex relationship with 0%. However, there were few statistically significant differences; the only ones observed were at 24 h, 2% nanoparticle-substituted GICs released less fluoride than the 0% control, and at the longest time point, 33 days, the 20% nanoparticle-substituted GICs released more fluoride than the 2% nanoparticle-substituted GICs. It is not clear why the most highly substituted GIC released the most fluoride, especially as the greater the proportion of CHX-HMP nanoparticles, the smaller the total mass of fluoride-containing filler present. It is possible that the presence of the CHX-HMP nanoparticles alters the setting reaction and this renders fluoride more mobile in the cement lattice, but this is a hypothesis that has yet to be tested, and should be considered in the context that only the 2%-20% comparison showed significant differences. Although the impact of fluoride release from restorative materials is still a source of some controversy \[[@B1]\], it is not of concern in this study since the nanofunctionalized GICs showed similar fluoride release profiles to the unmodified cements. This is in contrast to an earlier report in which GICs supplemented with CHX digluconate solution exhibited a reduction in fluoride release \[[@B18]\].
######
Cumulative fluoride release for specimens with differing levels of nanoparticle substitution at each of 5 time points
**Cumulative fluoride release \[ng.mm**^**-2**^**\] (standard deviation in parentheses)**
-------- ------------------------------------------------------------------------------------------- --------------------- ------------------ ----------------- ---------------------
**0** 7.38 (1.69)^a^ 27.65 (0.85)^a^ 85.82 (5.68)^a^ 119.9 (11.0)^a^ 154.5 (14.3)^a,\ b^
**1** 2.61 (1.11)^b^ 18.59 (3.12)^a,\ b^ 70.72 (8.40)^a^ 106.9 (14.1)^a^ 148.1 (16.6)^a^
**2** 1.68 (0.73)^b^ 17.10 (2.33)^b^ 64.30 (7.68)^a^ 101.1 (12.2)^a^ 147.5 (15.7)^a^
**5** 1.74 (0.70)^b^ 19.31 (4.32)^a,\ b^ 82.67 (7.49)^a^ 124.2 (9.2)^a^ 181.4 (14.0)^a,\ b^
**10** 1.83 (0.96)^b^ 18.73 (9.50)^a,\ b^ 81.73 (30.22)^a^ 122.8 (39.3)^a^ 184.4 (49.8)^a,\ b^
**20** 2.09 (0.34)^b^ 23.74 (5.89)^a,\ b^ 81.73 (13.15)^a^ 134.6 (19.2)^a^ 204.4 (23.4)^b^
Within each time point, superscript letters indicate statistically homogeneous groups, so figures with different superscript letters are statistically significantly different to a 95% confidence level at that time.
Tensile strength
----------------
Diametral tensile strength of the 6 specimen groups are shown in Table [3](#T3){ref-type="table"}. The ANOVA gave a p value of 0.054 indicating that, although there was a numerical trend towards lower tensile strength for 10 and 20% substitution cements, there was no statistically significant difference between these values and those of the other GICs. The fact that substitutions up to 5% appeared to have no significant deleterious effect on the tensile strength of the cements is encouraging.
######
Diametral tensile strength of GIC specimens
**NP substitution \[%\]** **Diametral tensile strength \[MPa\] (standard deviation in parentheses)**
--------------------------- ----------------------------------------------------------------------------
0 14.1 (3.7)
1 14.3 (4.9)
2 15.7 (4.3)
5 15.5 (1.1)
10 11.5 (2.8)
20 9.4 (2.6)
The ANOVA gave p = 0.054 indicating that there were no statistically significantly differences between tensile strengths of the different GICs to a 95% confidence level.
Nanoparticle characterization
-----------------------------
Dynamic light scattering (DLS) indicated that there were structures of mean diameter 196 nm, but it was observed that there was substantial polydispersity and the standard deviation was large (76 nm). The correlation functions observed during some DLS measurements suggested the presence of some much larger particles. This may be explained by observations made by atomic force microscopy (AFM) (Figure [3](#F3){ref-type="fig"}) which indicated that nanoparticles sometimes formed aggregates which could be as large as several micrometres. The individual nanoparticles which compose these aggregates were regularly shaped, globular and had typical diameters of 80--90 nm (Figure [3](#F3){ref-type="fig"}). This would not be observed in DLS since the signal is proportional to diameter to the 6th power, so the signal from these small nanoparticles would be masked by that from the larger aggregates.
{#F3}
Zeta potential measurements indicated that the nanoparticles had a mean surface charge of -55 mV (standard deviation 1.4 mV), indicating a net negative charge.
Morphology and structure
------------------------
Scanning electron micrographs of representative GIC specimens are shown in Figure [4](#F4){ref-type="fig"}. The appearances of the GIC specimens with different substitutions of nanoparticles were similar, with the glass filler particles and surrounding matrix clearly visible. Only the 20% nanoparticle substitution exhibited a slightly different appearance (Figure [4](#F4){ref-type="fig"}f), with textured areas depleted in glass particles and a smoother fracture plane which are suggestive of nanoparticle aggregates. It is possible that these could lead to a reduction in strength since they cannot be presumed to interact with the polyacid in the same way as the glass filler particles, and future studies will address this important question.
{#F4}
Overview
--------
By adding CHX-HMP nanoparticles to a commercial GIC it has proven possible to create a material which releases CHX for in a dose-dependent manner for longer than has been observed using some other approaches to CHX functionalization. The results of this study would suggest that substitutions of up to 20% nanoparticles for glass by mass using this approach may be suitable for further development as clinical materials. 10% and 20% substitutions showed a numerical reduction in strength but not a statistically significant one; subsequent follow-on studies will allow for further investigation of this observation. Higher substitutions, of 30% or more, are unlikely to find application without other changes to the cement as this resulted in a dry, crumbly cement with unacceptable handling properties. Options to use more dispersed nanoparticles, rather than the ground aggregates discussed here, are under investigation. Possibilities for CHX recharging and the microbiological impact of the CHX release are the subject of ongoing experiments, as well as larger studies to investigate the effect on different kinds of strength. Interestingly, it has recently been shown that resin-modified GICs can act as a short-term in vivo reservoir for topically applied CHX \[[@B19]\].
In this study, nanoparticles were substituted for powder by weight. An alternative approach is to substitute like-for-like by surface area, and this would account for the fact that the specific surface area of the nanoparticles is higher than that of the glass filler particles.
Conclusions
===========
Novel GICs have been created which contain antimicrobial CHX-HMP nanoparticles at a range of dopings. These GICs released soluble CHX over a period of at least 33 days, and the quantity of CHX released was dependent on the doping of nanoparticles in the cement. All cements released fluoride with a similar profile to the control, unmodified cement and moderate substitutions did not detrimentally affect the tensile strength of the material. These cements may find clinical application as dental biomaterials which prevent or reduce the incidence of secondary caries and protect the tooth and soft tissues from bacterial infection.
Methods
=======
Synthesis and preparation of CHX-HMP nanoparticles
--------------------------------------------------
Aqueous stock solutions of chlorhexidine digluconate and sodium hexametaphosphate (Sigma Aldrich, Gillingham, UK) were mixed in deionized water such that the final concentration was 4 mM CHX and 5 mM HMP. The resulting colloidal suspension of CHX-HMP nanoparticles was mixed thoroughly and then centrifuged at 21000 g for 60 min. The supernatant was removed and discarded and the nanoparticle pellet dried for at least 48 h at 40°C. The pellet was then removed from the centrifuge tubes and ground to a fine white powder composed of nanoparticle aggregates using an agate mortar and pestle. Further details regarding the properties of CHX-HMP nanoparticles and their antimicrobial efficacy can be found elsewhere \[[@B9]\].
Prototype nanofunctionalized glass ionomer cements
--------------------------------------------------
A commercially available GIC (Diamond Carve (TM), Kemdent, Purton, UK) was used as the starting material. This commercially available GIC comprises a powder, which consists of alumina-silica based glass filler particles containing calcium fluoride and other minor salts and freeze-dried poly (vinyl) phosphonic acid, and a liquid which contains polyacrylic and tartaric acids. It is mixed in a ratio of 1:4 liquid:powder by mass. Cylindrical GIC specimens with nominal dimensions of 6 mm diameter and 3 mm height were formed by mixing the GIC according to the manufacturers' instructions and packing into Perspex molds coated with a thin layer of petroleum jelly to aid removal. The mixing was carried out by one individual (OJO) with extensive experience of GIC mixing and handling. The precise dimensions of each specimen were measured using calipers and recorded. The nanoparticle powder created by grinding the nanoparticle pellet and thus yielding compacted clusters of nanoparticles was used to substitute for the GIC powder at fractions of 0, 1, 2, 5, 10, 20 and 30% by mass. Ten specimens of each substitution were created giving a total of 70 specimens. They were removed from the mold within 60 minutes and placed in individual small, sealed plastic vessels that contained wet tissue paper not in direct contact with the specimen, to achieve an atmosphere of 100% humidity but prevent the specimen being in contact with liquid water which could result in dissolution during the critical early phases of setting. These were stored at 37°C for 7 days.
After this the specimens were divided into two sets of 5 specimens each. One set of each substitution was set aside for tensile strength and fracture surface morphology analysis and the other set was used to investigate the CHX and fluoride leaching from the cement.
For the investigations of CHX and fluoride release, each specimen was immersed in 1 mL artificial saliva in individually labeled vials at 37°C. The artificial saliva was composed of CaCl~2~ · 2H~2~O 0.103 gL^-1^, MgCl~2~ 0.019 gL^-1^, KH~2~PO~4~ 0.544 gL^-1^, C~8~H~18~N~2~O~4~S (HEPES buffer acidic form) 4.77 gL^-1^, KCl 2.24 gL^-1^, 1.80 mL 1 M HCl, KOH titrated to obtain a pH of 6.8. Specimens were periodically removed and placed in duplicate tubes containing fresh artificial saliva so that the artificial saliva the specimen had been incubated in could be sampled for CHX and fluoride concentrations. A pilot study was conducted to establish the saturation limit of fluoride concentration within the vessels to ensure that the sampling periods were selected appropriately and erroneous readings owing to saturation of the eluent by a fluoride salt were not obtained by leaving too large a gap between readings. Using the findings from this pilot study, the sampling occurred at hourly intervals during the first day, followed by intervals of 4 hours, then daily and then weekly. Controls containing only artificial saliva without a GIC specimen were sampled in the same way.
Chlorhexidine measurements
--------------------------
CHX concentration in the artificial saliva was measured using ultraviolet spectrophotometry. The 1 mL artificial saliva was placed into a semi-micro cuvette transparent under ultraviolet wavelengths and absorption was measured at 255 nm using a spectrophotometer (Hitachi U1900, Tokyo, Japan). The reading was converted to CHX concentration with reference to calibration standards at 5--50 μmol.l^-1^. The concentration was converted to moles of CHX released per unit surface area of the GIC specimen with reference to the individual dimension measurements for each specimen and was normalized by subtracting the reading for 0% substitution to correct for any other molecules present in the system which absorbed at 255 nm such as the polyacrylic acid which is another component of the GIC.
Fluoride measurements
---------------------
Fluoride concentration in the artificial saliva was measured using an ion-selective electrode (9609BNWP, Thermo Fisher Scientific, Waltham, MA, USA) by mixing 0.5 mL artificial saliva with 0.5 mL TISAB solution (Thermo Fisher Scientific, Waltham, MA, USA). The data output was converted to mg/L fluoride ion with reference to calibration standards of 0.1, 0.5, 1, 2 and 5 mg/L F^-^, also diluted with equal quantities of TISAB.
Tensile strength measurements
-----------------------------
Indirect tensile strength (S~T~) was measured by applying a compressive diametric force to the curved sides of the cylindrical specimen (n = 5 per group) until fracture occurred, using a universal testing machine (LR5K, Lloyd Instruments, Ametek, FL, USA) recording the load at fracture (L) and using this with the specimen dimensions of height (h) and diameter (d) to calculate tensile strength according to the relationship:
$$\textit{Sr} = \frac{2\mathit{L}}{\pi\textit{dt}}$$
Statistical analysis
--------------------
Cumulative CHX and fluoride release at time points of 1 h, 24 h, 8 days, 15 days and 33 days, and indirect tensile strength, were compared using one-way ANOVAs with a Tukey honestly significant difference post-hoc test.
Characterization of CHX-HMP nanoparticles
-----------------------------------------
The size and zeta potential of the nanoparticles as prepared in colloidal suspension were measured using a Malvern Zetasizer (Malvern, UK). The size, morphology and aggregation of the nanoparticles were investigated when immobilized on glass coverslips. Coverslips were cleaned by ultrasonicating for 10 minutes in acetone followed by ultrasonicating for 10 minutes in ethanol then coated by dipping into a freshly prepared colloid of the nanoparticles for 30 seconds, then rinsing in running deionized water for 10 seconds, then allowing to dry in air. The nanoparticle-coated coverslips were imaged using AFM (Nanoscope IIIa, Digital Instruments, CA, USA).
Morphology and structure
------------------------
Specimens which had been tested for tensile strength were coated with a thin layer of gold-palladium (SC7620, Emitech, Taiwan) and examined using SEM (Phenom, Eindhoven, Netherlands). Images were obtained at nominal magnifications of 400, 1000 and 5000×.
Abbreviations
=============
CHX: Chlorhexidine; CHX-HMP: Chlorhexidine hexametaphosphate; GIC: Glass ionomer cement.
Competing interests
===================
The University of Bristol filed a patent application relating to the work presented in this manuscript in 2013.
Authors' contributions
======================
ERH, OJO and CAB were the students who carried out the laboratory work and statistical analysis. JAH was the industry supervisor of the study, providing guidance and input on study design from the industrial perspective. DJOS was the clinical supervisor of the study, providing a clinician's perspective on important material properties and helping to develop the study parameters and design. MEB was the principal, scientific supervisor of the study. She conceived the study, supervised the students in the laboratory, directed the analysis and wrote the manuscript. All authors read and approved the final draft of the manuscript.
| {
"pile_set_name": "PubMed Central"
} |
The whole GCH array dataset is deposited in GEO Database (accession number GSE53440).
Introduction {#sec001}
============
The yeast *Saccharomyces cerevisiae* can be found naturally in many niches in the environment, but it is most commonly known for its role as "baker's yeast" in either the traditional or industrial fermentative production of bread, beer or wine. It can be found in many dietary supplements as the main component and a part of these yeast cells are alive. It has also been used as a nutritional supplement and as an agent to treat antibiotic-related diarrhea, commercialized as *Saccharomyces boulardii* \[[@pone.0122382.ref001]\]. Classically, *S*. *cerevisiae* has been considered a non-pathogenic and safe organism. However, in the last few years, cases of diagnosed *S*. *cerevisiae* infections have increased, probably due to an increase in the numbers of immunocompromised patients and the progress made in diagnostic methodologies, including microbial identification by molecular techniques. *S*. *cerevisiae* has been identified in a wide variety of infections, ranging from cutaneous infections and vaginitis in healthy individuals, to systemic bloodstream infections and infections of vital organs in immunocompromised and critically ill patients \[[@pone.0122382.ref001]--[@pone.0122382.ref003]\]. Infected patients are premature children, elderly people or patients suffering from immunosuppression due to AIDS, treatment with immunosuppressive agents, or other conditions associated with an insufficient immune response. Moreover, severe infections with *S*. *cerevisiae* have been occasionally reported in patients with no obvious predisposing factors \[[@pone.0122382.ref004],[@pone.0122382.ref005]\]. All these data have changed the status of *S*. *cerevisiae*, which is now considered to be an emerging opportunistic pathogen \[[@pone.0122382.ref006]--[@pone.0122382.ref009]\]. Several studies have analyzed the potential virulence of this yeast species *in vitro* \[[@pone.0122382.ref010]--[@pone.0122382.ref013]\], while others have used *in vivo* models \[[@pone.0122382.ref013]--[@pone.0122382.ref017]\]. These reports suggest that some clinical, but also non-clinical, *S*. *cerevisiae* strains have the potential to cause disease in murine models regardless of the host's immune status.
To elucidate the genetic determinants that influence the ability of *S*. *cerevisiae* to cause infection, we selected strains that showed high levels of virulence attributes (pseudohyphal and growth at 42°C, phosphatases and proteases secretion, and the ability to infect and kill mice) independently of the isolation origin \[[@pone.0122382.ref013]\], \[[@pone.0122382.ref016]\], \[[@pone.0122382.ref018]\], and we used these strains for a comparative genetics approach. We observed increased copy number (CN) of the genes related to the *de novo* biosynthesis of the purine nucleotides pathway. One of the key genes of this pathway, which encodes an enzyme catalyzing the second pathway step, is *GUA1*. The importance of this pathway for virulence of *S*. *cerevisiae* was confirmed by experimental infections in immunodeficient murine models using a Δ*gua1* mutant of the clinical strain D14, isolated from a dietary product. We also revealed that exogenous guanine, an end product of the purine nucleotides pathway in its triphosphorylated form, increases the survival of yeast strains in *ex vivo* blood infections. Finally, we show the implication of the DNA damage checkpoint that activates dNTP biosynthesis in yeast cells during blood infections. We conclude that opportunistic yeasts may use an enhanced *de novo* biosynthesis of the purine nucleotides pathway to counteract the immune system of the host. An increased efficiency of dNTP biosynthesis may in turn promote DNA repair after DNA damage produced by neutrophil oxidative bursts.
Materials and Methods {#sec002}
=====================
Yeast strains and media {#sec003}
-----------------------
Several *S*. *cerevisiae* isolates were used in this work: a clinical vaginal isolate (isolate 60), an isolate from the respiratory tract (isolate 102) \[[@pone.0122382.ref010]\], a brewer's strain isolated from a commercial nutritional complement product (D14) \[[@pone.0122382.ref013]\], auxotrophic laboratory strains W303 (*MATa*; *ura3*-52; *trp1*Δ2; *leu2*-3,112; *his3*-11; *ade2*-1; *can1*-100), BY4741 (*MATahis3*Δ0 *leu2*Δ0 *met15*Δ0 *ura3*Δ, Open Biosystems), Δ*dun1* (*MATa his3*Δ0 *leu2*Δ0 *met15*Δ0 *ura3*Δ, Open Biosystems), W1588-4C (*MAT* **a** *ade2-1 can1-100 his3-11*,*15 leu2-3*,*112 trp1-1 ura3-1 RAD5*) (31), rnr1\* (*MAT* **a** *ade2-1 can1-100 his3-11*,*15 leu2-3*,*112 trp1-1 ura3-1 RAD5 rnr1-W688G*) \[[@pone.0122382.ref019]\] and prototrophic strain S288C (*MATα SUC2 gal2 mal mel flo1 flo8*-1 *hap1*). Isolates 102, 60 and D14 were chosen from a yeast collection of potential pathogenic isolates obtained mainly from hospitals, but also from commercial dietary products \[[@pone.0122382.ref016]\], \[[@pone.0122382.ref018]\]. These isolates were proposed to have a higher virulence potential since they showed increased pseudohyphal and 42°C growth \[[@pone.0122382.ref018]\], augmented phosphatases and proteases secretion \[[@pone.0122382.ref018]\] and were able to survive and to colonize in the brain and kidney of immunocompetent mice after blood infections, thereby causing morbidity and death among their hosts \[[@pone.0122382.ref013]\], \[[@pone.0122382.ref016]\]. In addition, strains 60 and D14 were less efficiently phagocytized by macrophages as compared to other strains \[[@pone.0122382.ref013]\]. *S*. *cerevisiae* strains were cultured in YPD (1.0% yeast extract, 1.0% Bacto-peptone and 2.0% glucose) and in SC medium (5.6% YNB without amino acids, 1.3% complete amino acid drop out (Formedium, UK)). The corresponding media were supplemented with CuSO~4~ (0--2.0 mM) and also mycophenolic acid (MPA) (7 μg/ml) or methyl methanesulfonate (MMS) (0.004%) or guanine (25 μM) based on previously reported methods \[[@pone.0122382.ref020]--[@pone.0122382.ref022]\].
The mutant D14Δ*gua1* was obtained by a PCR-based gene-replacement strategy using the subsequent deletion cassettes containing geneticin or hygromycin resistance \[[@pone.0122382.ref023]\]. The primers used had 40 bp homologous to the gene coding sequence at their 5' or 3' end. The amplified PCR products were used for gene disruption, performed by homologous recombination following yeast transformation using the lithium acetate procedure. The PCR reactions were as follows: 2 min at 94°C; 30 cycles of 15 s at 94°C, 30 s at 50°C, 2 min at 72°C and 5 min at 72°C. Cells were transformed with the PCR product following the LiAc protocol (<http://home.cc.umanitoba.ca/~gietz/website>). After heat shock, cells were incubated for 3 h in YPD liquid medium at 30°C. Finally, transformed cells were selected on YPD plates with geneticin (200 mg/L) or hygromycin. *GUA1* gene deletion was confirmed by a PCR analysis and guanine auxotrophy. Strain D14GUA1 was obtained by transforming strain D14Δgua1 with a plasmid containing a wild-type copy of *GUA1* (pGUA1) and by selecting in YPD with 200 μg/mL of G418 (Sigma). The plasmid obtained from Iglesias-Gato *et al*. \[[@pone.0122382.ref021]\] was marker-swapped for KANMX by homologous recombination using a cassette obtained by PCR and a pUG6 plasmid as a template.
Yeast growth in the presence of copper salt (0, 0.5, 1 and 2 mM of CuSO~4~) was quantified in microtiter plates on a reader model POLARstar Optima (BGM Labtech, Offenburg, Germany). Wells were filled with the appropriate number of yeast cells and 0.5 ml of YPD medium (with or without copper salt) to reach an initial OD of approximately 0.2 (corresponding to a starting cell number of ∼10^6^ cells/ml). Non-inoculated wells were included for each experimental series to determine, and consequently subtract, noise signals. Growth was monitored by optical density (OD) changes at a wavelength of 600 nm. Measurements were taken every 30 min for 72 h at 28°C (until yeast cells reached the stationary phase) after a 20-second pre-shaking. All experiments were carried out in triplicate and maximal OD was obtained for each condition.
DNA labeling and competitive genome hybridization {#sec004}
-------------------------------------------------
DNA was extracted from yeast strains grown from single colonies in YPD medium following the procedure described by Belloch *et al* \[[@pone.0122382.ref023]\]. Before the labeling reaction, 4 μg of DNA were treated with RNAse A (1 mg/mL, Roche Diagnostics) for 1 h at 37°C and with Proteinase K (1 mg/ml, Quiagen) for 1.5 h at 37°C with agitation. Phenol extracted DNA was digested with MseI or RsaI (New England Biolabs, Inc.), according to the manufacturer\'s instructions, and digestions were mixed. The fragmented sample was purified using the QIA-quick gel extraction kit (Qiagen, Germany), heat denatured for 5 min at 100°C, and then cooled on ice. DNA was labeled in 21-μl reaction mixtures by random priming using the BioPrime Array CGH Genomic Labeling System (Invitrogen). Unincorporated nucleotides are removed by filtration through a Qiaquick PCR Purification column (Qiagen). Equal amounts of labeled DNA (1μ) were used as probes to hybridize to the PCR-amplified open reading frames (ORFs) of homoploid *S*. *cerevisiae* S288C DNA spotted onto microarrays (Eurogentec S.A., Belgium). Microarrays were pre-hybridized with 5 ml of pre-hybridization solution (3× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate), 0.1% SDS, and 0.1 mg/ml BSA) at 42°C for 1 h. After washing slides with isopropanol and water, samples were centrifuged at 1300 rpm for 10 min at room temp. Then 10 μl of samples and 50 μl of hybridization solution (50% (v/v) formamide, 5x SSC, 0.1% SDS and 0.1 mg/ml DNA from salmon sperm) were denatured (95°C, 1 min) and added to the slides that were incubated for 16 h at 42°C in a Gene Machines Hyb Chamber (Genomic Solutions). Slides were recovered and washed 5 min at 42°C in 1 (2x SSC, 0.1% SDS), 2x10 min in solution 2 (0.1 x SSC, 0.1% SDS), 5x1min in solution 3 (0.1 x SSC) and 5 s in solution 4 (0.01 x SSC). Slides were centrifuged for 10 min at 1300 rpm.
Microarray scanning and data normalization {#sec005}
------------------------------------------
Images were acquired using GenePix 4100A (Axon Instruments, Molecular Devices Corp., USA) with a 10-μm resolution using the GenePix Pro 6.1 software (Axon Instruments, Molecular Devices Corp., USA). Raw data were processed using Acuity 4.0 (Axon Instruments, Molecular Devices Corp., USA). All the slide data were normalized with a log~2~(ratio) average and filtered (signal of each channel \> 350 and signal/noise ratio \> 2.5). Poor or inconsistent signals were not considered for further analysis. Hybridization signals were depicted as the log2 hybridization signal ratio. Genomic material obtained from three biological independent replicates was used to perform hybridizations two or three times, including a dye-swap, and the genes with signal intensities ± 0.5 were considered for further analysis. FDR\<0.05 was used to select significant data. Heat map was generated in MeV 4.9 software using developers instructions were each row represents a specific gene and each column represents each strain. The lightest green boxes are the most underrepresented genes and the brightest red being the most overrepresented genes.
Flow cytometry {#sec006}
--------------
Yeasts strains were grown in 10 ml of YPD liquid medium under agitation. Early-stationary-phase cells (10^6^ cells ml^−1^) were harvested by centrifugation, washed and fixed in 70% ethanol at 4°C for 5 min. Fixed cells were harvested by centrifugation and resuspended in 0.01 M phosphate-buffered saline buffer (pH 7.2) containing 400 μl of RNase (10 mg ml^−1^). After incubation at 37°C for 30 min, cells were harvested by centrifugation and re-suspended in 950 μl of 0.01 M phosphate-buffered saline buffer (pH 7.2) containing 50 μl of propidium iodide (0.005%). Samples were analyzed using a flow cytometer FACScan analyzer (Becton Dickinson). DNA content values were scored on the basis of the fluorescence intensity of the first peak by comparing with the haploid W303 and diploid S288c strains.
qRT-PCR {#sec007}
-------
To validate the microarrays results, the DNA from the five strains was isolated and the copy number of the three genes was quantified. The genes selected were *IMD2*, *IMD3* and *SAM3* with criteria explained below. qRT-PCR was performed with gene-specific primers (200 nM) (IMD2-F: `AACATTCCTGTCAAGACATCGG`, IMD2-R: `TCAGTTATGTAAACGCTTTTCGTAA`, IMD3-F: `CGGTTTACAACATTCTTGTCAAGAC`, IMD3-R: `AAACGCTTTTCGTAAGAATGTAAGT`, SAM3-F: `GCGGGTGAAATATACGTATCGG`, SAM3-R: `ATTCTCCACGGAACACCCAGTA`, ACT1F: `TACAACTCCATCATGAAGTG` and ACT1-R: `GCCAAAGCGGTGATTTCC`) in a 20 μl reaction using Light Cycler Fast Start DNA Master PLUS SYBR green (Roche Applied Science, Germany) in a LightCycler 2.0 System (Roche Applied Science, Germany) device. All the samples were processed for the melting curve analysis, amplification efficiency and DNA concentration determination using the LightCycler 2.0 System. A mix of all the samples and serial dilutions (10^-1^ to 10^-5^) was utilized as the standard curve. Data were normalized with the *ACT1* gene to correct for the DNA content in each sample. Data are presented as the average ± standard deviations of three biological replicates.
Experimental infection {#sec008}
----------------------
*S*. *cerevisiae* strains were grown at 30°C on YPD. After 24 h, cells were harvested, washed twice with sterile phosphate-buffered saline (PBS) and diluted to the desired density. Healthy mice were infected by intravenous inoculation into the tail vein with 2 × 10^7^ viable CFU of yeast in a 0.2-ml volume of PBS, as previously described \[[@pone.0122382.ref015]\]. Twenty-four DBA/2 mice, aged 7--8 weeks, were respectively used (8 mice per *S*. *cerevisiae* strain). Mice were housed in groups of five in individually ventilated cages and cared for in strict accordance with the principles outlined in the *European Convention for the Protection of Vertebrate Animals Used for Experimental and Other Scientific Purposes* (<http://conventions.coe.int/Treaty/en/Treaties/Html/123.htm>). Experiments were approved by the governmental ethics committee (Comité Ético de Experimentación Animal, Universitat de Valencia). Neutropenia in mice was produced by an intraperitoneal injection of 150 mg/kg of cyclophosphamide (Sigma-Aldrich) 1 day before inoculation with each *S*. *cerevisiae* strain. Five days later, all the animals received a second dose of cyclophosphamide (150 mg/kg) to maintain neutropenia until the end of the assay. Similar treatments have demonstrated that severe neutropenia can be achieved and maintained \[[@pone.0122382.ref024]\]. The infected animals were visually monitored at least twice daily and they were humanely euthanized by cervical dislocation when any apparent signs of suffering were found. Mouse survival was analyzed by using Kaplan-Meyer curves and the log-rank test. No adverse events occurred during the assay.
Survival assay in human blood {#sec009}
-----------------------------
To investigate the survival of yeast strains in blood we used the method previously described \[[@pone.0122382.ref025]\]. Yeast strains were grown overnight in YPD medium at 30°C. Cells were washed once and suspended in PBS buffer (phosphate-buffered saline: 150 mM NaCl, 16 mM Na~2~HPO~4~, 4 mM NaH~2~PO~4~, pH 7.4) at a density of 5 x 10^7^ cells/ml. Human peripheral venous blood was collected from healthy volunteers by venipuncture using ammonium heparin syringes (Monovette, Sarstedt) as described \[[@pone.0122382.ref025]\]. Blood was obtained from healthy human donors with written informed consent. The blood donation protocol and use of blood for this study were approved by the institutional ethics committee (Comité Ético del Insituto de Agroquímica y Técnología de los Alimentos, CSIC). After leukocyte count with Neubauer chamber, yeast cells were inoculated (1:1 ratio of yeast:leukocytes) in blood and incubated for 90 min at 37°C as described \[[@pone.0122382.ref025]\]. Then, 100 μl of the 10^-3^ to 10^-4^dilutions were spread on YPD plates and incubated for 48 h at 30°C. Colony-forming units (cfu) were counted and relative percentages of survival were determined as follows: (((cfu/cfu~t\ =\ 0~))/((cfu/cfu~t\ =\ 0~)~control~)\*100)-100. To avoid effects of individual blood differences, survival was normalized to the reference strain included in each set of experiments. Each strain was tested 3 times. Data were represented as averages ± standard deviations.
Results {#sec010}
=======
Virulent strains of *S*. *cerevisiae* must have specific attributes which contribute to their potential to cause opportunistic infections. To search for specific features of virulent strains of *S*. *cerevisiae*, we studied their genome structures by microarray CGH. This technique allowed us to identify those genes with copy number variations (CNV) in the opportunistic strains compared with non-virulent strains. We selected three different opportunistic *S*. *cerevisiae* strains (D14, 60, 102) which have been shown to exhibit high levels of virulence *in vitro* and *in vivo*, presented enhanced survival rates in blood, increased burdens in brain and kidney, and were able to kill mice \[[@pone.0122382.ref013]\], \[[@pone.0122382.ref016]\], \[[@pone.0122382.ref018]\]. Since the ploidy of the different strains is important for interpreting CGH data, we analyzed ploidy by flow cytometry. The results suggested that strains 102 (2.00 ± 0.07) and D14 (2.16 ± 0.07) are diploids (considering 20% deviation plausible due to experimental error), whereas the result of strain 60 (2.41 ± 0.05) indicate that it is a diploid strain with additional copies of some chromosomes. The labeled DNA from the four strains was hybridized with DNA from control strain S288C to a standard cDNA microarray containing cDNA of the S288C genome. The signals obtained were filtered and normalized. The whole dataset is deposited in GEO Database (accession number GSE53440) using standard MIAME criteria.
Comparative genomic hybridization of opportunistic strains {#sec011}
----------------------------------------------------------
When we analyzed the CGH data in a genome-wide graph with all the strains, including control strain W303 ([Fig. 1A](#pone.0122382.g001){ref-type="fig"}), we observed a significant amount of CNV in the subtelomeric regions (22.8%). When we analyzed the absolute signal intensity of the genes in relation to the distances to telomeres ([Fig. 1B](#pone.0122382.g001){ref-type="fig"}), we observed that CNV significantly accumulated in the first 30 kb, which matches the subtelomeric region. The sliding window analysis, which averaged signal intensity (1 Kb), also revealed a significant increase in the subtelomeric region. These observations are consistent with previous CGH studies which compared the genome of *S*. *cerevisiae* strains of different origins \[[@pone.0122382.ref026]\], \[[@pone.0122382.ref027]\]. As expected, W303, whose genome is 85% isogenic to S288C \[[@pone.0122382.ref028]\], was the strain that showed less CN (50) as compared to the opportunistic strains D14 (180), 102 (334) and 60 (430),with differences found mainly in the number of increased CN ([Fig. 1C](#pone.0122382.g001){ref-type="fig"}). To confirm these data, we tested for the presence of W303 in the sequenced genome ([www.yeastgenome.org](http://www.yeastgenome.org/)) of the first 25 genes with less relative intensity data, and most of them \[[@pone.0122382.ref028]\] were absent. These results also confirmed that some genes, previously described as highly variable among strains of different sources \[[@pone.0122382.ref026]\], \[[@pone.0122382.ref027]\], for example, members of the *PAU*, *MAL*, *AAD* and *FLO* gene families, the *ASP3* and *ENA* clusters, the *SNO/SNZ* and *SOR/MPH* genes and other genes, i.e., *DAK2* or *AGP3*, showed altered CN in opportunistic strains. Among these genes, whose numbers had increased in opportunistic strains, we found several genes related to DNA damage repair, such as *DDR48*, *CAC2* and *RRD1* in strain 102 and *RAD18*, *POL4*, *MSH2* and *IMP2* in strain 60. One interesting result was the strong increase in the CNV of *PRB1*, which encodes a protease whose protein abundance increased after DNA damage in strains 60 and 102. Another observation was the increase in the CNV for gene *YAR068W*. An overexpression of this uncharacterized gene has been seen causing enhanced resistance to antifungal drugs, such as ketoconazole, miconazole and benomyl \[[@pone.0122382.ref029]\].
{#pone.0122382.g001}
We searched for genes with CNV in all investigated strains and found that 45 genes were common for the opportunistic yeasts ([Fig. 1D](#pone.0122382.g001){ref-type="fig"}). A hierarchical clustering of this core set of CNV-enriched genes ([Fig. 2A](#pone.0122382.g002){ref-type="fig"}) revealed a clear difference between opportunistic and laboratory strains. All except 14 transposon genes and two other genes, the genes showed opposite or different CNV values in the opportunistic strains as compared to W303. Twenty-seven of these 45 genes presented decreased CN, and four increased and 14 variable values among the different strains. Interestingly, 31.1% of the core genes were transposons, a fact that reflects one of the most important differences observed when comparing genomes of different strains. Next, we focused on those genes showing an altered CN in the opportunistic strains when compared to laboratory strains. Among this group of genes, we focused our study on the *CUP1* cluster, which displayed a decreased CN in the opportunistic strains when compared with laboratory strains. The SAM cluster and *IMD2* had an increased CN in the opportunistic strains *vs*. laboratory strains.
{#pone.0122382.g002}
Genetic variability in the *CUP1* cluster {#sec012}
-----------------------------------------
The *CUP1* cluster contains a duplication of the *CUP1* gene flanking an uncharacterized ORF (*CUP1-1*, *YHR054C* and *CUP1-2*). This cluster shows a decreased CN in the opportunistic strains when compared with laboratory strains ([Fig. 1E](#pone.0122382.g001){ref-type="fig"}). High variability in CN has been described for this cluster, which is probably the result of recombination events between duplicates \[[@pone.0122382.ref027]\]. Since the gene copy number correlates with Cu resistance \[[@pone.0122382.ref030]\], we studied the growth of both, the opportunistic strains and W303, in minimal media with a Cu concentration range in order to validate the CGH analysis with phenotypical data. The results obtained ([Fig. 2B](#pone.0122382.g002){ref-type="fig"}) confirm the CGH data as none of the opportunistic strains was able to grow, not even at the lowest Cu concentration. Conversely, lab strain W303, which has a larger number of *CUP1* copies, was able to grow with 1 mM of Cu.
Alteration in S-adenosylmethionine metabolic genes {#sec013}
--------------------------------------------------
Two of the genes with an increased CN in the opportunistic *S*. *cerevisiae* strains were *SAM3* and *SAM4*, which encode a high-affinity S-adenosylmethionine (SAM) permease and an S-adenosylmethionine-homocysteine methyltransferase, respectively. We studied the genomic content of *SAM3* in the different strains with qPCR in more detail ([Fig. 2C](#pone.0122382.g002){ref-type="fig"}). The results confirm the CGH data and suggest at least one more copy of the gene in strain D14 and two more copies in strains 60 and 102, whereas no variation was observed in haploid strain W303. These enzymes are involved in the utilization of SAM as a sulfur source, and also in glutathione (GSH) biosynthesis, which is the main antioxidant molecule in yeast cells. As the oxidative stress response is very important for opportunistic yeasts to survive in human blood, we studied if these strains were capable of increasing the GSH content using SAM as a sulfur source. However, no differences were observed (results not shown). In fact, *SAM3* and *SAM4* are located in a subtelomeric cluster that has been shown to have higher CNs in other strains \[[@pone.0122382.ref031]\]. Therefore, we cannot rule out that this genomic alteration is not a specific feature of the opportunistic strains.
Genetic changes in the *de novo* purine nucleotides biosynthetic pathway {#sec014}
------------------------------------------------------------------------
We focused our study on *IMD2*, a gene showing increased CNs in the three opportunistic strains compared with laboratory strains ([Fig. 2A](#pone.0122382.g002){ref-type="fig"}). This gene belongs to the *IMD* gene family (*IMD1-4*), which code for Inosine monophosphate dehydrogenases (IMD), enzymes which catalyzes the rate-limiting step in the *de novo* GTP biosynthetic pathway. We confirmed the increased CNs of *IMD2* by qPCR, and showed at least two more copies of the gene in strain 102 and three more copies in strains 60 and D14 ([Fig. 2C](#pone.0122382.g002){ref-type="fig"}). Similar alterations were monitored for its closely related gene, *IMD3*, where increased CN was observed in strains 102 and D14 ([Fig. 2A](#pone.0122382.g002){ref-type="fig"}), but not in strain 60. These results were also confirmed by qPCR, showing at least one more copy of *IMD3* in strains 102 and D14 ([Fig. 2C](#pone.0122382.g002){ref-type="fig"}). No signal was detected for *IMD3* amplification in strain 60.
We wondered if the increased CNs observed for the *IMD* genes in the opportunistic strains could be a result of their adaptation or ability to survive in human environments and whether this could be associated with their ability to cause infections. *IMD2* is a subtelomeric gene, which has been correlated with altered CNs in strains of different origins. In fact, other genes located next to *IMD2* exhibit CNVs ([Fig. 1E](#pone.0122382.g001){ref-type="fig"}). In contrast, *IMD3* is localized at the right arm of chromosome XII and is not close to any ribosomal genes, transposons or any other genomic feature that are known to favor recombination or genome rearrangements. Furthermore, no other genes close to *IMD3* have been shown to have altered CNs ([Fig. 1E](#pone.0122382.g001){ref-type="fig"}). We studied the CNVs of the other genes implicated in the *de novo* biosynthesis of the purine nucleotides pathway ([Fig. 2D](#pone.0122382.g002){ref-type="fig"}) and we observed increased CN for other genes, such as *GUK1* in strain 60 and *RNR4* in 102. No significant increased CN was observed in the W303 strain and no significant decrease in any gene related to the purine nucleotides pathway was found. All these data suggest that this pathway might be of special importance for the opportunistic strains to develop their specific abilities.
In order to test the role of the *de novo* biosynthesis of the purine nucleotides pathway in *S*. *cerevisiae* virulence, we wanted to construct a knock out mutant that is unable to use this pathway within the genetic background of the opportunistic strain D14. Deleting the *IMD* genes, which are redundant among them and showed increased CN, can be a complex approach (around 10--14 copies can be found, 2 of each IMD1-4 gene and extra copies depending on the strain). Instead of we selected gene *GUA1* of D14, because Gua1p catalyzes the subsequent step of the Imd enzymes, and because no increased copy numbers of *GUA1* exist in D14. In fact, and as expected, the deletion of two *GUA1* copies in the diploid strain D14 led to a *GUA1* knock out mutant (D14Δgua1). We also constructed a reconstituted *GUA1* (D14GUA1) strain as a control by transforming the D14Δgua1 mutant with an episomal plasmid containing *GUA1*.
Next, we evaluated the virulence of these two strains and the wild-type D14 in murine infection models ([Fig. 3](#pone.0122382.g003){ref-type="fig"}). Since *S*. *cerevisiae* infections have been described mainly in immunodeficient patients, we used a mice model of systemic infection with cyclophosphamide-induced neutropenia, which has been previously been used in *S*. *cerevisiae* and in other pathogenic organisms such as *C*. *albicans* \[[@pone.0122382.ref016]\], \[[@pone.0122382.ref024]\], \[[@pone.0122382.ref032]\]. The results revealed that only 15% of mice infected with strain D14 survive until day 22, whereas D14Δgua1 was unable to cause death in the mice. The *GUA1*-reconstituted strain caused a similar pattern (no significantly different, *p*-value\<0.05) of mice survival as compared to the parental strain, but was moderately reduced in virulence (50% survival of mice). On the contrary, The *GUA1*-reconstituted strain and D14Δgua1 strain were significantly different (*p*-value\<0.05) to the wild type strain.
{#pone.0122382.g003}
Decreased survival in D14Δgua1 mice can be a consequence of low guanine levels in the host. However, an increased dNTP pool which could be used for DNA repair. An increased DNA repair may be required during systemic infections since oxidative damage of yeast DNA is one of the key events which occur after blood infection. To test this hypothesis, we studied yeast growth in the presence of mycophenolic acid (MPA), an inhibitor of Imd enzymes. In fact, the expression of *IMD* genes is induced by mycophenolic acid to result in resistance to the drug \[[@pone.0122382.ref023]\], \[[@pone.0122382.ref033]\], \[[@pone.0122382.ref034]\]. The results ([Fig. 4A](#pone.0122382.g004){ref-type="fig"}) reveal that the opportunistic strains are able to grow in the presence of 7 μg/ml of MPA, whereas the growth of laboratory strains S288C and W303 was strongly affected. These results phenotypically confirm the predictions made based on the CGH and qPCR data. Next, we tested the growth of the strains in the presence of the DNA damaging agent methyl methanesulfonate (MMS), which produces alkylated DNA that is poorly replicated by DNA polymerases and must be efficiently repaired. We observed that the laboratory yeast strains S288c or W303 were highly sensitive to MMS (0.004%) treatment, whereas the opportunistic yeasts strains were much less affected ([Fig. 4](#pone.0122382.g004){ref-type="fig"}).
{#pone.0122382.g004}
In order to study the role of the *novo* purine nucleotide biosynthesis pathway when yeast cells face human phagocytes, we quantified fungal survival after *ex vivo* human blood infection when neutrophils attack yeast cells and produce reactive oxygen species. We studied yeast survival after 90 min of blood incubation by colony counting on rich media. This experimental set up was already used to show increased survival of opportunistic yeasts compared to other strains in blood infections \[[@pone.0122382.ref025]\]. Here we observed that addition of guanine to rich media increased survival in the opportunistic strain D14, but also in the laboratory wild-type strains BY4741 and W1588 (selected to compare with their isogenic mutants, see below) ([Fig. 4B](#pone.0122382.g004){ref-type="fig"}). In order to ensure that increased survival was not due to an additional and more efficient nitrogen source, we performed the same experiment adding the equivalent amount of nitrogen. However, no increased survival of yeast cells was found (data not shown). In this assay, we also analyzed the behavior of two mutants with lower dNTP levels due to mutation of genes which influence the expression of the last enzyme of the dNTP biosynthetic pathway, ribonucleotide reductase (RNR). One mutant tested was Δdun1, which lacks the *DUN1* gene. *DUN1* encodes a key activator of RNR in response to DNA damage. The Δ*dun1* mutant was previously shown to have dysfunctions in RNR activation and decreased dNTP pools after DNA damage \[[@pone.0122382.ref034]\], \[[@pone.0122382.ref035]\]. The other mutant tested was rnr1\*, containing a point mutation that disrupts RNR activation in response to DNA damage \[[@pone.0122382.ref019]\]. The results ([Fig. 4B](#pone.0122382.g004){ref-type="fig"}) indicate that both mutants were less able to survive in human blood as compared to their isogenic parental wild type strains. Especially rnr1\* had a strongly (more than 70%) reduced ability to survive the challenge of blood cells. These data suggest that *de novo* dNTP biosynthesis plays a role in yeast survival after DNA damage produced by neutrophils or other blood cells.
Discussion {#sec015}
==========
To elucidate the genetic determinants that enable certain strains of *S*. *cerevisiae* to cause infection, we used a comparative genetics approach. We compared the genomes of strains showing high level of virulence attributes (pseudohyphal and 42°C growth, phosphatases and proteases secretion, and the ability to infect and kill mice) independently of the isolation origin \[[@pone.0122382.ref013]\], \[[@pone.0122382.ref016]\], \[[@pone.0122382.ref018]\] with the gene content of non-virulent laboratory strains using microarrays. Many strains of *S*. *cerevisiae* which have been isolated in clinical settings present very low levels of virulence in mice infection models and other strains, such as strain D14, isolated from a dietetic supplement, show a relative high level of virulence in different infection models. Thus, we propose that the term "opportunistic strain" is more accurate and useful to categorize these strains rather than "clinical strains", since not all clinical strains cause infections. We define opportunistic *S*. *cerevisiae* strains as those strains which can, in contrast to most other strains, cause infections in mice and can kill mice. These strains also survive better in human blood infection models than other strains, which may enable these strains to disseminate across the body and to reach organs with adequate propagation conditions under certain circumstances \[[@pone.0122382.ref025]\]. Thus, increased blood survival can be a key feature to distinguish these opportunistic yeasts strains from others. Previous studies by our group have demonstrated that opportunistic strains show a specific transcription pattern after human blood infection, which indicate a specific oxidative stress response and DNA damage repair \[[@pone.0122382.ref025]\]. DNA damage is apparently produced by reactive oxygen species generated mainly by neutrophils and a proper DNA repair may be a specific characteristic of opportunistic strains. In this study, we aimed at further characterizing these strains by using a genomic approach.
A common mechanism described for human microbial pathogens is to increase dNTP pools in order to repair DNA damage caused by the oxidative burst of phagocytes. For example, the importance of a *de novo* nucleotide biosynthetic pathway for phagocyte survival has been demonstrated for bacterial pathogens such as *Salmonella enterica* \[[@pone.0122382.ref036]\], *Bacteroides fragilis* \[[@pone.0122382.ref037]\], *Pseudomonas aeruginosa* \[[@pone.0122382.ref038]\], *Staphylococcus aureus* \[[@pone.0122382.ref039]\], *Streptococcus pyogenes* \[[@pone.0122382.ref040]\], *Bacillus anthracis* and *Escherichia coli* \[[@pone.0122382.ref041]\]. The relevance of *de novo* nucleotide biosynthetic pathways has also been implicated for fungal pathogens like *Cryptococcus neoformans* \[[@pone.0122382.ref042]\] or *Candida albicans* \[[@pone.0122382.ref043]\], \[[@pone.0122382.ref044]\] and nucleotide biosynthetic pathways have been discussed as a target for antifungal compounds \[[@pone.0122382.ref045]\]. The results obtained in our study highlight a specific genomic fingerprint of opportunistic strains that increase the gene copy number of genes involved in the *de novo* biosynthesis of purine nucleotides. We conclude that *S*. *cerevisiae* opportunistic strains, like other human pathogens, have the enhanced ability to produce dNTPs, the substrates that are used by DNA repair machineries. This is supported by the observation that mutantsΔdun1 and rnr\*, which showed lower dNTP pools and defects in RNR activation after DNA damage \[[@pone.0122382.ref019]\], \[[@pone.0122382.ref034]\], \[[@pone.0122382.ref035]\], display poorer survival in human blood as compared to wild type strains. Other genes involved in DNA repair are also highlighted by our CGH study. One interesting example is *IMP2*, a transcription factor gene, which is activated in response to DNA damage produced by oxidative stress and which regulates the expression of DNA repair genes. Thus, an enhanced dNTP production pathway combined with efficient DNA repair machinery may be a key feature to define opportunistic *S*. *cerevisiae* yeasts. Our study opens up the possibility the *de novo* purine biosynthetic pathway is a possible alternative target to treat *S*. *cerevisiae* infections, as previously proposed for other fungal pathogens such as *C*. *albicans* \[[@pone.0122382.ref043]--[@pone.0122382.ref045]\].
In general terms, increased dNTP biosynthesis to supply DNA repair seems to be a general strategy used by some pathogens to survive human macrophages oxidative burst. However, those genes that have been shown to be associated with virulence differ among the different organisms. The role of Imd activity, which is the rate-limiting step in the *de novo* purine biosynthetic pathway, has been stressed in the fungal pathogen *C*. *neoformans* \[[@pone.0122382.ref042]\]. In *C*. *albicans*, other genes, or genes associated with the same pathway, for example *ADE2*, *ADE7* or *GUA1*, have also been implicated in virulence \[[@pone.0122382.ref043]\], \[[@pone.0122382.ref044]\]. In contrast, studies dealing with bacterial pathogens have proposed that RNR activity is most important, since it is the key rate-limiting step in deoxyribonucleotides formation \[[@pone.0122382.ref037]--[@pone.0122382.ref039]\]. Here we propose that both *IMD* genes and RNR are important for opportunistic yeasts to cause infections. The fact that opportunistic yeasts have increased *IMD* gene copy numbers suggests that the level of such activity is crucial, although we also observed that diminished RNR activity dramatically affects survival in blood infections. Therefore, both activities, Imd and RNR, influence the ability of opportunistic *S*. *cerevisiae* strains to overcome an oxidative burst attack by phagocytes and to survive at least transiently in blood. This property may enable these strains to reach other organs and to proliferate.
New knowledge on opportunistic yeast is essential to understand the intrinsic differences between these strains and the standard strains used in the food industry considered safe. This will help in the selection of strains that are suitable to produce foods. Some criteria has been previously suggested to avoid opportunistic yeast in foods as growth above 37°C \[[@pone.0122382.ref046]\], \[[@pone.0122382.ref047]\] or increased oxidative stress resistance \[[@pone.0122382.ref025]\]. Here we propose new criteria to discard opportunistic strains for the food industry: increased resistance to DNA damage agents or increased resistance to *de novo* dNTP biosynthesis inhibitors. A combination of all this criteria will be useful to select safer strains for the production of human foods.
We thank Mercedes Tamame for kindly sending us pGUA1 plasmid. We thank Sergi Puig for kindly giving us dun1 strain and Rodney Rothstein for the rnr1\* strain. S. Llopis was a recipient of a FPU fellowship from the Spanish Ministry of Education and Science (MEC). R. Pérez-Torrado was supported by the JAEDOC postdoctoral program (CSIC) cofounded by ESF. This work was supported by PROMETEO grant (Project PROMETEOII/2014/042) from the Generalitat Valenciana.
[^1]: **Competing Interests:**The authors declare that they have no competing interests.
[^2]: Conceived and designed the experiments: RPT AQ. Performed the experiments: RPT BP RGP SLL. Analyzed the data: RPT AQ. Contributed reagents/materials/analysis tools: AQ BH. Wrote the paper: RPT AQ BH.
[^3]: Current address: Duke University School of Medicine. Department of Pharmacology and Cancer Biology. Durham, North Carolina, United States of America
| {
"pile_set_name": "PubMed Central"
} |
All relevant data are within the paper and its Supporting Information files.
Introduction {#sec005}
============
In clinical practice, computed tomography (CT) has become an essential imaging modality for the diagnosis, sequential follow-up, and assessment of treatment response of intracranial diseases. Particularly, CT is commonly performed in the field of neuroradiology as an initial screening tool to evaluate acute intracerebral haemorrhage (ICH) \[[@pone.0186024.ref001]\]. Although most primary ICHs originate from chronic hypertension or cerebral amyloid angiopathy, secondary ICHs are associated with intracranial vascular abnormalities (e.g., arteriovenous malformations and aneurysms), tumour, coagulopathy, or trauma \[[@pone.0186024.ref002]\]. Therefore, it is important to identify the underlying causes of ICH for a prediction of prognosis on the basis of neuroimaging studies for appropriate patient management. However, it may be difficult to assess the exact cause of ICH because high attenuation from acute ICH can mask true contrast enhancement or active extravasation of contrast media \[[@pone.0186024.ref003]\].
In many previous studies, dual-energy CT (DECT) was used to differentiate acute ICH and iodinated contrast media using the attenuation difference between two different X-ray spectra \[[@pone.0186024.ref004]--[@pone.0186024.ref007]\]. However, DECT with two different X-ray spectra has an inherent limitation that it does not demonstrate authentic spectral data at all energy levels of a polychromatic X-ray source. In contrast to DECT, dual-layer detector CT obtains the spectral data simultaneously during a single scan. It provides spectral data without any change to the radiologic workflow, and reduces the influence of patient motion and cross-scatter and consequently improve overall image quality. Previous studies have the use of virtual monoenergetic reconstructions to image iodine inserts using dual-layer detector spectral CT with phantoms \[[@pone.0186024.ref008], [@pone.0186024.ref009]\]. To the best of our knowledge, there is no objective study that has evaluated acute ICH on dual-layer detector spectral CT with spectral analyses using commercial software (a spectral diagnostic suite \[SpDS\]). Therefore, the purpose of this study was to investigate the clinical feasibility of spectral analyses using dual-layer detector spectral CT in the evaluation of acute ICH.
Materials and methods {#sec006}
=====================
Study population {#sec007}
----------------
The Institutional Review Board (IRB) of Gyeongsang National University Changwon Hospital approved this retrospective study and waived the need for written informed consent from the participants. Retrospective data collection and analysis were performed according to our IRB guidelines. The patient's record and information were anonymized and de-identified prior to analysis. A review of a database at our institution identified consecutive patients with acute ICH who were admitted to the emergency department (ED) and underwent intracranial CT angiography (CTA) on a spectral CT scanner between June 2016 and October 2016. We then selected thirty-six consecutive patients with acute ICH using electronic medical records (EMRs) and a picture archiving and communicating system (PACS). We enrolled 30 of the 36 patients; six were excluded because of inadequate medical records (n = 3) or poor image quality from uncontrolled motion artefacts (n = 3). The final 30 patients who were included in this study comprised four women and 26 men (age range, 25--83 years; mean age, 59.3 years).
Imaging acquisition and spectral data processing {#sec008}
------------------------------------------------
CT imaging was performed on a dual-layer detector CT unit (IQon Spectral CT, Philips Healthcare, Best, The Netherlands). A true noncontrast (TNC) image and CTA in the arterial phase with spectral data were obtained for all patients. Conventional 120 kVp images and spectral images (spectral level 3) were reconstructed from the same dataset. All images were spatially registered and linked to each other in three dimensions because a single acquisition of dual-layered spectral CT was used to reconstruct both polychromatic and virtual monoenergetic (MonoE) datasets. The following acquisition parameters were applied: 120 kVp; 200 mAs; collimation, 64 × 0.625 mm; pitch factor, 0.985; rotation time, 0.5 seconds; field-of-view (FOV), 250 mm; slice thickness, 0.8 mm; slice increment, 0.4 mm; and scan time, 2.7 seconds. For the enhancement of intracranial arteries, 80 mL of contrast medium (Pamiray 370; Dongkuk Pharm., Seoul, Korea) was injected intravenously at a flow rate of 4 mL/s using a power injector (Covidien CT OptiVantage DH injector; Mallinckrodt Pharm., Dublin, Ireland). A bolus tracking method was used routinely to achieve optimal synchronization of the contrast medium flow and scanning. Once the injection is started, the bolus tracking software measures attenuation values within one internal carotid artery and a spiral scan is automatically started as soon as the threshold of 100 Hounsfield units (HU) is exceeded.
A diagram of the data acquisition and post-processing pathway for spectral CT images is shown in [Fig 1](#pone.0186024.g001){ref-type="fig"}. Conventional anatomical images were generated by the summation of raw data from the lower and upper layers for all patients. Raw data from the lower and upper layers were separated into photoelectric and Compton scattering sinograms and reconstructed into a spectral base image (SBI) from which all spectral results can be derived. Spectral results are acquired within a single scan without the need for special modes because the acquisition of spectral data is dependent on the dual-layer detector rather than the X-ray tube. All spectral results were displayed in the same manner as conventional CT images using commercial SpDS software. The spectral image reconstruction for the retrospective spectral analysis included MonoE, virtual noncontrast (VNC) and iodine overlay fusion images. On MonoE images, each image series was obtained at an energy level represented by a value in the range from 40 keV to 200 keV. The appearances of MonoE images change as keV values change, even when the window and level settings are unchanged. The different appearance of the MonoE images is caused by the approximation to the iodine k-edge at 33 keV because attenuation coefficients of materials decrease as monoenergy increases to values above their k-edge energies. The pixels in these images represent HU. Principles of the material decomposition mechanism are schematically demonstrated in [Fig 1](#pone.0186024.g001){ref-type="fig"} \[[@pone.0186024.ref010]\]. VNC images were obtained for comparison with TNC images. The iodine contrast material detected by the spectral analyses was displayed in red on iodine overlay fusion images that were obtained from arterial-phase images. Volume-rendered (VR) and maximum intensity projection (MIP) images of the intracranial arteries were obtained using a bone removal method.
{#pone.0186024.g001}
Imaging analyses {#sec009}
----------------
An attending neuroradiologist (H.J.B. with 7 years of experience in brain, head, and neck imaging) interpreted the images while blinded to the clinical history of each patient. For all patients, the initial assessment was performed using conventional anatomical CT images. Then, an additional spectral analysis was performed to detect contrast enhancement in or adjacent to the acute ICH. All spectral data acquired during the scan were easily accessed from the PACS for the retrospective spectral data analysis. To compare with TNC images, we evaluated VNC images, and image quality of VNC images was graded according to the following items by visual assessment: (a) differentiation of grey-white matter at the level of the lateral ventricles; (b) demarcation of basal ganglia at the level of foramina of Monro; (c) demarcation of the cerebral sulci; and (d) visualization of intracranial vessels. For each of these items, the dichotomization simplified the image quality assessments to inadequate versus sufficient. In addition, we analysed spectral plots to evaluate material differentiation ([S1 File](#pone.0186024.s001){ref-type="supplementary-material"}). To acquire a spectral plot, the same neuroradiologist manually drew region of interest (ROI) defining the enhancing tumour portion or active extravasation on a selected representative MonoE image while avoiding traversing vessels and tumour necrosis. In addition, intracranial arteries were reviewed on VR and MIP images to evaluate the vascular cause of the acute ICH.
Results {#sec010}
=======
Study population {#sec011}
----------------
The clinical data of the study patients are summarized in [Table 1](#pone.0186024.t001){ref-type="table"}. CTA was obtained successfully for all patients without any side effect or complication. In the 30 patients included in the study, various neurologic symptoms were noted, including headache (n = 12, 40%), abrupt mental status change (n = 8, 26.7%), motor weakness (n = 6, 20%), sensory change (n = 3, 10%), and dizziness or vertigo (n = 1, 3.3%). The locations of ICH were as follows: the thalamus (10/30, 33.3%); basal ganglia (9/30, 30%); lobar pattern (7/30, 23.3%); pons (2/30, 6.7%); and cerebellum (2/30, 6.7%). The most common cause of acute ICH was chronic hypertension (18/30, 60%) followed by trauma (5/30, 16.7%), brain tumour (3/30, 10%), Moyamoya disease (2/30, 6.7%), and haemorrhagic diathesis by anticoagulation therapy (2/30, 6.7%). A diagnosis of hypertensive ICH or haemorrhagic diathesis-related ICH was based on the clinical information and resolution of the lesions on follow-up CT or magnetic resonance (MR) imaging. The CT angiographic spot sign within the ICH was detected in four patients (13.3%, [Fig 2](#pone.0186024.g002){ref-type="fig"}). All three tumour cases were diagnosed as metastatic tumours from lung cancer (n = 2) or hepatocellular carcinoma (n = 1) on the basis of the clinical background and additional or follow-up imaging, including CT, MR, and positron emission tomography (PET)-CT imaging. The ICHs from metastatic tumours showed a vividly enhancing solid tumour portion on spectral CT images ([Fig 3](#pone.0186024.g003){ref-type="fig"}).
{#pone.0186024.g002}
{#pone.0186024.g003}
10.1371/journal.pone.0186024.t001
###### Clinical characteristics of patient population.
{#pone.0186024.t001g}
Characteristics
------------------------- ------------------------------------
Age (years) 59.3 ± 13.7
Gender Female: Male = 4 (13.3): 26 (86.7)
Location of ICH
Lobar 7 (23.3)
Basal ganglia 9 (30)
Thalamus 10 (33.3)
Pons 2 (6.7)
Cerebellum 2 (6.7)
Cause of ICH
Hypertension 18 (60)
Trauma 5 (16.7)
Moyamoya disease 2 (6.7)
Tumour 3 (10)
Haemorrhagic diathesis 2 (6.7)
Spot sign
Presence 4 (13.3)
Absence 26 (86.7)
Note.---Data presented in parentheses are percentage of each item. AVM, arteriovenous malformation; ICH, intracerebral haemorrhage.
Image analyses {#sec012}
--------------
Of these 30 patients, the results of dichotomized assessment regarding to VNC images as follows: (a) differentiation of grey-white matter at the level of the lateral ventricles; inadequate (7, 23.3%) vs. sufficient (23, 76.7%), (b) demarcation of basal ganglia at the level of foramina of Monro; inadequate (9, 30%) vs. sufficient (21, 70%), (c) demarcation of the cerebral sulci; inadequate (12, 40%) vs. sufficient (18, 60%) and (d) visualization of intracranial vessels; inadequate (13, 43.3%) vs. sufficient (17, 56.7%). Twelve cases in the item (c) revealed inconspicuous delineation of cerebral sulci, which reduced overall image quality and interpretability. According to item (d), 13 inadequate cases showed suboptimal iodine suppression in the subcalvarial spaces on VNC images, which can make distinguishing blood vessels from extraaxial haemorrhages difficult. However, these blood vessels were easily identified during assessment because of their serpentine continuity along the predictable vascular course.
In the analyses of MonoE images, they tended to enhance visualization of the contrast-enhanced tumour portion or active contrast extravasation at low keV values and reduced beam hardening at high keV values; thus, the conspicuity of the lesions was increased. In particular, use of MonoE images acquired at keV values less than 70 keV allowed better visualization of contrast-enhancing structures such as blood vessels, contrast extravasation, or enhancement of the tumour portion on the visual assessment. We also found that overall image noise increased as keV of MonoE image decreased. In addition, spectral plots allowed discrimination between haemorrhage and iodine in patients with brain tumours and active extravasation. On the spectral plots, iodine showed the highest HU values on 40 keV MonoE images and decreased HU values with a steep slope on higher keV MonoE images. However, the haemorrhage and brain parenchyma demonstrated low HU values with an unchanged slope (Figs [2](#pone.0186024.g002){ref-type="fig"} and [3](#pone.0186024.g003){ref-type="fig"}). At 40 KeV, iodine has more than a two-fold higher HU value compared to that at 120 keV. At higher values above 100 keV, all materials such as iodine, cerebrospinal fluid, the brain parenchyma, and haemorrhage had similar HU values. In addition, the intracranial arteries also became more obvious with the effective suppression of bone beam hardening by the skull vault on MonoE images at higher keV values (100--120 keV) during the spectral analysis.
Discussion {#sec013}
==========
The overall worldwide incidence of ICH ranges from 10 to 20 cases per 100,000 in the population and increases with age \[[@pone.0186024.ref002], [@pone.0186024.ref009]\]. ICH can be classified as either primary or secondary depending on the underlying aetiology of the bleeding \[[@pone.0186024.ref001], [@pone.0186024.ref002]\]. Therefore, it is important to evaluate the underlying causes of ICH for appropriate patient management. Although unenhanced CT is the standard imaging modality for the detection of ICH in clinical practice, contrast-enhanced CT, CTA, and MR imaging may be needed to further detect information regarding ICH \[[@pone.0186024.ref001], [@pone.0186024.ref011], [@pone.0186024.ref012]\]. It is well known that the presence of active contrast extravasation within an ICH on CTA has been associated with haematoma expansion and patient mortality \[[@pone.0186024.ref011], [@pone.0186024.ref012]\]. In addition, approximately 10% of ICH occurs from tumour bleeding, and it can be difficult to detect tumour enhancement because of the masking effect of the coexisting high attenuation from acute ICH \[[@pone.0186024.ref003]\].
In a previous study, DECT was used to differentiate high ICH attenuation from that of iodine by the attenuation difference between the two materials at different energies, which is based on the photoelectric effect and Compton scattering. While the photoelectric effect is strongly energy-dependent, Compton scattering is not. CT attenuation depends on the photon's energy, the atomic number, and the concentration of the material being scanned. Therefore, materials in the body can be differentiated by using different X-ray spectra, an effect that is particularly useful for materials with a large atomic number because of the photoelectric effect. Iodine has a high atomic number and shows a large attenuation difference between high- and low-energy X-rays, whereas haematomas do not. This was the basic mechanism used to differentiate iodine from acute ICH in previous studies using DECT \[[@pone.0186024.ref004]--[@pone.0186024.ref006]\].
However, to the best of our knowledge, there is no study that evaluated ICH using dual-layer detector spectral CT with spectral data analyses. In this study, we were able to investigate the clinical feasibility of dual-layer detector spectral CT for evaluating acute ICH. This technique is the most up-to-date commercial technical approach, which is based on an energy-resolving layer detector with a single polychromatic spectrum. In the dual-layer detector, the sensitivities of the two detectors are determined by the scintillator materials such as ZnSe or CsI in the upper layer and Gd~2~ O~2~S in the lower layer, which are related to the spectral resolution. In this model, spectral data were reconstructed at different virtual monoenergies, resulting in virtual MonoE images. These MonoE images reflect the HU values of the scanned region at one specific energy, in contrast to the HU values derived from a conventional polychromatic spectrum.
In the present study, all three brain tumours were correctly diagnosed based on the spectral analysis, and CTA spot signs suggesting active bleeding were more easily detected with increased conspicuity in the spectral analysis, especially on MonoE images taken at lower keV values (\< 70 keV) than on conventional images. In addition, spectral plots showed a simple visualization of material decomposition in patients with brain tumours and active extravasation. The result of the spectral plot is based on MonoE images and the spectral plot represents a change in HU for the ROI according to the change in keV. Therefore, the slope of the ROI in the spectral plot reveals material composition because the X-axis represents keV and the Y-axis represents HU. We suggest that the spectral plot can be a time-efficient analysis to differentiate materials without reviewing all MonoE images according to the predefined keV levels. Similar to the material characterization capabilities of DECT that have been previously studied \[[@pone.0186024.ref006], [@pone.0186024.ref011], [@pone.0186024.ref012]\], in this pilot study, we found that it is possible to differentiate ICH from coexisting iodinated contrast materials using spectral CT with a dual-layer detector. However, in this study, all CT images were evaluated by a single reader, and visual qualitative assessment was done with a small proportion of tumour and active bleeding. These factors may have affected the results of the current study in terms of objectivity. Therefore, additional studies are needed to clarify these issues.
Although we used visual assessment for image analyses, we also found that overall image noise increased as MonoE keV decreased which was documented in the previous studies \[[@pone.0186024.ref013], [@pone.0186024.ref014]\]. This is an inverse relation between improved visualization of the contrast enhancing structures and increased overall image noise at low keV, which can have an adverse effect on the image interpretations. However, this issue may be overcome by changing the keV value, i.e., MonoE images should be taken at 40 or 50 keV for evaluation of blood vessels or contrast-enhancing lesions, and higher keV MonoE or polychromatic images should be used for evaluation of other structures of interest.
In the current study, bone beam hardening artefacts were reduced in images taken at higher keV values. Therefore, it may provide additional information about intracranial lesions at the level of the skull base or intraosseous segments of the intradural and extradural arteries than does conventional CTA. In addition, CTA with dual-layer detector spectral CT provided anatomical information as well as the ability to characterize structures based on material content within a single scan. Although image post-processing with DECT is only conducted at a workstation, spectral data can be accessed easily from a PACS for retrospective, on-demand, spectral analysis to further investigate a particular ROI. Because of this benefit, there is no need to recall the patient for additional imaging in daily clinical practice. Therefore, we believe that these advantages of dual-layer detector spectral CT can be helpful in determining a time-efficient diagnosis in a clinical setting, especially pediatric, non-cooperative, and emergent patients.
Although we assessed overall image quality of VNC and TNC images using dichotomized scoring, the results were not sufficient to validate the image quality of VNC images as an alternative to TNC images owing to the small sample size and visual assessment by a single radiologist. In this pilot study, suboptimal image quality of VNC images was observed in 13 patients (43.3%) because of insufficient iodine suppression of leptomeningeal or bridging vessels in the subcalvarial spaces, which can mimic extraaxial haemorrhages in the acute trauma setting. While suboptimal iodine suppression usually occurred in the subcalvarial blood vessels, these were readily recognizable and were not confused with pathological condition because of their serpentine continuity along the predictable vascular course. The suboptimal demarcation of the cerebral sulci was another drawback of VNC images, which reduced overall image quality. This issue can decrease the accuracy during image interpretation in the clinical practice and may necessitate additional TNC images to confirm the diagnosis of subcalvarial pathologies such as subarachnoid haemorrhage or cortical microhaemorrhage. Further, additional studies having larger sample sizes and involving quantitative image assessment should be carried out in the near future to validate our results.
In this study, the radiation dose from a single spectral CT scan is approximately 1.2 mSv, which is lower than that from a typical conventional CTA scan and similar to that of a single set of DECT-CTA images \[[@pone.0186024.ref011], [@pone.0186024.ref015], [@pone.0186024.ref016]\]. The average radiation dose in this study was about 2.3 mSv because we routinely acquired TNC images for comparison with VNC images. Unfortunately, we could not show that TNC images were replaced with VNC images to reduce the spectral CT radiation dose. Therefore, we expect that CTA with spectral CT can be an ideal imaging modality for patients with acute ICH as it reduces both scanning time and radiation dose in addition to achieving further improvement of VNC image quality.
There are several limitations to this study. First, there was an unavoidable selection bias because of the study's retrospective nature. Second, a relatively small number of patients were included in this pilot study with a high proportion of hypertensive ICH. It may be influenced by the particular situation regarding to the early period of the official opening of our hospital. Therefore, our study had a weakness for generalization. Third, we did not investigate diagnostic accuracy by statistical analysis and perform a quantitative assessment because this pilot study was focused on early clinical experience with dual-layer detector spectral CT. Lastly, we used the image review of only one attending neuroradiologist. Therefore, further studies with larger sample sizes, various ICH aetiologies and multi-reader assessment of image quality are required for the validation of our results.
Conclusions {#sec014}
===========
In conclusion, it is feasible to evaluate acute ICH in the clinical settings using dual-layer detector spectral CT even though the image quality of VNC images is perceived to be inferior yet. In this study, the MonoE images taken at lower keVs were beneficial for depicting contrast enhancement; further, spectral plot might be helpful for material differentiation in patients with acute ICH. Spectral data analyses offer additional diagnostic information based on material decomposition with a time-efficient workflow because the appearance of the synthesized MonoE images changed as the predefined keV values changed. We expect that further research may reveal additional applications for dual-layer detector spectral CT and also validate its clinical performance.
Supporting information {#sec015}
======================
###### This is the minimized data set regarding to imaging analyses of this study.
(XLSX)
######
Click here for additional data file.
The authors would like to thank Minjae Park for his advice regarding to technical background of the dual-layer detector spectral CT into this work, and 'Elsevier Language Editing Service' for the English language review and editing; <http://webshop.elsevier.com/languageservices/languageediting/>.
[^1]: **Competing Interests:**The authors have declared that no competing interests exist.
| {
"pile_set_name": "PubMed Central"
} |
Background {#Sec1}
==========
Systemic sclerosis (SSc), also known as scleroderma, is a systemic autoimmune disease characterized by early inflammation and vascular injury followed by fibrosis of the skin and visceral organs \[[@CR1]--[@CR3]\]. Early SSc is dominated by immune cell activation, followed by secreting pro-inflammatory cytokines that activate endothelial cells and causes upregulation of adhesion molecules. The adhesion molecules in turn recruit immune cells. Continuously activated immune cells secrete profibrotic cytokines to activate fibroblasts and transform myofibroblast, which are caused by a complex series of signal pathways initiated by a number of profibrotic molecules, including transforming growth factor-β (TGF-β), connective tissue growth factor (CTGF), endothelin-1 (ET-1), interleukin (IL)-4, IL-6, and IL-13 \[[@CR4], [@CR5]\]. Currently, the scant treatment method is a serious challenge for SSc.
Macrophages and T cells (especially T helper cells) are involved in the process of SSc \[[@CR6]--[@CR8]\]. Monocyte and macrophage phenotypes are highly heterogeneous and can be dynamically regulated by the microenvironment in SSc \[[@CR9], [@CR10]\]. Monocytes can differentiate into macrophages in response to various stimuli dependent on the tissue microenvironment after infiltrating the tissue. The differentiated macrophages can be classified as classically activated inflammatory macrophages (M1) and alternatively activated tissue profibrotic macrophages (M2) \[[@CR11]\]. M1 macrophages can promote tissue inflammation by producing pro-inflammatory cytokines, including tumor necrosis factor (TNF)-α, IL-6, and IL-12. M2 macrophages can induce or maintain tissue fibrosis by producing profibrotic cytokines, including TGF-β, IL-4, and IL-13. T helper (Th) 1 cells and their secreted interferon-γ (IFN-γ) are effective anti-fibrosis factors \[[@CR12]\]. Th17 cells promote inflammation by secreting IL-17A, IL-21, and IL-22, resulting in the development and progression of SSc \[[@CR13]\]. In reverse, in SSc, Th2 cells promote the polarization of M2 by producing IL-4 and IL-13, while M2 macrophages produce IL-13 to promote Th2 differentiation, creating a positive feedback loop \[[@CR14]\]. Tregs can produce TGF-β1 and IL-10, which are essential for maintaining self-tolerance and preventing autoimmunity \[[@CR15]\]. Overall, regulating Th1/Th2/Th17 cell balance and inhibiting macrophage polarization are effective strategies for the treatment of SSc.
It is reported that intercellular adhesion molecule 1 (ICAM-1) and vascular cell adhesion molecule 1 (VCAM-1) elevated in SSc \[[@CR16]--[@CR18]\]. It has been demonstrated that ICAM-1 expression by fibroblasts and endothelial cells were induced by a number of the cytokines important in SSc pathogenesis (TNF-α, IFN-γ, IL-1β, and IL-17). Similarly, VCAM-1 expression can be induced by TNF-α in a dose-dependent manner \[[@CR19]\]. Once induced, ICAM-1 and VCAM-1 subsequently recruit and activate monocytes to repair damaged endothelial cells \[[@CR20], [@CR21]\]. Monocytes can then boost the co-stimulation and transmigration of inflammatory cells into the extracellular matrix (ECM) and contribute to dysregulated angiogenesis \[[@CR20], [@CR21]\]. When overexpressed, these adhesion molecules can be detected in a circulating soluble form and are considered markers of underlying endothelial activity and damage. Hence, ICAM-1 and VCAM-1 expression are associated with SSc disease activity and severity.
SAHH and its substrate S-adenosyl-l-homocysteine (SAH) are deeply involved in the process of transmethylation mediated by S-adenosylmethionine (SAM). Immune cells are especially prone to methylation \[[@CR22]\], and modifications of DNA and protein by methylation are key factors in immune responses in the progression of inflammatory disease \[[@CR23], [@CR24]\]. Till now, three types of SAHH inhibitors have been described: the irreversible type I, type II inhibitors, and the reversible type III inhibitors. Because of the relatively long turnover rate of SAHH, the irreversible inhibitors manifest significant toxicity, whereas toxicity of type III inhibitors show greatly lowered yet still retain a parallel ability to block the enzyme's activity \[[@CR25]\]. The reversible type III SAHH inhibitor DZ2002 \[methyl-(adenin-9-yl)-2-hydroxybutanoate\] has been found to have an immunomodulatory function and to alleviate disease in several inflammatory and autoimmune animal models \[[@CR25]--[@CR29]\]. It is reported that DZ2002 prevented lupus-like disease from developing in both BXSB and MRL-Faslpr mouse models and DZ2002 ameliorated imiquimod-induced psoriasis-like skin lesions in mice via suppression of T cell-derived IL-17 \[[@CR24], [@CR30]\]. These hinted that methylation inhibition might be an approach for the treatment of autoimmune diseases.
Given that DZ2002 presents a prominent activity of anti-inflammatory and immunosuppression, DZ2002 may have potential therapeutic effects on SSc. To assess this hypothesis, we here investigated the effects of DZ2002 on dermal fibrosis of a BLM-induced mice model of SSc by focusing on the three major pathologic characteristics of SSc: fibrosis, immune abnormalities, and vasculopathy.
Methods {#Sec2}
=======
Mice {#Sec3}
----
Wild-type C57BL/6 mice were purchased from SLRC Laboratory (Shanghai, China). Female mice, 8 weeks old at the beginning of the BLM treatment, were used. All the experiments were conducted in accordance with the National Institutes of Health Guide for Care and Use of Laboratory Animals and were approved by the Bioethics Committee of the Shanghai Institute of Materia Medica **(**IACUC protocol \# 2018-10-TW-12 for C57BL/6 mice).
BLM-induced mice model of SSc {#Sec4}
-----------------------------
BLM (TargetMol, Boston, USA) was dissolved in phosphate buffer saline (PBS, Sigma-Aldrich, St. Louis, MO, USA) at the concentration of 300 μg/ml and sterilized by filtration. BLM (1.5 mg/kg) was injected subcutaneously into a single location on the shaved backs of C57BL/6 mice with an insulin needle. The injection was carried out consecutively for 4 weeks depending on the purpose of experiments.
Experimental design {#Sec5}
-------------------
DZ2002 was synthesized at Shanghai Institute of Materia Medica (Shanghai, China). For therapy, 50 female C57BL/6 mice were randomly divided into five groups as follows: normal (*n* = 10), vehicle (ddH2O, *n* = 10), prednisolone (PNS) 2 mg/kg (*n* = 10), DZ2002 100 mg/kg (*n* = 10), and DZ2002 50 mg/kg (*n* = 10). The immunosuppressive drug PNS was used as a positive control to evaluate the pharmacodynamic effects of DZ2002. Mice were treated with a once-daily dosing regimen by oral intragastric administration from 8 weeks to 12 weeks of age. The dosage (milligrams per kilogram) listed is the dose per administration, and the doses were administered each day at 10:00 a.m. DZ2002 was dissolved and PNS was dispersed in ddH2O (Fig. [1](#Fig1){ref-type="fig"}a). At the end of the experiment, mice were euthanized to obtain skin for analysis. Fig. 1DZ2002 attenuated skin fibrosis and TGF-β signaling activation in a BLM-induced SSc mice model. **a** Experimental design of BLM-induced SSc mice model and drug treatment. **b** Total skin thickness of the back of each group of mice (*n* = 10 per group). **c** Representative skin sections stained with hematoxylin and eosin stain (top, original magnification, × 100), Masson's trichrome stain (middle, original magnification, × 100), and α-SMA immunohistochemical staining (bottom, arrows represent α-SMA, original magnification, × 100) (scale bar = 100 μm). **d** Dermis thickness and subcutaneous fat layer thickness were measured on H&E stained images (*n* = 10 per group). **e** mRNA levels of molecules related to extracellular matrix metabolism, such as TGF-β, Col1a1, Col1a2, CTGF, VEGF, and MMP-13, in the lesional skin were assessed by quantitative real-time reverse transcription-PCR. **f** TGF-β, Smad3, p-Smad3, Smad4, Smad7, STAT1, and p-STAT1 proteins in the skin of C57BL/6 mice from the normal group, vehicle-treated group, and DZ2002-treated group (DZ2002: 50 mg/kg, 8 mice skin tissue mixed proteins). Mean ± SEM. HPF, high-power field, \**P* \< 0.05, \*\**P* \< 0.01, \*\*\**P* \< 0.001, \*\*\*\**P* \< 0.0001. PNS, prednisolone; d, day; ns, no significance
Histologic assessment and immunohistochemistry {#Sec6}
----------------------------------------------
All skin sections were taken from the paramidline, lower back region. Skin sections from BLM-induced SSc mice were taken together with hypodermal tissues. Sections were stained with hematoxylin and eosin and Masson's trichrome stain. We examined dermal thickness, which was defined as the thickness of the skin from the top of the granular layer to the junction between the dermis and subcutaneous fat. Then we examined subcutaneous fat thickness. Immunohistochemistry was performed using antibodies directed against a-SMA (1:200, Cell Signaling Technology, Beverly, MA, USA), then with peroxidase-labeled secondary antibody (1:100, R&D systems, Minneapolis, MN, USA), followed by color development with the aminoethylcarbazole system. α-SMA (brown part) was counted under a high-power microscopic field. These pictures were analyzed by using Image-Pro Plus software (Media Cybernetics, Silver Springs, MD, USA). Each section was examined independently by two investigators in a blinded manner.
Cell culture {#Sec7}
------------
Human dermal fibroblast cell line (BJ), human dermal microvascular endothelial cell line (HMEC-1) \[[@CR31]\], and human acute monocytic leukemia cell line (THP-1) were purchased from ATCC (Manassas, VA, USA). BJ cells were cultured in DMEM (Gibco, Grand Island, NY, USA) containing 10% fetal bovine serum (FBS) (Gibco, Grand Island, NY, USA), 100 U/ml penicillin and 100 μg/ml streptomycin (Corning Incorporated, Corning, NY, USA) at 37 °C in a humidified 5% CO~2~ atmosphere. After adhering the cells, they were starved for 24 h with 0.1% FBS in DMEM, then stimulated with human TGF-β1 (PeproTech, Rocky Hill, NJ, USA) for 24 h. HMEC-1 cells were cultured in MCDB131 (without L-glutamine) as base medium (Sigma-Aldrich, St. Louis, MO, USA), supplemented with 10 ng/ml epidermal growth factor (Sigma-Aldrich, St. Louis, MO, USA), 1 μg/ml hydrocortisone (Sigma-Aldrich, St. Louis, MO, USA), 10 mM glutamine (Sigma-Aldrich, St. Louis, MO, USA),100 U/ml penicillin, 100 μg/ml streptomycin, and 10% FBS in 70 cm^2^ flasks until confluence at 37 °C, 5% CO~2~ atmosphere. After the cells were attached, TNF-α is stimulated for 24 h. THP-1 cells were cultured in RPMI 1640 (Gibco, Grand Island, NY, USA) containing 10% FBS, 100 U/ml penicillin, and 100 μg/ml streptomycin at 37 °C in a humidified 5% CO~2~ atmosphere.
Cell Counting Kit-8 (CCK-8) cytotoxicity assay {#Sec8}
----------------------------------------------
Dispense 100 μl of cell suspension (BJ 2000 cells/well, HMEC-1 5000 cells/well) in a 96-well plate. Pre-incubate the plate for 24 h at 37 °C, 5% CO~2~ atmosphere. Add 100 μl of various concentrations of toxicant into the culture media in the plate. Incubate the plate for 24 h in the incubator. Add 20 μl of CCK-8 solution (Sigma-Aldrich, St Louis, MO, USA) to each well of the plate. Incubate the plate for 1--4 h in the incubator. Measure the absorbance at 450 nm using a microplate reader.
BMDMs culture {#Sec9}
-------------
Bone marrow-derived macrophages (BMDMs) were generated from bone marrow by macrophage colony-stimulating factor (M-CSF) induction, as described previously \[[@CR32]\]. Bone marrow was aseptically flushed from the tibiae and femurs of euthanized mice and depleted of red blood cells using Red Blood Cell Lysis Buffer (Beyotime, Shanghai, China). Cells were incubated in IMDM (Gibco, Grand Island, NY, USA) in a cell culture dish at 37 °C for 2 h to remove macrophages. Nonadherent cells were resuspended in IMDM supplemented with 10% heat-inactivated FBS, 100 U/ml penicillin, 100 mg/ml streptomycin, 2 mM L-glutamine (Thermo Fisher Scientific, Pittsburgh, PA, USA), and 10 ng/ml M-CSF (PeproTech, Rocky Hill, NJ, USA) and cultured for 7 days. Nonadherent cells were removed, and M-CSF-conditioned medium was changed on day 3. To acquire M1 macrophages, macrophages were stimulated with 10 ng/ml IFN-γ (PeproTech, Rocky Hill, NJ, USA) and 1 μg/ml lipopolysaccharide (LPS, Sigma-Aldrich, St Louis, MO, USA) for 48 h for mRNA analysis or protein assays. To acquire M2 macrophages, macrophages were stimulated with 40 ng/ml IL-4 (PeproTech, Rocky Hill, NJ, USA) for 48 h for mRNA analysis or protein assays.
Cell adhesion assay for HMEC-1 and THP-1 {#Sec10}
----------------------------------------
HMEC-1 (5 × 10^4^ cells/ml) were seeded in 24-well plates and were activated with TNF-a (40 ng/ml) in the presence or absence of DZ2002 for 6 h. THP-1 were labeled with 5 μM Calcein AM (BD Pharmingen, San Diego, CA, USA) in RPMI 1640 medium with 10% FBS for 30 min. The labeled THP-1 cells were washed with PBS three times, resuspended in endothelial cell medium and then added (1 × 10^5^ cells) onto HMEC-1. The co-culture system was incubated for 30 min at 37 °C in a humidified atmosphere of 5% CO~2~. The cells were washed with PBS three times to remove unadhered cells and were fixed with 3.7% formaldehyde for 15 min. The cells were visualized and imaged under an inverted fluorescent microscope (Olympus IX-73) and were counted in three microscopic fields of each well. These pictures were analyzed by using Image-Pro Plus software. Each section was examined independently by two investigators in a blinded manner.
Preparation of skin cell suspension {#Sec11}
-----------------------------------
A 2 × 2 cm^2^ piece of depilated back skin was minced and then digested in 6 ml of RPMI 1640 medium containing 10% FBS and 2 mg/ml crude collagenase (Sigma-Aldrich, St. Louis, MO, USA), 1.5 mg/ml hyaluronidase (Sigma-Aldrich, St. Louis, MO, USA), and 0.03 mg/ml DNase I (Roche Applied Science, Indianapolis, IN, USA) at 37 °C for 120 min \[[@CR33]\]. Samples were passed through a 70-μm Falcon cell strainer (BD Biosciences, San Jose, CA, USA) to obtain single-cell suspensions. After centrifugation at 1200 rpm for 5 min, the cell pellet was resuspended in a RPMI 1640 medium. The cells were passed through a 40-μm Falcon cell strainer. The harvested cells were washed with ice-cold PBS and used for flow cytometric analysis and real-time PCR analysis.
Western blot analysis {#Sec12}
---------------------
Primary skin cells were homogenized in sodium dodecyl sulfate sample buffer containing proteinase and phosphatase inhibitor. Stimulated BJ cells or HMEC-1 cells were lysed with RIPA lysis buffer (Beyotime, Shanghai, China). Homogenate were collected, and protein concentrations were then measured using the BCA assay kit (Thermo Fisher Scientific, Pittsburgh, PA, USA). The total protein samples were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes (Amersham Pharmacia Biotech, Buckinghamshire, UK). After blocking, the membranes were incubated with anti-TGF-β, anti-Smad3, anti-phospho-Smad3, anti-Smad4, anti-Smad7, anti-STAT1, anti-phospho-STAT1, anti-inducible NO synthase (iNOS), anti-Arginase 1 (Arg-1), anti-ICAM-1, and anti-VCAM-1 (all from Cell Signaling Technology, Beverly, MA, USA) (Mouse specific ICAM-1, Abcam, Cambridge, UK). After washing with TBS with Tween-20, the secondary antibodies (1:10000, Bio-Rad, Richmond, CA, USA) were added, and HRP-conjugated monoclonal mouse anti-GAPDH (1:10000, Kangcheng, Shanghai, China) was used as a control for normalization. Signals were detected with an ECL system (Amersham Bioscience, Buckinghamshire, UK) and exposed to classic autoradiography film.
Cytokine analysis by ELISA {#Sec13}
--------------------------
Skins from mice in each group were homogenized with lysis buffer to extract total protein. The concentration of total protein was determined by the BCA protein assay kit, collecting supernatant after BMDMs culture for 48 h. The levels of IL-1β, IL-4, IL-6, IL-10, IL-12p40, IL-17A, TNF-α, and IFN-γ in skin homogenates and cell culture supernatants were quantified by ELISA kits according to the manufacturer's protocol (BD Pharmingen, San Diego, CA, USA).
Real-time PCR {#Sec14}
-------------
In order to obtain gene expression of certain proteins in tissues and cells, mRNA was isolated from the skin of tested mice with a Total RNA kit (Tiangen, Shanghai, China) and reverse transcribed with an RT Master Mix kit (Takara, Dalian, China). QRT-PCR was performed with the SYBR Premix Ex Taq kit (TOYOBO, Osaka, Japan) on an Applied Biosystems 7500 Fast Real-Time PCR System (Applied Biosystems, Foster city, CA, USA). Mouse β-actin and human GAPDH were used as a housekeeping gene. The sequences of primers were summarized in Additional file [1](#MOESM1){ref-type="media"}: Table S1.
Flow cytometric analysis {#Sec15}
------------------------
Surface marker and intracellular transcription factor staining were conducted and analyzed using our previously reported methods \[[@CR29]\]. Briefly, for surface marker and intracellular transcription factors, cells were collected and stained with FITC-, PE-, PerCP-Cy5.5-, APC-, BV421-, or BUV395-conjugated monoclonal antibodies (mAbs) for membrane molecules or intracellular staining after being blocked with anti-mouse CD16/CD32 (Fc Receptor Block, eBioscience, San Diego, CA, USA). For intracellular staining of cytokines, cells were cultured in the presence of GolgiStop (10 μg/ml, eBioscience, San Diego, CA, USA) for 6 h and then collected for the following staining. The immunofluorescent Abs used in this study were all from BD Biosciences (San Jose, CA, USA). Flow cytometric analysis was performed on a 2-laser/4-color FACS Calibur analytical cytometer or a 4-laser/13-color BD LSR II analytical cytometer (BD Biosciences, San Jose, CA, USA), and the data were analyzed with FlowJo software (Tree Star, Ashland, OR, USA).
Immunofluorescence cytochemistry {#Sec16}
--------------------------------
BJ cells on coverslips were fixed in 4% paraformaldehyde (PFA) for 30 min and were permeabilized with 1% Triton X-100 for 10 min. After blocking with 3% BSA for 1 h, cells were incubated with rabbit anti-α-SMA (1:200) and anti-phosphorylated Smad3 (1:50; Santa Cruz Biotechnology, Santa Cruz, CA, USA) at 37 °C for 2 h. After washing with 1% PBS-Tween, Alexa Fluor 647-conjugated anti-rabbit secondary antibodies (BD Biosciences, San Jose, CA, USA) were added. Negative control reactions were included in each experiment and carried out by replacing primary antibodies with PBS. The cells were counterstained with 4′, 6-diamidino-2-phenylindole (DAPI, Abcam, Cambridge, UK). All images were captured using a Leica TCS SP8 STED confocal microscope. Skin tissues were embedded in OCT compound and sectioned on a cryostat (8 μm thick). After fixation in PFA, sections were blocked with 5% BSA for 60 min and then incubated with FITC-anti-CD3 (1:100, R&D systems, Minneapolis, MN, USA) or FITC-anti-CD11b (1:100, R&D systems, Minneapolis, MN, USA) at 4 °C overnight. The sections were counterstained with DAPI. Fluorescent cells were visualized and digital images were captured using a Leica TCS SP8 STED confocal microscope.
Statistical analysis {#Sec17}
--------------------
One-way ANOVA followed by Dunnett's multiple comparison test was used to compare parameters involving multiple groups, with GraphPad Prism 7.0 (GraphPad Software, San Diego, CA, USA) statistical software. *P* values less than 0.05 were considered significant.
Results {#Sec18}
=======
DZ2002 alleviates skin fibrosis in BLM-induced SSc mice model {#Sec19}
-------------------------------------------------------------
We evaluated the effect of DZ2002 on BLM-induced dermal fibrosis. Compared to vehicle, DZ2002 significantly decreased skin thickness and dermal thickness in BLM-induced mice (Fig. [1](#Fig1){ref-type="fig"}b--d). As well, BLM-dependent decrease in subcutaneous fat thickness, which is usually seen in BLM-induced mice, was extraordinarily attenuated by DZ2002 administration (Fig. [1](#Fig1){ref-type="fig"}c, d). Consistent with this finding, DZ2002 treatment significantly reduced collagen accumulation and α-SMA expression in the dermis compared with the vehicle group (Fig. [1](#Fig1){ref-type="fig"}c, d). Meanwhile, DZ2002 treatment significantly reduced collagen content and mRNA expression of the Col1a1 and Col1a2 while promoting that of the matrix metalloproteinase-13 (MMP-13) in the lesional skin of BLM-induced mice (Fig. [1](#Fig1){ref-type="fig"}e). We also found that DZ2002 suppressed the mRNA expression of vascular endothelial growth factor (VEGF) in mice skin tissue (Fig. [1](#Fig1){ref-type="fig"}e). As hallmarks of SSc dermal fibroblasts elevated the expression, TGF-β and CTGF were further examined in experimental SSc mice \[[@CR34], [@CR35]\]. Remarkably, the decrease in mRNA expression of the TGF-β and CTGF was noted in BLM-induced mice exposed to DZ2002 (Fig. [1](#Fig1){ref-type="fig"}e), which was also confirmed at protein levels by western blot (Fig. [1](#Fig1){ref-type="fig"}f). Moreover, the downstream proteins of TGF-β such as p-Smad3, Smad4, and Smad7 also changed accordingly (Fig. [1](#Fig1){ref-type="fig"}f). We also evaluated the effects of DZ2002 on IFN-γ/STAT1 inflammatory activation signaling pathway. Notably, DZ2002 was able to strongly inhibit phosphorylation of STAT1 (Fig. [1](#Fig1){ref-type="fig"}f). These findings suggest that DZ2002 exerts a potent anti-fibrotic effect on dermal fibrosis by reducing the production of collagen, facilitating its degradation and regulating expression of various soluble factors in SSc mice model.
DZ2002 lightens TGF-β-induced collagen accumulation in BJ cells {#Sec20}
---------------------------------------------------------------
Here, we explored if DZ2002 directly affected the profibrotic phenotype of dermal fibroblasts in vitro. To evaluate the molecular mechanism, we stimulated BJ cells with recombinant active TGF-β1, which directly acts on TGF-β receptors. CCK-8 assay shows that DZ2002 at 500 μM and below had almost no toxicity to BJ cells (Fig. [2](#Fig2){ref-type="fig"}a). We mainly focused on the expression of type I collagen and matrix metalloproteinase-1 (MMP-1). DZ2002 significantly reversed the stimulatory effect of TGF-β on Col1a2 and MMP-1 mRNA levels dose-dependently in BJ cells (Fig. [2](#Fig2){ref-type="fig"}b). Additionally, DZ2002 also significantly reduced TGF-β1-induced mRNA expression of CTGF (Fig. [2](#Fig2){ref-type="fig"}c). Fig. 2DZ2002 inhibited TGF-β dependent activation collagen accumulation of BJ cells. **a** Survival rate of BJ cells at different concentrations of DZ2002. **b**, **c** Col1a2, MMP-1, and CTGF mRNAs in the BJ cells were assessed by quantitative real-time reverse transcription-PCR. **d** Itgav, Itgb3, and Itgb5 mRNAs, encoding integrin αV, integrin β3, and integrin β5, were determined by quantitative real-time reverse transcription-PCR in BJ cells. **e** Immunofluorescence staining of α-SMA (green) in BJ cells (DAPI, blue, 24 h). **f** Western blot of α-SMA in BJ cells (24 h). **g** Immunofluorescence staining of p-Smad3 (green) in BJ cells (DAPI, blue, 2 h). **h** Western blot of p-Smad3 in BJ cells (2 h). **e**--**h** TGF-β = 25 ng/ml, DZ2002 = 200 μM. Mean ± SEM. \**P* \< 0.05, \*\**P* \< 0.01, \*\*\**P* \< 0.001. ns no significance
Since αVβ3 and αVβ5 integrins, latent receptor forms of TGF-β increase the sensitivity of dermal fibroblasts to TGF-β by recruiting and activating TGF-β on the cell surface \[[@CR36]--[@CR40]\]. Integrin αVβ3 and αVβ5 are upregulated by binding with latency-associated peptide on dermal fibroblasts of SSc patients \[[@CR36]\]; therefore, we assessed the expression of these molecules in BJ cells. DZ2002 could be reduced to some extent the mRNA expression of integrin αV, β3, and β5 subunits (Fig. [2](#Fig2){ref-type="fig"}d), thus raising the possibility that DZ2002 may attenuate the establishment of TGF-β signaling in response to dermal fibroblasts. Additionally, under the same condition, DZ2002 reduced TGF-β1-induced expression of α-SMA protein (Fig. [2](#Fig2){ref-type="fig"}e, f). At the same time, the fibrin protein phosphorylated Smad3 of TGF-β downstream also had been inhibited (Fig. [2](#Fig2){ref-type="fig"}g, h). These results indicate that DZ2002 augments exogenous TGF-β-dependent CTGF induction through αVβ3 integrin- and αVβ5 integrin-dependent activation of TGF-β in dermal fibroblasts. Overall, these consequences reveal that DZ2002 inhibits fibrogenic genes, including Col1a2 and CTGF, at least partly through the integrin-mediated inhibition of TGF-β signal pathway.
DZ2002 suppresses T cell activation and reduces inflammatory cell infiltration {#Sec21}
------------------------------------------------------------------------------
To investigate the impact of DZ2002 on BLM-induced inflammatory conditions, we next examined the expression of pro-inflammatory cytokines, T helper cytokines and chemokines, including TNF-a, IFN-γ, IL-1β, IL-4, IL-6, IL-10, IL-17A, and monocyte chemotactic protein 1 (MCP-1) in the lesional skin of BLM-induced mice. Notably, a significant reduction of mRNA expression due to DZ2002 treatment was observed in all the genes (Fig. [3](#Fig3){ref-type="fig"}a). To further elucidate the expression of these cytokines in skin tissue, we extracted the skin tissue homogenate. Consistently, DZ2002 inhibited the secretion of cytokines in the skin tissue (Fig. [3](#Fig3){ref-type="fig"}b). These indicated that DZ2002 widely suppressed the expression of pro-inflammatory, Th1, Th2, and Th17 cytokines, and chemokines in the context of BLM-induced mice skin. Immunofluorescence analysis demonstrated that the CD11b^+^ and CD3^+^ fluorescence intensity (green) in the skin were reduced after DZ2002 treatment compared with the vehicle control (Fig. [3](#Fig3){ref-type="fig"}c). To further assess if DZ2002 diversely affects the differentiation of T helper cells, we carried out flow cytometric analysis with skin cells. Of note, DZ2002 restored the BLM-dependent increase in the CD4^+^ T lymphocyte and CD25-activated T lymphocyte number in the skin (Fig. [3](#Fig3){ref-type="fig"}d). At the same time, DZ2002 inhibited the differentiation of Th1, Th2, and Th17 cells as well as the infiltration of neutrophils and macrophages (Fig. [3](#Fig3){ref-type="fig"}e, f). These data suggest that DZ2002 supports the inhibition of an SSc-like immune response in the lesional skin of BLM-induced mice. Fig. 3DZ2002 inhibited the activation of T cells and infiltration of neutrophils and macrophages in BLM-induced mice. **a** TNF-α, IFN-γ, IL-1β, IL-4, IL-6, IL-10, IL-17A, and MCP-1 mRNAs were determined by quantitative real-time reverse transcription-PCR in the BLM-induced SSc mice skin single-cell suspension. **b** Determination of TNF-α, IFN-γ, IL-1β, IL-4, IL-6, IL-10, IL-12, and IL-17A proteins in skin tissue homogenate supernatant of BLM-induced SSc mice by ELISA (*n* = 7 per group). **c** Immunofluorescence assay for CD3^+^ and CD11b^+^ cells in skin frozen sections of BLM-induced SSc mice. **d** Left: The percentages of CD3^+^ CD4^+^ cells and CD4^+^ CD25^+^ cells were measured by flow cytometry; Right: Percentage statistics of CD3^+^ CD4^+^ cells and CD4^+^ CD25^+^ cells (2 groups and *n* = 5 per group). **e** Left: Representative staining analysis of intracellular IL-17, IL-4, and IFN-γ in CD4^+^ T cells by flow cytometry; Right: Percentage statistics of IL-17A^+^ CD4^+^ cells, IL-4^+^ CD4^+^, and IFN-γ^+^ CD4^+^ cells (2 groups and *n* = 5 per group). **f** Left: The percentages of neutrophils (CD11b^+^ Gr-1^+^) and macrophages (CD11b^+^ F4/80^+^) were measured by flow cytometry; percentage statistics of CD11b^+^ Gr-1^+^ cells and CD11b^+^ F4/80^+^ cells (2 groups and *n* = 5 per group). Mean ± SEM, PNS prednisolone. \**P* \< 0.05, \*\**P* \< 0.01, \*\*\**P* \< 0.001, ns no significance
DZ2002 suppresses M1 and M2 polarization of macrophages {#Sec22}
-------------------------------------------------------
We evaluated the impact of DZ2002 on macrophages in the context of BLM-induced mice. DZ2002 significantly reduced mRNA expression of iNOS and IL-12p40; established polarization markers for M1 macrophages, also to chitinase 3-like 3 (Ym-1) and Arg-1; and established polarization markers for M2 macrophages, in the skin lesions of BLM-induced mice (Fig. [4](#Fig4){ref-type="fig"}a, b). The expression of iNOS and Arg-1 proteins also supported the findings (Fig. [4](#Fig4){ref-type="fig"}c). We used normal mouse BMDMs to examine the effect of DZ2002 on macrophage differentiation in vitro. DZ2002 also significantly reversed the IFN-γ and lipopolysaccharides-trigged iNOS and IL-12p40 mRNA expression and IL-4-induced Fizz1, Ym-1, and Arg-1 mRNA expression (Fig. [4](#Fig4){ref-type="fig"}d, e). iNOS and Arg-1 proteins are also inhibited by DZ2002 (Fig. [4](#Fig4){ref-type="fig"}g). And in the supernatant of BMDMs, DZ2002 also significantly inhibited TNF-α, IL-6, and IL-12p40 (M1 makers) production (Fig. [4](#Fig4){ref-type="fig"}f). Similarly, we also used flow cytometry to verify that DZ2002 reduced the proportion of MHCII (Fig. [4](#Fig4){ref-type="fig"}h). And the proportion of F4/80^+^ iNOS^+^ (M1 makers) and F4/80^+^ CD301^+^ (M2 makers) cells was significantly reduced (Fig. [4](#Fig4){ref-type="fig"}i). Taken together, these results indicate that DZ2002 inhibits the differentiation of M1 macrophages and M2 macrophages in vivo and in vitro to contribute to anti-inflammatory and anti-fibrosis. Fig. 4DZ2002 inhibited macrophage activation in vivo and in vitro. **a** IL-12p40 and iNOS mRNAs of M1 makers were determined by quantitative real-time reverse transcription-PCR of single-cell suspension in the BLM-induced SSc mice skin. **b** Ym-1 and Arg-1 mRNAs of M2 makers were determined by quantitative real-time reverse transcription-PCR of single-cell suspension in the BLM-induced SSc mice skin. **c** Western blot of iNOS and Arg-1 of single-cell suspension in the BLM-induced SSc mice skin (DZ2002 = 50 mg/kg). **d** IL-12p40 and iNOS mRNAs of M1 makers were determined by quantitative real-time reverse transcription-PCR in the BMDMs. **e** Fizz1, Ym-1, and Arg-1 mRNAs of M2 makers were determined by quantitative real-time reverse transcription-PCR in the BMDMs. **f** TNF-α, IL-6, and IL-12p40 cytokines in the supernatant secreted by BMDMs were determined by ELISA (*n* = 4 per group). **g** Western blot of iNOS and Arg-1 in the BMDMs (DZ2002 = 200 μM). **h** The percentages of MHCII of M1 were measured by flow cytometry. **i** The percentages of CD11b^+^ iNOS^+^ (M1 makers) and CD11b^+^ CD301^+^ (M2 makers) were measured in the BMDMs by flow cytometry. Mean ± SEM. \**P* \< 0.05, \*\**P* \< 0.01, \*\*\**P* \< 0.001; LPS lipopolysaccharide, PNS prednisolone, ns no significance
DZ2002 reduces adhesion molecule expression in vivo and in vitro {#Sec23}
----------------------------------------------------------------
We examined the effect of DZ2002 on the expression of ICAM-1 and VCAM-1 in BLM-induced mice. As expected, ICAM-1 and VCAM-1 protein expression in the skin lesions of BLM-induced mice was significantly reduced by DZ2002 treatment (Fig. [5](#Fig5){ref-type="fig"}a). In vitro, we used TNFα-stimulated HMEC-1 to study the effects of DZ2002 on vascular injury. CCK-8 assay shows that DZ2002 at 2000 μM and below had almost no toxicity in HMEC-1 cells (Fig. [5](#Fig5){ref-type="fig"}b). Notably, DZ2002 obviously reduced the adhesion molecule markers ICAM-1, VCAM-1, VEGF, basic fibroblast growth factor (bFGF), and ET-1 mRNA expression in a dose-dependent manner (Fig. [5](#Fig5){ref-type="fig"}c). And ICAM-1 and VCAM-1 proteins were also alleviated in HMEC-1 (Fig. [5](#Fig5){ref-type="fig"}d). Fig. 5DZ2002 reduced ICAM-1 and VCAM-1 in vivo and in vitro. **a** Western blot of ICAM-1 and VCAM-1 in BLM-induced SSc mice skin (DZ2002 = 50 mg/kg). **b** Survival rate of HMEC-1 at different concentrations of DZ2002 treatment. **c** Evaluation of mRNA levels of molecules associated with the upregulation of endothelial adhesion molecules in HMEC-1 by quantitative real-time reverse transcription-PCR, such as ICAM-1, VCAM-1, VEGF, bFGF, and ET-1. **d** Western blot of ICAM-1 and VCAM-1 in the HMEC-1 cells (TNF-α = 40 ng/ml, DZ2002 = 200 μM). **e** Representative photomicrographs showing the stained THP-1 cells on HMEC-1 cell layer (× 40, scale bar = 100 μm). The HMEC-1 cells were stimulated with TNF-a (40 ng/ml) with or without DZ2002 (200 μM). After 6 h incubation, those stained THP-1 cells adhered on HMEC-1 cell layer were counted. Mean ± SEM. \**P* \< 0.05, \*\**P* \< 0.01, \*\*\**P* \< 0.001, \*\*\*\**P* \< 0.0001; ns no significance
Vascular endothelial cells recruit circulating monocytes to regions of vascular injury and/or infection \[[@CR41]\]. Following their entry into the vessel wall, monocytes differentiate into macrophages, which drive an inflammatory response to neutralize invading pathogens, repair tissue damage or activate other immune cells \[[@CR42]\]. Here, we used THP-1 and HMEC-1 to co-culture in vitro to simulate this environment. It could be seen that THP-1 adhered to HMEC-1 damaged after TNF-α stimulation and DZ2002 reversed THP-1 adhesion (Fig. [5](#Fig5){ref-type="fig"}e). These results indicated that DZ2002 repaired vascular endothelial injury and reduced endothelial cell adhesion molecule expression.
Discussion {#Sec24}
==========
SSc is a chronic autoimmune disease that is characterized by diffuse fibrosis in the skin and internal organs. It is widely accepted that excessive deposition of ECM and vascular endothelial cell activation in SSc are highly associated with the inflammatory abnormalities \[[@CR43]\]. Our previous research showed that DZ2002 had therapeutic effects on a variety of autoimmune diseases \[[@CR25]--[@CR28]\]. We hypothesized that its immunosuppressive activity might have a therapeutic effect in SSc. As hypothesized, DZ2002 inhibited the development of dermal fibrosis in BLM-induced mice. DZ2002 prevented BLM-induced dermal fibrosis by affecting pathological processes that broadly influenced the basis of tissue fibrosis, such as fibroblast activation, activation of TGF-β/Smad signaling pathway, Th1/Th2/Th17-skewed immune polarization, and M1 and M2 macrophages activation. In addition, DZ2002 inhibited the upregulation of microvascular endothelial cell adhesion molecules. These results signify that DZ2002 has the potential to modulate different SSc-related pathological features (Fig. [6](#Fig6){ref-type="fig"}). Fig. 6The pharmacodynamic mechanism of DZ2002 on experimental systemic sclerosis models
TGF-β has been implicated to play a critical role in initiating and sustaining the fibrotic evolution in SSc, a function mediated through both Smad-dependent and Smad-independent pathways \[[@CR3]\]. Fibroblasts of SSc patients show constitutive Smad2/3 phosphorylation and nuclear localization, and various anomalous Smad signals are over activated. Thus, targeting TGF-β/Smad signal pathway is an effective strategy for the treatment of SSc. Our data showed that DZ2002 inhibited the phosphorylation of Smad3 protein in BLM-injected SSc mice skin and human dermal fibroblasts. The inhibition of Smad3 phosphorylation, thus interfering with its binding to Smad-binding elements, might lead to reduced transcription of fibrogenic genes such as Col1a1, Col1a2, and CTGF, as seen in human fibroblasts in vitro. Additionally, our limited data suggested that DZ2002 suppressed STAT1 phosphorylation in BLM-injected mice skin.
BLM-injection characteristically induces Th2/Th17-skewed immune polarization in fibrotic tissues, leading to fibroblasts activation \[[@CR44]\]; nevertheless, DZ2002 treatment diversely reduced mRNA and protein levels of cytokines produced by Th1, Th2, and Th17 cells in the lesional skin of BLM-induced mice. Under the same condition, the proportion of Th1, Th2, Th17, and activated T cells was decreased, suggesting that DZ2002 attenuated the inflammatory state in response to BLM-induced tissue injury by inhibiting the differentiation of T cells such as Th1, Th2, and Th17 cells. Our previous study also showed that DZ2002 prevented T cell differentiation by affecting the function of bone marrow-derived dendritic cells (BMDCs) in BMDC-T cell co-culture systems \[[@CR45]\]. In addition to the altered T cell subsets, the proportion of inflammatory cells, including neutrophils and macrophages, was attenuated in response to DZ2002 in the lesional skin of BLM-induced mice, which was probably to be at least partially credited to the DZ2002-dependent inhibition of MCP-1. DZ2002 also inhibited macrophages differentiating to a pro-inflammatory M1-like phenotype and a profibrotic M2-like phenotype in the lesional skin of BLM-induced mice partly by directly acting on macrophages, and these were proved by in vitro experiments on BMDMs. DZ2002 significantly inhibited M1 macrophage markers (IL-12p40 and iNOS) and M2 macrophage markers (Fizz1, Ym-1, Arg-1) in BMDMs, also proved by flow cytometry. Considering the effects of DZ2002 on the macrophage phenotype in BLM-treated mice and BMDMs, we draw a conclusion that DZ2002 might exert anti-fibrotic effects on pathological fibrosis partly by inhibiting macrophage differentiation into M1 and M2 phenotypes. Also, DZ2002 ameliorated the upregulation of human microvascular endothelial cell adhesion molecules, including ICAM-1, VCAM-1, and VEGF. This was partly due to the inhibition of inflammatory cytokines by DZ2002, which reduced the recruitment of monocytes to repair vascular damage. Moreover, we confirmed the direct anti-fibrotic effect of DZ2002 on dermal fibroblasts stimulated with TGF-β1.
To summarize, this study confirmed the exclusive anti-fibrotic effects of DZ2002 in the context of BLM-induced dermal fibrosis. Of note, the multi-faceted inhibition effects of this agent on the TGF-β-induced human dermal fibroblasts fibrosis process have not been reported previously. Also, DZ2002 improved vasculopathy through inhibiting adhesion molecule expression. These data in vivo and in vitro demonstrate that DZ2002 exerts an effective therapeutic effect on the experimental SSc models by preventing the three main components of the pathogenesis of SSc.
Conclusions {#Sec25}
===========
The present study demonstrates that DZ2002 has anti-inflammatory, anti-fibrotic, and anti-angiogenic effects on experimental SSc models by acting on BLM-induced mice and various types of cells, including fibroblasts, endothelial cells, and immune cells. These indicate DZ2002 is expected to become a potential new drug for the treatment of SSc.
Supplementary information
=========================
{#Sec26}
**Additional file 1: Table S1.** Specific primers used in real-time PCR analysis.
SAHH
: S-adenosyl-l-homocysteine hydrolase
BLM
: Bleomycin
SSc
: Systemic sclerosis
TGF-β
: Transforming growth factor-β
IL
: Interleukin
CTGF
: Connective tissue growth factor
ET-1
: Endothelin-1
ECM
: Extracellular matrix
α-SMA
: α-Smooth muscle actin
Col1a1
: Collagen type I α 1
TNF-α
: Tumor necrosis factor-α
IFN-γ
: Interferon-γ
Th
: T helper
ICAM-1
: Intercellular adhesion molecule-1
VCAM-1
: Vascular cell adhesion molecule 1
SAH
: S-adenosyl-l-homocysteine
SAM
: S-adenosylmethionine
PNS
: Prednisolone
BJ
: Human dermal fibroblast cell line
HMEC-1
: Human dermal microvascular endothelial cell line
THP-1
: Human acute monocytic leukemia cell line
FBS
: Fetal bovine serum
CCK-8
: Cell Counting Kit-8
BMDMs
: Bone marrow-derived macrophages
M-CSF
: Macrophage colony-stimulating factor
LPS
: Lipopolysaccharide
Arg-1
: Arginase 1
Ym-1
: Chitinase 3-like 3
iNOS
: Inducible NO synthase
PFA
: Paraformaldehyde
DAPI
: 4′,6-Diamidino-2-phenylindole
MMP
: Matrix metalloproteinase
MCP-1
: Monocyte chemotactic protein 1
bFGF
: Basic fibroblast growth factor
Fizz1
: Found in inflammatory zone 1
HPF
: High-power field
**Publisher's Note**
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Zongwang Zhang and Yanwei Wu contributed equally to this work.
Supplementary information
=========================
**Supplementary information** accompanies this paper at 10.1186/s13075-019-2074-9.
Not applicable
ZWZ, YWW, JPZ, LLN, and WT designed the research. ZWZ, YWW, BW, QQ, HL, HML, CF, and CLF performed the experiments. ZWZ and WT analyzed data. ZWZ and WT drafted the manuscript. All authors contributed to the interpretation of the data and reviewing the manuscript. All authors approved the final draft for submission.
This work was supported by the "Personalized Medicines-Molecular Signature-based Drug Discovery and Development", Strategic Priority Research Program of the Chinese Academy of Sciences (grant no. XDA12020113), Science & Technology Commission of Shanghai Municipality, China (grant No. 18431907100).
Not applicable
Not applicable
Not applicable
The authors declare that they have no competing interests.
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#s1}
============
Along with surgery and chemotherapy, radiation therapy (RT) is an important modality in cancer treatment, with about half of cancer patients receiving RT as a curative or palliative treatment. However, there are still many challenges to improve its efficiency by minimizing normal tissue toxicity while maximizing tumor control. In this perspective, an approach combining RT with radiosensitizers has been long investigated in order to allow for better dose modulation ([@B1]). With the emergence of small molecules, macromolecules, and nanomaterials, recent progresses in the field have been significant ([@B2]). However, given the severe limitations of radiosensitizers currently administered in the clinic, there is an urgent need for new targeted RT-compatible anticancer treatments.
*Withania somnifera* is a plant originated from the dry regions of Asian countries and that possesses diverse biological activities, such as anti-inflammatory, anti-stress, antioxidant, immunomodulatory, anti-angiogenic, and anticancer activities. These numerous effects are known to be due to the presence of withanolides, a class of steroidal lactones, present in the roots, and leaves of this plant ([@B3]). Thus, several studies have shown that withanolides exert their antitumor activity by inducing ROS production, cell cycle arrest, cytoskeleton destabilization, etc. ([@B4]). Interestingly, the most studied *W. somnifera*-derived compound, withaferin A (WFA), has also shown to have a radiosensitizing effect ([@B5]). Yang et al. confirmed that WFA enhanced radiation-induced apoptosis in renal and leukemia cell lines through down-regulation of the anti-apoptotic protein Bcl-2 ([@B6], [@B7]). This effect has also been demonstrated *in vivo*, where WFA treatment when associated with fractionated RT synergistically increased the complete response of mouse melanoma, fibrosarcoma, and Ehrlich ascites carcinoma ([@B8], [@B9]). Therefore, these studies suggest the great interest of withanolides as candidate cancer drugs and especially as potential radiosensitizers.
In this study, the effect of withanolide D (WD) in combination to radiation on several cancer cell lines was investigated. Structurally, WD is an isomer of WFA that feature 2([@B3])-en-1-one, 4β-hydroxy, and 5β,6β-epoxy moieties in their A and B rings. They differ from each other only in the position of the additional hydroxy group present in their side chain (C-27 in WFA and C-20 in WD) ([Figure 1](#F1){ref-type="fig"}).
{#F1}
Materials and Methods {#s2}
=====================
Withanolides
------------
WFA and WD were obtained from the aerial parts of aeroponically grown *W. somnifera* ([@B10]) as reported previously ([@B11]). Their purities were confirmed to be \>98% by HPLC analyses ([Supplementary Figures 1, 2](#SM1){ref-type="supplementary-material"}) and proton NMR spectroscopy ([Supplementary Figures 3, 4](#SM1){ref-type="supplementary-material"}). Compounds were dissolved in DMSO to obtain 10 mM solutions.
Cell Culture
------------
The human ovarian SKOV3 (HTB-77), intestinal Caco-2 (HTB-37), prostate DU145 (HTB-81) and 22Rv1 (CRL-2505), lung A549 (CCL-185), and breast MCF7 (HTB-22) cancer cells, were obtained from ATCC. SKOV3 were maintained in McCoy\'s 5A medium, Caco-2 in Dulbecco\'s Modified Eagle\'s Medium, DU145 in Minimum Essential Medium, 22Rv1 in RPMI-1640 medium, A549 in F-12K medium, MCF7 in Minimum Essential Medium with 0.01 mg/mL insulin, all supplemented with 10% of Fetal Bovine Serum (FBS) and 1% of streptomycin/penicillin. They were incubated at 37°C in 5% CO~2~ atmosphere. Medium was changed twice a week.
Proliferation Assay
-------------------
Cell proliferation was assessed by modified MTT assay according to the manufacturer\'s recommendations (CellTiter 96® Non-Radioactive Cell Proliferation Assay, Promega). Briefly, cells were seeded at a concentration of 5,000 cells/well in a 96-well plate and allowed to attach for 4 h. Cells were then exposed to a range of concentrations from 0.156 to 80 μM of WD for 48 h. DMSO (0.8%) served as vehicle and control. After adding reagents according to the manufacturer\'s recommendations, absorbance was recorded at 570 nm using an Epoch Microplate Spectrophotometer (Biotek, Vermont, USA). The concentration of drugs that resulted in 50% of cell death (IC~50~) was determined from dose-response curve by using PRISM 7.0 (GraphPad Software, San Diego, CA, USA). Experiments were repeated three times, and data represented as the mean of quadruplicate wells ± SEM.
Irradiation
-----------
Irradiation was performed using a cabinet X-ray machine (X-RAD 320, Precision X-Ray Inc.) at 320 kVp and 12.5 mA with a 2 mm Al filter. The source-to-axis distance was 42 cm. The beam was calibrated using a UNIDOS E PTW T10010 electrometer and TN30013 ionization chamber, with measurement done in air, for a 15 cm × 15 cm field size. The dose rate was 3 Gy/min.
Clonogenic Assay
----------------
Clonogenic assay was performed as previously described ([@B12]). Briefly, cells were seeded in six well plates (100, 200, 1,000, and 2,000 cells/well for DMSO-treated cells irradiated at 0, 2, 4, and 6 Gy, respectively, and 200, 400, 4,000, and 8,000 cells/well for WFA- or WD-treated cells irradiated at 0, 2, 4, and 6 Gy, respectively), allowed to attach for 4 h and then exposed to WFA, WD or DMSO at a concentration of 0.7 μM, 1 h before irradiation. They were then irradiated at 0, 2, 4, or 6 Gy. The medium was removed immediately after irradiation and replaced by a drug-free medium, then changed once a week. Cells were fixed with 6% of glutaraldehyde and stained with 0.5% of crystal violet approximatively after 14 days. Colonies \>50 cells were counted manually using a binocular loupe. Experiments were repeated three times, and data represented the mean of duplicate wells ± SEM.
Alkaline Comet Assay
--------------------
SKOV3 cells were seeded in T-25 flasks (200,000 cells/mL) and exposed to DMSO or WD when reached 70% confluence at a concentration of 0.7 μM 1 h before being irradiated at 2 Gy. The medium was removed right after irradiation and replaced by a drug-free medium. Alkaline comet assay was then performed 1 or 24 h after irradiation as previously described ([@B13]). Briefly, cells were harvested and resuspended in PBS at a density of 20,000 cells/mL. Successively, 400 μL sample of this cell suspension was mixed with 1.2 mL of 1% low melting agarose, and 1 mL of this mixture was disposed on microscope slides previously coated with 1% low melting agarose. Slides were treated with alkaline lysis solution (1.2 M NaCl, 100 mM Na~2~EDTA, 0.1% sodium lauryl sarcosinate, 0.26 M NaOH) for 20 h at 4°C. Slides were successively washed three times in 0.03 M NaOH, 2 mM Na~2~EDTA (alkaline buffer) for 20 min each, followed by electrophoresis in alkaline buffer for 25 min at 0.6 V/cm. Then, slides were washed and neutralized in distilled water and finally stained with sytox green (1 mM) immediately before imaging. Olive tail moment (defined as the percent DNA in the tail multiplied by the distance between the means of the head and tail distributions) was analyzed and quantified using CaspLab software.
Immunofluorescence
------------------
SKOV3 cells were seeded on collagen-coated coverslip at a density of 10,000 cells/well. Cells were exposed to DMSO or WD at a concentration of 0.7 μM 1 h before being irradiated at 2 Gy. The medium was removed right after irradiation and replaced by a drug-free medium. Cells were successively fixed for 15 min in 4% PFA, and permeabilized with 0.1% Triton X-100 for 5 min at room temperature at 1, 6, and 24 h post-irradiation for γH2Ax/53BP1 foci formation assay and 24 h only for mitotic catastrophe assay. After permeabilization, cells were blocked for 30 min with 1% BSA at room temperature before being incubated for 1 h with antibodies diluted in blocking buffer and raised against 53BP1 (Abcam, ab21083, 1:400), phospho-histone H2AX (Merck Millipore, 05-636, 1:800), α-tubulin (Santa Cruz Biotechnology, sc-5286, 1:150), or pericentrin (Abcam, ab4448, 1:4,000). After thorough washes with PBS, cells were incubated for 30 min in the dark with Cy™ 3-conjugated AffiniPure Goat Anti-Mouse IgG (Jackson ImmunoResearch, 115-165-062, 1:400) and Alexa Fluor® 647-conjugate AffiniPure Goat Anti-Rabbit IgG (Jackson ImmunoResearch, 111-605-045, 1:200). The nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI) for 5 min. Coverslip were mounted on the slides with a drop of mounting medium and imaged using an Epifluorescence microscope (Zeiss Axio Imager M2).
Immunoblotting
--------------
SKOV3 cells were seeded in T-25 flasks (200,000 cells/mL) and exposed to DMSO or WD when they reached 70% confluence at a concentration of 0.7 μM 1 h before being irradiated at 2 Gy. The medium was removed right after irradiation and replaced by a drug-free medium. Cells were directly lyzed using the radioimmunoprecipitation assay (RIPA) buffer at 1, 6, and 24 h post-irradiation. Proteins were quantified in 96-well plates by using the Pierce BCA Protein Assay kit (Thermo Fisher Scientific) following manufacturer\'s recommendations. Five micrograms of total proteins were then eluted by adding Laemmli buffer, and heated at 95°C for 10 min. Samples were size-separated by electrophoresis on 4--20% gradient acrylamide Tris-HCl Protein precast gels (Biorad) at 90 V for 2 h in 25 mM Tris, 192 mM glycine, 0.1% SDS buffer, and transferred to PVDF membranes (Invitrogen) at 300 mA for 90 min in 25 mM Tris, 192 mM glycine, 0.1% SDS with 20% ethanol. Membranes were blocked with PBST (PBS plus 0.1% Tween-20) containing 5% non-fat milk at room temperature for 60 min. Blots were then incubated (4°C, overnight) with primary antibodies against ATM (Cell signaling Technology, 2873, 1:1,000), pATM (phospho S1981) (Abcam, ab81292, 1:50,000), DNA-PKcs (Santa Cruz Biotechnology, sc-5282, 1:200), pDNA-PKcs (phospho S2056) (Abcam, ab18192, 1:1,000), GAPDH (Santa Cruz Biotechnology, sc-365062, 1:200), Rad51 (Abcam, ab63801, 1:100), and XRCC4 (Santa Cruz Biotechnology, sc-271087, 1:100). After five washes (5 min/each) with PBST, blots were incubated with a horse radish peroxidase (HRP) conjugated goat secondary antibody against rabbit IgG (Jackson ImmunoResearch, 111-035-003, 1:5,000) or mouse IgG (Sigma, A 4416, 1:10,000) at room temperature for 1 h. After washing five times with PBST (5 min/each), blots were developed with Clarity™ Western ECL substrate and imaged using a ChemiDoc™ XRS+ system (BioRad).
Statistical Analysis
--------------------
Data are expressed as the mean ± SEM. Statistical analysis were performed by non-parametric Wilcoxon signed-rank test or Wilcoxon--Mann--Whitney test. Welch\'s *t*-test was used when sample size was too small for the non-parametric tests. *P*-value (*p*) \< 0.05 (two-sided) was considered statistically significant for all of the statistical calculations. All statistical analyses and graphs were performed with PRISM 7.0 (GraphPad Software, San Diego, CA, USA).
Results {#s3}
=======
WD Inhibited Cell Proliferation of Various Cancer Cell Types
------------------------------------------------------------
Cell proliferation in prostate (22Rv1, DU145), lung (A549), intestinal (Caco2), breast (MCF7), and ovarian (SKOV3) cancer cells exposed to different concentrations of WD for 48 h was determined by modified MTT assay. Cell viability decreased significantly in all cell lines after WD exposure in a concentration-dependent manner ([Supplementary Figure 5](#SM1){ref-type="supplementary-material"}). To estimate the IC~50~ concentration, we normalized cell viability from 100% in DMSO and performed a nonlinear regression using dose response curve fitting \[log(inhibition) vs. normalized response (variable slope)\] in GraphPad Prism. Results showed that WD had an IC~50~ \< 3 μM in all cell lines, with Caco-2 being the most sensitive (IC~50~ = 0.63 μM) and SKOV3 the most resistant (IC~50~ = 2.93 μM) among the tested cell lines ([Supplementary Table 1](#SM1){ref-type="supplementary-material"}). For further molecular studies and protein expression profiling, a sub-cytotoxic concentration of 0.7 μM was selected, which is therapeutically relevant and effective in cancer xenograft studies *in vivo* ([@B14], [@B15]).
WD Showed a Higher Radiosensitizing Effect Than WFA in Several Cancer Cell Lines
--------------------------------------------------------------------------------
A colony formation assay was then performed to evaluate the radiosensitizing effect of WFA and WD. As shown in [Figure 2](#F2){ref-type="fig"}, pre-irradiation treatment of human cancer cells with WD for 1 h significantly decreased the surviving fraction after X-ray irradiation for Caco-2 (*p* = 0.021), DU145 (*p* = 0.0025), MCF7 (*p* = 0.0002), A549 (*p* = 0.0159), and SKOV3 (*p* = 0.001) compared to cells only exposed to DMSO. Interestingly, the sensitizing enhancement ratio (SER~0.1~) of WFA and WD calculated at the D10 value was superior for WD than WFA for all cell lines, suggesting that WD has a higher radiosensitizing effect ([Table 1](#T1){ref-type="table"}). In order to determine the mechanisms by which WD achieved its radiosensitizing effect, SKOV3 cells, which displayed the highest radiosensitivity to WD + irradiation (SER~0.1~ = 2.22, [Table 1](#T1){ref-type="table"}) were selected.
{#F2}
######
Sensitizing enhancement ratio (SER~0.1~) of WFA and WD drugs was calculated at the D10 value as the radiation dose needed for radiation alone divided by the dose needed for WFA/WD plus radiation at a survival fraction of 10%.
**Cell line** **SER**~****0.1****~
--------------- ---------------------- ------
22Rv1 0.92 0.99
A549 1.12 1.21
Caco2 1.17 1.29
DU145 1.21 1.37
MCF7 1.49 1.75
SKOV3 1.97 2.22
WD Induced a Persistence of DNA Damages, and Increased the Expression of γH2AX and 53BP1 Foci in Irradiated Cells
-----------------------------------------------------------------------------------------------------------------
Promotion of radiation-induced DNA damages and inhibition of DNA double-strand breaks (DSBs) repair by WD was first assessed in SKOV3 cells. Alkaline comet assay was performed to ascertain the combination of radiation-induced DNA single-strand breaks, double-strand breaks and alkali-labile sites after 1 h pre-treatment with WD. Results showed that 2 Gy-irradiation induced formation of comet tail at 1 h which returned to baseline level at 24 h ([Figure 3A](#F3){ref-type="fig"}). Interestingly, olive tail moment normalized with non-irradiated cells was significantly higher at 1 and 24 h (*p* = 0.0086 and *p* \< 0.0001, respectively) post-irradiation in cells exposed to WD compared to those exposed to vehicle ([Figure 3B](#F3){ref-type="fig"}). As DSBs are the most lethal radiation-induced damages, this type of lesion was further investigated and the formation of γH2AX and 53BP1 foci, two markers of DSBs, was monitored in SKOV3 cells irradiated with X-rays with or without WD pre-treatment. The number of co-localized γH2AX and 53BP1 foci highly increased 1 h after 2 Gy-irradiation compared to sham-irradiated cells. At 6 h post-irradiation, the number of foci started to decrease to return to normal level at 24 h ([Figure 3C](#F3){ref-type="fig"} and [Supplementary Figures 6A,B](#SM1){ref-type="supplementary-material"}). However, SKOV3 cells pre-treated with WD presented more γH2AX foci at 1 and 6 h post-irradiation (*p* = 0.0095 and *p* = 0.019, respectively) and more 53BP1 foci at 6 and 24 h post-irradiation (*p* = 0.026 and *p* = 0.041, respectively) than sham-irradiated cells compared to DMSO-treated cells ([Figures 3D,E](#F3){ref-type="fig"}). Together, these data suggest that WD induced the persistence of DNA damages and especially DSBs.
{#F3}
WD Down-Regulated DNA Damage Repair Proteins in Irradiated SKOV3 Cells
----------------------------------------------------------------------
In order to further investigate how WD prolonged DNA damages, the expression of proteins involved in DNA damage repair was assessed by western blot. SKOV3 cells were pretreated with WD for 1 h before sham or 2 Gy-irradiation. WD was immediately replaced after irradiation by drug-free medium for 1, 6, and 24 h. ATM and DNA-PK kinases, and their phosphorylated forms were overexpressed from 1 h after irradiation and maintained at a high expression level for at least 24 h in DMSO-treated cells ([Figure 4](#F4){ref-type="fig"}). However, in WD-treated cells, their expression level decreased rapidly and almost returned to baseline level after 24 h compared to sham-irradiated samples. XRCC4 protein was overexpressed after irradiation and maintained high level up to 24 h in DMSO-treated cells whereas its expression stayed stable post-irradiation compared to sham-irradiated cells in WD-treated cells. Conversely, Rad51 expression level was identically increased after irradiation between DMSO- and WD-treated cells.
{#F4}
WD Induced Mitotic Catastrophe (MC) in Irradiated SKOV3 Cells
-------------------------------------------------------------
Recent evidences showed that the combined loss of ATM and p53 promoted cellular death due to MC after DSBs ([@B16], [@B17]). SKOV3 is a p53-deficient cell line ([@B18]) and recent study showed that cisplatin-induced DNA damage triggered MC ([@B19]). Since our results also demonstrated a loss of ATM expression after irradiation when combined with WD exposure, we anticipated that WD radiosensitizing effect could promote cellular death by MC. Immunofluorescence assays for α-tubulin and pericentrin stains showed that control cells displayed normal bipolar mitotic spindles, whereas WD-treated cells exhibited an abnormal mitotic spindle with large defects in chromosomal alignment, such as the existence of multi-polar spindle morphology 24 h after 2 Gy-irradiation ([Figure 5A](#F5){ref-type="fig"}). The proportion of multipolar metaphases in WD-treated cells exposed to irradiation increased compared to non-exposed cells and WD-exposed or irradiated cells alone ([Figure 5B](#F5){ref-type="fig"}). In addition, cells displaying aberrant spindles did not progress through anaphase, and the percentage of cells in anaphase significantly decreased for cells exposed to WD and irradiation (*p* = 0.0439) ([Figure 5C](#F5){ref-type="fig"}).
{#F5}
Discussion {#s4}
==========
Although the radiosensitizing effect of WFA has already been demonstrated, this study confirmed the previous observation and also showed, for the first time, the higher radiosensitizing effect of WD. This radiosensitizing effect was obtained by performing a 1 h WD pretreatment at a pre-determined concentration of 0.7 μM. However, since the effect of WD toxicity on cell proliferation varied for each cell line, different concentrations should be tested to optimize this radiosensitizing effect. Our results suggest that WD prolonged radiation-induced DNA damages by inhibiting DNA damage repair. DNA-PKcs and ATM are the two main kinases involved in response to DSBs ([@B20]) and both were shown to be inhibited by the presence of WD after irradiation. DNA-PKcs is usually considered as a core regulator and biomarker of non-homologous end joining pathway (NHEJ) repair pathway while ATM as the initiator of DNA repair via homologous recombination (HR) ([@B21], [@B22]). However, recent studies demonstrated that this separation is not as simple and that the two proteins play an important role in both pathways ([@B23], [@B24]). Indeed, while HR is elevated in DNA-PKcs null cells, DNA-PKcs kinase inhibitors has been reported to suppress HR in normal cells, highlighting possible role of DNA-PK in HR pathway fate ([@B25]). Moreover, ATM levels are regulated by DNA-PKcs ([@B26]) and ATM also phosphorylates DNA-PKcs after irradiation ([@B27]), suggesting that these two proteins cooperate to regulate DNA DSBs repair by NHEJ and HR. Therefore, effector proteins specifically involved in NHEJ and HR pathways such as XRCC4 and Rad51, respectively, were further investigated ([@B28]). Results showed that the expression level of XRCC4 and, interestingly, ^S2056^pDNA-PKcs, decreased in WD pre-treated irradiated cells. Phosphorylation of DNA-PKcs is required to induce conformational changes at broken ends and allows access to XRCC4 complex at these sites to ligate both loose ends ([@B29]). Therefore, both WD-induced XRCC4 and ^S2056^pDNA-PKcs inhibition could result in much reduced fidelity and efficiency of NHEJ. Rad51 expression remained unchanged when irradiated cells were pre-treated with DMSO or WD, suggesting that the compound did not inhibit HR pathway. However, in addition to this immunoblot analysis, supplementary investigations would be necessary to certainly confirm NHEJ pathway inhibition.
This delay in radiation-induced DNA damage repair lead the cells to MC. This result is consistent with previous studies which reported that mutant p53 and inhibition of G2 checkpoint proteins such as ATM, promote radiation-induced MC ([@B30]--[@B32]). A recent study showed that WD induces apoptosis in p53-wild type cells through Bax/Bak dependent pathway, but that Bak can compensate against loss of Bax in p53-null cells and thus still induces apoptosis ([@B33]). However, we did not observe an increase of apoptosis level in SKOV3 p53-null cells when analyzed by TUNEL assay (data not shown). This may be explained by the low dose (0.7 μM) and short incubation time (1 h) that we voluntary selected in this study to induce a radiosensitizing effect while limiting the cytotoxic effect of WD. In addition, since the combined loss of ATM and p53 inhibit cell cycle checkpoints ([@B16]), we did not investigate cell cycle in SKOV3 cells. However, WD has been shown to induce G2/M phase arrest in pancreatic adenocarcinoma cells ([@B34]). Thus, additional investigations should be performed to assess if cell cycle arrest could play a role in WD radiosensitizing effect, in particular in p53-wild type cells. Indeed, because of their highest SER~0.1~, mechanistic analyses in this study focused on p53-null SKOV3 cells, however, WD also radiosensitized other cancer cell lines to a lesser degree, including p53-functional cell lines (i.e., A549 and MCF7). This suggests that the radiosensitizing effect of WD is independent of p53 status and the fate of radiation-induced cell death in p53-effective cells would mainly depend on WD-induced ATM (or DNA-PKcs) inhibition. WD has been previously shown to increase JNK and p38MAPK phosphorylation ([@B35]), two pathways strongly responsive to ionizing radiation and that influence tumor cell radiosensitivity because of their activity associated with radiation-induced DNA damage response ([@B36]). Radiation-induced JNK activation is dependent on ATM and its activation activates caspases and regulates proteins implicated in apoptosis regulation, including p53 ([@B36], [@B37]), while MAPKK-p38γ cascade is required for radiation--induced G2 arrest ([@B38]). These mechanisms could explain the radiosensitizing effect of WD that we observed in p53 wild-type cell lines. In addition, JNK and p38MAPK activation enhances the ceramide accumulation, an important messenger for radiation response that triggers radiation-induced apoptosis ([@B39]--[@B42]). The effect of WD on the ceramide production and MAPKK pathways activation upon irradiation needs to be investigated further to assess if it plays a role on the radiosensitizing effect of WD, independently, or not, of p53. Further studies involving additional cell lines, with different p53 status, and *in vivo* experiments would be required to surely affirm radiosensitizing effect of WD and its cellular effect.
After WFA, WD highlighted the potential of withanolides to act as promising sources of radiosensitizers. Several studies investigated the relationship between the structure and the diverse bioactivities reported for withanolides to better understand their mechanism of action and modify the withanolide scaffold accordingly to enhance its effect ([@B43]--[@B46]). Although WD-treated cells displayed a lower expression level of DNA damage repair proteins after irradiation than non-treated cells (maximal at 24 h), this inhibition is not complete. Thus, additional studies are required to identify the key structural components of withanolides responsible for this inhibition in order to optimize their radiosensitizing effect.
Data Availability Statement {#s5}
===========================
All datasets generated and analyzed for this study are included in the article/[Supplementary Material](#SM1){ref-type="supplementary-material"}.
Author Contributions {#s6}
====================
JL and FZ designed the research. JL, TC, LM, and EW conducted the experiments. JL, TC, and LM analyzed the data. JL drafted the manuscript. J-LV, MC, AG, and FZ revised the manuscript. All authors read and approved the final manuscript.
Conflict of Interest
--------------------
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Authors would like to greatly thank Colleen\'s Dream Foundation for its financial support. We also thank Dr. Kurt Gustin from the Biomedical Imaging Core at the UA College of Medicine---Phoenix for providing epifluorescence imaging service.
**Funding.** Colleen\'s Dream Foundation funded part of this work. The funders had no role in the study design, data collection, data analysis, decision to publish, or manuscript preparation.
Supplementary Material {#s7}
======================
The Supplementary Material for this article can be found online at: <https://www.frontiersin.org/articles/10.3389/fonc.2019.01468/full#supplementary-material>
######
Click here for additional data file.
DSBs
: double-strand breaks
HR
: homologous recombination
MC
: mitotic catastrophe
NHEJ
: non-homologous end joining
RT
: radiation treatment
WD
: withanolide D
WFA
: withaferin A.
[^1]: Edited by: Gaspar Kitange, Mayo Clinic, United States
[^2]: Reviewed by: Xi Yang, Fudan University, China; Jinping Liu, University of Pennsylvania, United States
[^3]: This article was submitted to Radiation Oncology, a section of the journal Frontiers in Oncology
| {
"pile_set_name": "PubMed Central"
} |
{#sp1 .73}
{#sp2 .74}
{#sp3 .75}
{#sp4 .76}
{#sp5 .77}
{#sp6 .78}
{#sp7 .79}
{#sp8 .80}
{#sp9 .81}
| {
"pile_set_name": "PubMed Central"
} |
**Conflict of interests**: MA, YL, EG, NE, JL, MT, ED are employees of GenomeDx Biosciences. Dr. Lotan has received research funding for other projects from GenomeDx. Other authors declare no potential conflicts of interest.
Adeno
: prostatic adenocarcinoma
ADT
: androgen deprivation therapy
AUC
: area under the curve
CCLE
: Cancer Cell Line Encyclopedia
DRS
: drug response score
FFPE
: formalin‐fixed paraffin‐embedded
GDSC
: Genomics of Drug Sensitivity in Cancer
GG
: grade group
GRID
: Genomic Resource Information Database
HDAC
: histone deacetylases
IC50
: half maximal inhibitory concentration
JHMI
: John Hopkins Medical Institute
mCRPC
: metastatic castration resistance prostate cancer
PARP
: poly (ADP‐ribose) polymerase
PD
: poorly differentiated
RP
: radical prostatectomy
SC/NE
: small cell neuroendocrine
SCGScore
: small cell genomic score
SCLC
: small cell lung carcinoma
Introduction {#ijc32430-sec-0001}
============
Genomic diversity and clinical relevance of small cell/neuroendocrine (SC/NE) are well established in the castration‐resistant prostate cancer (CRPC) setting,[1](#ijc32430-bib-0001){ref-type="ref"}, [2](#ijc32430-bib-0002){ref-type="ref"}, [3](#ijc32430-bib-0003){ref-type="ref"} but it is less characterized in treatment‐naïve primary tumors due to its rarity. It is well established that some aggressive primary prostate tumors share features with late‐stage therapy‐resistant disease and unsurprisingly confer a poor prognosis.[4](#ijc32430-bib-0004){ref-type="ref"} Certain genomic alterations that are enriched in SC/NE, including the *RB1* and *TP53* pathway aberration, can often be found in primary tumors, albeit at a lower overall frequency than metastatic disease.[5](#ijc32430-bib-0005){ref-type="ref"} It is therefore conceivable that some localized adenocarcinomas are inherently more capable of "transdifferentiating" to SC/NE once under the concerted influence and selective pressure of AR‐targeted therapy.
Most prior studies comparing SC/NE and high‐grade adenocarcinoma that have identified SC/NE signatures and candidates for targeted therapy were based on few SC/NE patients due to disease rareness.[6](#ijc32430-bib-0006){ref-type="ref"} To accurately identify the wide spectrum of SC/NE and characterize its biology in a large cohort, we leveraged our previously reported meta‐NE signature[6](#ijc32430-bib-0006){ref-type="ref"} and genome‐wide expression data of more than 25,000 primary tumor samples from the Decipher Genomic Resource Information Database (GRID) registry. The meta‐NE signature was identified to predict histologically SC/NE tumors, but we found genomic heterogeneity within histologically SC/NE tumors. In this work, we refined the meta‐NE signature and modeled it as a single score to predict tumors that are genomically similar to SC/NE tumors. We hypothesize that some primary adenocarcinomas harbor features of SC/NE and that patients with such tumors are at higher risk of progression under the influence of AR‐targeted therapy. Therefore, our objective is to develop a genomic tool to identify and characterize primary tumors with SC/NE‐like features and differentiate them from poorly differentiated (PD) adenocarcinoma with the goal of understanding their biology and identifying potential targeted therapy.
Materials and Methods {#ijc32430-sec-0002}
=====================
Patient cohorts {#ijc32430-sec-0003}
---------------
Our initial discovery cohort (John Hopkins Medical Institute \[JHMI\]‐SC) consisted of 33 formalin‐fixed paraffin‐embedded (FFPE) tumors retrieved from John Hopkins Registry.[6](#ijc32430-bib-0006){ref-type="ref"} This cohort included six morphologically diagnosed "pure" SC/NE specimens (pure SC), 11 high‐grade adenocarcinomas (mostly Grade Group \[GG\] 5), 1 adenocarcinoma with NE differentiation, as well as tumor foci from 15 specimens harboring concurrent small cell and adenocarcinoma histology. In these 15 specimens, either the predominant adenocarcinoma foci were sampled (termed mixed‐prostatic adenocarcinoma \[Adeno\], *n* = 5), or the small cell foci (termed mixed‐SC, *n* = 10). Additionally, we used 97 FFPE GG5 adenocarcinoma samples from Johns Hopkins natural history cohort[7](#ijc32430-bib-0007){ref-type="ref"} for model development.
We retrieved external gene expression data from eight publicly available datasets obtained from patients with SC/NE (as well as from more typical AR‐positive metastatic CRPC \[mCRPC\]): Beltran *et al*.[2](#ijc32430-bib-0002){ref-type="ref"} (mCRPC:34, mCRPC‐SC/NE:15), Kumar *et al*.[1](#ijc32430-bib-0001){ref-type="ref"} (mCRPC:151, mCRPC‐SC/NE:20), LTL331R system(GSE59984),[8](#ijc32430-bib-0008){ref-type="ref"} GSE66187 with typical SC/NE and atypical SCC with AR‐positive, GSE43346 with small cell lung carcinoma (SCLC) and normal tissues, Cancer Cell Line Encyclopedia (CCLE) cell lines[9](#ijc32430-bib-0009){ref-type="ref"} with SCLC or non‐SCLC and Genomics of Drug Sensitivity in Cancer (GDSC) cell lines[10](#ijc32430-bib-0010){ref-type="ref"} with SCLC or non‐SCLC, and mCRPC cohort (mCRPC:107, NE:12) from Aggarwal *et al*.[3](#ijc32430-bib-0003){ref-type="ref"} Finally, we used independent cohorts of low‐grade adenocarcinoma prior to androgen deprivation therapy (ADT) treatment, adenocarcinoma, PD and NE post‐ADT, and *de novo* SC/NE from the University of Calgary. Tumors slides were reviewed by one of the study pathologists (T.A.B.) to characterize SC/NE features. Tumors were stained with *SYP*, chromogranin, *CD56*, *PSA*, *PSAP* and *AMACR*.
To identify primary prostatic tumors that are genomically similar to SC/NE, we used the expression profile of prospective 17,967 radical prostatectomy (RP) and 6,697 biopsies deidentified and anonymized cases from the Decipher Genomic Resource Information Database (GRID) with basic demographic and pathological data obtained through clinical use of the Decipher test. Furthermore, we used 283 neuroblastoma samples (GSE85047) run on the same Human Exon Array technology used for the Decipher GRID comparative analysis. Neuroblastoma samples were used as a positive control for neuroendocrine biomarkers and signatures validation.
Tumor transcriptome profiling {#ijc32430-sec-0004}
-----------------------------
Tumor specimens for the 33 JHMI‐SC, 97 JHMI and GRID samples were obtained from archived paraffin blocks and RNA extraction was performed using the Decipher prostate cancer classifier assay as previously described.[11](#ijc32430-bib-0011){ref-type="ref"} Briefly, RNA was amplified, labeled and hybridized to Human Exon 1.0 ST microarrays (Affymetrix, Santa Clara, CA) covering 46,050 genes. The SCAN algorithm was used for individual patient profile preprocessing and normalization.[12](#ijc32430-bib-0012){ref-type="ref"} ComBat[13](#ijc32430-bib-0013){ref-type="ref"} algorithm was used for adjusting expression for dataset effect. Neuroblastoma cohort was normalized with SCAN algorithm[12](#ijc32430-bib-0012){ref-type="ref"} as it was profiled on the same Human Exon 1.0ST array.
Development of SC genomic fingerprint and score {#ijc32430-sec-0005}
-----------------------------------------------
We started with 306 neuroendocrine and small cell related genes recently reported from meta‐analysis.[6](#ijc32430-bib-0006){ref-type="ref"} Since we hypothesized that there is molecular heterogeneity within samples that are not captured by histological annotations, we conducted consensus clustering, using ConsensusClusterPlus R package,[14](#ijc32430-bib-0014){ref-type="ref"} on the 33 samples with pam clustering algorithm, Pearson correlation distance, ward linkage, 1,000 iterations and 80% sampling rate parameters identifying three clusters (SC/NE, mixed and Adeno). Genes differentially expressed between SC/NE and Adeno clusters were used to define better segregation. For model development, we pooled 97 GG5 samples from natural history cohort with the 33 JHMI‐SC cohorts to define a training set. We grouped genes into modules; genes were ranked based on their expression change in the LTL331R system upon NE differentiation, and grouped into 10 modules based on gene expression fold change (Supporting Information Table [S1](#ijc32430-supitem-0003){ref-type="supplementary-material"}). For each module of genes, a logistic regression with ridge penalty was fitted and the penalty parameter was selected *via* 10‐fold cross‐validation, which created 10 models. The prediction probabilities from these 10 models were further averaged with weights proportional to their area under the curve (AUC) on the training data, such that the more powerful model received a higher weight in the final prediction. The final averaged prediction probability is called SCGScore. Flow chart of gene reductions, model developments and overview of model\'s evaluation are detailed in [Supporting Information Methods](#ijc32430-supitem-0001){ref-type="supplementary-material"}.
Drug response score {#ijc32430-sec-0006}
-------------------
Using *in vitro* drug sensitivity and microarray data from the GDSC panel, we generated gene signatures predicting lung cancer cell lines (154 cells) sensitivity to 265 drugs from prostate cancer clinical trials. For each drug, we identified gene signature (drug response related genes and their correlations to the half maximal inhibitory concentration \[IC50\] value). Most significantly correlated genes were selected and the expression of the corresponding genes in the Decipher GRID was extracted for drug response score (DRS) calculations. A patient‐specific DRS was calculated using these correlation coefficients (Cor) as weighting factors of the corresponding gene expression normalized by the sum of Cor. DRSs were calculated for 265 drugs for every patient in the Decipher GRID characterize their associations with SCGScore.
Statistical analysis {#ijc32430-sec-0007}
--------------------
Statistical analyses were performed in R version 3.0. All statistical tests were two‐sided using a significant level of 0.05. Chi‐square test was used for statistical associations between categorical variables (GG) and the Wilcoxon test was used for continuous variables (DRSs, Decipher score).
Results {#ijc32430-sec-0008}
=======
Development of a prostatic small cell genomic fingerprint {#ijc32430-sec-0009}
---------------------------------------------------------
To develop a molecular classifier to identify SC/NE prostate cancer in the localized, treatment‐naïve setting, we first selected 306 genes associated with NE prostate cancer as previously reported by our group.[6](#ijc32430-bib-0006){ref-type="ref"} Since we hypothesized that there is molecular heterogeneity underlying the histological annotations, the 306 genes were used to guide the consensus clustering of the 33 prostate samples from JHMI‐SC cohort revealing three clusters with distinct biological and histological characteristics. The first cluster was enriched with histologically pure SC and mixed‐SC (SC/NE cluster), the second was enriched with histological adenocarcinoma (Adeno cluster) and the third showed atypical small cell histology with overexpression of AR‐signaling (mixed cluster; Fig. [1](#ijc32430-fig-0001){ref-type="fig"} *a*, Supporting Information Fig. [S1](#ijc32430-supitem-0002){ref-type="supplementary-material"}). When the cluster memberships were evaluated with respect to histology, we noted a small number of tumors with discordant genotypes relative to their histology (i.e., a mixed‐SC tumor was found in the Adeno cluster).
![Small cell genomic fingerprint discovery and evaluation. (*a*) Consensus clustering of the 306 genes revealing three clusters: one enriched with histologically small cell (SC/NE), one enriched with Adeno (Adeno) and one mixed. (*b*) t‐SNE of the 212 genes (212signature) differentially expressed between SC/NE cluster and adenocluster showing clear discrimination between the groups. (*c*--*e*) Identified 212signature is evaluated in four public cohorts with NE. (*c*) Genes ordered based on their fold change in the LTL331R system. (*d*) mCRPC‐NE samples were clustered together in mCRPC cohorts with few samples exceptions. (*e*) 212signature also discriminated small cell lung carcinoma (SCLC) from non‐SCLC in Cancer Cell Line Encyclopedia (CCLE) and Genomics of Drug Sensitivity in Cancer (GDSC). \[Color figure can be viewed at <http://wileyonlinelibrary.com>\]](IJC-145-3453-g001){#ijc32430-fig-0001}
To circumvent some of the challenges posed by the molecular heterogeneity in histological annotation, we compared the SC/NE and Adeno clusters finding 216 genes were differentially expressed (120 upregulated and 96 downregulated) after multiple testing adjustment (adjusted *p* \< 0.05) within the SC cluster (Supporting Information Table [S1](#ijc32430-supitem-0003){ref-type="supplementary-material"}). These 216 genes represented a set of candidate NE‐specific genes suitable for model development. Only 19 genes were in common with the 69 SC/NE signature presented by Aggarwal *et al*.[3](#ijc32430-bib-0003){ref-type="ref"}
These 216 genes were found to readily distinguish the three tumor types, with the mixed cluster demonstrating intermediate behavior between the well‐separated SC/NE and Adeno clusters (Fig. [1](#ijc32430-fig-0001){ref-type="fig"} *b*). The SC/NE and mixed clusters had genomic properties consistent with typical small cell biology, including high cell cycle activity, high NE biomarker expression (i.e., *PCSK1*, *SCG2* and *CHGB*) and expression of low *AR*, *REST* and *RB1* (Supporting Information Fig. [S2](#ijc32430-supitem-0002){ref-type="supplementary-material"}). To confirm the association of these 216 genes with neuroendocrine differentiation, we downloaded gene expression data generated from the LTL331R system.[8](#ijc32430-bib-0008){ref-type="ref"} This is a mouse model where an initial adenocarcinoma tumor, upon host castration, relapses as terminally differentiated NE making this an ideal platform to further refine the 216 gene set. This cross‐platform analysis found the majority (212/216) of genes were covered by the Agilent platform used by the authors of the initial LTL331R study.
We next ranked the 212 genes based on their fold change (FC) upon NE differentiation in the LTL331R system (Fig. [1](#ijc32430-fig-0001){ref-type="fig"} *c*). The most upregulated genes included *PCSK1*, *CHGA*, *SEZ6L* and *RNF182* with FC \> 9 folds, while AR‐regulated genes like *FOLH1* and *FOLH1B* were the most downregulated (FC \< 8 folds). To reduce the bias of the individual genes, we used the fold‐change ranking to categorize the 212 genes into 10 distinct modules, including seven upregulated (total *n* = 119) and three downregulated (total *n* = 93) groups (Supporting Information Table [S1](#ijc32430-supitem-0003){ref-type="supplementary-material"}). The vast majority of these genes were consistent with the established molecular landscape of SC/NE, including notable SC/NE biomarkers (e.g., *CHGA*) or drivers of SC/NE progression (e.g., *PEG10*). AR‐regulated genes (i.e., *KLK2* and *PSMA*) were also included, presumably due to their considerable downregulation in AR‐negative SC/NE. The 212 genes also included genes associated with cell proliferation (*TOP2A*, *MCM*, *TPX2* and *EZH2*) and those downstream of the *E2F1*‐*RB1* axis (*CLSPN*). Based on these data, we termed these 212 genes, the "212signature."
Evaluation of the 212signature in treatment‐induced small cell prostate carcinoma and SCLC {#ijc32430-sec-0010}
------------------------------------------------------------------------------------------
To evaluate the 212signature, we first applied it to eight publicly available transcriptome datasets containing histologically confirmed NE in background of mCRPC and SCLC in a background of non‐SCLC (Figs. [1](#ijc32430-fig-0001){ref-type="fig"} *c*--[1](#ijc32430-fig-0001){ref-type="fig"} *e*, Supporting Information Fig. [S3](#ijc32430-supitem-0002){ref-type="supplementary-material"}). In prostate, SC/NE specimens generally grouped together based on the 212signature, although there were some outliers. Across the three mCRPC cohorts tested (Kumar *et al*., Beltran *et al*. and Aggarwal *et al*.), we found similar patterns of clustering. Even though the 212 genes were selected from *de novo* SC/NE, the signature remains effective in mCRPC setting differentiating treatment‐induced NE. In a cohort of SCLC and normal samples from 42 tissues (GSE43346), SCLC showed to have unique profiles compared to normal tissues from various sites[15](#ijc32430-bib-0015){ref-type="ref"} (Supporting Information Fig. [S3](#ijc32430-supitem-0002){ref-type="supplementary-material"}). However, central nervous system tissues (brain, cerebellum, cerebral cortex, hippocampus and thalamus) clustered with the SCLC. We also selected SCLC and non‐SCLC cell lines from the CCLE database and GDSC finding the 212signature discriminated SCLC and NSCLC cell lines (Fig. [1](#ijc32430-fig-0001){ref-type="fig"} *e*). Here it is worth mentioning that we used clustering based on the 212signatures to assess the feasibility of the genes in the signatures to assess the biological signal. Taken together, these data suggest that the 212signature can be applied to other tumor types with NE features, regardless of the tissue of origin.
Development of a small cell genomic score to identify SC/NE tumors {#ijc32430-sec-0011}
------------------------------------------------------------------
We used the 212signature to develop a single sample genomic classifier or small cell genomic score (SCGScore) model using a training cohort of 97 GG5 and 33 JHMI‐SC cohorts. When we generated new clusters with this merged cohort, we found the majority of SC/NE and mixed cases clustered together (Fig. [2](#ijc32430-fig-0002){ref-type="fig"} *a*, Supporting Information Fig. [S4](#ijc32430-supitem-0002){ref-type="supplementary-material"}), but three samples from the natural history cohort also clustered with the SC/NE group. These three cases had developed metastasis within 2 years and had developed CRPC after subsequent hormonal therapy. Upon pathological rereview, one case was found to have small cell morphology. These cases were therefore included in the final SC/NE group (*n* = 23) for comparison to the Adeno group (*n* = 107). Using the SCGScore model, we found significantly higher scores in the SC/NE group and mixed group compared to the Adeno group (Fig. [2](#ijc32430-fig-0002){ref-type="fig"} *b*). Additionally, the SC/NE cases had high scores based on prognostic signatures[16](#ijc32430-bib-0016){ref-type="ref"} indicating that they are very high‐risk tumors, but these signatures had lower sensitivity compared to the SCGScore which had a much higher rate of false positives (Fig. [2](#ijc32430-fig-0002){ref-type="fig"} *c*, Supporting Information Fig. [S5](#ijc32430-supitem-0002){ref-type="supplementary-material"}). This suggests that prognostic signatures are not suitable for identifying small cell carcinoma.
![Development and validation of small cell genomic score (SCGScore). (*a*) Pooling 97 GG5 adenocarcinoma from JHU with the 33 samples to define 23 SC and 107 adenocarcinomas set for modeling. (*b*) SCGScore is significantly higher in SC/NE compared to mixed and adenocarcinoma. (*c*) Existing prognosis signature (Decipher) is not ideal for discriminating between SC/NE and adenocarcinoma with low sensitivity. (*d*) Independent validation of SCGScore in Calgary cohort showing that post‐ADT NE tumors have high scores compared to PD and high‐grade adenocarcinoma. Also, *de novo* NE with NE biomarker IHC positivity has higher scores that histologically NE tumors with negative IHC. (*e*) Validating the model in public mCRPC cohorts showing superior performance predicting SC/NE in the setting of mCRPC. \[Color figure can be viewed at <http://wileyonlinelibrary.com>\]](IJC-145-3453-g002){#ijc32430-fig-0002}
Validating SCGScore in four cohorts of *de novo* NE and treatment‐induced NE {#ijc32430-sec-0012}
----------------------------------------------------------------------------
Next, we evaluated the SCGScore model on four independent cohorts from the University of Calgary (Fig. [2](#ijc32430-fig-0002){ref-type="fig"} *d*) with patients presenting with either *de novo* SC/NE or post‐ADT NE, and Kumar *et al*., Beltran *et al*. and Aggarwal *et al*. cohorts. For comparison, we used post‐ADT PD samples and pre‐ADT adenocarcinoma. The post‐ADT NE group scored higher for SCGScore compared to pre‐ADT Adeno group, post‐ADT Adeno and post‐ADT PD achieving an AUC of 0.98 suggesting that SCGScore could be used to discriminate between histologically challenging cases (i.e., PD *vs*. small cell). Tumors with *de novo* NE, tumors with NE+/*AR*− had higher scores than those with NE+/*AR*+. The model also achieved superior performance (average AUC 0.95) when applied to mCRPC cohorts distinguishing mCRPC‐NE from mCRPC (Fig. [2](#ijc32430-fig-0002){ref-type="fig"} *e*). Also, among 180 patients with GG1--4 from a case--cohort[7](#ijc32430-bib-0007){ref-type="ref"} (GG5 were excluded as they were used in model training), two patients had relatively high SCGScore with metastasis‐free survival rate less than 2 years. Finally, we characterized the clinical impact of the SCGScore in retrospective RP tissues (*n* = 70) treated with adjuvant ADT and have GG5. Patients with the highest SCGScore had a shorter median metastasis‐free survival than patients with low SCGScore (18 *vs*. 54 months, *p* = 0.006).
Evaluating SCGScore in prospective prostate RP and biopsy tissues from the Decipher GRID {#ijc32430-sec-0013}
----------------------------------------------------------------------------------------
To define the prevalence of primary prostate tumors with SC/NE‐like transcriptomic profiles, we measured the SCGScore in 17,967 RP and 6,697 biopsy prospective tissues collected from the clinical use of the Decipher test (Fig. [3](#ijc32430-fig-0003){ref-type="fig"} *a*). Additionally, as neuroblastoma is histologically similar to small cell carcinoma and expresses NE markers, we included a neuroblastoma cohort profiled on the same Human Exon microarray platform as an internal control. The SC/NE samples from discovery cohort and neuroblastoma control samples were found to have very high SCGScores (Fig. [3](#ijc32430-fig-0003){ref-type="fig"} *a*). In both the RP and biopsy settings, 1% of sample (151 in RP, 58 in biopsy) was classified as SC/NE‐like tumors by the model using a cutoff of 0.25 as defined from the discovery cohort. However, only a few tumor samples had SCGScores as high as the SC/NE tumors and these were all GG5. This indicates that the predicted SC/NE‐like tumors based on the model are not fully NE differentiated or that perhaps they are molecularly heterogeneous compared to typical NE. Comparing the SCGScore with previously developed model by Kumar *et al*., we found good correlation between the two signatures in the 33 SC/NE samples (*R* ^2^ = 0.56, *p* \< 0.001) and mCRPC samples (*R* ^2^ = 0.66, *p* \< 0.001)[1](#ijc32430-bib-0001){ref-type="ref"}, [2](#ijc32430-bib-0002){ref-type="ref"} (Supporting Information Figs. [S6A and S6B](#ijc32430-supitem-0002){ref-type="supplementary-material"}), with weaker correlation (*R* ^2^ = 0.37, *p* \< 0.001) in the primary tumors from prospective RP and biopsy samples (Supporting Information Figs. [S6C and S6D](#ijc32430-supitem-0002){ref-type="supplementary-material"}) even though there is minimal gene overlap between the two signatures. Similarly, we found stronger correlation between SCGScore and the model developed by Tsai *et al*. in the 33SC/NE samples (*R* ^2^ = 0.85, *p* \< 0.001) but not in the prospective cohort (*R* ^2^ = 0.18, *p* = 0.01) (Supporting Information Fig. [6E](#ijc32430-supitem-0002){ref-type="supplementary-material"}).
![Evaluating SCGScore in a large prospective cohort. (*a*) Evaluating the SCGScore in prospective prostate RP (*n* = 17,967) and BX (*n* = 6,697) and neuroblastoma (*n* = 283) compared to SC/NE tumors. (*b*) SCGScore across pathological GG in RP samples. (*c*) Frequency of predicted SC/NE‐like across GG showing higher frequency in GG5. (*d*) Predicted SC/NE‐like patients have distinct genomic fingerprint compared to GG5 (*n* = 1,679) and distinct pathway activity. \[Color figure can be viewed at <http://wileyonlinelibrary.com>\]](IJC-145-3453-g003){#ijc32430-fig-0003}
Characterizing the clinical and molecular features of patients with high SCGS {#ijc32430-sec-0014}
-----------------------------------------------------------------------------
To gain further insight into the clinical and molecular characteristics of the predicted 151 RP patients with SC/NE‐like features from the prospective cohort, we found the SCGScore tended to trend with GGs (Figs. [3](#ijc32430-fig-0003){ref-type="fig"} *b* and [3](#ijc32430-fig-0003){ref-type="fig"} *c*, Supporting Information Fig. [S7A and S7B](#ijc32430-supitem-0002){ref-type="supplementary-material"}, Table [S2](#ijc32430-supitem-0003){ref-type="supplementary-material"}). GG 5 was enriched with SC/NE‐like patients (2.7%) compared to GG 4 (1.5%), GG 3 (0.53%), GG 2 (0.5%) and GG 1 (0.5%; Fig. [3](#ijc32430-fig-0003){ref-type="fig"} *c*), with a similar trend observed for the biopsy samples (Supporting Information Fig. [S7B](#ijc32430-supitem-0002){ref-type="supplementary-material"}). Interestingly only 31% of the SC/NE‐like cases were GG 5 and about 40% were GG 1--3 suggesting that NE differentiation could begin in early prostate cancer development. In terms of metastatic potential, 80% of the SC/NE‐like patients had high Decipher compared to 31% for Adeno patients. A similar frequency of high Decipher risk scores (74%) was seen in the prospective biopsy SC/NE‐like patients.
To gain a deeper understanding of the pathway activity in SC/NE‐like tumors, we compared the 151 SC/NE‐like patients to 1,679 GG5 RP patients. The SC/NE‐like tumors had a distinct pathway activity profile (Fig. [3](#ijc32430-fig-0003){ref-type="fig"} *d*, Supporting Information Figs. [S8 and S9](#ijc32430-supitem-0002){ref-type="supplementary-material"}). SC/NE‐like tumors had higher activity of KRAS, hypoxia and immune pathways, with lower activity of xenobiotic metabolism, PI3K‐signaling, AR‐signaling and ER‐signaling. This data suggests SC/NE‐like tumors as a biologically distinct, aggressive subclass of PCa with distinctive therapeutic targets.
Characterizing the drug response profile of patients with high SCGScore {#ijc32430-sec-0015}
-----------------------------------------------------------------------
As the initial 212signature was able to discriminate between SCLC and NSCLC in both cell lines (CCLE & GDSC) and in tissues (Fig. [1](#ijc32430-fig-0001){ref-type="fig"} *e*, Supporting Information Fig. [S3](#ijc32430-supitem-0002){ref-type="supplementary-material"}). When we applied SCGScore to the GDSC cohort, we found higher SCGScore values in SCLC compared to squamous, NSCLC (Supporting Information Fig. [S10](#ijc32430-supitem-0002){ref-type="supplementary-material"}). For prostate cancer cell lines in GDSC, NCIH660 neuroendocrine prostate cancer cell line had a SCGScore of 0.93 (Supporting Information Fig. [S10](#ijc32430-supitem-0002){ref-type="supplementary-material"}), but, this cell line lacked IC50 values. Next, we correlated the SCGScore with IC50 in 154 lung cancer cell lines from GDSC where a negative correlation would suggest a sensitivity to a given drug, while a positive score would suggest resistance. We found the SCGScore was positively correlated with the IC50 of Erlotinib, Trametinib, Dasatinib, Lapatinib and Afatinib, but was negatively correlated (better response) with Navitoclax, Vorinostat and Belinostat (Supporting Information Table [S3](#ijc32430-supitem-0002){ref-type="supplementary-material"}).
Unfortunately, none of the six prostate cancer cell lines with combined IC50 and expression data were SC/NE‐like, and therefore we were unable to directly correlate SCGScore to IC50 in the prostate space. As a surrogate for this analysis, we built a drug response signature for each of the 265 drugs from lung carcinoma cell lines. For each drug, we defined a signature that was grouped into a score representing the drug response; high scores are indicative of potential response while low scores would suggest a lack of potential response. We applied these drug response models to our prospective prostate cohort comparing SC/NE‐like to GG5 patients. DRS of histone methyltransferase (SGC0946), histone deacetylases (HDAC) inhibitor (CAY10603, Belinostat) and poly (ADP‐ribose) polymerase (PARP) inhibitor (Rucaparib) showed the highest positive correlation with SCGScore (Figs. [4](#ijc32430-fig-0004){ref-type="fig"} *a*--[4](#ijc32430-fig-0004){ref-type="fig"} *c*, Supporting Information Table [S4](#ijc32430-supitem-0002){ref-type="supplementary-material"}) suggesting potential response. To confirm some of these predictions in neuroendocrine prostate cancer cells, we used drug response profiles from the CCLE experiments that have neuroendocrine prostate cancer cell line NCI‐H660. Drug response profiles showed high sensitivity (low IC50) to Olaparib and Vorinostat supporting predicted DRS results observed in the prospective cohort (Figs. [4](#ijc32430-fig-0004){ref-type="fig"} *d* and [4](#ijc32430-fig-0004){ref-type="fig"} *e*). These data suggest that SC/NE‐like are distinct from GG5 tumors and potentially share common therapeutic targets with SCLC.
![Therapeutic implications of SCGScore. SC/NE‐like are predicted to be more sensitive to (*a*) PARP inhibitors, (*b*) HDAC inhibitors and (*c*) methylation inhibitors. (*d*--*e*) NCIH660 (prostatic NE cell line) showed to respond to both PAPR and HDAC inhibitors. \[Color figure can be viewed at <http://wileyonlinelibrary.com>\]](IJC-145-3453-g004){#ijc32430-fig-0004}
Discussion {#ijc32430-sec-0016}
==========
Herein, we identify a subset of treatment‐naïve pPCa that is molecularly analogous to mCRPC‐NE and SCLC using a 212 gene signature. These genes were grouped these into 10 modules spanning several key pathways including AR‐signaling, NE differentiation, *RB*‐loss and cell proliferation, with NE differentiation pathway having the largest weight. In what we believe this is the largest ever study that uses transcriptomics to identify SC/NE‐like prostatic tumors.
The data presented here offer several novel insights. First, we demonstrate that prostatic SC/NE are molecularly analogous to SCLC suggesting that prostatic SC/NE suggesting their shared common pathways and potentially common therapeutic targets.
Second, *de novo* SC/NE and treatment‐induced NE have a similar molecular profile in terms of expressing NE biomarkers and low AR‐activity. Third, about 0.5--1% of GG1 and GG2 have high SCGScore suggesting that NE differentiation can occur in very early stages of cancer. Fourth, the SCGScore might be useful for identifying patients with a NE‐like genomic profile who may not be suitable for AR‐targeted therapy and may be candidates for novel therapeutics. Thus, implementing SCGScore as part of commercially used prognostic test allows for the identification of patients for enrollment in such clinical trials.
Our study did have some noteworthy limitations. Lack of IHC staining of NE markers and lack of outcome in the prospective cohort limited assessing our model in the prospective cases. Also, lack of drug response data in prostatic cell lines prevented us from building prostate specific drug response models.
In conclusion, evaluating SCGScore in very large cohorts of primary tumors provides insights on the early differentiation of neuroendocrine carcinoma. SCGScore can be used together with prognostic models to identify high‐risk tumors that are more likely to develop neuroendocrine disease, and thus not suitable candidates for hormonal therapy.
Supporting information
======================
######
**Appendix S1**: Supporting Information
######
Click here for additional data file.
######
**Figure S1**: T‐SNE scatter plot of the 33 JHMI‐SC samples based on the 306 meta genes from Tsai *et al*. 2017.
**Figure S2**: Genomic‐based groups of Adeno, mixed and SC/NE are representing the expecting biology where SC/NE have high expression of cell cycle, lower AR, REST, and higher neuroendocrine biomarkers.
**Figure S3**: Evaluating the 212signature in mCRPC cohorts. A. mCRPC cohort (Aggarwal *et al*.) with NE tumors, B. mCRPC cohort (GSE66187) with NE‐mCRPC samples stained with CHGA, SYP and AR. C. Also, 212signature is able to discriminate SCLC from various normal tissues.
**Figure S4**: T‐SNE plot of the 33 JHMI‐SC samples with the 97 GG5 from John Hopkins natural history cohort.
**Figure S5**: Penney signature (prognostic signature) showing high sensitivity to predict SC/NE where more than 50% of the Adeno had high scores similar to SC/NE.
**Figure S6**: Correlation plots between existing NE signature (Kumar NE) and SCGScore in mCRPC setting (A‐B) and GRID primary data (C‐D). E. Correlation plot between Tsai *et al*. signature and SCGScore in GRID primary data.
**Figure S7**: SCGScore distribution across Grade Groups (GG). A. Patients with high SCGScore have GG5 tumors in biopsy. B.GG5 in biopsy showed to be more enriched with SC/NE‐like tumors (3%) compared to GG1‐3 that have only 0.5%.
**Figure S8**: T‐SNE of the 151 predicted SC/NE‐like from GRID RP and GG5 patients from GRID RP. Results showing that predicted SC/NE‐like tumors are distinct from GG5 based on the 212signature.
**Figure S9**: Pathway activity associating pathways with SC/NE‐like tumors.
**Figure S10**: SCGScore across cell lines from GDSC showing SCLC have high scores while other tumors not. Prostate cancer NE cell line (NCIH600) have high score of SCGScore.
######
Click here for additional data file.
######
**Table S1** 212genesignature with their *p*‐value in discovery JHMI‐SC cohort and FC in 331R system
**Table S2**: Pathological characteristics of SC/NE‐like tumors and Adeno samples from prospective RP cohort
**Table S3**: Correlation between calculated SCGScore and IC50 for lung cancer sample in GDSC cohort
**Table S4**: Correlation between calculated Drug response scores and SCGScore in prospective RP patients. Higher correlation indicates higher potential response.
######
Click here for additional data file.
This work was supported in part by the Prostate Cancer Foundation Young Investigator Award (to T.A.B). This work was also supported by Prostate Cancer Canada and is proudly funded by the Movember Foundation Grant \#B2013‐01.
| {
"pile_set_name": "PubMed Central"
} |
{#sp1 .325}
| {
"pile_set_name": "PubMed Central"
} |
Like human immunodeficiency virus type 1 (HIV-1), human immunodeficiency virus type 2 (HIV-2) causes AIDS-defining events, yet immunodeficiency and disease progression is far slower \[[@CIT0001], [@CIT0002]\]. Indeed, based on the definition of long-term nonprogressors used in HIV-1 infection, the proportion of people living with HIV-2 (PLWH2) meeting this profile was reported to be up to 6%, as compared to \<1% among people living with HIV-1 (PLWH1) \[[@CIT0003]\]. Moreover, HIV-2 is characterized by a lower replication capacity than HIV-1, that is, approximately 30 times less as shown by modeling \[[@CIT0004]\], yielding 9% of patients meeting definition criteria for elite controllers \[[@CIT0003]\].
In terms of treatment, HIV-2 is naturally resistant to 2 antiretroviral drug classes commonly used to treat HIV-1 infection, namely, nonnucleoside reverse transcriptase inhibitors and fusion inhibitors \[[@CIT0008]\]. Furthermore, phenotypic studies suggest that HIV-2 susceptibility to protease inhibitors (PIs) varies according to the specific drug tested \[[@CIT0009]\]. Thus, the limited armamentarium of drugs available to manage HIV-2 infection calls for evaluation of every innovation when it becomes available in HIV-1 infection.
Geographically, the HIV-2 epidemic is restricted mainly to West Africa, Angola, Mozambique, and Europe (mostly Portugal or France, due to migratory flows from affected African regions) as HIV-2 is less efficiently transmitted than HIV-1, essentially through sexual or mother-to-child transmission \[[@CIT0010]\]. In this context, high level of evidence for the treatment of HIV-2 remains scarce and powerful designs such as randomized clinical trials are difficult to implement.
In addition, while disease profile includes traditional determinants of progression such as clinical stage, CD4 lymphocyte count, and plasma viral load (pVL), the latter is much lower on average than in HIV-1 infection and below the detection threshold in very large proportions of untreated PLWH2 \[[@CIT0010], [@CIT0013]\]. Therefore, evaluation of treatment innovations cannot predominantly rely on the measurement of pVL for eligibility and outcome, and CD4 cell response should also be considered within a composite endpoint \[[@CIT0016]\].
Given the lack of randomized trials, observational studies currently provide the best available evidence for treatment guidelines, including a combination of 2 nucleoside reverse transcriptase inhibitors (NRTIs) and 1 PI as recommended first-line combination antiretroviral therapy (cART) since 2012 \[[@CIT0017]\].
Current first-line cART yields lower response in terms of CD4 cell count in PLWH2 than in PLWH1, even with regimens including recently approved PIs \[[@CIT0014], [@CIT0019], [@CIT0022]\].
More recently, 2 series of observations raised a specific interest for the evaluation of raltegravir in antiretroviral-naive PLWH2. First, phenotypic virological studies showed that the optimal inhibitory concentration 50% (IC~50~) for HIV-2 were those of lopinavir, darunavir, and raltegravir \[[@CIT0009], [@CIT0027]\]. Second, raltegravir was associated with dramatic immune recovery in some heavily pretreated patients experiencing virological failure \[[@CIT0028]\].
We postulated that a combination of raltegravir plus 2 NRTIs could yield a better CD4 response than the average response reported with PI-containing regimens, along with a good safety profile.
METHODS {#s6}
=======
Trial Design and Participants {#s7}
-----------------------------
The French National Agency for Research on AIDS and Viral Hepatitis (ANRS) 159 HIV-2 study was a noncomparative multicenter, nationwide, 1-step, phase 2 clinical trial, conducted in 18 hospital centers in France, from July 2012 to December 2015.
We used the setting of the French ANRS CO5 HIV-2 cohort, involving \>1000 adults infected with HIV-2 only in 120 clinical sites, to identify potentially eligible patients, and investigators invited them to participate in the study. Eligibility criteria for the trial were the following: antiretroviral-naive PLWH2 with a history of Centers for Disease Control and Prevention group B or C--defining event, CD4 count \<500 cells/μL, or CD4 decrease \>50 cells/μL/year over the last 3 years, with a last value of ±10% of CD4 nadir or confirmed HIV-2 RNA pVL ≥100 copies/mL. Pneumocystis prophylaxis, combined with toxoplasmosis prophylaxis in the presence of specific antibodies, was an additional eligibility criteria for participants with CD4 count \<200 cells/μL. By April 2013, the scientific committee of the trial recommended that women who had started a antiretroviral therapy (ART) other than an integrase inhibitor (INSTI)--containing combination during pregnancy for prevention of mother-to-child transmission and stopped it at the time of delivery could be further included if they met other eligibility criteria.
Noneligibility criteria were absence of effective contraception method in women of childbearing age; pregnancy or planning a pregancy; breastfeeding; progressive opportunistic infection with curative treatment not compatible with that of the trial; malignancy requiring chemotherapy or radiotherapy; chronic hepatitis C with Metavir score F2 or greater, decompensated cirrhosis; uncontrolled insulin-dependent diabetes mellitus; corticosteroid treatment \>3 weeks; hemoglobinemia \<7 g/dL, polynuclear neutrophil count \<500/µL, platelet count \<50 000/µL, creatinine clearance \<50 mL/minute, aminotransferases or alkaline phosphatases or bilirubin \>2.5 times the upper limit of normal laboratory range; contraindication to 1 of the excipients of study treatments; judicial protection; or participation in any other medication trial.
The protocol was approved according to French regulation by an Ethics Committee (Comité de Protection des Personnes de Paris Ile de France XI). The trial was performed in accordance with good clinical practices and the ethical principles of the Declaration of Helsinki. All participants provided signed informed consent. Data were regularly reviewed by an independent data monitoring committee.
Study Treatment, Design, and Procedures {#s8}
---------------------------------------
Study treatment consisted of emtricitabine 200 mg/day, tenofovir disoproxil fumarate 300 mg/day, and raltegravir 400 mg twice daily from week (W) 0 to W48.
The primary endpoint was a composite criterion, defined as the proportion of participants in therapeutic success, that is, surviving at W48 without any of the following events from trial start: CD4 gain \<100 cells/μL at W48 compared to the baseline CD4 count (mean W --4, W0), pVL ≥40 copies/mL from W24 confirmed within the next 4 weeks, raltegravir permanent discontinuation, or new B or C event validated by a dedicated independent data monitoring committee.
By April 2013, as a pVL quantification assay with a detection threshold of 40 copies/mL became available, the scientific committee of the trial recommended that the detection threshold of pVL used for the virological endpoint of the composite outcome be lowered from 100 copies/mL to 40 copies/mL.
The component related to discontinuation of raltegravir was included to reflect the potential of durability of this first-line regimen, as the number of potential third antiretroviral agents to be associated to NRTIs is very limited in HIV-2 infection.
The main secondary endpoint was the mean change in CD4 lymphocyte count between baseline (mean of W --4 and W0 count) and W12. Other endpoints included evolution of the rate of participants with undetectable ultrasensitive pVL (uspVL), with undetectable total HIV-2 DNA from baseline to W48, respectively; change in CD4 percentage, pVL, and total DNA values from baseline to W48; description of the resistance mutations selected (number and type of mutations in the reverse transcriptase \[RT\] and the integrase genes) compared to W0, in case of virological failure; rate of clinical progression (from stage A to B, C, or death and from stage B to C or death); adherence; rate of treatment switch or discontinuation (overall and for each study drug); safety (number, nature, severity and time to adverse event \[AE\]).
Clinical examination and laboratory sampling, from which CD4 cell counts and pVL were performed, were performed at screening (W --4), baseline (W0), W4, W8, W12, W18, W24, W36, and W48.
Ultrasensitive pVL quantification was performed at W0, W24, and W48 using real-time reverse-transcription PCR assay with a threshold between 5 and 10 copies/mL depending on the volume of plasma available (Biocentric, Bandol, France) \[[@CIT0033]\].Total HIV-2 DNA quantification was performed at the same time points using a real-time PCR assay with a threshold of 6 copies developed by the ANRS AC-11 HIV Quantification group \[[@CIT0034]\].
Resistance-associated mutations to NRTIs and INSTIs were assessed at screening and at time of virological failure, defined as the first pVL value ≥40 copies/mL (confirmed within the following 4 weeks) after a value \<40 copies/mL, by sequencing the RT and integrase genes from plasma specimens using the consensus technique of the ANRS AC11 Resistance Group ([www.hivfrenchresistance.org](http://www.hivfrenchresistance.org)). HIV-2 resistance mutations were identified using the list generated by the Collaborative HIV and Anti-HIV Drug Resistance Network \[[@CIT0033], [@CIT0035]\].
Adherence to treatment was assessed by the ANRS self-administered questionnaire of adherence at W4, W12, W24, W36, and W48, by trough plasma concentrations of antiretroviral drug measurement in case of virological failure, and by checking the number of treatment trial pills at each dispensing visit.
Clinical and laboratory AEs were graded using the ANRS Scale for Grading Adult Adverse Events (grade 1, mild; grade 2, moderate; grade 3, severe; grade 4, life-threatening).
Statistical Methods {#s9}
-------------------
The primary efficacy analysis was an intention-to-treat analysis, which included all participants who received at least 1 dose of emtricitabine/tenofovir and raltegravir, regardless of treatment changes during the trial. Missing data were considered as failure. We also performed the analysis on available data. Data were analyzed with SAS software (version 9.2 and higher).
RESULTS {#s10}
=======
Among the 32 participants included in the study, 30 received at least 1 dose of the study drugs and were included in the intention-to-treat analysis. Twenty-eight participants (93%) completed the W48 follow-up: 1 participant withdrew consent (while reporting a permanent discontinuation of raltegravir) and 1 participant was lost to follow-up ([Figure 1](#F1){ref-type="fig"}).
{#F1}
Baseline Characteristics {#s11}
------------------------
Two-thirds of participants were women and most of them originated from West Africa. They had a long history with HIV-2 infection while remaining at the asymptomatic stage with rather high baseline CD4 nadir. Baseline median CD4 count was \>400 cells/µL and reached at least 500 cells/µL in one-quarter of the participants. Median pVL was 2.5 log~10~ copies/mL in 20 participants with pVL ≥40 copies/mL ([Table 1](#T1){ref-type="table"}).
######
**Baseline Characteristics of the Participants in the ANRS (France REcherche Nord&Sud Sida-Hiv Hépatites) 159 HIV-2 Trial (N = 30**)
Characteristic No. (%)
------------------------------------------------------------ --------------------
Female sex 20 (66.7)
Age, y, median (IQR) 49 (46.0--52.6)
\<35 1 (3.3)
35--44 5 (16.7)
45--54 17 (56.7)
55--64 4 (13.3)
≥65 3 (10.0)
Place of birth
West Africa 26 (86.7)
France 4 (13.3)
HIV-2 transmission mode
Heterosexual contact 23 (76.7)
Transfusion 4 (13.3)
Unknown 3 (10.0)
Time since HIV-2 diagnosis, y, median (IQR) 11 (7.5--13.9)
CDC stage (week --4)
A 26 (86.7)
B 2 (6.7)
C 2 (6.6)
Nadir CD4 count, cells/µL, median (IQR) 351 (286.0--455.0)
CD4 count, cells/µL (week --4, week 0), median (IQR) 436 (314--507)
\<200 3 (10.0)
200--349 6 (20.0)
350--499 13 (43.3)
≥500 8 (26.7)
CD4/CD8 ratio (week 0), median (IQR) 0.9 (0.6--1.2)
Plasma HIV-2 RNA ≥40 copies/mL (week 0) 20 (66.7)
Plasma HIV-2 RNA ≥5 copies/mL (week 0) 23 (92.0)
Plasma HIV-2 RNA (week 0), log~10~ copies/mL, median (IQR) 2.5 (1.8--3.2)
Total HIV-2 DNA (week 0) \>6 copies (by PCR) 8 (32.0)
Data are presented as No. (%) unless otherwise indicated.
Abbreviations: CDC, Centers for Disease Control and Prevention; HIV-2, human immunodeficiency virus type 2; IQR, interquartile range; PCR, polymerase chain reaction.
Outcomes {#s12}
--------
The proportion of participants declaring moderate to good adherence between W4 and W48 ranged from 76% to 83%.
At W48, the composite endpoint of success was reached in 12 of 30 participants (40% \[95% confidence interval {CI}, 22.7%--59.4%\]). Failure was mainly due to not reaching the immunological criterion of the endpoint, that is, CD4 gain \<100 cells/µL from baseline (n = 15), then to virological failure, that is, pVL ≥40 copies/mL (n = 1) or to withdrawal before W48 (n = 2).
In terms of immunological response, the median CD4 count change in the 15 patients with a gain \<100 cells/µL was +38 cells/µL; a decrease in CD4 cell count was observed in 5 of them (overall in 16.7% of participants).
Among the 22 participants with baseline CD4 count \<500 cells/µL and the 8 with CD4 count ≥500 cells/µL, 36% (95% CI, 17.2%--59.3%\]) and 50% (95% CI, 15.7%--84.3%\]) experienced treatment success, respectively. Overall, median CD4 count change at W12 and W48 was +73 cells/µL (IQR, +2 to +143 cells/µL; n = 30) and +87 cells/µL (IQR, +38 to +213 cells/µL; n = 28), respectively ([Figure 2](#F2){ref-type="fig"}).
![Median CD4 cell count from week (W) 0 to week 48, ANRS (France REcherche Nord&Sud Sida-Hiv Hépatites) 159 HIV-2 trial. Median CD4 change at W12 and W48: +73/µL (interquartile range \[IQR\], +2 to +143) (n = 30) and +87/µL (IQR, +38 to +213) (n = 28), respectively. At W48, median CD4 change: +115/µL and +70/µL in those with baseline plasma viral load ≥40 and \<40 copies/mL, respectively. Mean CD4 count between W --4 and W0 was 437 cells/µL and is represented by the dotted horizontal line.](ciy24502){#F2}
At W48, median CD4 count change was +115 cells/µL and +70 cells/µL in participants with baseline pVL ≥40 and \<40 copies/mL, respectively. Median baseline CD4 count was 465 cells/µL (IQR, 406--658 cells/µL) in participants gaining at least 100 CD4 cells/µL at W48 vs 414 (IQR, 292--480 cells/µL) in those with immunologic failure. Furthermore, among the 28 participants followed up to W48, the CD4 cell count increased to ≥500 cells/µL in 8 of the 21 with baseline count \<500 cells/µL at baseline, and to ≥350 cells/µL in 4 of the 9 with baseline CD4 count \<350 cells/µL. Median CD4 percentage was 33% (IQR, 24.5%--37%) at baseline and 36% (IQR, 29%--42%) at W48, resulting in a median change of +3% (IQR, +4.5% to +5%).
In terms of virological response, at W48, 27 of 28 participants who completed the 48-week follow-up had pVL \<40 copies/mL ([Figure 3](#F3){ref-type="fig"}). Virological failure was reported in 1 patient, with a pVL of 3.3 log~10~ copies/mL at W18 who reported good adherence in dedicated self-questionnaire. Antiretroviral plasma concentrations measured on W18 and W20 samples were in the expected range, but the number of pills brought back to hospital was smaller than expected, suggesting incorrect adherence between W8 and W12. At time of virological failure, drug resistance mutations were evidenced in the integrase region (E92Q, T97A, and Y143C/G/H/R), but no resistance mutation was detected in the RT region.
{#F3}
In the analysis considering only participants with available data at W48 (n = 29), treatment success was reported in 12 of 29 (41.4% \[95% CI, 23.5%--61.1%\]).
At W48, among those participants with available data for these specific secondary endpoints, uspVL was \<5 copies/mL in 13 of 15 participants (87%) and HIV-2 total DNA was \>6 copies by PCR in 3 of 26 (12%). In the sole case of virological failure, DNA was unchanged at the time of failure (W24) as compared to W0.
Fifty-six AEs were reported in 20 participants and were mainly mild to moderate; 50% of these AE were reported in 3 participants. Among the 5 grade 3 clinical AEs (4 participants; hysterectomy \[1\]; treatment trial overdosing without toxicity \[1\]; bronchopulmonary carcinoma \[1\]; intestinal functional disorder \[1\]; gastrointestinal disorder \[1\]), only 1 AE was related to study treatment: gastrointestinal disorder at day 1, leading to a switch from emtricitabine/tenofovir to lamivudine/abacavir, followed by complete resolution. A third of clinical AEs (n = 17 \[30.4%\]) were gastrointestinal disorders: nausea (5), abdominal pain (2), others (10) (1 each). None of the participants experienced grade 4 AEs. Four severe AEs occurred in 3 participants: lacunar ischemic stroke (n = 1), *Helicobacter pylori* gastritis/*Escherichia coli* cystitis/functional colopathy (n = 1), treatment trial overdosing without toxicity (n = 1), and suspected recurrence of bronchopulmonary cancer (n = 1), but none was related to antiretrovirals. There was no death nor new B or C event reported over the trial duration. One pregnancy was reported in a participant who never received the study treatments.
DISCUSSION {#s13}
==========
In this noncomparative trial, a first line cART-containing raltegravir yielded a complete success rate of 40% at 1 year, comparable to PI-containing regimens, and was well tolerated. Ultrasensitive pVL and total DNA were undetectable in most participants.
Improving evidenced-based treatment of HIV-2 infection is very challenging in terms of trial designs. First, the best design to inform evidence-based guidelines is controlled randomized trials. Nevertheless, due to the limited number of people living with HIV-2 in western and northern regions of the globe, such trials can only be conducted in West Africa \[[@CIT0016], [@CIT0036]\], and it is therefore useful to also consider observational studies or pilot trials to inform guidelines for clinical decision or further research. Our pilot trial was indeed the first step of designing an ongoing randomized trial in West Africa within the ANRS network (ANRS 12294, ClinicalTrials.gov identifier NCT02150993).
Second, plasma HIV-2 RNA level above the routine detection limit cannot reasonably be used as the sole or main inclusion criterion as it pVL is quantifiable in fewer PLWH2 than in PLWH1; for example, in the ANRS nationwide HIV-2 cohort, no more than 19% of patients present with pVL \>100 copies/mL at enrollment (Sophie Matheron, personal communication). Therefore, other markers of disease progression should be considered, such as clinical staging, or CD4 count. Immunological progression remains slower than that of HIV-1, thus relevant decline of CD4 slope in the context of HIV-2 seems more appropriate \[[@CIT0017]\]. Third, a composite endpoint is useful to optimize the power of trials. Therefore, our composite primary endpoint included clinical progression and CD4 gain in addition to virological outcome. Retention in the first-line cART regimen also seemed highly relevant as the number of potent antiretroviral drugs is far smaller than for HIV-1; moreover, a high level of NRTI cross-resistance impacts this drug class more rapidly \[[@CIT0037]\].
Using a similar composite endpoint, PI-containing cART in antiretroviral-naive patients yielded a global success rate of 45% (95% CI, 31%--60%) at 12 months in an European retrospective study, the immunological component of the composite endpoint being the main cause of failure, with a median increase of 76--88 cells/µL/year after at least 3 months of treatment \[[@CIT0024]\].
Based (1) on available data on raltegravir-containing cART in PLWH1, with a mean CD4 cell increase of 144 cells/µL at W48 \[[@CIT0021]\], (2) on median values of raltegravir IC~50~ and IC~90~ reported in 14 PLWH2 (2.4 nM and 12.5 nM, respectively) \[[@CIT0027]\], and (3) on case reports on immunovirological outcomes of this cART regimen in 3 PLWH2 with an history of multiple treatment failure \[[@CIT0027], [@CIT0029], [@CIT0030], [@CIT0032]\], we prioritized this antiretroviral combination to be evaluated in a pilot trial before consideration as experimental arm in a randomized trial.
The median CD4 cell change in response to treatment (+87 cells/µL at 1 year), even if lower than in HIV-1, was close to that recently reported with another first-line INSTI-containing cART in Senegal \[[@CIT0038]\]. In addition, we showed a trend for higher success rate among PLWH2 starting with baseline CD4 count ≥500 cells/µL than with CD4 \<200 cells/µL, pleading for an early start of cART. Our results also demonstrate the virological strength of a raltegravir-containing regimen to reach high levels of undetectable pVL and HIV DNA. These results, along with absence of any serious treatment-related side effects, are current arguments for recommending INSTI-containing regimen as first-line cART, at the same level than PI-containing ones. So far, the randomized trial comparing these regimens is ongoing in West Africa and will provide more information on the preferred one.
***Acknowledgments.*** We thank A. Diallo, A. Dumont, V. Petrov-Sanchez (INSERM-ANRS); B. Autran (Scientific Committee); D. Costagliola, H. Fischer, A.-G. Marcelin, F. Raffi (Independent Committee); M.-L. Chaix, P.-M. Girard, O. Lortholary, A.-M. Taburet (Independent Data Monitoring Committee); N. Chaghil, L. Pinoges, E. Rouch (Clinical Trials Unit). We also thank all of the patients who agreed to participate in this trial, and all the investigators involved in the ANRS CO5 VIH2 cohort.
***Financial support.*** This work was supported by INSERM-ANRS, sponsor of the trial; Merck Sharp Dohme-Chibret; and Gilead Sciences.
***Potential conflicts of interest.*** All authors: No potential conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.
***ANRS 159 HIV-2 Trial Study Group.*** Hôpital Bichat, Paris: S. Matheron; Hôpital Antoine Béclère, Clamart: F. Boue; Hôpital Bicêtre, Paris: C. Goujard; Hôpital Européen Georges Pompidou, Paris: L. Weiss; Hôpital Lariboisière, Paris: A. Rami; Hôpital Louis Mourier, Colombes: E. Mortier; Hôpital Pitié-Salpêtrière, Paris: R. Tubiana; Hôpital Saint Antoine: P. Campa; Hôpital Saint Louis, Paris: D. Ponscarme; Hôpital Bocage, Dijon: L. Piroth; Hôpital la Croix Rousse, Lyon: P. Miailhes; Hôpital Gui de Chauliac, Montpellier: J. Reynes; Centre Hospitalier René Dubos, Pontoise: L. Blum; Hôpital Delafontaine, Saint Denis: M.-A. Khuong; CHI Villeneuve Saint Georges: O. Patey; CHI Créteil: B. Elharrar; C. H. Mulhouse: G. Beck-Wirth; CHU Angers: P. Fialaire. INSERM-ANRS: I. Amri; F. Cardon; L. Marchand.
[^1]: Members of the ANRS 159 HIV-2 Trial Study Group are listed in the Notes.
| {
"pile_set_name": "PubMed Central"
} |
*Vibrio vulnificus* is a highly invasive human pathogen and presents a food safety issue worldwide. Human infections caused by *V. vulnificus* are primarily caused by contaminated seafood consumption or contaminated skin wounds, which can lead to septicemia, wound infections, and high hospitalization and fatality rates ([@R1],[@R2]). *V. vulnificus* strains are biochemically divided into 3 biotypes (BTs). BT3, found in Israel, is associated with infections caused by contaminated fish ([@R3]). Until now, only 3 BT3 isolates had been isolated from the environment in direct contrast with their large clinical numbers ([@R3],[@R4]).
BT3 is a clonal group, which various molecular methods have failed to differentiate among its strains, ([@R4],[@R5]) with the exceptions of rep-PCR ([@R6]), simple-sequence repeats (SSR) ([@R7]), and recently pulsed-field gel electrophoresis (PFGE) ([@R8]). SSR analysis of *V. vulnificus* was highly discriminative among BT3 strains ([@R7]). These SSR markers have been used for typing and for epidemiologic studies in many bacterial species ([@R9],[@R10]). We present results from a 3-year monitoring program of clinical and environmental *V. vulnificus* using SSR as an epidemiologic genotyping tool.
The Study
=========
A total of 414 *V. vulnificus* isolates were studied, including a reference panel of 32 strains previously studied ([@R7]). A total of 360 environmental *V. vulnificus* isolates were successfully retrieved from September 2004 through October 2006 from artificial fish ponds and stores in the western Galilee region of Israel (from 21 samplings), and 22 clinical isolates were retrieved from nearby hospitals during matching years ([Tables 1](#T1){ref-type="table"}, [2](#T2){ref-type="table"}). Fish samples were collected and gills and fins/scales were pooled from ≈10 *Tilapia* spp. (300--400 g). Each sample was incubated in 0.5 L modified alkaline peptone water with 4% NaCl, pH 6.9, at 37^o^C for 16 to 18 h. Samples were diluted in saline and streaked on thiosulfate-citrate-bile-salts-sucrose (TCBS) agar. Suspected colonies were further grown on chromogenic agar (CHROMagar Microbiology, Paris, France), and validated by amplification of *V. vulnificus*--specific gene vvh ([@R7]). All *V. vulnificus* colonies were green on TCBS agar. Notably, not all bacterial isolates showed the expected turquoise-colonies on CHROMagar but rather pale white colonies. The latter colonies were further identified as BT3. All other isolates, which were BT1, showed the expected turquoise-colony phenotype. No BT2 isolates were found. However, 6 previously studied isolates ([@R7]) showed the expected turquoise-colony.
###### *Vibrio vulnificus* isolates from Israel that were genetically analyzed, 2003--2006\*
Isolate identification Biotype Date Origin/hospital name IMH no.†
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------- --------- ---------- ---------------------------------------------------------------- ----------
Environmental
VVyb1 3 2004 Store: yb1-yb53
Pond: yb54-yb58
VVyb2--VVyb58, (VVyb38, VVyb45 missing) 1
VVyb63, VVyb66, VVyb67 VVyb71--VVyb73, VVyb83, VVyb86--Vvyb93,
VVyb95--Vvyb109, VVyb111--VVyb126 3 2005 Store: yb87-yb126, yb158-yb193
Pond: yb62-yb86
VVyb59--VVyb62, VVyb64,VVyb65, VVyb68--VVyb70, VVyb74--VVyb82,
VVyb84, VVyb85, VVyb159, VVyb162--VVyb164, VVyb167--VVyb170,
VVyb172--VVyb182, VVyb187, VVyb189--VVyb193 1
VVyb94, VVyb110, VVyb158, VVyb160, VVyb161, VVyb165, VVyb166,
VVyb171, VVyb183--VVyb186, VVyb188 ND
VVyb127--Vyb133, VVyb137--VVyb157 3 2006 Store: yb127-yb134, yb194-yb206
Pond: yb135-yb157, yb207-yb221
VVyb134--VVyb136, VVyb195--VVyb208, VVyb210--VVyb216,
VVyb218--VVyb221, (VVyb201, VVyb212 missing) 1
VVyb209, VVyb217 ND
v232‡ 3 2003 Dec Fish 8/03e
Clinical§
v233 3 2006 May Rivka Ziv 1/06
v234 3 2006 Sep HaEmek 2/06
v235 ND 2006 Oct Western Galilee 3/06
v236 3 2005 Feb HaEmek 1/05
v237 3 2005 Jun Western Galilee 2/05
v238 3 2005 Jun Western Galilee 5/05
v239 3 2005 Aug Rambam 6/05
v240 3 2005 Oct Western Galilee 7/05
v241 3 2005 Nov Rambam 8/05
v242 3 2005 Nov Rambam 9/05
v243 3 2005 Dec Carmel 10/05
v244 3 2005 Nov Western Galilee 11/05
v245 3 2005 Jun Carmel 3/05
v246 ND† 2005 Jun Rambam 4/05
v247 3 2004 Jun HaEmek 2/04
v248 3 2004 Jun HaEmek 3/04
v249 3 2004 Jun HaEmek 4/04
v250 3 2004 Jul Rambam 5/04
v251 3 2004 Aug Western Galilee 6/04
v252 1 2004 Aug Western Galilee 7/04
v253 3 2004 Oct Carmel 9/04
v254 3 2003 Dec Barzilai 8/03
\*IMH, Israeli Ministry of Health; ND, not determined.
†All clinical isolates are part of the IMH collection.
‡Previously studied ([@R7]).
[§]{.ul}All clinical isolates are associated with fish.
###### Environmental *Vibrio vulnificus* isolates obtained in Israel from artificial fish ponds and fish stores, 2004--2006
Dates No. samples No. isolates No. *V. vulnificus*\* Biotype 3, % Biotype 1, %†
--------------- ------------- -------------- ----------------------- -------------- ---------------
2004 Sep--Oct 7 58 58 2 98
2005 May--Oct 5 280 166 28 72
2006 Mar--Oct 9 251 136 21 79
Total 21 589 360
\*Tested by specific amplification of *vvh* gene.
†Including isolates not determined.
SSRs were used to genetically characterize 254 clinical and environmental *V. vulnificus* isolates ([Table 1](#T1){ref-type="table"}), including 32 previously studied isolates ([@R7]). DNA extraction, PCR and primers, SSR sizing, and statistical analysis were conducted as previously described ([@R7]). Capillary electrophoresis was performed by using a 3130 Genetic Analyzer and analyzed with GeneMapper-v4.0 (Applied-Biosystems Inc., Foster City, CA, USA). Two to 34 alleles were detected at the 12 SSR loci among the isolates. Environmental isolates were selected from 21 samples with an average 8.5 isolates per enrichment. We removed 71 isolates that had identical SSR genotypes and originated from the same enrichment from the analysis because they were probably clones. Thus, 183 isolates were discriminated to 170 SSR types. SSR variation data was used to calculate genetic relationships among isolates. A genetic distance matrix was generated followed by cluster analysis ([@R7]). The resulting dendrogram ([Figure 1](#F1){ref-type="fig"}) showed clear separation between BT3 isolates and the others (average genetic distance of 0.825 ± 0.101). Genetic distances among BT3 isolates were rather low (average 0.369 ± 0.174) relative to high genetic distances (average 0.804 ± 0.149) found among isolates of the other biotypes, in accordance with our previous analysis of 32 isolates ([@R7]). The new studied isolates showed a variety of SSR genotypes and were spread throughout the dendrogram ([Figure 1](#F1){ref-type="fig"}). Further analysis using eBURST ([@R12]) showed similar grouping results (data not shown) ([@R7]).
![A) Genetic relationships based on simple-sequence repeat (SSR) variation data among 183 *Vibrio vulnificus* isolates including 135 new environmental, 22 new clinical, and 26 previously studied isolates. B) A subtree enlargement of panel A displaying a set of 65 *V. vulnificus* biotype 3 isolates. Similar clinical and environmental isolates, showing an epidemiologic connection, are indicated by arrows. The genetic-distance matrix was generated based on 212 polymorphic points (the sum of alleles across 12 SSR loci). Genetic relationships are based on unweighted pair group method with arithmetic mean cluster analysis of SSR variation using MEGA4 software ([@R11]). Scale bar represents genetic distance.](08-0839-F1){#F1}
Differentiation of SSR alleles results at locus VV0401 into environmental types (E-types, [\>]{.ul}12 repeats), and clinical types (C-types, \<10 repeats) was tested ([@R13]). Of the clinical isolates, 44 isolates (98%) exhibited the expected C-type allele/repeat but only 69 isolates (33%) of the environmental isolates showed the E-type allele/repeat, rejecting the null hypothesis (p\<0.0001), using Pearson χ^2^ test. Notably, 97% of 110 BT3 isolates (clinical/environmental) showed the C-type allele/repeat, in contrast to BT1 isolates (72 isolates, 2%). If, C-type allele/repeat at VV0401 is an indication of potential pathogenicity of *V. vulnificus* strains, then our results further support the high virulence of BT3. However, additional studies are needed to confirm the relationship of this locus to pathogenicity.
Three clinical BT3 isolates exhibited identical SSR genotypes and 2 clinical BT3 isolates had a genotype related to 5 environmental isolates sampled on related dates from nearby regional areas ([Figure 1](#F1){ref-type="fig"}, panel B). One clinical BT3 isolate v239 (August 2005) showed SSR genotypes identical to the environmental isolate VVyb95 obtained in the same month (August 2005) and to VVyb194 (obtained in March 2006). A second BT3 clinical isolate v240 (October 2005) and the fish-pond isolate VVyb132 (March 2006) had identical SSR genotypes. Moreover, these 2 results suggest survival of *V. vulnificus* strains through the winter season, either in a viable nonculturable state ([@R14]) or as viable cells in sediment that can serve as a shelter for some subpopulations ([@R15]). The latter scenario is more probable in artificial fish ponds because water circulation is high throughout the growth period and pond sediment remains untouched. A third clinical isolate was analyzed earlier and showed an epidemiologic connection: v254 isolate was obtained from an injured woman (injured by a fish) in December 2003. Analysis of microbial flora on the fish (found in the woman's freezer) identified a *V. vulnificus* BT3 isolate, v232. These 2 isolates showed identical SSR genotypes, confirming the fish as the origin of infection ([Figure 1](#F1){ref-type="fig"}, panel B). Additionally, clinical BT1 isolate v252 (August 2004), showed similar SSR genotypes, 1 repeat difference in 1 locus, to environmental isolate VVyb50 (October 2004).
To strengthen our typing results, we compared the epidemiologic SSR results to those of PFGE in 12 representative BT3 isolates. PFGE was performed and analyzed as described previously ([@R8]). Results for PFGE were generally similar to results for SSR ([Figure 2](#F2){ref-type="fig"}). PFGE patterns were similar ([\>]{.ul}85%) between isolates. Identical PFGE patterns and SSR genotypes were seen in isolates v239 (clinical) and VVyb95, VVyb194 (environmental), as were isolates v232 (fish) and v254 (clinical). Identical PFGE patterns and single-locus variants in SSR genotypes were seen in clinical isolate v237 and the environmental isolates VVyb89 and VVyb1. Notably, VVyb1, which was isolated a year earlier, differentiated from VVyb89 by another single repeat in an additional single locus, confirming higher resolution of SSR method. Finally, identical SSR genotypes and PFGE patterns that differed by 1 band were seen in v240 and VVyb132.
{#F2}
Conclusions
===========
The developed isolation and enrichment procedures obtained large numbers of BT3 and BT1 from the environment. Results showed that although BT3 makes up only ≈21% of the *V. vulnificus* isolates from fish, BT3 accounts for ≈86% of the clinical cases and thus could imply high pathogenicity for this group ([@R4]). Genetic analysis of this large survey confirms the distinctness (clonality) ([@R5]) of BT3 and the high resolution power of the SSR ([@R7]).
SSR genotyping of *V. vulnificus* was used to determine the genetic relatedness between clinical and environmental isolates and identify the source of contamination. SSR can serve as an epidemiologic tool to indicate the infection source of pathogens such as *V. vulnificus*, and can potentially provide knowledge for preventive steps in terms of public health.
*Suggested citation for this article*: Broza YY, Danin-Poleg Y, Lerner L, Valinsky L, Broza M, Kashi Y. Epidemiologic study of *Vibrio vulnificus* infections by using variable number tandem repeats. Emerg Infect Dis \[serial on the Internet\] 2009 Aug \[*date cited*\]. Available from <http://www.cdc.gov/EID/content/15/8/1282.htm>
We thank Zinaida Korenman for technical assistance.
This research was supported by The Environment and Health Fund, RGA0804; the Institute for Future Security Study, Technion; the Grand Water Research Institute, Technion; the Israeli Water Commission; and NATO, Project CBD.MD.SFP 981456.
Dr Broza was a PhD student in the laboratory of Applied Genomics at the Faculty of Biotechnology and Food Engineering, Technion-Israel Institute of Technology, at the time of this study. This work was part of his dissertation on the human pathogen *V. vulnificus.* His current research is studying nanomaterial-based sensors for detection of disease biomarkers through breath samples.
| {
"pile_set_name": "PubMed Central"
} |
**TO THE EDITOR:** Myelodysplastic syndromes (MDS) are a group of heterogeneous hematological malignancies which demand personalized and risk-adapted clinical management \[[@B1]\]. Current therapeutic approaches are rather limited for patients unsuitable for allogeneic stem cell transplantation (SCT), the only realistic and potentially curative treatment measure that exists \[[@B1]\]. With regard to patients with high risk MDS, the standard of care is currently represented by treatment with hypomethylating agents (HMAs), such as decitabine and azacitidine. The latter is used as initial therapy in most cases, and induces responses in 40--50% of treated patients \[[@B2][@B3]\]. Obstacles to azacitidine administration as well as recommendations for the optimization of treatment with this agent have been reported \[[@B2][@B4]\]. However, despite optimal management of azacitidine treatment, the duration of its clinical benefit, although variable, is usually transient and almost all patients ultimately experience loss of response to the drug, disease progression, and therefore very poor outcomes \[[@B1][@B2][@B5][@B6]\]. After this loss of response or disease progression despite treatment, there are no standard care regimens available \[[@B5]\]. Rescue strategies including intensive chemotherapy (ICT) only provide minor benefits, whereas allogeneic SCT is feasible only in a minority of cases. With these results in mind, especially the catastrophic outcome of azacitidine-failed patients, typical concerns about decision making and clinical management in these settings can be summarized by an unusual case we observed which is reported herein. A 59-year-old woman was admitted for profound malaise due to pancytopenia on March 2015. The bone marrow (BM) and trephine biopsy revealed refractory anemia with an excess of blasts-2 (RAEB-2), remarkable multilineage dysplasia, and 18% of BM infiltrating blasts; the karyotype analysis and molecular study for typical abnormalities found in MDS were negative. She was diagnosed as having an Inter-2 MDS, according to the International Prognostic Scoring System \[[@B7]\]. On the basis of the patient\'s overall fitness level, and given the lack of a suitable familiar donor to proceed to immediate allogeneic SCT, we recommended therapy with azacitidine (75 mg/m^2^, schedule 5+2+2). Therapy was started on April 2015 without significant adverse effects. Meanwhile, a matched unrelated donor (MUD) was fruitlessly sought. After six cycles (September 2015), a partial remission (according to Cheson\'s criteria) was achieved \[[@B8]\]. Because of this, the same treatment was continued for another three cycles until December 2015, when a progressive pancytopenia unveiled progression to secondary acute myelogenous leukemia (AML). At the time of evolution, standard cytogenetic tests, FISH analyses, and mutational studies which are usually performed in the AML diagnostic work-up (such as BCR/ABL P190, BCR/ABL P201, RUNX1/RUNXT1, CFBbeta/MYH11, DEK/CAN,FLT3-ITD and NPM1) were found to be negative; therefore, as the patient was considered eligible for an anthracycline-based induction ICT, she received one course of standard "3+7" consisting of daunorubicin 45 mg/m2 daily (days 1--3) and cytarabine 100 mg/m^2^ daily (continuous IV infusion days 1--7). Unfortunately, the patient was resistant to this induction ICT; her BM, which was revaluated 14 and 28 days after the induction treatment, remained severely dysplastic with ---20% of leukemic infiltration (December 2015). In addition, the course of therapeutic aplasia was complicated by a severe pulmonary aspergillosis, which was successfully treated with voriconazole. The patient complained of painful dysesthesia of the lower limbs, and a magnetic resonance imaging (MRI) scan of the spine revealed a massive osteolytic lesion at the D11 vertebral body without neural compression. A percutaneous biopsy of D11 revealed the AML localization of the involved vertebral body, and a vertebloplasty was performed (April 2016). At that time, the patient was properly informed of the seriousness of her clinical situation, as well as the absence of effective standard therapeutic options, and that some available measures were only for palliative purposes. Despite this, she asked us to continue the anti-leukemic therapy, while evaluating any form of potentially applicable causal options. After the approval from the Institutional Board of our hospital, the patient consented to therapy with decitabine at the daily dosage of 20 mg/m^2^ for five days every four weeks (July 2016); she received four courses without any side effects \[[@B9][@B10]\]. Prior the start of decitabine treatment, a BM exam was performed revealing 20% of BM infiltrating blasts, whereas karyotype and molecular findings were normal. Meanwhile, the patient received stereotactic radiation therapy on the D11 vertebral body up to a total dose of 24 Gy, given in 3 fractions (8 Gy per day, September 2016) without any adverse reaction. Given the progressive improvement of blood counts and the significant reduction in transfusion requirements achieved after the fourth course of decitabine (November 2016), we performed a comprehensive BM reassessment; this showed a complete remission (CR) with incomplete hematological recovery \[[@B8]\]. In the light of her good clinical condition as well as the therapeutic response to decitabine (certainly better than we could have expected in an AML patient refractory to multiple treatment lines and complicated by extramedullary localizations), she was considered a fit candidate for haploidentical SCT. The patient underwent haploidentical SCT with her daughter as donor in January 2017 \[[@B11]\]. The conditioning regimen consisted of thiotepa 5 mg/kg on days −6 and −5, fludarabine 50 mg/m^2^ on days −4−3−2, and intravenous busulfan 3.2 mg/kg on days −4−3. The stem cell source was unmanipulated bone marrow. Graft versus host (GvHD) prophylaxis consisted of Post-Transplant Cyclophosphamide (PTCy) 50 mg/kg given on days +3 and +5 and cyclosporine A 1.5 mg/kg given as a continuous i.v. infusion from days 0 to +20, adjusted for blood levels (200 to 400 ng/mL), and then orally until day +180. The patient achieved a neutrophil count of 0.5×10^9^/L on day +17 and a platelet count of 30×10^9^/L on day 28; chimerism was full donor, by microsatellites, from the first evaluation on day +30. In particular, no acute graft versus host disease (GvHD) or other clinically significant side effects occurred. During her follow-up, a second vertebroplasty was performed on August 2018, due to the osteopenic collapse of the D12 vertebral body in the absence of any histological finding of AML localization. To date (October 2019), 43 and 23 months from the MDS primary diagnosis and allogeneic SCT, respectively, the patient has maintained a stable and long lasting CR and is well and active. In conclusion, in this case, decitabine achieved the CR of a secondary, pretreated and refractory AML, allowing for a bridge to successful allogenic SCT. Although the favorable clinical course of our patient has to be considered as unusual in contrast to what we unfortunately observe in most of the patients with high-risk MDS after azacitine-failure (or its transformation into secondary AML), it offers some interesting insights to consider. The achievement of a CR using decitabine in a patient who had previously received azacitidine is quite rare; it is well known that decitabine therapy is typically of little benefit after azacitidine failure \[[@B12]\]. In our case, as a mere speculation, we believe that the long period of time (about one year) that elapsed between the administration of decitabine from the first hypomethylating treatment with azacitidine may have contributed to re-establishing a good enough sensitivity to epigenetic therapy. This reported experience demonstrates the efficacy and applicability of haploidentical SCT, bridged by decitabine in our case, even for cases of clinically complex and pretreated patients with a long disease history. Also, the availability of novel agents able to induce a significant clinical response in patients with refractory AML could increase the number of patients who could benefit from allogeneic SCT (in its various practices) as an effective consolidation strategy, even after a long history of disease.
**Authors\' Disclosures of Potential Conflicts of Interest:** No potential conflicts of interest relevant to this article were reported.
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#s1}
============
Ecological theory suggests that predictions about population dynamics in planktonic systems have limited skill if intraspecific trait variation and heterogeneous resource distributions are ignored. Intraspecific trait variation within microalgal and bacterial populations has been shown to non-linearly affect intra- and interspecific interactions and population responses to resource availability, pathogens, and predators (Hellweger and Kianirad, [@B23]; Bolnick et al., [@B6]; Bucci et al., [@B8]). Such trait variations can arise from any of the following: physiological history of individual cell lines, niche plasticity, natural selection, mutation, genetic drift, or recombinant events (Bolnick et al., [@B6]). Even clonal populations (isogenic) in apparently homogenous cultures exhibit varying genetics, biochemistry, physiology, and behavior, all of which can produce a range of "growth phenotypes" (Lidstrom and Konopka, [@B36]; Damodaran et al., [@B12]; Kopf et al., [@B33]). This recognition has given rise to agent-based or individual-based models which don\'t assume that cell attributes within a population are uniformly or normally distributed around their mean values (e.g., Hellweger and Kianirad, [@B23]). In addition, aquatic ecologists and biogeochemists now recognize that the planktonic realm has fine-scale structure imposed by heterogeneous distributions of nutrients, oxygen, particles, colonial microbes, and polymeric gels, as well as by metazoan behavior, and symbiotic associations (Azam, [@B2]; Simon et al., [@B43]; Wagner et al., [@B52]; Stocker, [@B45]; Gemmell et al., [@B17]). Historically however, the vast majority of measurements have overlooked this microspatial heterogeneity and cryptic material exchanges (e.g., Canfield et al., [@B9]), and have provided net biogeochemical transformations at best. They also fail to unequivocally link key players to particular processes. These limitations hamper a deeper mechanistic understanding of planktonic dynamics. Single-cell techniques are the only way forward to empirically determine how intraspecific phenotypic/genotypic variations and microspatial architecture (the aquascape) determine individual ecophysiologies and translate into collective population responses.
Several single-cell technologies have enabled examination of intra-population variability in genotype, phenotype, activity and cellular growth. Single-cell genomics and transcriptomics provide unique information on genetic variations and are gaining popularity in the aquatic sciences (Stepanauskas, [@B44]; Wu et al., [@B55]). The synchrotron X-ray microprobe has provided insights into intraspecific variability in individual cell elemental stoichiometries (Twining et al., [@B49]). Highly sensitive serial resonant mass sensor arrays have enabled measurement of changes in buoyant masses of bacterial and mammalian cells as they transit through microfluidic channels, thereby tracking somatic growth of individual cells through time (Cermak et al., [@B10]). Combining fluorescent *in situ* hybridization (FISH) with microautoradiography (MAR-FISH) (Lee et al., [@B34]) or with secondary ion mass spectrometry (nano-SIMS-FISH) (Orphan et al., [@B39]) provides single-cell resolution for linking identity to ecophysiology in complex microbial assemblages. Nano-SIMS-FISH has been combined with SIP to detect nutrient assimilation by individual cells (Musat et al., [@B38]; Orphan et al., [@B40]; Foster et al., [@B14]). Using deuterated water (D~2~O) and ^15^$\text{NH}_{4}^{+}$ as tracers and SIP-Nano-SIMS analysis enabled Kopf et al. ([@B33]) to demonstrate that intra-population variability in growth rates and ammonium assimilation could be measured in chemostat-grown bacterial cells. However, these techniques generally have low sample throughput, demanding sample preparation requirements, and can be costly in terms of time and/or money invested per cell, all of which can limit the scale of population surveys.
Raman microspectroscopy is amenable to single-cell applications and is complementary to MAR-FISH and nano-SIMS-FISH. Raman microspectroscopy has the advantages of non-destructively yielding intracellular molecular information, of requiring minimal sample preparation, and enabling rapid interrogation of many preserved or live cells. Recent advances in Raman microspectroscopic technology have dramatically broadened its microbiological applications (Brehm-Stecher and Johnson, [@B7]; Wagner, [@B51]; Huang et al., [@B29]; Wang et al., [@B53]). For example, Raman spectra of single cells have revealed metabolic histories and species identity, through characterization of an organism\'s macromolecular composition (e.g., Huang et al., [@B28], [@B27]; Hermelink et al., [@B24]; Hall et al., [@B21]). Huang et al. ([@B30]) demonstrated that intensity ratios of specific wavenumbers within Raman spectra from individual bacteria varied quantitatively with amount of ^13^C-glucose available. Furthermore, those cells were phylogenetically identifiable by FISH probing (SIP-Raman-FISH). Li et al. ([@B35]) recently demonstrated that assimilation of ^13^C-enriched dissolved inorganic carbon (DIC) by individual photoautotrophic cells can be accurately quantified from wavenumber shifts in resonance Raman (SCRR) spectral peaks emanating from carotenoid pigments. Carotenoids are excellent target analytes because all photoautotrophic microbial taxa produce at least one form as accessory light-harvesting pigments or as protection against reactive oxygen species (Garcia-Asua et al., [@B16]). They are easily resolved by resonance Raman scattering, which increases photon scattering efficiency over spontaneous Raman scattering by at least a 1,000-fold by using laser excitation within the electronic transition frequency band of the analyte (e.g., Taylor et al., [@B48]; Robert, [@B42]).
We present a refinement of the SIP-SCRR-FISH approach (Li et al., [@B35]) that now enables quantitative Raman spectrometric measurement of growth rates in individual photoautotrophic cells. We use this tool to examine growth as the ultimate expression of inter- and intraspecific trait variability. Cells from replicate isogenic *Synechococcus* sp. cultures provided with varying concentrations of ^13^C-bicarbonate were interrogated by SCRR through time course experiments to determine their degree of labeling from which single-cell growth rates were calculated and compared to independent measurements of population growth. ^13^C-labeled populations of *Synechococcus* sp. and the diatom, *Thalassiosira pseudonana*, growing at known, but different rates were mixed to demonstrate that SCRR could distinguish among them. Optimal experimental design, underlying assumptions, analytical precision, and intra-population heterogeneity are evaluated. This technique has the advantages of relatively high sample throughput, low analytical costs, and potential application to natural phytoplankton communities in field experiments.
Materials and methods {#s2}
=====================
Media preparation and cultivation conditions
--------------------------------------------
Phytoplankton cultures were grown in f/2 media (Guillard and Ryther, [@B20]) in which total inorganic carbon (C~T~ = CO~2~ + H~2~CO~3~ + $\text{HCO}_{3}^{-}$ + $\text{CO}_{3}^{2 -}$) was either replaced or augmented with varying proportions of ^13^C-enriched bicarbonate \[$\text{HCO}_{3}^{-}$; *see Supplementary Material (SM) [1](#SM1){ref-type="supplementary-material"}*\]. The ^12^C- and ^13^C-bicarbonate solutions (Cambridge Isotope Laboratories, Inc. Andover, MA; 99% ^13^C, 97% chemical purity) were prepared as 0.4 M working stocks. Nutrient and bicarbonate solutions were aseptically added to autoclaved filtered seawater (\<0.22 μm = FSW) in 200-ml sealed septum bottles. C~T~ concentrations were then measured using a flow injection analysis system and compared to known C~T~ standards (Hall and Aller, [@B22]).
For the SIP experiment examining isotopic end-members, cultures of the cyanobacterium, *Synechococcus* sp. (RS9916), were grown at natural ^13^C abundances and at conditions under which 96% of the C~T~ was replaced with ^13^C-bicarbonate by pH manipulation (see *SM [1](#SM1){ref-type="supplementary-material"}*). For the SIP calibration experiment, complete f/2 media was augmented with sterile bicarbonate solutions with varying ^13^C abundances (*f*~media~ = ^13^C~media~/(^12^C~media~ + ^13^C~media~)) and a mean final C~T~ of 3.78 ± 0.10 mM C (2 mM added). To minimize variations in growth conditions, C~T~ was kept constant while *f*~media~ was manipulated. Nominal (gravimetric) *f*~media~ determinations (0.011, 0.10, 0.20, 0.30, 0.40, and 0.50) were corrected to 0.011, 0.11, 0.22, 0.32, 0.43, and 0.54, respectively, based on actual *f*~media~ measured by isotope ratio mass spectrometry (IRMS) at the UC Davis Stable Isotope Facility. After inoculation, each septum bottle was attached to its own venting system permitting atmospheric exchange of all gases, except CO~2~. Modifications of Li et al.\'s ([@B35]) venting system were as follows: Ascarite II™ (Thomas Scientific®) replaced NaOH pellets as our CO~2~ trap in open 30 ml syringe barrels, capped with polyester fiber, and connected to 0.2 μm in-line filters that were joined to septum bottles by 1/8" (O.D.) stainless steel swan neck tubes. The swan neck replaced vertical tubes to prevent condensate from the CO~2~ trap leaking into septum bottles during extended incubations and contaminating samples. Cultures were subsampled through time while incubating at 20°C on a rotating platform that assured uniform light exposure (48--63 μmol quanta m^−2^ s^−1^) during a 12:12 h light/dark cycle.
Directly evaluating the SIP-SCRR method\'s ability to measure varying *f*~cell~ or growth rate in natural phytoplankton assemblages is difficult, if not impossible, without an independent method to measure non-uniform growth rates among the assemblage members. Therefore, a *Synechococcus* sp. assemblage was constructed by mixing equal volumes of 12-days cultures from six different *f*~media~ ratios. These cells were ostensibly growing at the same rate, but had different *f*~cell~ signatures due to exposure to different *f*~media~, essentially mimicking a mixed assemblage with varying growth rates. In a second experiment to compare contemporaneous slow-growing (*Synechococcus* sp.; mean g = 4.12 days) and fast-growing (*T. pseudonana*; mean g = 1.44 days) populations, cultures were grown in parallel at natural (0.011 *f*~media~) and elevated ^13^C abundances (0.48 *f*~media~) and subsampled through time. SCRR filters were prepared from mixtures of these cells as a mimic of a heterogeneous environmental sample.
Population growth in SIP experiments
------------------------------------
Daily change in *in vivo* chlorophyll *a* fluorescence was used to measure population growth in all treatments. After gentle agitation, triplicate 200-μl subsamples were assayed in a Turner Designs® Aquafluor™ fluorometer, calibrated according to manufacturer\'s instructions. When required, subsamples were diluted with sterile f/2 to remain below 80% detector saturation. Population growth rates (μ~pop~) were calculated from arbitrary fluorescence units (AFU) either as the regression slope of ln AFU vs. time for the entire exponential growth phase (mean μ~pop~) or as the difference between neighboring ln AFUs within the time course (instantaneous μ~pop,inst~ = \[ln AFU~t+1~--ln AFU~t~\]/\[(t+1)--t\]). Direct microscopic cell counts in the control sample confirmed that AFU values were highly correlated with cell concentrations (*r*^2^ = 0.97, *p* \< 0.001) during exponential growth phase, signifying that ΔAFU/Δt is a reliable proxy for population growth under our experimental conditions.
SCRR sample preparation from SIP experiments
--------------------------------------------
For SCRR microspectrometry sampling, volumes removed (0.25---5.00 ml) were adjusted to obtain cell densities of 20--50 cells per microscope field and replaced with N~2~ gas to prevent a partial vacuum within the incubation bottles. After vortexing subsamples, cells were collected on 25 mm 0.2 μm polycarbonate membranes (Millipore® GTTP™) and rinsed with phosphate-buffered saline and filtered to dryness. Membranes were then air-dried in Millipore® PetriSlides™ and stored at −20°C.
In early experiments, potential interference of Cy3 fluorescent oligonucleotide probes with SCRR data acquisition was evaluated on membrane subsamples subjected to FISH. Membrane wedges were hybridized with a general bacterial probe (EUBMIX: Amann et al., [@B1]; Daims et al., [@B11]) in 35% formamide for 2 h and washed with buffer for 30 min and stored at −20°C until analysis (Pernthaler et al., [@B41]).
Although polycarbonate membranes are widely used for FISH and other fluorescence microscopies, these and most other types of filters and membranes produce contaminating Raman emissions (Raman-active). Likewise, borosilicate glass slides and cover slips and most mounting fluids (Citifluor™, Vectashield™, Cargille® Type A™ immersion oil, and glycerol) are Raman-active, water being the exception. Therefore, we developed a technique to transfer cells from polycarbonate membranes to mirror-finished 304 stainless steel slides (1 × 3 × 0.0235″) supplied by Stainless Supply® (Monroe, NC USA). Prior to use, slides were sequentially washed in an ultrasonic bath using acetone, isopropanol, and methanol, followed by MilliQ™ water rinsing at each step, and then air-dried.
To prepare both probe-hybridized (FISH) and non-hybridized cells for SCRR, samples were transferred to cleaned metal slides using a modified filter-transfer-freeze (FTF) technique (Hewes and Holm-Hansen, [@B25]). Briefly, each membrane wedge was placed sample-side down on a small droplet of sterile, particle-free MilliQ™ water (2--5 μl depending on wedge size) and the slide was placed on an aluminum cooling (−80°C) block. The membrane was peeled from the steel surface immediately after freezing, leaving most cells frozen to the slide, which was then air-dried in the dark (Suter, [@B47]). Transfer efficiency of *Synechococcus* cells was determined microscopically on filter wedge replicates; one set analyzed prior to freeze transfer and the other afterwards.
Single-cell resonance Raman microspectrometry
---------------------------------------------
SCRR measurements were performed using a Renishaw® inVia™ confocal Raman microspectrometer configured with a modified upright Leica® DM2700™ fluorescence microscope, a computer-controlled motorized XYZ stage (0.1 μm step size), 514 nm Ar^+^ ion laser, and 1,040 × 256 CCD Peltier-cooled detector. After cells were manually targeted by mouse clicks on a Cy3-fluorescent or autofluorescent digital image, the automated stage centered each cell under the laser beam for Raman data acquisition. For each sample, 25--40 cells (0.5--1.8 μm equivalent spherical diameters) were individually interrogated at 10% laser power (92 ± 7 μW at sample) through a Leica® dry 100 × (NA = 0.90) objective lens which produced a spot diameter of 0.7 μm. Spectra were obtained using a 65 μm slit centered at 1,752 μm, a 1,800 line/mm diffraction grating, aligned to wavenumber region between \~400 and 2,000 cm^−1^ (1,350 cm^−1^ center) and acquisition of two 1 s CCD detector exposures.
Spectral analysis
-----------------
Spectra were processed using Renishaw\'s® Wire 4.1™ software by first subtracting baselines using the software\'s standard best fit polynomial algorithm, and its intensity normalization function to standardize all spectra to a maximum peak value of 1, because cell-to-cell Raman scattering intensities are variable. Effects of ^13^C enrichment on behavior of dominant resonance Raman (RR) peaks were evaluated by two methods; peak-picking and curve-fitting. For peak-picking, center positions and widths of dominant spectral peaks were simply tabulated by the software and recognition thresholds were adjusted manually when required.
For curve-fitting analyses, two of the dominant RR peaks, usually positioned at \~1,521 and 1,157 cm^−1^, were deconvolved using Wire 4.1\'s function with our own empirical scripts. Peak shape and position are controlled by relative contributions of each isotopologue bond (e.g., ^12^C^12^C, ^12^C^13^C, ^13^C^13^C) to total peak area (*see SM [2](#SM1){ref-type="supplementary-material"}*). Center positions ($\left\langle {\Delta\overset{\sim}{\nu}} \right\rangle$ = wavenumber, cm^−1^) for pure ^12^C^12^C and ^13^C^13^C isotopologues were estimated from the regression intercept and slopes of calibration curves \[e.g., ν (C = C) vs. *f*~*cell*~\], respectively. Center positions for the intermediate isotopologue peaks were estimated as the midpoint between the end-members. After local baseline correction, contribution of each isotopologue to triplet peak area was determined by allowing peak positions to float ± 1 cm^−1^ and peak widths to float ± 10 cm^−1^ and by running a full Voigt fit routine (floating Lorentzian and Gaussian contributions to peak form) for 5,000 iterations or a \<0.00001 tolerance (difference between reduced chi-square values of two successive iterations), whichever came first. The curve fit function continuously compared the sum of the isotopologues\' areas (synthetic peak) to observed data and reported the reduced Chi square values (goodness of fit).
Determination of cell labeling and single cell growth rates
-----------------------------------------------------------
Since CO~2~ in the septum bottles is not allowed to exchange with the atmosphere, the isotopic signature of autotrophic cells (*f*~cell~) must predictably approach that of the media\'s DIC (*f*~media~, scaled for isotopic fractionation, α) with every cell division. Fractional labeling of populations was thus computed as follows.
To estimate α, an average δ^13^C value of −23%0 for marine plankton and a mean δ^13^C value of +1.5%0 for the ocean\'s mixed layer C~T~ pool were selected from global summaries (Goericke and Fry, [@B18]; Hoefs, [@B26]). Cellular ^13^C fractionation, α~cell−DIC~, was calculated from the isotope ratios (R = ^13^C/^12^C) of phytoplankton and seawater DIC using Equation (1).
α
c
e
l
l
\-
D
I
C
=
R
c
e
l
l
R
D
I
C
=
δ
13
C
c
e
l
l
\+
1000
δ
13
C
D
I
C
\+
1000
=
\-
23
\+
1000
\+
1
.
5
\+
1000
=
0
.
976
Values for *f*~*cell*~ in specific samples were calculated via Equation (2), with uncertainties propagated from individual terms (*see SM [3](#SM1){ref-type="supplementary-material"} and [4](#SM1){ref-type="supplementary-material"} for derivations of equations*).
f
c
e
l
l
=
α
f
m
e
d
i
a
1
\+
(
α
\-
1
)
f
m
e
d
i
a
\+
(
α
f
o
1
\+
(
α
\-
1
)
f
o
\-
α
f
m
e
d
i
a
1
\+
(
α
\-
1
)
f
m
e
d
i
a
)
e
\-
n
l
n
(
2
)
where *f*~o~ = ancestral fractional isotopic signature of the media (0.011 ^13^C) and *n* = number of generations completed in the presence of enriched DI^13^C, which was calculated by Equation (3) using the population\'s specific growth (μ~*pop*~, day^−1^) measured by *in vivo* fluorescence.
n
=
t
g
=
t
μ
p
o
p
ln
(
2
)
where *t* = time elapsed (days) and *g* = generation time (days).
Single-cell growth (μ~sc~) can be calculated independently of population growth data if the *f*~cell~ of individual cells can be measured. SCRR enables such determinations if a spectral feature, such as mean peak wavenumber ($\left\langle {\Delta\overset{\sim}{\nu}} \right\rangle$) for the ν(C = C) or ν(C-C) bonds in carotenoids varies predictably with *f*~cell~. This relationship can be utilized according to Equation (4).
〈
Δ
ν
\~
〉
≈
b
0
\+
b
1
f
c
e
l
l
After determining the values of b~0~ and b~1~ experimentally, the number of generations (n) completed after spiking samples with DI^13^C can be calculated from measurements of $\left\langle {\Delta\overset{\sim}{\nu}} \right\rangle$ (Raman), *f*~media~, and *f*~o~ with Equation (5). Once *n* is derived, *g* or μ~sc~ is calculated from Equation (3).
n
≈
1
ln
(
2
)
ln
(
f
m
e
d
i
a
\-
f
o
f
m
e
d
i
a
\-
(
1
\+
(
α
\-
1
)
f
m
e
d
i
a
)
〈
Δ
ν
\~
〉
\-
b
o
α
b
1
)
Statistics
----------
Descriptive statistics, linear regressions, and analyses of variance (Kruskal-Wallis, Tukey, Dunn\'s and Holm-Sidak pair-wise comparison methods) were performed using SigmaPlot™ 13.0 software (Systat Software Inc.®). Graphics were produced by either Renishaw® Wire 4.1™ or SigmaPlot™ 13.0 software.
Results {#s3}
=======
SIP experimental design
-----------------------
For SIP experimental design, it was initially critical to establish whether DIC replacement or augmentation is preferable, to determine how much ^13^C-bicarbonate tracer is required, and then to accurately establish the value of *f*~media~. In principle, a known amount of the C~T~ pool must be removed before replacing with sufficient ^13^C-bicarbonate to return media to the original C~T~ pool size. Among published DIC replacement strategies, such as microwaving (Li et al., [@B35]), N~2~ purging, and pH manipulation, only pH manipulation adequately constrained the C~T~ and *f*~media~ terms (*see SM [1](#SM1){ref-type="supplementary-material"}*).
DIC replacement may be suitable for experiments with cultivated populations in synthetic media, but is clearly inappropriate for experiments with natural field assemblages. DIC augmentation requires less manipulation and therefore potentially introduces fewer artifacts. More than doubling C~T~ depressed the pH of filtered seawater (FSW) by only 0.23 units (Table [S1](#SM1){ref-type="supplementary-material"}). In this study, C~T~ concentrations established gravimetrically from DIC additions closely agreed with those determined by flow injection analysis and by mass spectrometry, so the C~T~ and *f*~media~ terms were very well-constrained in experiments presented below (Table [S2](#SM1){ref-type="supplementary-material"}), which is absolutely essential for accurate growth calculations. Accordingly, experiments reported here used media augmented with the same C~T~ but varying *f*~media~, and pH was titrated back to 8.0 prior to inoculation with 0.1N NaOH (SM1), unless otherwise noted.
SIP-Raman-FISH spectra
----------------------
While the identities of our cultures are known, phylogenetic identification of target cells by FISH enables application of SIP-SCRR to field samples. Cells hybridized against the Cy3-EUBMIX probe fluoresce brightly (Figure [1A](#F1){ref-type="fig"}) and can also be observed under bright-field illumination (Figure [1B](#F1){ref-type="fig"}). Cy3 fluorescence did not interfere with SCRR spectral acquisition when exciting within the carotenoid spectral absorption band (370--530 nm) at 514 nm (Figure [1C](#F1){ref-type="fig"}). Spectral responses to ^13^C enrichment of cellular biomass are evident in spectra *i--v* (Figure [1C](#F1){ref-type="fig"}). Spectrum *i* obtained in 2 s from a single cell grown in control media (*f*~media~ = 0.011) was dominated by three major peaks attributed to carotenoids. The 1,521 cm^−1^ peak has previously been assigned to ν (C = C) in-phase stretching, the 1,156 cm^−1^ to ν (C-C) in-phase stretching, and 1,004 cm^−1^ to a combination of δ (C = CH~3~) methyl deformation and δ (C-H) out-of-plane bending modes in β-carotene (Marshall and Marshall, [@B37]). Evolution of red-shifted peaks is evident as cells are grown in 54% DI^13^C for 3, 6, 12, and 24 days (spectra *ii--v*). Thus, cells actively growing photoautotrophically in DI^13^C are easily recognized by red-shifted positions of any of these three SCRR peaks.
{#F1}
Quantitative application of single-cell resonance Raman microspectrometry
-------------------------------------------------------------------------
To evaluate whether SCRR microspectrometry can accurately quantify fractional labeling of individual photoautotrophic cells (*f*~cell~), parallel time course incubations of *Synechococcus* sp. were conducted under varying *f*~media~. Populations maintained exponential growth for the first 18 days at a mean rate of μ~pop~ = 0.220 ± 0.002 day^−1^ (g = 3.20 ± 0.03 days) for all treatments as measured by *in vivo* fluorescence (Figure [S1](#SM1){ref-type="supplementary-material"}) and thus were unaffected by *f*~media~ values between 0.011 and 0.54 (*p* \> 0.98; Kruskal-Wallis one-way ANOVA). Thus, the range of *f*~media~ used did not appear to introduce artifacts to the SIP-SCRR experiments and theory dictates that all populations would incorporate ^13^C into intracellular pools in proportion to their particular *f*~media~.
SCRR spectra were subjected to Renishaw\'s® Wire 4.1™ curve-fitting routine to illustrate how resonances from ^12^C^12^C, ^12^C^13^C, and ^13^C^13^C bonds (isotopologues) contribute to the total area, position, and shape of peaks as cells become ^13^C-enriched. As an example, cells grown in *f*~media~ = 0.011 (Figure [2A](#F2){ref-type="fig"}) and 0.54 for 3--9 days illustrate the effects of ^13^C enrichment on the 1,521 cm^−1^ (ν (C = C)) peak (Figures [2B--D](#F2){ref-type="fig"}). During early stages of labeling (Figures [2B,C](#F2){ref-type="fig"}), the triplet peak widens with a distinct shoulder that disappears as ^12^C=^13^C replaces ^12^C=^12^C (Figure [2D](#F2){ref-type="fig"}). This transition causes red-shifting of the triplet peak\'s mean resonance ($\left\langle {\Delta\overset{\sim}{\nu}} \right\rangle$) and eventually leads to bandwidth narrowing (Figures [1](#F1){ref-type="fig"}, [2D](#F2){ref-type="fig"}). Curve-fitting skill is illustrated by the coherence between fit curve solutions (gray lines; Figure [2](#F2){ref-type="fig"}) and observed data (green lines; Figure [3](#F3){ref-type="fig"}). The mean reduced Chi Square was 1.00 ± 0.25 (S.D.) for the entire 0.54 *f*~media~ time course (*N* = 172 cells).
{ref-type="fig"}. Isotopologue peak widths and symmetries were allowed to float to optimize curve fits.](fmicb-08-01449-g0002){#F2}
{ref-type="supplementary-material"} and SM [3](#SM1){ref-type="supplementary-material"}, particularly Equations (S11, S20).](fmicb-08-01449-g0003){#F3}
Curve-fitting analysis of spectra from the entire *f*~media~ = 0.32 time course illustrate that the proportion of underlying peak areas attributed to ^12^C=^12^C bonds decreased asymptotically to \~50%, while those of ^12^C=^13^C and ^13^C=^13^C bonds increased asymptotically to \~44 and \~6%, respectively, as ^13^C was incorporated into cellular biomass (non-linear least-squares regression, solid lines, Figure [3](#F3){ref-type="fig"}). Observed changes are consistent with our model for ^13^C-incorporation at random positions in a representative carotenoid molecule (β-carotene) (dashed lines, Figure [3](#F3){ref-type="fig"}; see SM [2](#SM1){ref-type="supplementary-material"} and SM [3](#SM1){ref-type="supplementary-material"} for derivation). In the present example, fractionation would theoretically limit *f*~cell~ to a maximum value of 0.31, at which the expected proportions of ^12^C=^12^C, ^12^C=^13^C, and ^13^C=^13^C bonds and their associated peak areas would approach 48, 43, and 10%, respectively (Equations S11, S20). However, unlike ^12^C=^12^C and ^12^C=^13^C bonds, the observed and predicted proportions of ^13^C=^13^C bonds were significantly different after 24 days (Figure [3C](#F3){ref-type="fig"}). We attribute this discrepancy to lower curve-fitting skill caused by the baseline correction method employed, which could undoubtedly be improved upon. The same time/growth-dependent trends in isotopologue contributions to the (ν (C = C)) peak area were observed at all levels of DI^13^C augmentation (not presented). Curve-fitting skills for the isotopologues contributing to the 1,156 cm^−1^ (ν (C-C)) and 1,004 cm^−1^ (δ (C = CH)) peak areas were less satisfactory due to baseline uncertainties and overlapping spectral features from other molecular bonds, particularly near 1,004 cm^−1^ (not presented).
As an alternative to curve-fitting peak areas as a function of time (e.g., Figure [3](#F3){ref-type="fig"}), we simply compared software-determined center positions ($\left\langle {\Delta\overset{\sim}{\nu}} \right\rangle$) of major peaks in randomly selected cells (*N* = 25 cells sample^−1^) for all DI^13^C-augmented cultures to the control (0.011 *f*~media~). Figure [4](#F4){ref-type="fig"} illustrates typical asymptotic responses of the three major peak wavenumbers as functions of growth in the *f*~media~ = 0.32 treatment. Declines in peak wavenumber (red shifts) are steep over the first 2 generations, slowing as they approach an asymptote in about 5--7 generations. Clearly, growth-dependent changes in mean Raman peak wavenumbers ($\left\langle {\Delta\overset{\sim}{\nu}} \right\rangle$) of the three carotenoid signature peaks observed in single-cells (boxes and whiskers) agree well with theoretical predictions (dashed lines) of red-shifting resonances for a population isotopically equilibrating with its media (*see SM [2](#SM1){ref-type="supplementary-material"} for derivation of prediction*).
{ref-type="supplementary-material"} and incubation times (n = t/g). Box and whisker plot details are as described in Figure [3](#F3){ref-type="fig"}. Solid curves are hyperbolic fits to all observations (*N* = 325 cells). Dashed lines are theoretical predictions for a population isotopically equilibrating with its media based on wavenumber response to *f*~cell~ and generation time (*see SM [2](#SM1){ref-type="supplementary-material"} and SM [3](#SM1){ref-type="supplementary-material"} for derivation*). Solid horizontal lines are least square means of all observations in control samples. Peak assignments are **(A)** ν (C = C), **(B)** ν (C-C), and **(C)** δ (C = CH).](fmicb-08-01449-g0004){#F4}
The most useful, generalized trend is obtained by constructing calibration curves directly comparing $\left\langle {\Delta\overset{\sim}{\nu}} \right\rangle$ with *f*~cell~. Knowing *f*~media~ and population growth (μ~*pop*~ or *g*) from independent chemical and fluorescence measurements (Figure [S1](#SM1){ref-type="supplementary-material"}) and assuming an isotopic fractionation factor (α) of 0.976 Equation (1), *f*~cell~ can be calculated from Equation (2) and Equation (3). Mean SCRR wavenumbers ($\left\langle {\Delta\overset{\sim}{\nu}} \right\rangle$) of the three major carotenoid peaks declined linearly across all values of *f*~cell~ examined; from natural ^13^C abundances (*f*~cell~ = 0.0107) to an *f*~cell~ of \~0.53 (Figure [5](#F5){ref-type="fig"}). The regressions\' coefficients of determination (*r*^2^ = 0.88--0.97), the slopes\' standard errors (s.e. = 0.18--0.26) and 99% confidence intervals demonstrate well-constrained relationships between *f*~cell~ and ($\left\langle {\Delta\overset{\sim}{\nu}} \right\rangle$). Collectively, the robust linear regressions and their agreement with theory suggest that *f*~cell~ can be reliably approximated by simply measuring $\left\langle {\Delta\overset{\sim}{\nu}} \right\rangle$.
{ref-type="supplementary-material"}*). Solid lines are empirical least squares linear regressions through individual data points (*N* = 836 cells). Dotted lines define the 99% confidence intervals of each slope (*b*~1~); **(A)** ν (C = C), **(B)** ν (C-C), and **(C)** δ (C = CH).](fmicb-08-01449-g0005){#F5}
Analogous examination of the curve-fitting approach across the entire *f*~cell~ range yielded results consistent with peak position analysis, i.e., contributions of the individual isotopologues to triplet peak area varied predictably with cell ^13^C-content (not presented). However, scatter evident in isotopologue peak contributions, largely due to variable curve-fitting skill, renders isotopologue peak area analysis less suitable than the highly predictable relationship between $\left\langle {\Delta\overset{\sim}{\nu}} \right\rangle$ and *f*~cell~ for quantitative applications. Nevertheless, empirical curve-fitting results conform to theory and illustrate the mechanics behind red-shifting peak wavenumbers.
Variability and sensitivity of SCRR
-----------------------------------
Scatter in SCRR microspectrometric data emanate from both natural cell-to-cell variability in ^13^C-content inherent to different growth phenotypes and from analytical uncertainty. Our sample size (25--40 cells) was adequate to distinguish between time points and treatments and to represent intra-population variability (e.g., Figures [3](#F3){ref-type="fig"}--[5](#F5){ref-type="fig"}). If required, uncertainty could be reduced somewhat with larger sample sizes. Potential sources of analytical error in Raman microspectrometry include: improper spectrometer calibration, beam misalignment, varying laser intensity, poor beam focusing, variable target geometry, and contaminating materials associated with target cells. However, current instrumentation and best practices eliminate or minimize most of these issues. Concerns about small variations in focus and target geometry are irrelevant because they primarily affect emission intensity which is not critical to our analyses.
*Synechococcus* sp. cells perpetually grown on natural ^13^C/^12^C abundances serve as the best candidate for an authentic standard to assess analytical error. Table [1](#T1){ref-type="table"} summarizes spectral data obtained from all control cultures. Assuming ^13^C and ^12^C atoms are distributed at random positions in the carotenoid molecules, about 98% of all C = C bonds are predicted to be ^12^C=^12^C in populations grown in natural ^13^C abundances (*see SM [2](#SM1){ref-type="supplementary-material"}*). In the dataset most amenable to curve-fitting \[the ν (C = C) triplet\], however, the ^12^C=^12^C isotopologue in all control cultures accounted for an average of 89 ± 9% of the ν (C = C) peak area with ^12^C=^13^C and ^13^C=^13^C isotopologues contributing the remaining 9 and 1%, respectively (Figure [2A](#F2){ref-type="fig"}, Table [1](#T1){ref-type="table"}). Although within uncertainty of the predicted percentages, this large discrepancy likely results from our curve-fitting model\'s simplicity, uncertainties in absolute peak positions of the three isotopologues, baseline position, and peak broadening in condensed matter, all of which could contribute to curve-fitting inaccuracies.
######
Analytical precision of single-cell resonance Raman spectral features assessed from 150 *Synechococcus* sp. (RS9916) continuously cells grown at natural ^13^C abundances (*f*~media~ = 0.011).
**% of total area** **δ (C = CH) 〈Δν〉** **ν (C-C) 〈Δν〉** **ν (C = C) 〈Δν〉**
------------------------------------------ ------ --------------------- ------ ----------------------- -------------------- ----------------------
Minimum 52.9 0.0 0.0 1002.6 1155.1 1520.2
Maximum 100 47.1 4.6 1006.1 1157.2 1522.4
Mean 89.0 10.1 1.0 1004.4 1156.3 1521.1
*SD* 8.9 9.4 1.0 0.65 0.29 0.34
%CV[^a^](#TN1){ref-type="table-fn"} 9.9 93.0 97.9 0.06 0.03 0.02
99% C.I.[^b^](#TN2){ref-type="table-fn"} ±1.9 ±2.0 ±0.2 ±0.14 ±0.06 ±0.07
*%CV = coefficient of variation = standard deviation × 100/mean*.
*99% confidence intervals*.
In contrast, the mean wavenumbers ($\left\langle {\Delta\overset{\sim}{\nu}} \right\rangle$) of all three carotenoid signature peaks were practically invariant in natural ^13^C abundance control populations, and indicate the technique\'s analytical reproducibility (Table [1](#T1){ref-type="table"}). Precision for $\left\langle {\Delta\overset{\sim}{\nu}} \right\rangle$ peak position determinations was excellent; the 99% confidence intervals around the 1004.4, 1156.3, and 1521.1 cm^−1^ means were ±0.14, 0.06, and 0.07 cm^−1^, respectively for 150 *Synechococcus* cells. From these analyses, we conclude that $\left\langle {\Delta\overset{\sim}{\nu}} \right\rangle$ is a far more precise indicator of *f*~cell~ than isotopologue peak areas, and that the positions of the ν (C-C) and ν (C = C) peaks in control cultures are more reproducible than that of the δ (C = CH) peak.
Carotenoids in photoautotrophs have evolved a wide variety of chemical structures along diverging phylogenetic lines. Therefore, SCRR spectra from different species are likely to differ subtly. In the interest of developing a broadly applicable SCRR tool to ultimately determine growth rates in complex phytoplankton communities, available Raman spectral data for carotenoids were evaluated (Table [2](#T2){ref-type="table"}). Among a variety of photoautotrophic microorganisms, the ν (C = C) and δ (C = CH) peak positions in cells growing at natural ^13^C abundances spanned 10.2 and 8.1 cm^−1^, respectively, while the ν (C-C) peak spanned just 2.4 cm^−1^ (Table [2](#T2){ref-type="table"}). Furthermore, ν (C-C) peak wavenumbers observed in photoautotrophic protists were statistically indistinguishable from those of cyanobacteria (ANOVA *p*-value = 0.905), whereas the ν (C = C) and δ (C = CH) peak positions were more phylogenetically distinctive (Table [2](#T2){ref-type="table"}). Therefore, even though the ν (C = C) peak had the broadest dynamic range and highest precision within our *Synechococcus* cultures, we chose the more phylogenetically conserved ν (C-C) peak for further quantitative analyses.
######
Variability in major wavenumber positions (cm^−1^) for carotenoid peaks in photosynthetic microorganisms grown under natural stable isotope abundances.
**Source** **ν (C = C)** **ν (C-C)** **δ (C = CH)/δ (C-H)** **References**
------------------------------------------------- --------------- ------------- ------------------------ --------------------------------
β-carotene 1,518 1,157 1,010 Marshall and Marshall ([@B37])
**PROKARYOTES**
*Synechocystis* sp. (PCC 6803) 1,517 1,156 1,004 Li et al. ([@B35])
*Synechocystis elongatus* (PCC 7942) 1,522 1,158 1,006 Li et al. ([@B35])
*Synechococcus* sp. (RS9916) 1,521 1,156 1,006 This study
Prokaryote Mean 1519.9 1156.5 1005.3
*SD* 2.8 1.1 0.9
%CV 0.18 0.09 0.09
**EUKARYOTES**
Arctic "AMA" microalgae 1,524 1,157 1,003 Li et al. ([@B35])
*Thalassiosira weissflogii* (Bacillariophyta) 1,526 1,157 1,010 This study
*Thalassiosira pseudonana* (Bacillariophyta) 1,526 1,156 1,010 This study
*Heterocapsa triquetra* (Dinoflagellata) 1,525 1,156 1,011 This study
*Aureoumbra lagunensis* (Pelagophyte) 1,527 1,158 1,009 This study
Eukaryote Mean 1525.6 1156.6 1008.5
*SD* 1.2 0.9 3.2
%CV 0.08 0.08 0.32
ANOVA *p*-value[^a^](#TN3){ref-type="table-fn"} 0.006 0.905 0.250
*Kruskal-Wallis one way analysis of variance comparing prokaryotic and eukaryotic values. p-values exceeding 0.05 indicate no statistical difference between data sets*.
Theoretical sensitivity or limit of detection (LOD) of the mean SCRR wavenumber approach was calculated using the regression statistics from Figure [5](#F5){ref-type="fig"} and the formula, LOD = 3 *S*~P~*/b*~1~, where *S*~P~ is the pooled standard deviations (±0.305 cm^−1^) of $\left\langle {\Delta\overset{\sim}{\nu}} \right\rangle$ observed at each value of *f*~cell~ and *b*~1~ = 30.3 cm^−1^/*f*~cell~ (Winefordner and Long, [@B54]). This calculation yielded a LOD of 0.03 leading us to conclude that a Δ *f*~cell~ of \~3% is detectable by our technique.
Single-cell growth rates determined by SCRR
-------------------------------------------
Autotrophs will necessarily approach the isotopic signature of DI^13^C-augmented media (*f*~media~) as a function of time, growth rate, and isotopic fractionation \[Equation (2) and Equation (3)\]. Therefore, the empirical linear relationship between $\left\langle {\Delta\overset{\sim}{\nu}} \right\rangle$ and *f*~cell~ (Figure [5](#F5){ref-type="fig"}) can be used to estimate single-cell growth rates from time course (e.g., Figure [4](#F4){ref-type="fig"}) or brief end-point incubations. To illustrate, 25 single-cell growth rates (μ~sc~) throughout the 0.32 and 0.43 *f*~media~ incubations are compared to proximal daily population growth rates (μ~pop,inst~) determined at t~x−1~, t~x~, and t~x+1~ (open boxes and whiskers Figures [6A,B](#F6){ref-type="fig"}). While mean μ~pop~ from 0 to 18 days was \~0.21 day^−1^ for both populations (broken horizontal lines), these relatively slow-growing photoautotrophs (g ≈ 3.3 days) exhibited considerable day-to-day oscillations in μ~pop,inst~. (diamonds). We speculate that these oscillations result from intrinsic cell cycle periodicity and/or extrinsic variations in incubation conditions. Nonetheless, median growth rates computed over the entire exponential phase of *f*~media~ 0.32 and 0.43 cultures were 0.22 ± 0.07 and 0.20 ± 0.07 day^−1^ for single cells (Raman-derived) and 0.22 ± 0.07 and 0.22 ± 0.07 day^−1^ for populations (fluorescence-derived), respectively (Figures [6A,B](#F6){ref-type="fig"}). In fact, μ~sc~ computed from SCRR measurements between 0 and 18 d for the 0.11, 0.32, 0.43, and 0.54 *f*~media~ treatments were statistically indistinguishable from μ~pop,inst~ measured over the same period (*p* \> 0.05; ANOVA). Inexplicably, μ~sc~ rates derived from the 0.22 *f*~media~ treatment were lower than those from μ~pop,inst~ measurements in all but the 3 day sample (*p* \< 0.05 ANOVA; not presented).
{ref-type="supplementary-material"}) for cultures incubated in *f*~media~ = 0.32 **(A)** and 0.43 **(B)**, and Bland and Altman ([@B5]) plot comparing results from both methods and treatments **(C)**. See Materials and Methods for calculations. Box and whisker plot details are as described in Figure [3](#F3){ref-type="fig"}. Shaded boxes represent single-cell growth rates (μ~sc~) and open boxes represent daily population growth rates (μ~pop,inst~) at *t*~x−1~*, t*~x~, and *t*~x+1~. Horizontal broken line is mean μ~pop~ over entire exponential growth phase (0--18 days). **(C)** Differences between μ~sc~ and μ~pop,inst~ compared to their mean values to assess their agreement. Broken horizontal lines represent the 95% confidence intervals (±1.96 standard deviations).](fmicb-08-01449-g0006){#F6}
Both *in vivo* fluorescence and SIP-SCRR based measurements of specific growth rate are subject to errors that may contribute differing amounts to the observed variability. For comparison, the mean %CV for $\left\langle {\Delta\overset{\sim}{\nu}} \right\rangle$ values obtained by SCRR (± 0.05%; *N* = 861) within individual samples was two orders of magnitude smaller than that obtained by daily triplicate *in vivo* fluorescence measurements (± 3.7%; N = 144; S.E. = ± 0.23). The mean relative error for *f*~cell~ was ±2.0% (S.E. ± 0.25), which is controlled by uncertainties in *b*~0~ and *b*~1~ in Figure [5](#F5){ref-type="fig"} and measurements of *f*~media~. CVs for μ~sc~ tend to be higher at lower *f*~media~ and shorter incubations times *(see SM [4](#SM1){ref-type="supplementary-material"} for discussion of uncertainties)*. Of the observed variance for μ~sc~ through time ($\overline{CV}$ = 29% ±2.4 S.E.) illustrated in Figure [6](#F6){ref-type="fig"}, more than 93% appears to be due to cell-to-cell heterogeneity in physiological state within these isogenic populations. Likewise, most variance observed in μ~pop,inst~ through time ($\overline{CV}$ = 64%±24 S.E.) is likely due to actual variations in cell division, rather than analytical imprecision. We note that *in vivo* fluorescence measurements were made at approximately the same time every day (±1.5 h). Consequently, samples were undoubtedly withdrawn at different points within the 3.3 days cell cycle of these populations.
An approach routinely used in biomedical studies for evaluating agreement between two independent methods of measuring a single variable is to visualize the distributions of the differences between the two measurements plotted against their means (Bland and Altman, [@B5]). The Bland-Altman plot is more appropriate than regression analysis when those independent measurements have small dynamic ranges that are surpassed by the measurements\' observed variability. Distributions of observed μ~sc~---μ~pop,inst~ in the *f*~media~ = 0.32 and 0.43 at each time point are compared to (μ~sc~ + μ~pop,inst~)/2 in Figure [6C](#F6){ref-type="fig"}, where the μ~pop,inst~ term for each μ~sc~ was the mean determined at *t*~*x*−1~*, t*~*x*~, and *t*~*x*+1~. If the two measurement methods consistently produced identical results, all points would fall on the horizontal zero line. In the current case, the average difference is +0.003, meaning that on average the μ~sc~ method returns a 0.3% higher value than the μ~pop,inst~ method among all observations. The +1.96 and --1.96 boundaries (broken horizontal lines) represent the 95% confidence intervals, and 16 of the 17 outliers are positive, again illustrating that the μ~sc~ method tends to return slightly higher values than the μ~pop,inst~ method. Absolute accuracy of our μ~sc~ method cannot be determined from this analysis largely due to cell-to-cell variability and uncertainties over which of the daily μ~pop,inst~ measurements are most representative of a particular μ~sc~ time point. Nonetheless, the data distribution in Figure [6C](#F6){ref-type="fig"} illustrates that the vast majority of SIP-SCRR-derived growth rates are within the 95% confidence intervals of independently-measured population growth rates. Furthermore, the preceding error analysis suggests that most observed variability is attributable to real differences in growth activity and not to analytical error.
SIP-SCRR-derived growth rates in mixed assemblages
--------------------------------------------------
Our final objective was to demonstrate that SIP-SCRR can measure growth rates of individual cells in a mixed photoautotrophic assemblage whose members are assimilating ^13^C-bicarbonate at contrasting rates. In the absence of an independent method to measure single-cell growth in natural communities, we constructed an artificial assemblage by dispensing equal volumes of *Synechococcus* cell suspensions from the six *f*~media~ treatments (Figure [5](#F5){ref-type="fig"}) at a single time point. Figure [7A](#F7){ref-type="fig"} presents the SCRR spectra obtained from 13 individual cells in the mixed assemblage (Figure [7B](#F7){ref-type="fig"}). After interrogating multiple fields, the statistical distribution of *Synechococcus* sp. cells with six distinctive ^13^C signatures added to the mixture were indistinguishable (ANOVA; *p* = 1.00) from those recognized by SIP-SCRR (Figure [7C](#F7){ref-type="fig"}). Therefore, morphologically identical cells with distinctive ^13^C signatures can be recognized and reliably quantified within an isotopically mixed assemblage by this method.
{#F7}
To further evaluate application of this technique to phylogenetically diverse field assemblages, subsamples of parallel cultures of relatively slow-growing cyanobacteria and fast-growing diatoms were mixed at several time points. Grown in identical media (*f*~media~ = 0.011 and 0.48), the diatom, *T. pseudonana*, and cyanobacterium, *Synechococcus* sp., maintained exponential growth for 6 days, attaining population growth rates of 0.483 and 0.161 day^−1^, respectively. In Figure [8](#F8){ref-type="fig"}, the distribution of *f*~cell~ values computed by plugging SCRR observations of $\left\langle {\Delta{\overset{\sim}{\nu}}_{C - C}} \right\rangle$ into Equation (4) are compared to *f*~cell~ values predicted from α, *f*~media~, population growth measured by *in vivo* fluorescence of the cultures and Equation (2). The trend line computed from all cells (*N* = 253) spanning *f*~cell~ values of 0.0107 to 0.449 is not significantly different from the 1:1 line. These results demonstrate that within experimental error and intra-population variability, fractional ^13^C-labeling of single diatom and cyanobacterial cells computed from SIP-SCRR measurements closely agrees with fractional labeling of entire populations grown at natural and enhanced ^13^C-DIC abundances. Results presented in Figures [7](#F7){ref-type="fig"}, [8](#F8){ref-type="fig"} support the proposition that single-cell growth rates within mixed natural photoautotrophic assemblages can be computed from SIP-SCRR measurements.
{#F8}
Discussion {#s4}
==========
Single-cell productivity techniques
-----------------------------------
Individual or agent-based ecological models suggest that intra-population variability in genetics, cell line history and microspatial geochemical heterogeneity can influence population responses to environmental forcing along multiple pathways and lead to divergent outcomes (e.g., Hellweger and Kianirad, [@B23]; Bolnick et al., [@B6]; Bucci et al., [@B8]; Fredrick et al., [@B15]). Empirical testing of such models is challenging because the requisite single-cell productivity measurements have been limited historically to a few laboratory techniques with cultures. For example, division of individual cells encapsulated in mineral oil has been followed through time microscopically (Dewan et al., [@B13]; Damodaran et al., [@B12]) and increasing mass of individual cells has been measured through time in serial microfluidic mass sensor arrays (Cermak et al., [@B10]). MAR-FISH provides single-cell activity measurements, but does not yield growth rates *per se*. Recent evidence suggests that two complementary techniques, SIP-nano-SIMS-FISH and SIP-Raman-FISH, hold the most promise for providing single-cell growth rates in both the lab and field experiments.
SIP-nano-SIMS-FISH has been used to estimate growth of anaerobic phototrophic bacteria as well as for cells in syntrophic and symbiotic partnerships (Musat et al., [@B38]; Orphan et al., [@B40]; Foster et al., [@B14]). However, the growth computations in these studies depended upon several variables with unspecified uncertainties, such as substrate pools\' isotopic composition estimates, assumed assimilative/dissimilative relationships, and microscopic biovolume measurements used to derive elemental content based on allometry. Recognizing this limitation, Kopf et al. ([@B33]) recently demonstrated that accurate single-cell growth rates could be derived by applying appropriate mathematical models to data obtained from bacterial chemostat cultures labeled with ^2^H~2~O under well-constrained conditions and analyzed by SIP-nano-SIMS.
Previous studies applying SIP-Raman to aquatic microbiology measured cell activity levels by altered Raman spectral features and then cataloged them either morphometrically or by combining with FISH (e.g., Huang et al., [@B30]; Li et al., [@B35]; Berry et al., [@B4]). Huang et al. ([@B30]) were the first to demonstrate that red shift ratios in single-cell Raman spectra of cultivated bacteria varied predictably with ^13^C-content of the media. Li et al. ([@B35]) recently demonstrated that SIP-SCRR red shifts in the three major carotenoid peaks ($\left\langle {\Delta\overset{\sim}{\nu}} \right\rangle$) vary quantitatively with *f*~cell~ of *Synechocystis* sp. and *Synechococcus elongatus*, grown in ^13^C-bicarbonate enriched media. In fact, when their data from these two cyanobacteria are combined, we obtain functional responses that closely agree with those presented in Figure [5](#F5){ref-type="fig"}; e.g., $\left\langle {\Delta{\overset{\sim}{\nu}}_{C - C}} \right\rangle$ = 1156.5--31.0±1.4 *f*~cell~; *r*^2^ = 0.99 (computed from Figure [3](#F3){ref-type="fig"} in Li et al., [@B35]). The *f*~cell~ values derived from mass spectrometric analysis of extracted cellular protein (Li et al., [@B35]) yielded essentially the same functional responses as carotenoid $\left\langle {\Delta\overset{\sim}{\nu}} \right\rangle$. This is consistent with the postulates that turnover of major intracellular molecular pools are tightly coupled under optimal culture conditions and that the isotopic signatures are equivalent, i.e., α~protein~ ≈ α~carotenoids~, which supports the validity of our approach. Furthermore, it corroborates the proposition that as long as α and μ~pop~ are known and *f*~media~ is well-constrained, mean *f*~cell~ of an autotrophic culture can be reliably computed and precludes the need for mass spectrometry. To our knowledge, the present study is the first to extend SIP-SCRR data interpretation beyond fractional labeling and relative activity determinations to quantify absolute single-cell growth rates from photoautotrophic populations.
Single-cell growth rates by SIP-SCRR-FISH
-----------------------------------------
Our approach is well-suited to examine cell to cell heterogeneity in photoautotrophic growth in both cultures and field samples. Almost all carbon assimilated by aquatic photoautotrophs is drawn from the C~T~ pool. Therefore, dissolved ^13^C-bicarbonate, which equilibrates with C~T~ species, is an unsurpassed tracer of autotrophic production. Surface ocean values of C~T~ only vary between \~1.95 and 2.25 mM C globally (Key et al., [@B32]), and can be empirically measured in samples by several methods (e.g., Hall and Aller, [@B22]; Key et al., [@B32]; Kaltin et al., [@B31]). Therefore, the experimentalist can add precisely measured masses of H^13^$\text{CO}_{3}^{-}$ to samples and accurately calculate *f*~media~, which is ultimately needed to derive μ~sc~. In contrast, DIC replacement manipulations lead to uncertainties in C~T~ and calculation of *f*~media~ (Table [S1](#SM1){ref-type="supplementary-material"}), necessarily alter water chemistry, and are impractical for field experiments. Therefore, we conclude that spiking unaltered samples with a sterile H^13^$\text{CO}_{3}^{-}$ concentrate to achieve the desired *f*~media~ is preferable to any DIC replacement protocol (e.g., Li et al., [@B35]). In principle, our experiments demonstrated that H^13^$\text{CO}_{3}^{-}$ amendments as small as 10% are adequate to exceed our LOD (3% Δ*f*~cell~) within 0.5 generations (Figure [5](#F5){ref-type="fig"}) and enable recognition of active cells. However, if accurate and precise determination of μ~sc~ is the objective, then H^13^$\text{CO}_{3}^{-}$ amendments yielding 0.3--0.5 *f*~media~ and incubation durations of one generation are recommended to constrain the optimal propagated relative uncertainties (± σ~μ~/μ) to between 6.6 and 11.0% (see Figures [S2](#SM1){ref-type="supplementary-material"}--[S4](#SM1){ref-type="supplementary-material"}). These amendments do not appear to significantly alter photoautotrophic growth in seawater (Figure [S1](#SM1){ref-type="supplementary-material"}).
Sample processing requirements for SCRR or SCRR-FISH are relatively simple. Samples can be captured on membranes and frozen or preserved with formaldehyde or paraformaldehyde, but not glutaraldehyde (due to fluorescence interference). If FISH is required, samples are processed by standard membrane-based protocols (e.g., Pernthaler et al., [@B41]; Stoecker et al., [@B46]). In contrast, previous SIP-Raman-FISH studies required performing hybridizations on cell suspensions (liquid-FISH) concentrated by centrifugation or resuspension from a membrane, then spotting suspensions onto quartz or Al-coated glass slides, followed by drying, rinsing and DAPI counter-staining (Huang et al., [@B30]; Berry et al., [@B4]). Membrane-based approaches are advantageous because cell recovery exceeds that obtained by centrifugation or by resuspension from membranes. Furthermore, the risk of cell loss from the microscope slide during rinsing or counter-staining in the liquid-FISH technique is eliminated.
Hybridization of fluorophore-labeled oligonucleotide FISH probes into cells may introduce two challenges to Raman interrogation; unwanted fluorescence and isotope dilution. Persistent fluorescence from a sample illuminated by a laser within its electronic transition band can prevent acquisition of Raman spectra altogether. As demonstrated by Huang et al. ([@B30]), bacterial cells hybridized with FISH probes labeled with Cy3, Cy5, or Fluorescein (Fluos) were all amenable to single-cell Raman microspectroscopy using a 532 nm laser, but cells labeled with Cy3 required brief photobleaching prior to acquisition of spectra. We only tested a Cy3-labeled probe using our 514 nm laser and could obtain high quality SCRR spectra of carotenoids without significant fluorescent interference. Nonetheless, Cy5, Fluos and perhaps other yet-to-be tested fluorophores are probably preferable to Cy3 for most applications, but practitioners should test fluorophores with desired laser line.
Isotope dilution created by incorporation of a FISH probe is a concern with SIP-nano-SIMS because the instrument is measuring secondary ions generated from individual atoms ablated from a specimen, irrespective of source molecules. Thus, ionized carbon from isotopically-enriched biomass can be diluted with ionized light carbon from a FISH probe. In the SIP-Raman-FISH approach, however, isotopic content determinations arise from wavenumber shifts in signature molecular bond vibrations, in the present case from ν(C-C) in carotenoids. Therefore, aliasing *f*~cell~ determinations from diagnostic SCRR peaks by isotopically-light FISH probes is improbable in the present study and can be avoided in other Raman assays. For example, Huang et al. ([@B30]) determined ^13^C-content in FISH-hybridized cells from red shift ratios in diagnostic peaks emanating from the essential amino acid, phenylalanine, were indistinguishable from unhybridized cells. Cells hybridized by CARD-FISH (catalyzed reporter deposition) may not be amenable to Raman interrogation because required reagents may be Raman-active and contaminate the spectra, but this requires future evaluation.
While membrane-based cell collection, hybridization, and staining techniques present significant advantages for working with microbes in natural waters, all tested membranes and filters are Raman-contaminating and consequently can\'t be used directly in the instrument. Therefore, we were motivated to develop a rapid, simple method to efficiently transfer concentrated and stained cells from low sorption membranes (polycarbonate) to a Raman-neutral substrate for interrogation. With minimal optimization efforts, the filter-transfer-freeze (FTF) technique met those goals and achieved transfer efficiencies of 51--77% for *Synechococcus* to mirror-finished stainless steel slides. For most anticipated applications, complete cell transfer is not required to adequately subsample populations captured on membranes. However, evaluation of other membrane materials may be warranted to maximize transfer efficiencies and minimize selective retention on membranes. The steel slides have several advantages. Firstly, they are completely reusable after solvent cleaning. Secondly, they cost 1--3% that of the quartz and Al-coated slides used in previous Raman studies (e.g., Huang et al., [@B30]; Li et al., [@B35]). The third advantage is that potentially higher quantum efficiencies improve the method\'s sensitivity. This is explained as follows: Raman-shifted photons tend to scatter more or less in a 360° sphere (isotropic), depending on sample morphology, opacity, volume, and substrata properties. However, reflective opaque metal slides or Al-coated glass substrata limit scattering geometry to 180°, reflecting photons upward and thus potentially doubling photon capture efficiency compared to transparent substrata. Lastly, the highly reflective surface enables recognition of cells as small as bacteria under bright-field illumination without using any staining or FISH procedures.
Intra-population growth rate heterogeneity
------------------------------------------
In this study, computed growth rates within a given isogenic *Synechococcus* sp. population varied from cell to cell by \~27% (CV) around the mean at any particular time point after removal of the \~2% propagated analytical error. Similar intra-population variations have been observed in isogenic populations of photoautotrophic protists. For example, microscopic cell count measurements of μ~sc~ for *Chlorella vulgaris* varied between 0.55 and 1.52 day^−1^ ($\overline{x}$ = 1.16 ± 0.37 SD) among mineral oil-encapsulated droplets of media within a microfluidic device, yielding a CV of 32%, while μ~pop~ in bulk media was 1.12 day^−1^ (Dewan et al., [@B13]). Similarly, subpopulations in a synchronized isogenic *Chlamydomonas reinhardtii* culture grew at significantly different rates among oil-encapsulated droplets within a millifluidic sampler (Damodaran et al., [@B12]). Using SIP-nano-SIMS, Kopf et al. ([@B33]) reported CVs of 19--51% for μ~sc~ of chemostat cultures of the bacterium, *Staphylococcus aureus*. Interestingly, the heterogeneity in single cell growth rate (CV) varied inversely with chemostat dilution rate, i.e., behavior of individual cells in slow growing *S. aureus* populations was less uniform than that of fast growers. We observed a similar trend with fast-growing *T. pseudonana* (CV = 8--9%) and slower-growing *Synechococcus* sp. (CV = 23--34%) (Figure [8](#F8){ref-type="fig"}).
We are unaware of any comparable measurements of μ~sc~ in cyanobacteria. However, significant variations in elemental stoichiometry within *Synechococcus* field populations have been observed using synchrotron X-ray fluorescence microscopy. For example, an average CV of 120% in Si/P ratios was apparent among cells within discrete tropical and subtropical water samples (Baines et al., [@B3]). Similarly, quotas for P and Fe varied by 68 and 97% CV, respectively, among individual *Synechococcus* cells in the Sargasso Sea (Twining et al., [@B50]). Even individual *Synechococcus, Prochlorococcus*, and *Thalassiosira* cultures exhibit significant intra-population variability in intracellular P content (CV = 17--86%), which can affect the collective population growth rate (Fredrick et al., [@B15]). Thus, considerable cell-to-cell variation in elemental composition, activity, and growth of photoautotrophic populations can be expected whether they are isogenic and exposed to ostensibly homogenous conditions or are anisogenic and living in dynamic heterogeneous environments. Such findings and agent-based ecological models underscore the need for reliable single-cell measurements to better understand processes driving plankton population dynamics as well as biological cycling of key elements (Bolnick et al., [@B6]; Bucci et al., [@B8]; Fredrick et al., [@B15]).
Conclusions {#s5}
===========
We conclude that, compared to other technologies, our approach to single-cell growth measurement has some clear advantages. SCRR has minimal sample preparation requirements, provides relatively high sample throughput, and at low cost per analysis (excluding capital expenses). Specimens can be interrogated *in vivo, in vitro*, frozen, preserved, or dried, and don\'t require special analytical conditions, such as electrical conductivity, embedding, or high vacuum (as in nano-SIMS). Organisms can be recognized by bright-field, autofluorescence, fluorescent stains, or FISH illumination. All targeted cells in a field can then be immediately analyzed by Raman microspectrometry by altering the optical path with a few computer keystrokes and allowing the automated stage to center each target under the laser spot. Raman microspectrometry offers a unique and powerful set of capabilities to chemically interrogate individual cells and investigate biogeochemical processes at spatial scales relevant to microorganisms. Combining SIP with FISH and Raman microspectrometry allows identification of key players in cycling of a particular element or compound and thereby enables linking function with phylogeny. Furthermore, rates of material cycling can be determined from features within single-cell Raman spectra. If the stable isotope traces the sole source of that element into cells and *f*~media~ of that element can be measured, then single-cell growth rates can be determined.
As demonstrated in this study, a ^13^C-bicarbonate tracer and SCRR of carotenoids appear to be particularly well-suited to investigate intra- and inter-population variability in growth of cultured photoautotrophs. As documented by Goericke and Welschmeyer ([@B19]), carotenoid labeling rates closely match population growth rates during balanced growth which occurs during exponential phase in cultures. However, they also observed that nutrient and light stresses may alter this relationship, conditions which may be encountered in field samples. We recommend conducting SIP experiments over an entire diel cycle (24 h) to account for variability in light field responses and intrinsic biosynthetic periodicities and to produce sufficient signal. Resolution of the impact of light and nutrient limitation on broad application of the SIP-SCRR approach to examining intra- and inter-population growth variability in the field requires further systematic evaluation under a range of conditions.
Author contributions {#s6}
====================
All authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication.
Conflict of interest statement
------------------------------
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Authors are grateful to J. Collier for *Synechococcus* cultures, S. Zegers for *T. pseudonana* cultures, C. Heilbrun for DIC analyses and M. Pachiadaki for insightful comments on an earlier draft. IRMS analyses of DI^13^C were conducted by the UC Davis Stable Isotope Facility. Raman data were acquired in SoMAS\' NAno-Raman Molecular Imaging Laboratory (NARMIL), a community resource dedicated to environmental science applications and founded with NSF-MRI grant OCE-1336724. This research was also partially supported by NSF grants OCE-1335436 and OCE-1259110 and Gordon and Betty Moore Foundation Grant \#5064. School of Marine and Atmospheric Sciences contribution number 1436.
Supplementary material {#s7}
======================
The Supplementary Material for this article can be found online at: <http://journal.frontiersin.org/article/10.3389/fmicb.2017.01449/full#supplementary-material>
######
Click here for additional data file.
[^1]: Edited by: Chuanlun Zhang, Southern University of Science and Technology, China
[^2]: Reviewed by: Douglas Andrew Campbell, Mount Allison University, Canada; Ma Bo, Qingdao Institute of Bioenergy and Bioprocess Technology (CAS), China
[^3]: This article was submitted to Aquatic Microbiology, a section of the journal Frontiers in Microbiology
| {
"pile_set_name": "PubMed Central"
} |
1. Introduction
===============
G protein-coupled receptors (GPCRs) are a class of membrane proteins mediating important cellular signal transductions across the cell membrane. Structural studies for GPCRs have been conducted using X-ray crystallography and NMR spectroscopy \[[@B1-molecules-18-08579],[@B2-molecules-18-08579],[@B3-molecules-18-08579]\]. GPCRs have a structure fold containing an N-terminal domain, a seven-transmembrane core domain and a C-terminal domain. Although the membrane topology of the transmembrane domains of GPCRs is similar, the N- and C-terminal regions of different GPCRs have various length and structures. GPCRs comprise a protein superfamily consisting of Class A, B, C and Class Frizzled receptors (FZDs).The FZDs contain ten members and smoothened receptor (SMO) \[[@B4-molecules-18-08579]\]. The extracellular N-terminal domains of FZDs have a cystein-rich domain (CRD) followed by a hydrophobic linker with different lengths \[[@B5-molecules-18-08579]\]. The CRD binds Wnt molecules that are important in embryonic development and stem cell maintenance \[[@B5-molecules-18-08579]\].
The C-terminal regions of GPCRs interact with other proteins or help folding of the GPCRs. For class A rhodopsin-like family of GPCRs, the C-terminal region following the seventh transmembrane segment contains a helix which is amphipathic and referred as helix 8 \[[@B6-molecules-18-08579]\]. The helix 8 normally starts after a conserved NPXXY motif in the seventh transmembrane helix. Structures of the helix 8 from different GPCRs have been studied using NMR spectroscopy \[[@B7-molecules-18-08579],[@B8-molecules-18-08579],[@B9-molecules-18-08579]\]. In aqueous solutions, the peptide lacks secondary structures. In the presence of a membrane-mimicking environment such as detergent micelles, the peptide derived from the helix 8 forms a helical structure with hydrophobic residues interacting with detergent micelles \[[@B7-molecules-18-08579],[@B10-molecules-18-08579]\]. The conformational changes in the presence and absence of membrane-mimicking environment may represent the activation of the receptor upon ligand binding \[[@B7-molecules-18-08579],[@B11-molecules-18-08579]\]. The C-terminal domain following the transmembrane region of the FZDs have different length and sequence similarity ([Figure 1](#molecules-18-08579-f001){ref-type="fig"}). The domain has been shown to be necessary for localization of the receptors \[[@B12-molecules-18-08579]\]. Studies have also demonstrated that there is a short conserved motif (KTXXXW) present in the C-terminal domain \[[@B13-molecules-18-08579]\]. Mutations in this motif disrupted Wnt/β-catenin signaling \[[@B13-molecules-18-08579]\]. Further study has demonstrated that this motif interacted with the PDZ domain of the cytoplasmic protein-Dishevelled (Dvl) with a low binding affinity \[[@B14-molecules-18-08579]\]. By using a peptide scanning method, two motifs of the third intracellular loop and the C-terminal region of the FZD are identified to be important for forming a stable FDZ-Dvl complex \[[@B15-molecules-18-08579]\].
Due to the importance of the C-terminal domain of the FZDs in the Wnt/β-catenin signaling, it will be useful to understand their structures. Compared with class A rhodopsin-like family of GPCRs, there is no NPXXY motif present in the seventh transmembrane helix of the FZDs ([Figure 1](#molecules-18-08579-f001){ref-type="fig"}). If the C-terminal domain contains a helix 8 is still unknown. In this study, we conducted structural analysis of the C-terminal region of the FZD~1~ consisting of 25 residues ([Figure 1](#molecules-18-08579-f001){ref-type="fig"}). Our results indicated that the peptide was not structured in an aqueous solution, but formed a helical structure in detergent micelles such as sodium dodecyl sulfate (SDS) and dodecylphosphocholine (DPC). Sequence analysis and structural studies indicated that the C-terminal domain of FZD~1~ contains an amphipathic helix which is similar to other GPCRs. Residue such as Trp in this domain is interacting with micelles. Our result will be useful to understand the role of the C-terminal domains of FZDs in signal transductions.
2. Results and Discussion
=========================
The FZDs are unconventional GPCRs and classified as a novel family of GPCRs \[[@B4-molecules-18-08579]\]. For the conventional GPCRs, the C-terminal domain is important for signaling and G protein coupling. Examples include rhodopsin and CB1 receptors \[[@B16-molecules-18-08579],[@B17-molecules-18-08579]\]. The C-terminal domains of FDZs were also shown to involve in signal transduction, while its structure is still unknown.
2.1. Selection of the C-terminal Domain of FZD~1~
-------------------------------------------------
The C-terminal domains from the FZDs have different length and sequence diversity ([Figure 1](#molecules-18-08579-f001){ref-type="fig"}). To understand the structure, we focused on the FZD~1~ because it has a relatively short sequence which is suitable for NMR studies. Peptide corresponding to S623 to V647 was synthesized and purified ([Figure 1](#molecules-18-08579-f001){ref-type="fig"}). Further studies were carried out to understand its structure under different conditions.
{#molecules-18-08579-f001}
2.2. Circular Dichroism (CD) Experiment
---------------------------------------
Far-UV CD was first employed to examine the conformations of peptide derived from FZD~1~ under different conditions. CD result showed that when the peptide was prepared in an aqueous solution containing only 20 mM sodium phosphate, pH7.0, the peptide was unstructured ([Figure 2](#molecules-18-08579-f002){ref-type="fig"}A). In the presence of detergent micelles such as SDS and DPC that were used as membrane mimetics \[[@B18-molecules-18-08579]\], the CD spectra of the peptide in these two detergent micelles changed obviously with spectra characterized for an α-helix in which double minima at 222 and 208 nm were observed ([Figure 2](#molecules-18-08579-f002){ref-type="fig"}A). Further data analysis using the K2D2 server demonstrated that peptide acquired similar helical conformation in these two detergent solutions. It was predicted from the K2D2 server that the helical content was approximately 70% in these two solutions. Micelle-dependent conformational change has been demonstrated for several peptides derived from GPCRs and other membrane proteins \[[@B7-molecules-18-08579],[@B8-molecules-18-08579],[@B19-molecules-18-08579]\]. Our result suggested that the C-terminal region of the FZD~1~ could form an α-helix in the presence of detergent micelles mimicking conditions of cell membrane proximity.
{#molecules-18-08579-f002}
2.3. Fluorescence Analysis
--------------------------
To understand the interaction between the peptide and detergent micelles, fluorescence spectroscopy was carried out. It is known that the fluorescence emission spectrum of tryptophan residues is very sensitive to the environment, which can be used to measure residue and membrane interactions. In an aqueous solution, the emission maximum is close to 350 nm after excitation at 280 nm. If the tryptophan residue is buried in a hydrophobic environment such as in the presence of micelles, the emission maximum will be shifted to a lower wavelength \[[@B20-molecules-18-08579]\]. The peptide in the aqueous solution had an emission maximum of 350 nm, indicating that the W630 was exposed to the solvent ([Figure 2](#molecules-18-08579-f002){ref-type="fig"}B). In the presence of the DPC micelles, the emission maximum shifted to \~330 nm, indicating that the residue W630 was buried in the DPC micelles. Although there was no clear emission maximum observed when SDS micelles were present, the disappearance of maximum at 350 nm indicated that the tryptophan residue was buried in or affected by SDS micelles ([Figure 2](#molecules-18-08579-f002){ref-type="fig"}B).
2.4. NMR Study
--------------
SDS or DPC were used as a membrane mimetic in membrane protein structural studies \[[@B18-molecules-18-08579],[@B21-molecules-18-08579]\]. To study the structures of the peptide in the presence and absence of detergent micelles, NMR data for peptide dissolved in different solutions were collected. The peptide produced poorly dispersed spectra in DPC micelles which may due to the size of the peptide- DPC micelles (data not shown). Resonance assignments for the peptide in the aqueous solution and SDS micelles were obtained ([Figure 3](#molecules-18-08579-f003){ref-type="fig"}). Although peptide in aqueous solution exhibited dispersed cross-peaks in a ^15^N-HSQC spectrum, further NOE connection indicated that this peptide was not structured because only *d*~NN~ (*i*, *i*+1) connections were observed ([Figure 3](#molecules-18-08579-f003){ref-type="fig"}E). This result is consistent with the CD data. When the peptide was dissolved in SDS micelles, we conducted NMR experiments at 313 K because the signal was improved compared to 298 K. NOE connections indicated that residues L627 to N639 formed a helix, which was confirmed by the presence of αN (*i*, *i*+3) and αN (*i*, *i*+4) NOE connections ([Figure 3](#molecules-18-08579-f003){ref-type="fig"}). SDS micelles have been used as a membrane-mimicking system in structural studies of several membrane proteins \[[@B18-molecules-18-08579]\]. The data suggested that peptide may form a helix when it is close to the cell membrane.
2.5. Structure Determination
----------------------------
The structure of the peptide in SDS micelles was determined based on 104 NOEs ([Table 1](#molecules-18-08579-t001){ref-type="table"}). [Figure 4](#molecules-18-08579-f004){ref-type="fig"} shows the 20 superimposed structures with lowest energy. The root-mean-square deviation (RMSD) for the backbone atoms from the helix of the twenty structures was 0.4 Å ([Table 1](#molecules-18-08579-t001){ref-type="table"}). The restraints used in this study and the structural analysis using PROCHECK-NMR \[[@B22-molecules-18-08579]\] are shown in [Table 1](#molecules-18-08579-t001){ref-type="table"}. Residues L627 to N639 formed a helix and the C-terminus of this peptide was not structured ([Figure 4](#molecules-18-08579-f004){ref-type="fig"}). Helix wheel representation for this helix indicated that it behaves like an amphipathic helix ([Figure 4](#molecules-18-08579-f004){ref-type="fig"}). The hydrophobic residues such as W, L and Y are facing one side and charged residues such as R and N are facing the other side ([Figure 4](#molecules-18-08579-f004){ref-type="fig"}). These hydrophobic residues may be important for interacting micelles, for example, the Trp residue was buried in the micelles from our fluorescence experiment ([Figure 2](#molecules-18-08579-f002){ref-type="fig"}B).
Solvent-induced conformational changes have been observed in C-terminal domains of other GPCRs such as human β2 adrenergic receptor \[[@B8-molecules-18-08579]\] and CB1 cannabinoid receptor \[[@B7-molecules-18-08579]\]. In those receptors, their C-terminal domains contain a helix 8 following the seventh transmembrane helix. Helix 8 in GPCRs may serve as a membrane anchor because the helix is amphipathic and contains palmitoylation sites that can stabilize the interaction with membrane. The conformation of the helix may be regulated by allosteric modulators binding to the seven-transmembrane core domain. After activation by an allosteric ligand, the helix 8 may become exposed to form a disordered structure to expose its binding site for other proteins \[[@B7-molecules-18-08579]\]. In GPCRs such as rhodopsin, the hydrophobic residues in this motif may stack against each other to keep the receptor in a pre-receptive state \[[@B6-molecules-18-08579],[@B23-molecules-18-08579]\]. Unlike other GPCRs, there is no NPXXY motif present in the seventh transmembrane helix of FZDsFrom our studies, it was demonstrated that the C-terminal domain of FZD~1~ had similar characteristics to the helix 8 from other GPCRs such as CB1 receptor.
{#molecules-18-08579-f003}
molecules-18-08579-t001_Table 1
######
Experimental distance and statistics for the 20 structures determined in SDS micelles.
Restraints
---------------------------------------- --------
Intraresidue NOEs 71
Sequential NOEs (i to i+1) 65
Medium range NOEs (i to i + 2,3,4) 39
Dihedral angle restraints 0
Number of NOE violations \>0.5 Å 0
Ramachandran plot (%)
Residues in most favored regions 55.8
Residues in additional allowed regions 38.8
Residues in generously allowed regions 5.7
Residues in disallowed regions 0
Structural statistics
RMSD for backbone atoms (6-17) 0.4 Å
RMSD for heavy atoms (6-17) 0.86 Å
{#molecules-18-08579-f004}
The structure of the smoothened (SMO) receptor which shows sequence similarity to the FZDs was determined using X-ray crystallography. SMO also has a conserved KATXXXW motif at the C-terminus \[[@B24-molecules-18-08579]\] and this motif has been described to stabilize the receptor by packing parallel to the membrane ([Figure 5](#molecules-18-08579-f005){ref-type="fig"}). The tryptophan residue (W545) localized at the membrane interface, which is similar to what we observed in the C-terminal domain of the FZD~1~. The C-terminal domain of FZDs contains a conserved KTXXXW motif lacking an alanine residue for PDZ domain interactions \[[@B13-molecules-18-08579]\] ([Figure 1](#molecules-18-08579-f001){ref-type="fig"}). From our study, this motif had a tendency to form a helix. Although the KT residues within this motif are not helical under current conditions, this may arise from the fact that this peptide did not include the seventh transmembrane helix. The tryptophan residue (W630) is involved in the micelle interactions, which was confirmed from the fluorescence study ([Figure 2](#molecules-18-08579-f002){ref-type="fig"}B). Mutations in this motif may disrupt the helical structure or affect membrane interaction \[[@B13-molecules-18-08579]\], which in turn will affect the signal transduction or binding to Dvl \[[@B15-molecules-18-08579]\]. All these results suggest that the C-terminal domains of FZDs might have similar role to those of the conventional GPCRs. There is a cystein residue present in the helix 8 of GPCRs. This cystein can be modified by palmitoylation, which could serve as a lipid anchor. As there is no cystein present in the C-terminus of FZDs, this tryptophan residue may act as a lipid anchor to stabilize the interaction with membrane. Under different conditions, this domain interacts with Dvl or membrane, which will produce different signals for the downstream pathways. Previous studies have proposed that the helix 8 in GPCRs is involved in G-protein coupling and GPCR activation \[[@B10-molecules-18-08579]\]. The sequence similarity within FZDs suggested that the helix 8 may be a common structural feature within this family ([Figure 1](#molecules-18-08579-f001){ref-type="fig"}). This region might have the similar functional roles to the helix 8 of the classical GPCRs. Our study provided structural basis to understand its function.
![Role of the Trp residues in membrane binding. A. Crystal structure of the SMO receptor. B. Homology model of the FZD~1~ using the crystal structure of the SMO receptor as a template \[[@B25-molecules-18-08579]\]. The possible membrane interface is shown in dash line. The residues in both structures are shown in sticks.](molecules-18-08579-g005){#molecules-18-08579-f005}
3. Experimental
===============
3.1. Sample Preparation
-----------------------
The peptide sequence was SGKTLNSWRKFYTRLTNSKQGETTV and was derived from the C-terminal domain of the FZD~1~ corresponding to residue 623 to 647. This peptide was synthesized and purified from GL Biochem Ltd (Shanghai, China) with more than 95% purity. The peptide was prepared to a concentration of 5 mg/mL in a buffer containing 20 mM sodium phosphate, pH 7.0 or same buffer that contained 2% deuterated DPC or SDS. The samples with 10% D~2~O were transferred into 3-mm NMR tubes for data acquisition.
3.2. Circular Dichroism (CD) Spectroscopy
-----------------------------------------
For secondary structural analysis, samples were diluted to 0.2 mg of peptide per ml for CD analysis as described previously \[[@B19-molecules-18-08579]\]. Peptide was dissolved in sodium phosphate buffer in the present and absence of detergent micelles. All the CD experiments were conducted using a Chirascan™ Circular Dichroism Spectrometer at 25 °C. The CD spectra were recorded in a continuous mode with a 1-nm data pitch. The secondary structure was analyzed using the K2D2 server (<http://www.ogic.ca/projects/k2d2/>).
3.3. Fluorescence Spectroscopy
------------------------------
The effect of the solvent on the trytophan residue was measured using the similar method described \[[@B20-molecules-18-08579]\]. The fluorescence emission spectra were measured in a 96-well plate. Peptide was dissolved in the buffer to a concentration of 0.1 mg/mL. Samples with 100 μL volume in the presence and absence of micelles were subjected to analysis. Excitation wavelength was 280 nm and the emission was scanned from 305 to 405 nm. Experiment was carried out at 37 °C.
3.4. NMR Experiments
--------------------
The NMR data were collected on a Bruker Avance II spectrometer with a proton frequency of 700 MHz equipped with a cryoprobe. A total correlation spectroscopy (TOCSY) experiment was recorded with a mixing time of 80 ms \[[@B26-molecules-18-08579]\]. Two-dimensional (2D) nuclear overhauser effect spectroscopy (NOESY) experiments were recorded with mixing times of 100 ms, 200 ms and 300 ms, respectively. The experiments for peptide in sodium phosphate buffer were performed at 298 K. In the presence of micelles, the TOCSY and NOESY spectra were recorded at 313 K to increase the resolution. Proton chemical shift values were referenced directly to 4, 4-dimethyl-4-silapentane-1-sulfonic acid (DSS). All of the spectra were processed using Topspin 2.1 and analyzed using Sparky (<http://www.cgl.ucsf.edu/home/sparky/>).
3.5. Resonance Assignment and Structure Determination
-----------------------------------------------------
The resonance assignment for the peptide was obtained using the procedure including the identification of spin systems in a TOCSY spectrum and spin connections in a NOESY spectrum \[[@B26-molecules-18-08579],[@B27-molecules-18-08579]\]. The sequential connectivity of the residues was confirmed based on the connectivity in the HN-HN or Hα-HN region. A ^1^H-^15^N heteronuclear single quantum correlation spectroscopy (HSQC) in ^15^N natural abundance was acquired and processed to facilitate resonance assignment. The NOE peaks picked from a NOESY spectrum with a mixing time of 200 ms were integrated for structural determination. The peaks were assigned manually and calibrated to distances using CYANA 2.1 for structural determination \[[@B28-molecules-18-08579]\]. Structure determination for peptide in SDS micelles was conducted as described previously \[[@B26-molecules-18-08579]\]. The structure was calculated using torsion angle dynamics simulated annealing as implemented in CYANA using NOE restraints \[[@B26-molecules-18-08579],[@B28-molecules-18-08579]\]. One hundred structures were calculated and twenty of the structures with lowest energies were analyzed with MOLMOL \[[@B29-molecules-18-08579]\] and displayed using PyMOL (<http://www.pymol.org>).
4. Conclusions
==============
In summary, we carried out structural studies for a peptide derived from the C-terminal domain FZD~1~. Our results show that the peptide adopted different conformations under different solvent conditions. In the aqueous solution, it was not structured. It formed a helical structure when a membrane-mimicking detergent was present. The tryptophan residue (W630) in the KTXXXW motif may be important for membrane interaction.
CK appreciates the support from A\*STAR Investigatorship and A\*STAR JCO grant (10/03/FG/06/06).
*Sample Availability*: Commercial available.
The authors declare no conflict of interest.
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#Sec1}
============
Overexpression of somatostatin receptors on the neuroendocrine tumour cell surface is a very well-known phenomenon, but not all neoplasms demonstrate a high incidence of this receptor subtype. Therefore, there is a need to search for other peptide analogues for diagnostic and therapeutic purposes.
Insulinoma is the most common hormone-secreting neoplasm localized in the pancreas. In the majority of cases the tumours are benign. Their malignant form is observed in less than 10.0 % of insulinoma patients. Most often insulinoma appears as a singular lesion located in the pancreatic parenchyma (97.0 %), but rarely it might also be multifocal \[[@CR1]\]. A clinically accepted method for the diagnosis of insulinoma is the positive result of a 72-h fasting test with a high insulin level (over 6.0 μIU/ml), elevated C-peptide level (over 200.0 pmol/l) and absence of sulfonylurea in the plasma \[[@CR2]\]. Similar results also occur in cases of nesidioblastosis \[[@CR1]\]. Due to the usually small sizes of insulinoma lesions and the technical limitations of the majority of standard diagnostic modalities, its anatomic localization is sometimes difficult. In some cases not only multiphase contrast-enhanced computer tomography (CT), magnetic resonance imaging (MRI) or standard ultrasound (US), but even endoscopic ultrasound (EUS) is not sensitive enough to visualize the tumour. The sensitivity of EUS for the detection of insulinoma lesions in the pancreatic head, body and tail was reported to be 92.6, 78.9 and 40.0 %, respectively \[[@CR3]\]. However, other authors reported a higher sensitivity and specificity for the localization of intrapancreatic lesions by EUS (93.0 and 95.0 %, respectively) \[[@CR4]\]. Techniques such as selective arteriography, transhepatic peripancreatic venous blood sampling (TPVB), intra-arterial calcium stimulation test (ASVS) and intraoperative US revealed better sensitivity (95.0 %), but are much more time-consuming and of course invasive \[[@CR5]\]. Nuclear medicine imaging, such as somatostatin receptor scintigraphy (SRS), is positive in about 50.0--60.0 % of benign insulinomas \[[@CR1], [@CR2]\]. Surgery remains the only method for full recovery of insulinoma patients, but requires precise location of neoplasm foci. After a successful surgical therapy, a very good long-term survival outcome is being observed in patients with insulinoma \[[@CR2], [@CR6]\]. For these reasons more sensitive diagnostic options are required.
Glucagon-like peptide-1 (GLP-1) receptors have been found in normal organs such as pancreas, blood vessels, stomach or parafollicular C cells. GLP-1 receptor expression is also observed on different types of neoplastic cells (e.g. benign insulinoma, phaeochromocytoma, gastrinoma, paraganglioma, pulmonary neuroendocrine tumours and medullary thyroid carcinoma). According to Reubi and co-workers the density of GLP-1 receptors is extremely high especially on benign insulinoma cell surfaces. They have been found in almost 100 % of insulinoma cases---in contradistinction to ileal carcinoids, where GLP-1 receptor expression was found only in one third of cases with a heterogeneous distribution. Somatostatin receptors were detected in two thirds of examined insulinomas (mainly SSTR1, SSTR2 and SSTR5) \[[@CR7]\]. Comparable results were reported by Bertherat et al. \[[@CR8]\]. Therefore, in the near future GLP-1 receptor imaging may become a preferred diagnostic method for the localization of insulinomas hardly detectable by other examinations \[[@CR7]\].
The first study with a labelled GLP-1 analogue {\[Lys^40^(Ahx-DTPA-^111^In)NH~2~\]-exendin-4} was successfully performed by Wild et al. in two patients with insulinoma \[[@CR9]\]. Exendin-4 is a 39-amino acid peptide hormone originally found in the saliva of the Gila monster, which has been proved to be a long-acting potent agonist for GLP-1 receptors. A subsequent clinical study with the same ^111^In-labelled GLP-1 receptor agonist confirmed the high sensitivity of this method. In all examined patients benign insulinomas were detected (six of six) \[[@CR10]\]. The same group of scientists used \[Lys^40^(Ahx-HYNIC/EDDA)NH~2~\]-exendin-4 labelled with ^99m^Tc for the first time in a mouse model \[[@CR11]\]. The use of ^99m^Tc may improve the quality of images and radiation safety for patients and the staff by many procedural advantages related to the physical properties of this isotope.
The aim of this paper is to evaluate the clinical usefulness of the new radiopharmaceutical \[Lys^40^(Ahx-HYNIC-^99m^Tc/EDDA)NH~2~\]-exendin-4 injected into patients with suspicion of benign insulinomas that are hardly detectable or not diagnosed by other available methods.
Materials and methods {#Sec2}
=====================
Patients {#Sec3}
--------
Eleven patients (seven women and four men, mean age 48.0 ± 18.9 years, min. 16.0 years, max. 75.0 years) were enrolled in this study. There were eight patients with clinical and biochemical symptoms and signs of insulinoma, one patient with malignant insulinoma, one patient with suspected local recurrence of malignant insulinoma (MEN1) and one with nesidioblastosis. All patients were verified for a positive fasting test, elevated serum insulin and C-peptide level concentrations (fasting glycaemia 33.0--54.0 mg/dl, fasting insulinaemia 10.6--51.6 μIU/ml, C-peptide concentrations 4.7--6.3 ng/ml). None of the patients was treated with oral antidiabetic agents. Symptoms of hypoglycaemia have been observed for at least 1 year even up to 15 years in this group of patients. The Department of Endocrinology, Jagiellonian University Medical College in Cracow, Poland annually diagnoses up to 30 patients with suspicion of insulinoma. Because of difficulties in localizing the insulinoma focus by standard diagnostic methods, five to six of them were scheduled for the study with labelled GLP-1 analogue annually. Patients from other sites in Poland have also been diagnosed in our department during the time when this study was conducted.
Before GLP-1 receptor imaging, CT/MRI/EUS was performed in all patients; eight patients underwent SRS after the injection of ^99m^Tc-EDDA/HYNIC-TOC (Tektrotyd, Polatom, Poland) and two patients positron emission tomography (PET)/CT with ^68^Ga-DOTATATE. In all patients with suspicion of benign insulinoma all these scans were negative or equivocal. It should be stated that in this group of patients the negative or equivocal result of previous diagnostic studies was one of two main inclusion criteria to perform the GLP-1 examination (after stated clinical symptoms of insulinoma). In the first patient with malignant insulinoma the CT was positive in liver metastasis, but negative in the site of recurrence. In the second patient with malignant insulinoma ^68^Ga-DOTATATE PET/CT was positive for liver and lymph node metastases, but negative for the local recurrence. Detailed information about the patients is presented in Table [1](#Tab1){ref-type="table"}.Table 1Information about patients and examinations performedInitialsAgeSexDiagnosisHypoglycaemiaSRSCTGLP-1Subsequent treatmentJ.A.57FIns. susp.12 years−−+SurgeryR.G.54MIns. susp.1 year−−+Qualified for surgeryJ.W.16MIns. susp.1.5 years−+/−+SurgeryA.P.75MIns. susp.4 years+−+Disqualified from surgeryB.J.38FIns. susp.10 years^68^Ga-DOTATATE −−+SurgeryK.S.52FIns. susp.NA−+SurgeryM.K.62FIns. susp.NANot done−+SurgeryP.W.21MIns. susp.NA++/−+SurgeryA.K.65FMal. ins.4 years^68^Ga-DOTATATE + in liver meta and lymph nodes+ in liver meta, − in recurrence place−PharmacotherapyJ.J.57FMal. ins.15 years−+ in liver meta− in liver meta, + in site of recurrencePatient did not agree to surgeryA.U.31FNesidioNA+++Surgery*F* female, *M* male, *NA* not available, *−* negative result, *+* positive result, *+/−* equivocal result, *Ins. susp.* suspicion of insulinoma, *Mal. ins.* malignant insulinoma, *Nesidio* nesidioblastosis
After receiving detailed information about the study procedure, all patients gave their written consent. The study was approved by the local Ethics Committee.
Patients were placed on a liquid diet 1 day before the beginning of the examination and fasted on the day of the injection of the tracer. Each of them was carefully checked for any adverse reactions. Because of the natural disease course of insulinoma in fasting patients, blood pressure and glucose values were monitored before and after injection of the compound at several time points. Some of the patients required glucose infusion at the time of the GLP-1 receptor imaging procedure, because their level of glycaemia dropped below 40 mg/dl.
After GLP-1 receptor imaging patients without contraindications qualified for a surgical treatment. The histopathological confirmation of tumour presence and evaluation of its type was performed after surgery.
Preparation of \[Lys^40^(Ahx-HYNIC-^99m^Tc/EDDA)NH~2~\]-exendin-4 {#Sec4}
-----------------------------------------------------------------
The radiolabelling of \[Lys^40^(Ahx-HYNIC)NH~2~\]-exendin-4 with ^99m^Tc followed a two-vial kit formulation as previously published \[[@CR11]\] with some modifications \[[@CR12]\]. Briefly, 1 ml of a solution containing 30 mg (168 μmol) of tricine, 30 μg of \[Lys^40^(Ahx-HYNIC)NH~2~\]-exendin-4 and 33.6 μg of SnCl~2~ (12 μl of 22.2 mM SnCl~2~ × 2 H~2~O in 0.1 M HCl) was filtered into a glass vial strictly under air protection, and 1 ml of a solution containing 10 mg of EDDA (pH was adjusted to 7 with 1 M NaOH) was filtered into a second glass vial. The glass vials were immediately frozen in liquid nitrogen, lyophilized and closed afterward under nitrogen. For labelling, the EDDA vial was reconstituted with 1 ml of saline and 0.5 ml of it added to the \[Lys^40^(Ahx-HYNIC)NH~2~\]-exendin-4 vial, followed by 740--1,850 MBq of ^99m^TcO~4~^-^ in 0.3--1.5 ml eluate of ^99^Mo/^99m^Tc generator (Polatom, Poland) and incubated for 10 min at 100 °C. After cooling to room temperature, the reaction mixture was analysed by HPLC and thin-layer chromatography (TLC). The average radiochemical purity was more than 90 %. The mean injection activity for patients was 740 MBq.
Imaging technique {#Sec5}
-----------------
GLP-1 receptor imaging with the use of \[Lys^40^(Ahx-HYNIC-^99m^Tc/EDDA)NH~2~\]-exendin-4 was performed at the Nuclear Medicine Unit of the Endocrinology Department, University Hospital in Cracow. Firstly, images were acquired with a dual-head, large field of view E.CAM gamma camera with low-energy high-resolution (LEHR) collimators. From 2010 onwards all examinations were performed with the use of the Symbia TruePoint T16 hybrid system, also with LEHR collimators (Siemens Healthcare).
On the basis of the first patients' examinations our intention was to find the optimal acquisition protocol and to perform dosimetric calculations. Due to this, initial whole-body scans were carried out at six time points, from the first hour and in some cases even up to 30 h (1, 2, 3, 6, 24 and 30 h) after the injection of the compound. Standard camera settings (with scan speed 12 cm/min for each examination) were used for whole-body examinations with ^99m^Tc \[[@CR13]\].
Diagnostic single photon emission computed tomography (SPECT) studies were performed at two time points, between 3 and 4 h and between 5 and 6 h after the injection of the tracer. The SPECT examinations were done with standard camera settings like those used in the SRS study after labelled somatostatin analogue injection \[[@CR13]\]. The acquired data were reconstructed using the ordered subset expectation maximization (OSEM) FLASH 3-D iterative reconstruction method. After the installation of the new hybrid device in the Unit, SPECT/CT studies were carried out in all subsequent patients with the same settings for the SPECT part of the study and low-dose protocol for the CT part (130 kV, 13 mAs, slice thickness and reconstruction increment 5.0 mm).
In tumour tissue and kidneys, changes in the uptake were evaluated over time. Volumetric analysis was performed to assess the tumour to non-tumour ratio (T/nT ratio) and the kidney to non-tumour ratio (K/nT ratio). The total counts in defined regions of interest (ROI) were calculated. The background ROI was defined in the neighbouring normal tissue for each region. T/nT and K/nT ratios were assessed for SPECT scans performed between 2 and 3 h, 3 and 4 h, 5 and 6 h and 24 h (in two patients) after the injection of the compound. The obtained images were assessed by two experienced nuclear medicine specialists.
Absorbed dose assessment {#Sec6}
------------------------
We also analysed biokinetic characteristics of the new tracer and estimated the patient radiation dose. The organ radiation dose was assessed for five patients. The external conjugate view counting pair method was used for analysing whole-body images \[[@CR14]\]. Three source organs were chosen: kidneys, liver and lungs. The area under the time-activity curve for the whole body as well as the biological half-time and the effective half-time for the tracer were estimated. The organ absorbed dose coefficient D(r~T~, T~D~) was evaluated according to the Medical Internal Radiation Dose (MIRD) system \[[@CR15]\]\[[@CR16]\]. The total effective dose, using weighting factors defined in International Commission on Radiological Protection (ICRP) publication 103, was also evaluated \[[@CR17]\]. The blood clearance of the tracer and the radiation dose to the red marrow was established by the blood samples method \[[@CR18]\]. The plasma and blood fractions were measured separately. Urine samples were also collected (after the 1 h scan and before 2 h, 3 h and 6 h scans). The total volume of excreted urine and its activity was measured after each collection.
Results {#Sec7}
=======
The quality of the obtained \[Lys^40^(Ahx-HYNIC-^99m^Tc/EDDA)NH~2~\]-exendin-4 images was evaluated as very good.
In eight of eight cases with suspicion of insulinoma, focal uptake of the tracer in the pancreas was found (Fig. [1](#Fig1){ref-type="fig"}). In two patients GLP-1 receptor scintigraphy was repeated because of uncertain localization of the tracer in the pancreas. The second study confirmed the previous foci localization. So far, successful surgical excision of the tumour has been performed in six patients. Post-surgical resolution of symptoms was observed in all patients. Histopathological studies confirmed type G1 neuroendocrine tumours. One patient is awaiting surgery. One patient was disqualified from surgery due to cardiological disorders. In one patient (38-year-old woman) histopathological examination revealed the coexistence of insulinoma and nesidioblastosis.Fig. 1Patient J.A. (57 years old) with clinically overt hypoglycaemia. The examination with the GLP-1 analogue labelled with ^99m^Tc revealed focal accumulation of the tracer. The well-differentiated neuroendocrine tumour type G1 was found in the histopathological study after surgical excision of the part of the pancreas head selected during GLP-1 receptor imaging study. **a** Fusion of GLP-1 receptor imaging and CT sagittal slice. **b** Fusion of GLP-1 receptor imaging and CT axial slice. **c** Fusion of GLP-1 receptor imaging and CT coronal slice
In the group of patients with benign insulinoma both sensitivity and specificity of GLP-1 receptor imaging were assessed as 100 %.
The first patient with confirmed malignant insulinoma (65-year-old woman) was operated in 2007 due to a tumour of the pancreatic tail. She was qualified for peptide receptor radionuclide therapy (PRRT) in 2008, because of inoperable liver metastases. A complete response to the therapy was observed. However, in 2010 the patient presented again with hypoglycaemia (below 35 mg/dl) and hyperinsulinaemia (up to 58.8 μIU/ml). PET imaging with ^68^Ga-DOTATATE was positive for liver and lymph node metastases. The patient was also qualified for GLP-1 receptor imaging, which gave a negative result. The patient is being treated with verapamil and up to now symptoms of hypoglycaemia have not been observed any more. In the other patient (57-year-old woman), who has been suffering from MEN1 syndrome, the malignant insulinoma was resected in 1996. Liver and lymph node metastases were detected in 2010. GLP-1 receptor imaging confirmed the recurrence of malignant insulinoma (a focal uptake in the place of the removed pancreatic head), but uptake of neither GLP-1 nor somatostatin receptor analogues was observed in metastatic lesions. This patient did not agree to surgery.
Increased uptake of \[Lys^40^(Ahx-HYNIC-^99m^Tc/EDDA)NH~2~\]-exendin-4 in the upper abdomen (the result similar to SRS) was seen in the case of a patient with nesidioblastosis (31-year-old woman) (Fig. [2](#Fig2){ref-type="fig"}). In this patient, after three operations, each with partial pancreas resection, the observed focal accumulation of the tracer was necessary for further verification. The patient underwent surgical treatment, but died after the fourth surgery due to internal bleeding. The histopathological examination confirmed insulinoma coexisting with nesidioblastosis.Fig. 2Patient A.U. (31 years old) with nesidioblastosis. The examination with the GLP-1 analogue labelled with ^99m^Tc revealed focal accumulation of the tracer in the pancreatic head. Insulinoma with coexisting nesidioblastosis was confirmed in the histopathological examination after surgery. **a** CT study. **b** Fusion of GLP-1 receptor imaging and CT. **c** GLP-1 receptor imaging
No adverse reactions such as nausea, decreased blood pressure, bronchospasm, bradycardia, skin flushing or itching, hives or other effects after the tracer injection were reported by any of the patients at the time of the examination. Generally, we did not observe exacerbation of hypoglycaemia related to tracer injection, but this was difficult to verify in every patient because of the natural disease course of insulinoma and the necessity of a constant glucose infusion in most cases to prevent the drop of glucose to levels which would be dangerous for the patient.
Changes in the average uptake in tumour tissue and kidneys over time for the examined group of patients are presented in Fig. [3](#Fig3){ref-type="fig"}.Fig. 3Tumour to non-tumour ratio (*T/nT*) and kidney to non-tumour ratio (*K/nT*) over time (means ± SD). The T/nT remained relatively stable over time, while the K/nT decreased visibly (\~ 40 % from the time of injection to 6 h after injection)
As it was estimated for ^111^In-DOTA-exendin-4, the curve for the plasma samples revealed a biexponential clearance (a very fast alpha phase *T*~1/2~ = 12 min and a visibly slower beta phase *T*~1/2~ = 1 h 57 min) \[[@CR9]\]. A low activity concentration was also observed in blood samples, which probably was related to the noneffective centrifugation or the contamination of samples for the gamma counter measurements at the time of their collection. The cumulated activity in the total volume of excreted urine, which was measured up to 6 h post-injection, revealed that the saturation of the curve was not reached. Absorbed dose estimates are shown in Table [2](#Tab2){ref-type="table"}.Table 2Organ absorbed dose estimates for \[Lys^40^(Ahx-HYNIC-^99m^Tc/EDDA)NH~2~\]-exendin-4OrganMean estimated absorbed dose ± SD (mGy/MBq)Kidneys0.1124 ± 0.0526Liver0.0055 ± 0.0009Lungs0.0036 ± 0.0007Spleen0.0052 ± 0.0023Pancreas0.0043 ± 0.0017Intestine0.0066 ± 0.0026Red marrow0.0052 ± 0.0010Thyroid0.0001 ± 0.0000Muscles0.0009 ± 0.0003Ovaries0.0005 ± 0.0004Total body0.0027 ± 0.0010Effective dose (mSv/MBq)0.0024 ± 0.0003
Discussion {#Sec8}
==========
This paper presents our first experience with \[Lys^40^(Ahx-HYNIC-^99m^Tc/EDDA)NH~2~\]-exendin-4 in the detection of the pancreatic neuroendocrine tumour secreting insulin, insulinoma. To the best of our knowledge, it is the first application of this new compound in clinical practice. For most of the patients involved in this study, examination with this novel biomarker was the last hope for the localization of the insulinoma foci and further curative surgical treatment. We were able to detect tumour lesions (benign insulinoma foci) with an excellent sensitivity and specificity. The quality of images was evaluated as high by referring physicians.
Malignant insulinomas frequently lack GLP-1 receptors (positive GLP-1 receptor imaging only in 36 % of cases) but, in contrast to benign insulinomas, more often express SST2 receptors (positive SRS scan in 73 % of cases) \[[@CR19]\]. This finding is very interesting because it shows the ability of GLP-1 receptor imaging for predicting malignant or benign status of an insulinoma lesion. There were only two patients in our group with malignant insulinoma in whom GLP-1 scintigraphy was performed. In one of them, GLP-1 receptor imaging confirmed the local recurrence of malignant insulinoma, but no uptake of \[Lys^40^(Ahx-HYNIC-^99m^Tc/EDDA)NH~2~\]-exendin-4 was observed in liver metastases. This case might suggest the different biology of primary lesions and their metastases in terms of the expression of GLP-1 receptors. This phenomenon is well known for somatostatin receptors in neuroendocrine tumours \[[@CR20]\].
According to Wild et al., hypoglycaemia after ^111^In-DTPA-exendin-4 (10 ± 2 μg) injection is observed in patients with malignant insulinoma \[[@CR19]\]. The decrease of blood glucose level was more often seen in cases of patients with GLP-1 receptor-positive lesions than in those with negative ones. The mechanism of this phenomenon is unclear, as is the mechanism of possibly more severe hypoglycaemia occurring after cold somatostatin analogue injection in patients with malignant insulinoma. It is probably caused by the inhibitory effect of the somatostatin analogue on glucagon excretion \[[@CR21]\]. In the cases of insulinoma patients enrolled in our study it was difficult to assess the changes in glucose levels related only to the labelled GLP-1 injection because some of them received constant glucose infusion to prevent hypoglycaemia. Based on our experience, we believe that establishing this fact is a very difficult task, due to the natural disease course of insulinoma. Because of significant differences in the patients' conditions before the examination, each patient needed to be managed individually.
Nesidioblastosis is another clinical challenge. The pathogenesis of this disease still remains unclear. An increased expression of insulin-like growth factor 2 receptor (IGF2R), insulin-like growth factor 1 receptor alpha (IGF1Rα) and transforming growth factor beta receptor type 3 (TGFβR3) was observed in samples containing pancreatic islets from patients suffering from nesidioblastosis. It seems that certain growth factors and growth factor receptors may significantly contribute to the development of nesidioblastosis \[[@CR22]\]. Reubi et al. evaluated in vitro the GLP-1 receptor status of pancreatic tissues collected from patients suffering from hyperinsulinaemic hypoglycaemia, specifically after gastric bypass surgery. In this study the overexpression of the GLP-1 receptor was not found on the pancreatic islet cell surface. Thus, according to the authors, patients with post-gastric bypass hyperinsulinaemic hypoglycaemia should not be examined with GLP-1 receptor imaging in vivo \[[@CR23]\]. However, in clinical practice it should be expected that insulinoma may coexist with nesidioblastosis. Recurrent hypoglycaemias even after surgical resection of 70 % of the pancreas in patients with nesidioblastosis may suggest the presence of insulinoma \[[@CR24]\]. In the patient with nesidioblastosis examined with labelled GLP-1 receptor agonist in our department, accumulation of the tracer was observed in the pancreatic area. In this patient histopathological examination confirmed coexistence of insulinoma and nesidioblastosis. Moreover, in one patient suspected of having insulinoma (38-year-old woman), histopathological examination also revealed a microadenoma of the pancreatic head (in the localization of the focal uptake of GLP-1) and coexisting nesidioblastosis. These findings support the statement that, despite similar symptoms resulting from hyperinsulinaemia and hypoglycaemia, the pathogenesis of nesidioblastosis may vary significantly depending on the patient's clinical history and may cause lots of problems for clinicians. The other conclusion is that because of difficulties in the management of patients with nesidioblastosis and the potential coexistence of insulinoma and nesidioblastosis, physicians should include GLP-1 receptor imaging as a diagnostic possibility for these patients.
It should be highlighted that the use of image fusion was mandatory for proper diagnosis in all cases. In some cases the fusion of GLP-1 receptor scintigraphy and CT/MRI enables visualization of small lesions not reported in CT/MRI before. In our group of patients in two cases a focal lesion was found on repeat CT/MRI scans following a positive result of GLP-1 receptor scintigraphy.
Despite the fact that a high kidney uptake was visible in all cases, the image quality was excellent and tumour foci were very well visible with a high T/nT ratio. Background uptake over the whole body was low with the exception of the kidneys, which were strongly labelled owing to renal excretion of the compound. The same finding was reported for ^111^In-DOTA-exendin-4. Therefore, as it was suggested for GLP-1 receptor imaging with ^111^In, additional delayed images should be performed in patients with negative early scans \[[@CR10]\]. For optimizing the acquisition protocol, Gelofusine, fragmented albumin or other substances should be used for blocking tracer uptake by proximal kidney tubules \[[@CR11], [@CR25]\].
Recently, PET imaging radiotracers labelled with ^68^Ga and ^18^F, targeting GLP-1 receptors, have also been developed {\[Lys^40^(Ahx-DOTA-^68^Ga)NH~2~\]exendin-4, \[Lys^40^(^68^Ga-DOTA)\]exendin-3 and ^18^F-FBEM-EM3106B} \[[@CR11], [@CR26], [@CR27]\]. The potential of novel biomarkers based on GLP-1 analogues suitable for PET examinations has been tested only in animal models until now. Because of different advantages and disadvantages of all GLP-1 targeting tracers (labelled with ^99m^Tc, ^111^In, ^68^Ga or ^18^F) today it is a matter of argument which of them will be the best for detecting insulinomas in humans. Unquestionably, the GLP-1 receptor tracer labelled with ^99m^Tc is characterized by the general availability of the isotope, lower radiation exposure to patients and staff, and potential usefulness of this compound for intraoperative localization of insulinoma foci using a gamma probe \[[@CR11]\]. It is possible that all of them (no matter which radionuclide the GLP-1 analogue is conjugated with) will present the same very high sensitivity and specificity related to an extremely high density of GLP-1 receptors on benign insulinoma cell surface. Nowadays there is a lack of comparative studies with the use of different GLP-1 radiopharmaceuticals injected into the same patients. In the near future such examinations (based on exendin-4 labelled with ^99m^Tc, ^68^Ga and ^111^In) are planned. However, taking into account the very high sensitivity of this method the registration by national authorities of the labelled compound based on GLP-1 analogues might be considered in cases of patients with suspected benign insulinoma in the nearest future. Moreover, it is well known that there are other tumours overexpressing GLP-1 receptors; therefore, the next natural step should be the use of labelled GLP-1 analogues in cases of tumours other than benign insulinoma. In the clinical trial conducted in our department this tracer was used successfully for visualization of medullary thyroid cancer and phaeochromocytoma (unpublished data).
In this paper the biokinetics and initial dose assessment of \[Lys^40^(Ahx-HYNIC-^99m^Tc/EDDA)NH~2~\]-exendin-4 for SPECT/CT studies were also addressed. The average effective dose is comparable with the radiation dose to patients after SRS performed with ^99m^Tc-EDDA/HYNIC-Tyr^3^-octreotide \[[@CR28], [@CR29]\]. Wild et al. evaluated the pharmacokinetics and radiation dose to patients after injection of \[Lys^40^(Ahx-HYNIC-^99m^Tc/EDDA)NH~2~\]-exendin-4 in an animal study \[[@CR11]\]. Comparing the biokinetics, our study showed a faster retention of the tracer in human organs assumed to be radiation sources. For organ radiation doses and the effective dose, similar levels of these values were calculated, except for the kidneys (an about twofold higher absorbed dose was calculated in our study). These data should be considered only as the basis for further evaluation of radiation safety of this radiopharmaceutical.
Conclusion {#Sec9}
----------
GLP-1 receptor scintigraphy is a promising diagnostic tool for patients with clinical symptoms of insulinoma. It enables the localization of even very small tumours with excellent sensitivity and specificity, and proper imaging is the most important step for successful surgery and complete patient recovery. Despite some difficulties to overcome, the new compound seems to be an effective new tracer for clinical practice and appears to be safe for the patient.
This study was supported by the Polish Ministry of Science within Research Project N N402 445039 and the COST Action BM0607.
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#S0001}
============
Over the past 5 years, single-port techniques, such as laparoendoscopic single-site surgery (LESS), have been applied to perform a wide range of operations, both in adults and in children, including cancer resections and live donor nephrectomies \[[@CIT0001]--[@CIT0003]\].
There are numerous applications of LESS for urological indications involving either reconstructive or ablative procedures \[[@CIT0004], [@CIT0005]\]. Simple prostatectomies as well as diverticulectomies were performed transvesically via a single port placed intraperitoneally through an intra-umbilical incision \[[@CIT0004], [@CIT0006]\]. A few reports describe the LESS technique performed percutaneously, directly through the bladder wall for foreign body removal, bladder cuff excision or adenomectomy \[[@CIT0007]--[@CIT0010]\]. Because of the evolution of laparoscopy over recent years, however, a number of laparoscopic approaches for treating vesicovaginal fistula (VVF) have been published \[[@CIT0011], [@CIT0012]\]. Nevertheless, we have found only one report on LESS repair of VVF \[[@CIT0013]\].
The aim of the study was to o present our experience of percutaneous transvesical treatment of a VVF using the TriPort+ device inserted directly into the bladder. To the best of our knowledge, this is the first report of a percutaneous transvesical LESS approach for VVF repair.
Case report {#S0002}
===========
A 72-year-old woman presenting with continuous urinary leakage per vagina was admitted to our centre in August 2011. In May 2011, she had undergone hysterectomy because of a benign disease.
The exams demonstrated a 3-mm wide fistula between the bladder trigone and the upper part of vaginal vault ([Figure 1](#F0001){ref-type="fig"}). Microbiological examinations indicated typical chronic infection with *Escherichia coli*. All other laboratory parameters were within the normal range.
{#F0001}
Because of the localization of the VVF in the vaginal apex, we initially considered the laparoscopic approach. However, as our team had carried out several LESS transvesical operations over previous years, we decided to use this technique for VVF repair with a single port introduced directly into the bladder.
The patient was placed in the lithotomy position. Subsequently, both ureters, as well as the fistulous tract, were catheterized ([Figure 2 A](#F0002){ref-type="fig"}). We introduced an 18 F Foley catheter into the vagina and filled its balloon with water up to 30 ml. This manoeuvre plugged the vagina, and decreased gas leaks during the procedure.
{#F0002}
A 1.5-cm longitudinal skin incision was made 2 cm superior to the pubic symphysis. We used two stay sutures to facilitate the insertion of the port. The introducer with the inner ring of a TriPort+ access system (Olympus Winter&IBE GMBH, Hamburg, Germany) was inserted directly into the bladder via the skin incision under cystoscopic control. The rings of the TriPort+ were fixed to the abdominal wall area. The bladder was filled with carbon dioxide up to a pressure of 14 mm Hg.
We used a 10-mm, rigid, 30° videolaparoscope (Olympus Europa GmbH, Hamburg, Germany), which was introduced through the 10-mm channel of the TriPort+. The three 5-mm working channels were used as follows: 1) for a 5-mm standard laparoscopic dissector in the left hand, and 2) for a 5-mm monopolar rigid hook electrode or a 5-mm rigid needle-driver (Karl Storz, Tuttlingen, Germany) in the right hand. When needed, we introduced the suction tube through the third 5-mm channel of the TriPort+ or through the urethra.
The visualization inside the bladder was sufficient to recognize and control all anatomical structures. We separated the fistula from the bladder mucosa by a circular cut using a hook electrode ([Figures 2 B](#F0002){ref-type="fig"}, [C](#F0002){ref-type="fig"}). The superficial rim of the fistulous tract was resected, and either the vaginal wall or the bladder defect was closed tightly with an absorbable 3/0 running suture (The V-Loc™ 90 Absorbable Wound Closure Device, Covidien, Norwalk, CT, USA) ([Figures 2 D](#F0002){ref-type="fig"}, [E](#F0002){ref-type="fig"}). The integrity of the bladder was confirmed by filling it with 200 ml of saline. Betadine-soaked roller gauze was inserted into the vagina for one day. The skin incision was sutured with two stitches. The procedure was completed successfully with no extra port insertion. The operative time was 170 min, and the blood loss was insignificant. The postoperative period was uneventful. On the fifth postoperative day, ureteric catheters were removed and the patient was discharged home. She presented no vaginal leakage on discharge. An 18 F Foley catheter was retained for 2 weeks. Quinolone oral intake was administered for 3 weeks and pelvic rest was recommended for 2 months.
Five weeks after surgery, diagnostic scans (urethrocystography) revealed no presence of VVF ([Figure 3](#F0003){ref-type="fig"}). During a 4-month follow-up, the patient remained continent, and laboratory examination results were all within the normal range.
{#F0003}
Discussion {#S0003}
==========
To date, there is no consensus on what approach is the best for VVF repair. At present, laparoscopy and single-port procedures tend to replace open surgery, with comparable results \[[@CIT0011]--[@CIT0014]\]. These procedures mostly follow the O\'Conor technique \[[@CIT0015]\]. However, simpler techniques are also available \[[@CIT0016]\].
We also considered the extravesical approach for LESS repair of VVF, according to Abdel-Karim *et al*. \[[@CIT0013]\], but the abdominal access often requires extensive and time-consuming adhesiolysis, and, because we were familiar with different transvesical LESS procedures, we decided to use this approach in our patient.
After adequate dissection of the bladder from the vaginal wall, both openings of the fistula were closed with separate running V-Loc sutures, and a watertight closure was achieved. The advantages of our method are (i) the reduction of the number of ports, (ii) avoiding bivalving the bladder, and (iii) the possibility to use the transurethral access to facilitate intravesical dissection or suturing. The use of a barbed suture (V-Loc) also eliminates the need for tying knots and the number of instruments used. This also reduces the operative time.
The disadvantages are similar to those in other LESS procedures. This approach is probably not applicable for larger and complicated fistulae.
The authors realize that one case is insufficient to define more general conclusions. Nevertheless, based on our previous experience in the field on laparoscopy and transvesical laparoendoscopy, we find this approach to be a logical, minimally invasive, and simple access for either ablative or reconstructive procedures in the bladder.
Our positive experience in this approach shows that percutaneous transvesical LESS vesicovaginal fistula repair is a feasible, effective and safe procedure, but a larger number of patients is needed to thoroughly evaluate this treatment.
| {
"pile_set_name": "PubMed Central"
} |
Image in medicine {#S0001}
=================
A 64-year-old woman with past medical history of diabetes mellitus type 1, referred to our dermatology department for a 20-month nonpruritic erythematous ulcerated lesions of the two legs. Upon physical examination, a large oval erythematoviolaceous plaque covering more than the half of the right pretibial region was observed. The border was slightly elevated, indurated and infiltrative, the center was scattered by small necrotic, crusted ulcers and by small atrophic areas. Similar smaller plaques occurred on the left leg (A). No other systemic abnormalities were detected particularly regional lymphadenopathy. There was no history of trauma or insect bite. She denied having fever or any other systemic symptoms. Blood count, routine biochemical tests, urine analysis and chest radiography were normal except hyperglycemia at 11.8 mmol/l. Bacteriological tests were negative. Skin biopsy showed a dense lymphocytic infiltrate throughout the entire dermis admixed with multiple granuloma without necrosis. At higher magnification, epitheloid cells, histiocytes and multinucleated giant cells were observed within the granulomatous zones. These latter were outlined by a dense infiltrate of lymphocytes (B) Basophilic regular structures in the intracellular spaces corresponding to amastigotes belonging to *leishmania* were determined with the Giemsa stain (C). The patient was treated intramuscularly with 20 mg/Kg/day systemic meglumine antimoniate for 15 days. The clinical course was characterized by the healing of her ulcer but onset of some atrophic scars.
{#F0001}
| {
"pile_set_name": "PubMed Central"
} |
{.439}
{.440}
{.441}
{.442}
{.443}
{.444}
{.445}
| {
"pile_set_name": "PubMed Central"
} |
What Is Known About This Topic {#section1-2333393618816780}
==============================
- The practice of homecare nursing is changing due to demographic, health, and political developments.
- The concrete practice of homecare nursing is under researched, and we have little knowledge of the details of homecare nursing.
What This Article Adds {#section2-2333393618816780}
======================
- Homecare nursing practices take new forms in response to contextual changes, and new practices are becoming prominent.
- Homecare nurses' practice is becoming increasingly advanced but is complicated by the lack of knowledge transfer from hospitals to homecare and to the challenges of working alone in the context of a private home.
- Homecare nurses are frequently involved in negotiating care level across and within municipalities and specialist health care and, consequently, what kind of care the patient will receive.
- Patient trajectories are more complex than before, and homecare nurses play a key role in coordinating care among interdependent actors.
Introduction {#section3-2333393618816780}
============
During the last decades, the work of homecare nurses in Norway and other Western countries has undergone several changes, including changes in aging populations, changes in disease patterns, the decentralization of health care, nursing recruitment crises and the scarcity of public resources ([@bibr12-2333393618816780]; [@bibr17-2333393618816780]; [@bibr23-2333393618816780]; [@bibr46-2333393618816780]). However, few scholars have analyzed what these changes means for homecare nurses' work. Few studies on the performance of homecare nursing exist; therefore, the work of homecare nurses, specifically the less clinical parts of their work, has been largely invisible to outsiders ([@bibr5-2333393618816780]; [@bibr29-2333393618816780]; [@bibr39-2333393618816780]). This article aims to describe and discuss aspects of homecare nurses' work, with a specific focus on nurses "organising work" ([@bibr4-2333393618816780]). Based on empirical examples, we outline three responses of homecare nurses to changes in their work contexts, the specific organizational challenges they address and the skills that are increasingly needed to perform these activities. The article draws on material from participant observation and interviews with homecare nurses in two Norwegian studies conducted between 2011 and 2016.
Literature Review {#section4-2333393618816780}
-----------------
Regardless of where it is practiced, nursing is a multifaceted role in the sense that nurses must meet multiple professional demands and perform a wide range of tasks ([@bibr5-2333393618816780]; [@bibr28-2333393618816780]). summarizing the literature on nurses' tasks, interaction with patients, "bedside care," is most often described as the core of nursing work ([@bibr2-2333393618816780]; [@bibr8-2333393618816780]; [@bibr37-2333393618816780]; [@bibr44-2333393618816780]). Interacting with patients involves assessing the patient's situation, which includes diagnosing illnesses and planning and implementing proper interventions. Moreover, nurses must apply strategies and practices to improve care outcomes, such as implementing evidence-based practices, fulfill a professional commitment (the duty of care), and have an ethnical awareness of nursing practices ([@bibr37-2333393618816780]).
Caring for the patient also includes conducting what [@bibr4-2333393618816780], [@bibr5-2333393618816780]) calls "organizing work," which essentially entails planning and executing the individual patients' entire care trajectory. Organising work involves collaborating and coordinating with providers on different organizational levels and units and communicating and collaborating with family carers ([@bibr26-2333393618816780]), who in particular play an important role for home-dwelling patients. Organizing work also entails carrying out information management tasks, such as record keeping and information sharing, as well as "care space governance," which includes maintaining a focus on quality, safety, and cost effectiveness ([@bibr37-2333393618816780]).
The general features of nursing work are relevant as a background for our analysis of homecare nursing. However, homecare nursing has specific features that make it particularly challenging. First, homecare nursing takes place within the patient's home; thus, the home becomes both a workspace and a private, domestic space, which sets particular conditions under which care is provided ([@bibr15-2333393618816780]). Homecare nurses must adapt to patients' various values and lifestyles and establish trust and provide care in highly individualized contexts ([@bibr26-2333393618816780]; [@bibr42-2333393618816780]). Providing care in the patient's home affects social interaction and requires a balancing effort between providing care and maintaining the patient's autonomy, self-determination, and dignity (cf. [@bibr25-2333393618816780]). Providing care in the patient's home also creates a sense of intimacy, and nurses invest in emotional labor and act with sentiment despite a lack of financial compensation ([@bibr27-2333393618816780]). Second, homecare nurses mostly work alone. Although being autonomous and independent may allow homecare nurses to effectively provide individualized care to patients, homecare nurses work in isolation ([@bibr42-2333393618816780]) and have little access to professional support. Third, homecare nurses work with increasingly advanced cases ([@bibr11-2333393618816780]), which means that homecare nurses frequently need to contact other health care workers both for obtaining professional advice and for coordinating services. In particular, the discharge process from hospital to home has proven to be challenging, as it leaves homecare nurses with an insufficient understanding of the patient's situation ([@bibr18-2333393618816780]; [@bibr26-2333393618816780]; [@bibr40-2333393618816780]).
In the research literature, the therapeutic relationship with patients, whether it is exercising bedside care for hospital-bound patients or providing support at point of care for home-dwelling patients, is frequently portrayed as the core act in exercising nursing. Different types of organizing work are often depicted as inferior to direct patient contact, taking time away from care ([@bibr4-2333393618816780], [@bibr5-2333393618816780]; [@bibr9-2333393618816780]). [@bibr12-2333393618816780], however, describes these organizing efforts as a prerequisite for nurses' provision of safe and high-quality health care. In her study of homecare, she found that nurses spent much time creating and maintaining a net of assistance and support around the patients to help them to continue living in their own home. To build such nets around the patients requires competencies, skills, and formal systems for connecting organizations, providers, and family carers. Direct interaction with patients and organizing their care may be seen as two sides of the same coin, and both are necessary for providing good nursing.
To explore the enactment of homecare nursing and the intertwinement of homecare practice and contextual changes, we draw on insights from Translation Mobilization Theory (TMT). TMT ([@bibr7-2333393618816780]) explains the mechanisms through social actions are mobilized, the relationship between these practices, and the institutional contexts in which they are accomplished (p. 4). TMT draws on insights from ethnographic work on the social organization of health care work ([@bibr1-2333393618816780], [@bibr2-2333393618816780], [@bibr3-2333393618816780]) and on Normalization Process Theory (NPT; [@bibr33-2333393618816780]; [@bibr34-2333393618816780]), which aims to explain how implementation takes place and how practices become integrated into a wider sociotechnical context.
TMT has three core components: (a) *Project*: a sociotechnical ensemble of institutionally sanctioned strategic activity mobilized across a distributed action field. In our case, this means the activity of homecare nursing. (b) *Strategic action field*: the institutional context in which projects emerge and are progressed and which provide the normative and relational frame for collective action. Such a field further entails organizing logics, structures, materials/technologies, and interpretative repertoires. The strategic field in our case is the Norwegian health care system, health policies, the staffing of homecare, and the tools and technologies that homecare nurses use in their work. (c) *Mechanisms of mobilization and institutionalization*: processes through which agents operating within a strategic action field mobilize projects, drive action, and perform institutions. These processes are of particular relevance in our analysis of how homecare nurses respond to changes in their work context. Such processes include the following:
- Object formation: practices that fabricate and configure the objects of knowledge and practice and enroll them into an actor network.
- Articulation work: practices that assemble and align diverse actors (people, knowledge, materials, etc.) and through which object trajectories are mobilized (cf. [@bibr43-2333393618816780]).
- Translation: practices that enable objects to be shared and differing viewpoints, local contingencies, and multiple interests to be accommodated to enable concerted action.
- Reflexive monitoring: practices through which actors evaluate a field of action to generate situational awareness of project trajectories.
- Sensemaking: practices through which actors order, construct, and mobilize projects and enact institutions ([@bibr7-2333393618816780] p. 8).
The concepts are useful for understanding and articulating how homecare nurses make sense of their work and what strategies they apply when the context in which they work is constantly changing. The concepts link actors and actions with the overall structure of health care, emphasizing the relationship between practices and the institutional context, which is highly relevant in our analysis of homecare nursing in a changing health care landscape.
Setting: the Strategic Field of Norwegian Homecare {#section5-2333393618816780}
--------------------------------------------------
Community health care service in Norway is governed and financed by the municipalities, and municipalities are obliged by law to ensure that its inhabitants receive "necessary health and care services" that are well coordinated and safe ([@bibr30-2333393618816780]). Municipal care is divided into institutional care, such as nursing homes, and health care provided in the patient's home, such as homecare nursing. Overall, the main objective of homecare nursing is to ensure that home-dwelling patients' fundamental needs are met. An increasing number of Norwegian inhabitants are receiving homecare services, and the number is expected to increase ([@bibr38-2333393618816780]). The patients vary in terms of age. Circa 60% of them are above pension age (67 years). The largest patient group is between 80 and 89 years (31% of the patient population). Patients who are older than 81 to 89 years are likely to receive institutionalized care. Longitudinal statistics also indicate that the number of younger patients (which refers to patients who are below pension age) is increasing ([@bibr38-2333393618816780]). Many of these patients have chronic illness and are multimorbid, and some of these patients are physically disabled and consequently need advanced clinical care, including a multitude of medications and/or other care measures. In 2012, the Norwegian government implemented a health care reform (known as the Coordination Reform), in which responsibilities were transferred from the specialist health care service (hospitals) to the municipalities ([@bibr22-2333393618816780]). The reform encouraged early discharge from hospital; as a consequence, homecare patients are sicker than before ([@bibr41-2333393618816780]). Moreover, due to their complex conditions, homecare patients need contact with a range of other health care providers, such as general practitioners (GPs), out-patient clinics, rehabilitation institutions, and physiotherapists, in addition to homecare nurses.
The developments in Norwegian homecare are not unique to the Norwegian context, but represent general trends in Western health care. Thus, our study's findings should have relevance for understanding homecare practices in other countries.
Method {#section6-2333393618816780}
======
Study Design {#section7-2333393618816780}
------------
This article is based on data collected for two of our previous research studies. In Study 1, Bridging the Information Gap in Patient Transition (BIG; 2010--2014), we studied the interaction between homecare nurses, GPs, and hospital staff, as well as how homecare nurses used an electronic messaging system as a collaborative tool. Study 2, Interaction via ICT (SIKT) (2013--2016), was a follow-up to the first study, but focused more specifically on the use of collaboration to treat patients in need of both homecare and hospital treatment.
Methods and Setting Study 1 {#section8-2333393618816780}
---------------------------
Study 1 was conducted in three municipalities in Norway, which varied in size. We combined observations of homecare nursing work and semi-structured interviews, and data collection took place between May 2011 and January 2012. Access to and permission for observing and interviewing were granted by the head of the homecare services in the municipalities. After permission was obtained, a contact person approached and informed eligible nurses who met the inclusion criteria about the study. Nurses were included in the study if they had been in their position for more than 6 months; this ensured that the participating nurses had some experience. The contact person also organized the observations and interviews.
The observations covered 97 home visits, which resulted in a total of 85 hr of observation. The first author and the third author conducted the observations. By providing rich data on events in context, observations uncover the enactment of a practice or a project ([@bibr45-2333393618816780]). The observations were mostly organized in the same way across the three municipalities. We started the day with a morning meeting during which we observed all of the staff together, updated each other on the patients, coordinated work, and delegated tasks before heading out to the patients' homes. Together with each assigned nurse, we visited patients in their homes. During the visits, we asked for clarifications of what took place or discussed situations that occurred with the nurses we followed. Sometimes the patients or next of kin were involved in the discussions. After the home visits, we returned with the nurses to the homecare office. Each day ended with conducting an interview with the nurse we had observed and sometimes with other homecare nurses. Occasionally we would also observe nurses involved in office work; normally, one experienced nurse would use the day to coordinate the patients' appointments with other professionals, update medication lists, and order new medications. To record the observations, we took field notes. Our goal was to produce field notes that described the situations as accurately as possible.
The interviews included 43 persons and were conducted by the first author, the third author, and a PhD student. There were 23 homecare nurses, 11 general practitioners, five medical secretaries, and four project managers for the e-communication implementation. The interviews with homecare nurses would typically begin with an open-ended question: tell us about a typical day at your work. Concrete themes addressed how the nurses assessed the patients' needs, if they had experienced changes in their work over time---and, if so, what kind of changes---and how collaboration was experienced. All but one interview were conducted by two interviewers; one of the researchers led the interview and ensured that all themes in the interview guide were covered. The other researcher took notes and asked clarifying questions.
Methods and Setting Study 2 {#section9-2333393618816780}
---------------------------
In Study 2, we only conducted interviews, which took place in 2014. The setting was three municipalities (one of which was the same as in Study 1) and one hospital. Recruitment of interviewees followed the same procedure as in Study 1. In total, 41 persons were interviewed. These included 24 interviewees in homecare (the rest worked at the hospital): 12 worked as homecare nurses, whereas 12 worked as case handlers or other administrative staff. The interviews were conducted by the first author, the third author, and a researcher. As in Study 1, interviews were carried out by two researchers together.
Most municipalities today use a purchaser-provider-based system, which means that the case handlers in a municipality "buy" services from the providers (e.g., homecare) in the same municipality. The case handlers are thus the ones who formally decide what kind of municipal health and social care the inhabitants receive; consequently, case handlers have much knowledge of the needs of the patients and the services offered by the municipality. The purchaser-provider model requires good collaboration between purchasers and providers to continuously adapt services to the patients' needs.
Analysis {#section10-2333393618816780}
--------
The analytical themes presented in the results section illustrate the interplay between the overall development in the health care system and the practical accomplishment of homecare nursing. The findings presented in this article reflect only parts of the data collected in the studies, and other publications highlight other aspects of homecare in detail (see, [@bibr35-2333393618816780]; [@bibr36-2333393618816780]). We have selected data that allowed us to uncover and analyze *stories of change* to present in-depth insight regarding homecare nursing practices. Both in formal interviews and in conversations that occurred during the observations, nurses shared their experiences of the changed nature of homecare with us. The concept of "change" was therefore a guiding concept when going through the data material. After conducting discussions among the authors, three overall themes were identified. When presenting the themes, we highlight typical situations for each of the themes and give examples of caring/practice situations based on the field notes as well as excerpts from the interviews. Interviewees have been given pseudonyms.
Ethics {#section11-2333393618816780}
------
Both projects were reported to the Norwegian Centre for Research Data (Projects 26,230/2011 and 37399/2014), and they were approved by the management in the municipalities. All the homecare nurses gave their written informed consent. Because observations were made in the patients' home, all patients were asked in advance to approve that a researcher would follow the nurse into the patients' home. All patients signed an agreement that granted us permission to observe the work of nurses in their homes.
Results {#section12-2333393618816780}
=======
In this section, we present the study's results, which are organized in three themes. Each theme represents activities that homecare nurses increasingly carry out:
- Homecare nurses are frequently involved in negotiating care level and, consequently, what kind of care the patient will receive. This theme shows how nurses try to understand the patient's needs, communicate those needs and match patients' needs to available services.
- Homecare nurses' practice has become increasingly advanced, and homecare nurses now perform procedures that previously took place only in hospitals. This theme points to the importance of systems for knowledge transfer from hospitals to homecare and to the challenges of working alone in the context of a private home.
- Patient trajectories are more complex than before, and homecare nurses play an important role in coordinating care between interdependent actors. This theme depicts the major role homecare nurses play in being mediators between actors who lack good systems for information transfer and communication.
Negotiating Care Level and Services {#section13-2333393618816780}
-----------------------------------
A main finding from our studies was that homecare nurses take active parts in negotiations regarding the patients' service level, both through intra-municipal negotiations and in communication with the hospitals concerning the transfer of patients from hospital to home. Vignette 1 shows an example of homecare nurse Tina's first visit to a patient recently discharged from a rehabilitation stay and how she assesses the patient's care needs:"***Vignette 1*:**After lunch, we go to Mrs. Jackson, a woman in her eighties who lives with her husband in an elegant two-storey house. Mrs. Jackson came home from a four-week rehabilitation stay in a nursing home only ten minutes ago. Tina is only slightly oriented about how the stay has been and the status of Mrs. Jackson's health situation. I get the impression that the visit is mainly for establishing how Mrs. Jackson is feeling, what kind of help from community care she has been granted, and what she considers her needs to be. We enter the house and are shown into the living room. Mrs. Jackson suffers from mild dementia, but seems to be in a good shape otherwise. She is in a very good mood. Her husband is cognitively sharp, but has severe back problems and can only walk very slowly. Mr. Jackson and the nurse have a long conversation about the needs of Mrs. Jackson. Mrs Jackson sits at the table, but does not participate in the conversation. One problem is that Mr. Jackson finds it difficult to take his wife to the GP when that is needed, and he wants the GP to come to their home. They agree that Tina will ask the GP if he can come and see Mrs. Jackson at home. They also agree upon an increase in the number of homecare visits. From having visits twice a day, Mrs. Jackson will now have visits four times a day. We leave the house and aim for the GP's office to ask if he can see Mrs. Jackson at home. (Observation B3)"
In the example, Tina and Mrs. Jackson's husband discuss and agree upon the number of visits per day needed for Mrs. Jackson. The nurse must also address the couple's difficulties of going to the GP. The nurse's work can be seen as a practice of "object formation," the configuration of "the case" of Mrs. Jackson ([@bibr7-2333393618816780]). By investigating both the needs of Mrs. Jackson's and her husband, the nurse tries to determine what services Mrs. Jackson needs, how often she needs help, what other actors need to be enrolled in the patient's care network, and how to connect the various actors.
Our studies also revealed that homecare nurses play a significant role in the shaping of trajectories by informing caseworkers (purchasers) about the patients' situation, which is another input for object formation. This happened on a regular basis through a function in the electronic patient record system, through which providers can notify case handlers about changes in patients' needs. It also happened when patients were about to be discharged and case handlers needed more information about a patient. Our studies indicate that homecare nurses are more influential in discussions with the caseworkers than with hospital staff. The negotiations regarding transfer from hospital to home can be particularly challenging. In these conversations, the hospital's need for discharge as soon as the patient is done with specialist treatment meets community care's need for preparations, including getting the necessary equipment in place in the patient's home (e.g., a hospital bed). It is important to note that legally it is the hospital's responsibility to decide when the patient is ready for discharge. However, for patients in need of homecare services, the service quality can be substantially lower if homecare nurses have not been able to prepare the homecoming. Therefore, discussions regarding patients' needs and an agreement regarding the time of discharge is therefore necessary ([@bibr21-2333393618816780]).
Frequently, negotiations were nevertheless not an option. One homecare nurse said the following: "We had a patient that we requested to stay in the hospital a little bit longer. And they \[the hospital\] insisted that he should be discharged. So, we think there is more room for negotiation in the other departments than in the XX department." We do not know the hospital's motivation for insisting on discharge, but Norwegian hospitals are increasingly experiencing productivity pressure, so they aim to discharge patients early. Interestingly, we also heard examples of the opposite: homecare arranged a service in the patient's home, but the hospital judged it as insufficient and refused to discharge the patient:"It seems like there is most tumults with patients with psychiatric diagnoses when it comes to discharges. And if there is one time where we were close to having a quarrel with the hospital, it was related to a psychiatric patient that we thought could be discharged to home---because the patient was cared for by one \[nurse\] during the day and one \[nurse\] during the evening. And community care nursing during the night. The hospital did not think it was justifiable. But I am unsure whether we managed to explain how much services the patient actually had at home. But they refused to discharge him on that Friday, and we were sitting here, fully staffed. (Nurse A, municipality B)"
However, interviewees also told several stories of more peaceful negotiations over services, in which the hospital would postpone discharge so that homecare nurses could properly arrange for the patient to come home. One nurse said, "We do some negotiations and reach an agreement. The hospital will then discharge the patient the following day instead of today."
Performing Increasingly Advanced Clinical Practices {#section14-2333393618816780}
---------------------------------------------------
Another main finding in our studies---and perhaps the most visible one---was that homecare nurses care for much more ill patients than they did earlier and consequently must perform more advanced clinical assessments and procedures. This can be related to the public efforts of keeping patients out of hospitals and providing care in the municipality. Because tasks that traditionally were performed by specialists in hospitals are now carried out by homecare nurses, the significance of knowledge transfer from hospitals to the municipalities has increased. Another complicating factor in performing advanced clinical practices is that homecare nursing always takes place in the patient's home and, consequently, work must be conducted within a setting which is not originally designed for exercising care. Vignette 2 describes a situation from one of the observations, in which a nurse must perform a procedure she has not performed before:"***Vignette 2*:**The forth visit this day is to Mrs. Hughes, a widow in her seventies. She was discharged from hospital the day before, and this is the first visit from homecare after she came home. Mrs. Hughes lives alone and is dependent on homecare to assist her with various tasks. The nurse, Carrie, is informed that Mrs. Hughes has had her gallbladder removed and her bile duct cut. Apart from that, she knows little about the patient's current condition and needs. When Carrie talks to Mrs. Jackson, it becomes clear that the patient has a drain tube to remove excess fluid from her wound. She informs Carrie that she needs to empty the drain and measure the amount of fluid in it. The patient explains in detail what the hospital physician told her needs to be done with the drain, but Carrie has not been directly informed by the hospital and seems quite puzzled. It is obvious that Carrie has never seen a drain like this before and is unsure of how to perform the task. Regardless, they go into the bathroom to do the job. (Observation A3)"
The vignette represents a not uncommon example in which a nurse enters a patient's home and must perform relatively advanced tasks for which she is not prepared. As the vignette indicates, situations like this typically occur because of a lack of information from the hospital to homecare nursing about how to follow up on the patient's needs and how to carry out a specific procedure.
The displacement of patients from hospital to municipal care leads to a need for increased competences of various kinds among staff. Such a change necessitates transfer of competence between hospitals and homecare. In other words, procedures traditionally conducted in hospitals must be translated and made doable in private homes. However, knowledge transfer and translation are not always conducted systematically, as reported in several interviews:
Nurse B:
"The communication concerning procedures for wound care is poor. Often, we do not have a procedure at all. Then we call \[the hospital nurses\] and complain a bit. If they cannot help us, we contact the physician who is responsible for prescribing the procedure."
Nurse C:
"We do have quite some experience ourselves, because there are many wounds. But it is not our responsibility to make the procedure. If the patient gets an infection, we are not responsible, because it is the \[hospital\] doctor's job to make the procedure." (Municipality C)
The data showed several examples of poor education of homecare staff and a lack of communication between hospital and homecare when preparing for discharge of patients from hospital to home.
Coordinating Care Between Interdependent Actors {#section15-2333393618816780}
-----------------------------------------------
The third overall finding from our studies was that homecare nurses coordinate patients' services and that the frequency of performing this job increases as the complexity of health care increases. Currently, there is no dedicated professional role for maintaining oversight and managing and coordinating services along the caring trajectory. Throughout data collection, we witnessed numerous examples of homecare nurses having to take on the coordinator role to compensate for the lack of coordinative mechanisms in the health care system. Vignette 3 shows a typical example:"***Vignette 3*:**The second visit this morning goes together with nurse Ellie to Mr. Andrews, a widower in his seventies. Homecare has taken over the responsibility for his medications. Mr. Andrews' blood sugar level is at the moment very unstable, and Ellie's job is to measure his blood sugar level and make sure he takes his medications, which are administered through a multi-dose drug dispensing system. Ellie is also going to cook him breakfast. Ellie and Mr. Andrews start to talk about his medications. He complains about side effects from Glucophage, a medication for his diabetes. Because he experienced nausea after taking the medication, his GP ended the medication. Now it turns out that Mr. Andrews is back on Glucophage, and he is unsure of how this has happened. He thinks perhaps it is homecare staff who have told the GP that he must start up with the medication again. Ellie is unaware of the whole thing, so she looks through the nursing documentation from the previous days, which are stored in the patient record system on her PDA. After scrolling back some days, she informs Mr. Andrews that it his GP that has reinstated Glucopage. She tells Mr. Andrews that she will call the GP afterwards and talk to him about Mr. Andrews' medications. We leave the house, and Ellie documents in the PDA in Mr. Andrews patient record. Ellie will go back to Mr. Andrews at 1 PM to give him a B12 shot, and she hopes she can inform him then. (Observation A1)"
In particular, it is difficult to coordinate patients' medications across multiple health providers (e.g., hospitals, GPs, and community care; [@bibr24-2333393618816780]; [@bibr32-2333393618816780]). In Vignette 3, only community care and the GP are involved, but practices can be even more complex when the hospital is involved as well. We found that the most critical situations concerning medications happened when patients were discharged from the hospital to home, and there were inconsistencies between the medication lists in the hospital, the GP, and homecare. The GP is responsible for home-dwelling patients' medications, but most often GPs are informed about hospitalization later than homecare, and thus in practice it becomes the homecare nurses' job to clear up inconsistencies so that the patient will receive the right medications:"We compare \[the new list from hospital\] with our medication list from before they were admitted to hospital \[. . .\] If we see that it seems utterly wrong, or the hospital has not been aware that the patient was on certain medications, we contact the GP. If everything looks fine, we will make the changes and send a copy to the GP, informing him/her about the changes. (Nurse D, municipality C)"
Another example described how patients were regularly discharged from the hospital without the necessary medications and how the pharmacy was closed at the time homecare came to see them. In these cases, which were described as very frustrating, nurses had to borrow medications from other patients (or from a nearby nursing home) and contact the GP for a prescription on the following day.
In general, homecare nurses described much of their activity as to "tie up the loose ends and collect information." One nurse said that because she works so closely with patients---in the patients' home---she felt that homecare nurses have "essential information" about patients.
The examples show that homecare nurses act as coordinators for ensuring that home-dwelling patients receive their medications and act as intermediaries between GPs, hospitals, and patients. Research has shown that nurses often orchestrate and manage the work of others, specifically doctors' work ([@bibr2-2333393618816780]), and this is visible in our studies as well. Transitions between hospitals and the home challenge the ideal of coherent and continuous trajectories and cause a great need for coordination. However, our studies also showed that coordination between actors *within* the municipality is demanding. Both sorts of coordination are articulated and mobilized as prominent practices in homecare nursing.
Discussion {#section16-2333393618816780}
==========
Our studies have highlighted three areas of homecare work that are increasingly performed. The themes were derived from "stories of change" in our data and show how "organising work" has gained a central place in homecare nursing. We argue that changes in work practice are responses to contextual changes in the overall health care system. Seen in relation to TMT, changes in the strategic action field lead homecare staff to apply various strategies and resources and mobilize new activities to cope with these changes (cf. [@bibr7-2333393618816780]). With the advent of new activities, there is also a need for increased competencies and new systems and routines for supporting nurses in their tasks. The processes that drive actions as outlined in TMT, will consequently increase in intensity in times of rapid changes. For example, will the need for articulation work for aligning actors and mobilizing care trajectories increase and become more challenging when the institutional context is changing. Likewise, will sensemaking---literally practicses of making sense of what is going on in day to day work---become more demanding when, for example, new forms of division of labor are initiated, making practice more complex. In the article, we have shown examples of such practices and strategies, while trying to relate them to the changes taking place in homecare.
Research on nursing practices has shown that nurses are active participants in shaping and managing patients' trajectories ([@bibr5-2333393618816780]; [@bibr29-2333393618816780]; [@bibr42-2333393618816780]; [@bibr43-2333393618816780]). In our studies, we found that homecare nurses negotiate with other providers as well as informal carers regarding the service level (e.g., home or hospital) and the amount of services required in the home. The negotiations take place both through their own initiative and the initiative of others. Our results indicate that homecare nurses can greatly influence the patient's service when at home, but have less influence on the time of discharge even though some hospital departments have good dialogue with homecare services and allow for joint decision making.
The influence of nurses on home-dwelling patients' services found in this study is in line with that found in [@bibr46-2333393618816780] studies. She concluded that even though the purchaser-provider model was introduced to secure a separation between those contracting services and those providing the services, in practice purchasers and care staff collaborate to manage the fluctuating needs and everyday dilemmas of care. Our studies support this finding. We found that homecare nurses regularly adjusted and updated decisions via the electronic patient record system based on their knowledge of the patients' needs. Today, when home-dwelling patients are sicker and have more complex conditions than before, homecare nurses' advice on the patients' service needs are increasingly important. The negotiation over service level can be seen as an act of object formation ([@bibr7-2333393618816780]), in which nurses use their knowledge of the patient and enroll them into an actor network that can provide the most appropriate service level. Our studies showed that negotiations and clarifications of patients' conditions between homecare and hospital staff happened frequently yet were not always sufficient for accommodating the home or providing homecare nurses with sufficient knowledge so patients could be properly cared for. When patients are transferred from the hospital to home, hospital staff need to know that the patients have adequate care services at home, and homecare nurses need information about the patient's status to accommodate care. We found great variety between cases and informants in terms of how well the communication between hospital and homecare functioned. Previous studies have underlined that this interface is a challenging area ([@bibr18-2333393618816780]; [@bibr20-2333393618816780]). In our study, we found that electronic communication tools, like e-messaging, may ease communication and patient care planning (see [@bibr35-2333393618816780]). The increase in rapid discharges from the hospital increases the chance that patients experience more readmissions ([@bibr17-2333393618816780]), which makes care trajectories discontinuous and fragmented. In turn, this means that the transfer and communication between the hospital staff and homecare nurses must happen more frequently and that homecare nurses' role in the transfer process becomes even more important. Research has shown that homecare nurses may experience the transfer process as demanding and that there is room for improvement in the collaboration across hospitals and municipalities ([@bibr26-2333393618816780]), which is also supported by our findings.
Nurses increasingly take on advanced clinical work that was previously performed in hospitals. This development is encouraged by the health authorities, and represents a form of task shifting that has occurred in other countries ([@bibr14-2333393618816780]). The question is how the development is accommodated in terms of educating homecare staff and knowledge transfer from hospitals to municipalities. Studies indicate that there is a discrepancy between the advanced competence expected in policy documents and the actual competence nurses possess ([@bibr10-2333393618816780]; [@bibr17-2333393618816780]). Our studies add knowledge to this gap. We found that competence varied among nurses, and there were several occasions in which nurses called for more training and information about a specific procedure. In particular, the nurses in our studies called for better communication and knowledge transfer from specialist health care to community care. The latter is obliged by law as a duty of the specialist health care system (Helse- og omsorgsdepartementet, 1999). [@bibr10-2333393618816780] concluded their study by stating that there is a general lack of opportunities for competence development in the health and care sector. Because nurses must perform increasingly advanced tasks, it is worrisome if nurses are not able to increase their competence, which, in turn, challenges nurses' professionalism. The challenges of performing advanced care work is reinforced by the context in which homecare is performed. A private home is not arranged for performing advanced care, which can make it difficult to prepare sterile areas and have enough space to perform procedures (cf. [@bibr13-2333393618816780]; [@bibr15-2333393618816780]). In addition, homecare staff mostly work alone, so they cannot discuss practices or procedures while trying to perform them, as exemplified in Vignette 2.
The interviews demonstrated how nurses continuously evaluated the information they received from their collaborating partners (e.g., the hospital) or the lack of such. This happened regularly due to the large number of patients being rapidly discharged from hospital. Contextual changes and a lack of good routines for information and knowledge transfer therefore made it necessary for nurses to develop a continuous awareness of patients' needs as well as if, and how, homecare should respond to those needs ([@bibr7-2333393618816780]; [@bibr33-2333393618816780]).
Our studies showed that homecare nurses spend a lot of time and effort on coordination work. Scholars have argued that hospital nurses are "conductors of care" through the acts of coordination and delegation (cf. [@bibr29-2333393618816780]). Nurses in homecare hold this role, and the coordinator role is increasingly important as patient's trajectories become more complex (e.g., by more frequent readmissions to the hospital). The original illness trajectory concept, as developed by [@bibr43-2333393618816780], was developed in a hospital context. Later, the concept was expanded to better address the situation of patients/persons living at home and patients with more fluctuating trajectories and/or longer trajectories, like the patients in our studies. These developments seek to address that the original trajectory concept excludes extra complexities and unpredictabilities that emerge when patients receive services from many providers during a complex course of illness ([@bibr6-2333393618816780]; [@bibr16-2333393618816780]; [@bibr19-2333393618816780]). [@bibr6-2333393618816780] introduce the concept "caring trajectory" to reflect the combined health and social care contributions made in hospitals and in community care, as well as by informal carers, while [@bibr16-2333393618816780] coin the term "problematic trajectories." These trajectories are characterized by the presence of, for example, multimorbidity and complex care regimes. Our studies have shown that there are no routine trajectories for patients receiving homecare; rather, trajectories are problematic and require much effort to manage.
Homecare nurses coordinate care both with other health care and social care professionals and informal carers, as well as the patient. Negotiations of services, as described in the first result theme, should also be seen as a part of the coordinator job. By doing so, the nurses take on a significant role in articulating care trajectories. They align actors and actions to progress the trajectory. The nurses in our studies emphasized the problem that patients were discharged from the hospital to home without having medications---or prescriptions---with them. This is a typical situation that requires a lot of coordination work from homecare nurses, which can include telephoning the hospital, updating the patient's multidose drug dispensing list, electronically informing the GP about the changes and asking him or her to update and prescribe new medications, going to the pharmacy and fetching medications, explaining the new medication regime to the patient, and so on. Ensuring that the patient gets the right medications thus requires coordinating and aligning actors' perceptions of a situation (what is the right medications for this patient?) as well as engaging in concrete patient--nurse encounters. In daily work, the three homecare nursing practices described in this article often appear simultaneously (and must be understood together), as all three practices are necessary to deliver good patient care.
Strengths and Limitations {#section17-2333393618816780}
=========================
The strength of this study is that we have analyzed data from two studies conducted at different points of time, which allowed us to obtain a comprehensive understanding of the challenges faced by homecare nurses. It also provided us with rich data that allowed us to analyze the changes in the health care sector. Despite its strengths, the study has several limitations. One limitation is that data originally were not collected with the purpose of exploring how nurses responded to contextual changes. If that had been our main research question, we may have obtained even more detailed data on changes in work practices. Another limitation is that we conducted a limited number of observations, and by expanding the number of observations, we may have gained better insight into the practical accomplishment of current homecare nursing. However, we have discussed findings with other researchers and with homecare nurses and their leaders to validate our findings. It should be noted that qualitative data usually is not suited for generalizations. However, in this study we have data from different homecare districts representing different sizes and regions in Norway. Our findings are similar across these settings, which indicates that the trends we identify in the article are relevant for other homecare districts as well.
Conclusions and Implications for Practice {#section18-2333393618816780}
=========================================
Changes in the health care system imply a shift in health care professionals' work. In this article, we provide examples of how nurses cope with and react to the changes in the system, and we have shown three types of work practice that homecare nurses are spending more time conducting. Another Norwegian study on homecare nursing concludes that the nurses' ambitions for excellent care are threatened by the context that stems from developments in the health care system, thus leading to the lack of cooperation, competence, and continuity ([@bibr26-2333393618816780]). Our studies raise some of the same concerns, but we also find that homecare nurses are skilled and flexible in the exercise of care and are adaptive to change. In particular, we learned that organizing work is increasingly conducted. Such work is a remedy for fragmented patient trajectories and for facilitating collaboration among the patients' carers.
In this article, we have tried to show that clinical work at point of care and organizing work must be seen as intertwined, and in order to provide safe care for the patient, the care services must be thoroughly organized. Nurses play a significant role in organizing patients' trajectories. This has implications for nursing education and for nursing practice development. More competence in organizing and collaboration is needed both in municipalities and hospitals, and it should be appreciated that such practices are crucial in the provision of high-quality nursing.
We would like to thank the healthcare workers who participated in the studies. We would also like to thank Merete Lyngstad and Berit Brattheim for taking part in the data collection. Finally, we thank Nina Olsvold for comments on an earlier draft on this paper.
**Declaration of Conflicting Interests:** The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
**Funding:** The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Line Melby and Ragnhild Hellesø received support from the Research Council of Norway, grants no: 196365/V50 (BIG) and 229623/H10 (SIKT).
**ORCID iD:** Line Melby  <https://orcid.org/0000-0002-4507-0198>
**Line Melby**, PhD, is a senior researcher at SINTEF, Department of Health in Trondheim, Norway and an associate professor at the Norwegian University of Science and Technology, Department of Health Research Sciences in Gjøvik, Norway
**Aud Obstfelder**, RN, PhD is a professor and research manager at the Norwegian University of Science and Technology, Department of Health Research Sciences in Gjøvik, Norway
**Ragnhild Hellesø**, RN, PhD is a professor at University of Oslo, Institute of Health and Society in Oslo, Norway
| {
"pile_set_name": "PubMed Central"
} |
Related literature {#sec1}
==================
For details of the pharmaceutical uses of the closely related compound ambroxol, systematic name 4-(2-amino-3,5-dibromobenzylamino)cyclohexanol, see: Felix *et al.* (2008[@bb3]); Gaida *et al.* (2005[@bb4]); Lee *et al.* (2004[@bb5]). For bond length data, see: Allen *et al.* (1987[@bb1]).
Experimental {#sec2}
============
{#sec2.1}
### Crystal data {#sec2.1.1}
C~21~H~24~Br~2~N~2~O~3~*M* *~r~* = 512.24Triclinic,*a* = 8.695 (2) Å*b* = 11.124 (3) Å*c* = 12.090 (2) Åα = 73.870 (3)°β = 78.226 (3)°γ = 67.031 (2)°*V* = 1028.2 (4) Å^3^*Z* = 2Mo *K*α radiationμ = 3.97 mm^−1^*T* = 298 K0.30 × 0.30 × 0.30 mm
### Data collection {#sec2.1.2}
Bruker SMART CCD area detector diffractometerAbsorption correction: multi-scan (*SADABS*; Sheldrick, 1996[@bb6]) *T* ~min~ = 0.382, *T* ~max~ = 0.382 (expected range = 0.304--0.304)8514 measured reflections4305 independent reflections3035 reflections with *I* \> 2σ(*I*)*R* ~int~ = 0.032
### Refinement {#sec2.1.3}
*R*\[*F* ^2^ \> 2σ(*F* ^2^)\] = 0.042*wR*(*F* ^2^) = 0.096*S* = 1.024305 reflections256 parametersH-atom parameters constrainedΔρ~max~ = 0.51 e Å^−3^Δρ~min~ = −0.41 e Å^−3^
{#d5e472}
Data collection: *SMART* (Bruker, 2002[@bb2]); cell refinement: *SAINT* (Bruker, 2002[@bb2]); data reduction: *SAINT*; program(s) used to solve structure: *SHELXS97* (Sheldrick, 2008[@bb7]); program(s) used to refine structure: *SHELXL97* (Sheldrick, 2008[@bb7]); molecular graphics: *SHELXTL* (Sheldrick, 2008[@bb7]); software used to prepare material for publication: *SHELXTL*.
Supplementary Material
======================
Crystal structure: contains datablocks global, I. DOI: [10.1107/S1600536809005182/sj2575sup1.cif](http://dx.doi.org/10.1107/S1600536809005182/sj2575sup1.cif)
Structure factors: contains datablocks I. DOI: [10.1107/S1600536809005182/sj2575Isup2.hkl](http://dx.doi.org/10.1107/S1600536809005182/sj2575Isup2.hkl)
Additional supplementary materials: [crystallographic information](http://scripts.iucr.org/cgi-bin/sendsupfiles?sj2575&file=sj2575sup0.html&mime=text/html); [3D view](http://scripts.iucr.org/cgi-bin/sendcif?sj2575sup1&Qmime=cif); [checkCIF report](http://scripts.iucr.org/cgi-bin/paper?sj2575&checkcif=yes)
Supplementary data and figures for this paper are available from the IUCr electronic archives (Reference: [SJ2575](http://scripts.iucr.org/cgi-bin/sendsup?sj2575)).
Financial support from the Third Affiliated Hospital of Soochow University is acknowledged.
Comment
=======
Ambroxol, 4-(2-amino-3,5-dibromobenzylamino)cyclohexanol, is an expectorant agent which leads to bronchial secretion due to its mucolytic properties (Felix *et al.*, 2008; Gaida *et al.*, 2005; Lee *et al.*, 2004). In this paper, the crystal structure of the new title compound, (I), derived from the condensation reaction of 3-methoxysalicylaldehyde with 4-(2-amino-3,5-dibromobenzylamino)cyclohexanol in a methanol solution, is reported.
In (I), Fig. 1, the dihedral angle between the two benzene rings is 76.4 (3)°. The cyclohexyl ring adopts a chair configuration. All the bond lengths are within normal ranges (Allen *et al.*, 1987). There is an intramolecular O2---H2···N2 hydrogen bond which affects the solid state conformation of the molecule. The crystal structure is stabilized by intermolecular O---H···O hydrogen bonds (Table 1), forming chains running along the *b* axis (Fig. 2).
Experimental {#experimental}
============
3-Methoxysalicylaldehyde (1.0 mol, 152.1 mg) and 4-(2-amino-3,5-dibromobenzylamino)cyclohexanol (1.0 mmol, 378.1 mg) were dissolved in a methanol solution (10 ml). The mixture was stirred at room temperature to give a clear colorless solution. Crystals of the title compound were formed by gradual evaporation of the solvent for a week at room temperature.
Refinement {#refinement}
==========
H atoms were constrained to ideal geometries, with C--H = 0.93--0.97 Å, O--H = 0.82 Å, and with *U*~iso~(H) set to 1.2*U*~eq~(C) and 1.5*U*~eq~(O and C21).
Figures
=======
{#Fap1}
{#Fap2}
Crystal data {#tablewrapcrystaldatalong}
============
------------------------- ---------------------------------------
C~21~H~24~Br~2~N~2~O~3~ *Z* = 2
*M~r~* = 512.24 *F*(000) = 516
Triclinic, *P*1 *D*~x~ = 1.655 Mg m^−3^
Hall symbol: -P 1 Mo *K*α radiation, λ = 0.71073 Å
*a* = 8.695 (2) Å Cell parameters from 2079 reflections
*b* = 11.124 (3) Å θ = 2.3--24.6°
*c* = 12.090 (2) Å µ = 3.97 mm^−1^
α = 73.870 (3)° *T* = 298 K
β = 78.226 (3)° Block, colorless
γ = 67.031 (2)° 0.30 × 0.30 × 0.30 mm
*V* = 1028.2 (4) Å^3^
------------------------- ---------------------------------------
Data collection {#tablewrapdatacollectionlong}
===============
--------------------------------------------------------------- --------------------------------------
Bruker SMART CCD area detector diffractometer 4305 independent reflections
Radiation source: fine-focus sealed tube 3035 reflections with *I* \> 2σ(*I*)
graphite *R*~int~ = 0.032
ω scans θ~max~ = 27.0°, θ~min~ = 1.8°
Absorption correction: multi-scan (*SADABS*; Sheldrick, 1996) *h* = −11→10
*T*~min~ = 0.382, *T*~max~ = 0.382 *k* = −14→14
8514 measured reflections *l* = −15→15
--------------------------------------------------------------- --------------------------------------
Refinement {#tablewraprefinementdatalong}
==========
------------------------------------- -------------------------------------------------------------------------------------------------
Refinement on *F*^2^ Primary atom site location: structure-invariant direct methods
Least-squares matrix: full Secondary atom site location: difference Fourier map
*R*\[*F*^2^ \> 2σ(*F*^2^)\] = 0.042 Hydrogen site location: inferred from neighbouring sites
*wR*(*F*^2^) = 0.096 H-atom parameters constrained
*S* = 1.02 *w* = 1/\[σ^2^(*F*~o~^2^) + (0.0423*P*)^2^ + 0.0168*P*\] where *P* = (*F*~o~^2^ + 2*F*~c~^2^)/3
4305 reflections (Δ/σ)~max~ = 0.001
256 parameters Δρ~max~ = 0.51 e Å^−3^
0 restraints Δρ~min~ = −0.41 e Å^−3^
------------------------------------- -------------------------------------------------------------------------------------------------
Special details {#specialdetails}
===============
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Geometry. All e.s.d.\'s (except the e.s.d. in the dihedral angle between two l.s. planes) are estimated using the full covariance matrix. The cell e.s.d.\'s are taken into account individually in the estimation of e.s.d.\'s in distances, angles and torsion angles; correlations between e.s.d.\'s in cell parameters are only used when they are defined by crystal symmetry. An approximate (isotropic) treatment of cell e.s.d.\'s is used for estimating e.s.d.\'s involving l.s. planes.
Refinement. Refinement of *F*^2^ against ALL reflections. The weighted *R*-factor *wR* and goodness of fit *S* are based on *F*^2^, conventional *R*-factors *R* are based on *F*, with *F* set to zero for negative *F*^2^. The threshold expression of *F*^2^ \> σ(*F*^2^) is used only for calculating *R*-factors(gt) *etc*. and is not relevant to the choice of reflections for refinement. *R*-factors based on *F*^2^ are statistically about twice as large as those based on *F*, and *R*- factors based on ALL data will be even larger.
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Fractional atomic coordinates and isotropic or equivalent isotropic displacement parameters (Å^2^) {#tablewrapcoords}
==================================================================================================
------ ------------- ------------- ------------- -------------------- --
*x* *y* *z* *U*~iso~\*/*U*~eq~
Br1 1.17738 (5) 0.24247 (4) 0.32321 (4) 0.04726 (14)
Br2 1.28346 (5) 0.54977 (4) 0.58131 (3) 0.04402 (14)
O1 0.2331 (3) 0.9422 (3) 0.3052 (2) 0.0418 (7)
H1 0.1581 0.9883 0.2635 0.063\*
O2 0.9413 (3) 0.8807 (3) −0.1082 (2) 0.0430 (7)
H2 0.8961 0.8692 −0.0416 0.065\*
O3 1.1435 (4) 0.9116 (3) −0.2940 (2) 0.0510 (7)
N1 1.0518 (4) 0.5182 (3) 0.1542 (2) 0.0305 (7)
H1A 1.0527 0.4432 0.1469 0.037\*
N2 0.9133 (3) 0.7617 (3) 0.1103 (2) 0.0249 (6)
C1 1.1053 (4) 0.6460 (3) 0.2638 (3) 0.0265 (8)
C2 1.1080 (4) 0.5243 (3) 0.2507 (3) 0.0263 (7)
C3 1.1674 (4) 0.4110 (3) 0.3381 (3) 0.0287 (8)
C4 1.2199 (4) 0.4172 (3) 0.4357 (3) 0.0317 (8)
H4 1.2586 0.3405 0.4931 0.038\*
C5 1.2143 (4) 0.5386 (4) 0.4468 (3) 0.0302 (8)
C6 1.1591 (4) 0.6523 (3) 0.3620 (3) 0.0306 (8)
H6 1.1576 0.7334 0.3702 0.037\*
C7 1.0414 (4) 0.7698 (3) 0.1692 (3) 0.0259 (7)
H7A 0.9929 0.8481 0.2027 0.031\*
H7B 1.1349 0.7799 0.1128 0.031\*
C8 0.7481 (4) 0.7705 (3) 0.1797 (3) 0.0269 (8)
H8 0.7560 0.6812 0.2263 0.032\*
C9 0.6954 (4) 0.8659 (4) 0.2601 (3) 0.0377 (9)
H9A 0.6977 0.9525 0.2156 0.045\*
H9B 0.7751 0.8324 0.3168 0.045\*
C10 0.5210 (5) 0.8823 (4) 0.3222 (3) 0.0410 (10)
H10A 0.5213 0.7973 0.3722 0.049\*
H10B 0.4908 0.9464 0.3705 0.049\*
C11 0.3920 (4) 0.9293 (3) 0.2392 (3) 0.0322 (8)
H11 0.3880 1.0175 0.1918 0.039\*
C12 0.4406 (4) 0.8330 (4) 0.1607 (3) 0.0414 (10)
H12A 0.3600 0.8662 0.1045 0.050\*
H12B 0.4380 0.7468 0.2062 0.050\*
C13 0.6155 (4) 0.8160 (4) 0.0976 (3) 0.0378 (9)
H13A 0.6452 0.7509 0.0504 0.045\*
H13B 0.6143 0.9006 0.0463 0.045\*
C14 0.9912 (4) 0.6381 (3) 0.0655 (3) 0.0255 (7)
H14 0.9030 0.6289 0.0329 0.031\*
C15 1.1251 (4) 0.6565 (4) −0.0355 (3) 0.0296 (8)
C16 1.0875 (4) 0.7752 (4) −0.1185 (3) 0.0298 (8)
C17 1.1986 (5) 0.7917 (4) −0.2178 (3) 0.0367 (9)
C18 1.3514 (5) 0.6900 (5) −0.2301 (3) 0.0454 (11)
H18 1.4274 0.7007 −0.2951 0.055\*
C19 1.3921 (5) 0.5718 (4) −0.1457 (4) 0.0476 (11)
H19 1.4954 0.5038 −0.1541 0.057\*
C20 1.2793 (4) 0.5554 (4) −0.0496 (3) 0.0368 (9)
H20 1.3068 0.4759 0.0063 0.044\*
C21 1.2282 (7) 0.9219 (5) −0.4076 (3) 0.0797 (17)
H21A 1.3418 0.9132 −0.4048 0.120\*
H21B 1.1715 1.0074 −0.4549 0.120\*
H21C 1.2288 0.8521 −0.4400 0.120\*
------ ------------- ------------- ------------- -------------------- --
Atomic displacement parameters (Å^2^) {#tablewrapadps}
=====================================
----- ------------- ------------- ------------- --------------- --------------- ---------------
*U*^11^ *U*^22^ *U*^33^ *U*^12^ *U*^13^ *U*^23^
Br1 0.0577 (3) 0.0246 (2) 0.0575 (3) −0.00942 (19) −0.0134 (2) −0.00831 (19)
Br2 0.0490 (3) 0.0547 (3) 0.0320 (2) −0.0205 (2) −0.01256 (18) −0.00628 (18)
O1 0.0277 (14) 0.0408 (16) 0.0439 (16) −0.0053 (13) 0.0025 (12) −0.0038 (13)
O2 0.0370 (16) 0.0440 (16) 0.0302 (14) −0.0040 (13) 0.0040 (12) −0.0011 (12)
O3 0.0602 (19) 0.0570 (18) 0.0303 (15) −0.0259 (16) 0.0092 (13) −0.0038 (14)
N1 0.0405 (18) 0.0227 (15) 0.0294 (16) −0.0106 (14) −0.0055 (14) −0.0073 (13)
N2 0.0216 (15) 0.0262 (15) 0.0263 (15) −0.0077 (12) −0.0019 (12) −0.0068 (12)
C1 0.0208 (18) 0.0288 (19) 0.0291 (18) −0.0088 (15) −0.0031 (15) −0.0048 (15)
C2 0.0186 (18) 0.0276 (18) 0.0313 (19) −0.0066 (15) 0.0006 (14) −0.0094 (15)
C3 0.028 (2) 0.0219 (18) 0.036 (2) −0.0074 (16) −0.0032 (16) −0.0082 (16)
C4 0.027 (2) 0.029 (2) 0.031 (2) −0.0038 (17) −0.0074 (16) −0.0005 (16)
C5 0.0239 (19) 0.039 (2) 0.0297 (19) −0.0106 (17) −0.0051 (15) −0.0097 (17)
C6 0.0249 (19) 0.032 (2) 0.037 (2) −0.0127 (16) 0.0005 (16) −0.0098 (17)
C7 0.0232 (18) 0.0268 (18) 0.0297 (18) −0.0096 (15) −0.0042 (15) −0.0073 (15)
C8 0.0257 (19) 0.0221 (17) 0.0311 (19) −0.0081 (15) −0.0016 (15) −0.0047 (15)
C9 0.030 (2) 0.048 (2) 0.035 (2) −0.0078 (18) −0.0039 (17) −0.0168 (18)
C10 0.036 (2) 0.048 (2) 0.032 (2) −0.0031 (19) −0.0032 (18) −0.0155 (18)
C11 0.028 (2) 0.0266 (19) 0.036 (2) −0.0076 (16) 0.0033 (16) −0.0049 (16)
C12 0.027 (2) 0.050 (2) 0.050 (2) −0.0111 (19) −0.0040 (18) −0.018 (2)
C13 0.027 (2) 0.054 (3) 0.039 (2) −0.0119 (19) −0.0026 (17) −0.0247 (19)
C14 0.0221 (18) 0.0295 (19) 0.0270 (18) −0.0080 (15) −0.0046 (14) −0.0098 (15)
C15 0.027 (2) 0.039 (2) 0.0253 (18) −0.0134 (17) −0.0032 (15) −0.0096 (16)
C16 0.025 (2) 0.037 (2) 0.0259 (18) −0.0096 (17) −0.0037 (15) −0.0072 (16)
C17 0.038 (2) 0.049 (2) 0.028 (2) −0.020 (2) 0.0007 (17) −0.0109 (18)
C18 0.035 (2) 0.073 (3) 0.037 (2) −0.026 (2) 0.0066 (19) −0.024 (2)
C19 0.026 (2) 0.062 (3) 0.054 (3) −0.003 (2) −0.001 (2) −0.032 (2)
C20 0.026 (2) 0.042 (2) 0.041 (2) −0.0053 (18) −0.0057 (17) −0.0149 (18)
C21 0.131 (5) 0.086 (4) 0.029 (2) −0.063 (4) 0.027 (3) −0.013 (2)
----- ------------- ------------- ------------- --------------- --------------- ---------------
Geometric parameters (Å, °) {#tablewrapgeomlong}
===========================
------------------ ----------- ------------------- -----------
Br1---C3 1.900 (3) C9---C10 1.514 (5)
Br2---C5 1.895 (3) C9---H9A 0.9700
O1---C11 1.423 (4) C9---H9B 0.9700
O1---H1 0.8200 C10---C11 1.499 (5)
O2---C16 1.364 (4) C10---H10A 0.9700
O2---H2 0.8200 C10---H10B 0.9700
O3---C17 1.363 (4) C11---C12 1.509 (5)
O3---C21 1.416 (4) C11---H11 0.9800
N1---C2 1.381 (4) C12---C13 1.520 (5)
N1---C14 1.448 (4) C12---H12A 0.9700
N1---H1A 0.8600 C12---H12B 0.9700
N2---C14 1.476 (4) C13---H13A 0.9700
N2---C7 1.482 (4) C13---H13B 0.9700
N2---C8 1.490 (4) C14---C15 1.535 (4)
C1---C6 1.389 (4) C14---H14 0.9800
C1---C2 1.396 (4) C15---C16 1.385 (5)
C1---C7 1.517 (4) C15---C20 1.388 (5)
C2---C3 1.395 (5) C16---C17 1.397 (5)
C3---C4 1.376 (4) C17---C18 1.381 (5)
C4---C5 1.376 (5) C18---C19 1.391 (6)
C4---H4 0.9300 C18---H18 0.9300
C5---C6 1.374 (5) C19---C20 1.379 (5)
C6---H6 0.9300 C19---H19 0.9300
C7---H7A 0.9700 C20---H20 0.9300
C7---H7B 0.9700 C21---H21A 0.9600
C8---C9 1.513 (5) C21---H21B 0.9600
C8---C13 1.518 (4) C21---H21C 0.9600
C8---H8 0.9800
C11---O1---H1 109.5 H10A---C10---H10B 107.9
C16---O2---H2 109.5 O1---C11---C10 107.9 (3)
C17---O3---C21 117.3 (3) O1---C11---C12 112.8 (3)
C2---N1---C14 120.0 (3) C10---C11---C12 109.8 (3)
C2---N1---H1A 120.0 O1---C11---H11 108.8
C14---N1---H1A 120.0 C10---C11---H11 108.8
C14---N2---C7 107.0 (2) C12---C11---H11 108.8
C14---N2---C8 112.1 (2) C11---C12---C13 110.9 (3)
C7---N2---C8 116.3 (2) C11---C12---H12A 109.5
C6---C1---C2 120.2 (3) C13---C12---H12A 109.5
C6---C1---C7 121.4 (3) C11---C12---H12B 109.5
C2---C1---C7 118.4 (3) C13---C12---H12B 109.5
N1---C2---C3 121.7 (3) H12A---C12---H12B 108.1
N1---C2---C1 120.4 (3) C8---C13---C12 112.7 (3)
C3---C2---C1 118.0 (3) C8---C13---H13A 109.1
C4---C3---C2 121.9 (3) C12---C13---H13A 109.1
C4---C3---Br1 118.5 (3) C8---C13---H13B 109.1
C2---C3---Br1 119.6 (3) C12---C13---H13B 109.1
C5---C4---C3 118.9 (3) H13A---C13---H13B 107.8
C5---C4---H4 120.6 N1---C14---N2 113.6 (3)
C3---C4---H4 120.6 N1---C14---C15 113.8 (3)
C6---C5---C4 121.1 (3) N2---C14---C15 109.0 (3)
C6---C5---Br2 119.3 (3) N1---C14---H14 106.7
C4---C5---Br2 119.6 (3) N2---C14---H14 106.7
C5---C6---C1 119.9 (3) C15---C14---H14 106.7
C5---C6---H6 120.0 C16---C15---C20 118.9 (3)
C1---C6---H6 120.0 C16---C15---C14 118.9 (3)
N2---C7---C1 111.6 (3) C20---C15---C14 122.1 (3)
N2---C7---H7A 109.3 O2---C16---C15 122.2 (3)
C1---C7---H7A 109.3 O2---C16---C17 116.8 (3)
N2---C7---H7B 109.3 C15---C16---C17 121.0 (3)
C1---C7---H7B 109.3 O3---C17---C18 126.0 (3)
H7A---C7---H7B 108.0 O3---C17---C16 114.9 (3)
N2---C8---C9 112.9 (3) C18---C17---C16 119.1 (4)
N2---C8---C13 108.8 (3) C17---C18---C19 120.2 (4)
C9---C8---C13 109.3 (3) C17---C18---H18 119.9
N2---C8---H8 108.6 C19---C18---H18 119.9
C9---C8---H8 108.6 C20---C19---C18 120.1 (4)
C13---C8---H8 108.6 C20---C19---H19 120.0
C8---C9---C10 112.0 (3) C18---C19---H19 120.0
C8---C9---H9A 109.2 C19---C20---C15 120.6 (4)
C10---C9---H9A 109.2 C19---C20---H20 119.7
C8---C9---H9B 109.2 C15---C20---H20 119.7
C10---C9---H9B 109.2 O3---C21---H21A 109.5
H9A---C9---H9B 107.9 O3---C21---H21B 109.5
C11---C10---C9 112.0 (3) H21A---C21---H21B 109.5
C11---C10---H10A 109.2 O3---C21---H21C 109.5
C9---C10---H10A 109.2 H21A---C21---H21C 109.5
C11---C10---H10B 109.2 H21B---C21---H21C 109.5
C9---C10---H10B 109.2
------------------ ----------- ------------------- -----------
Hydrogen-bond geometry (Å, °) {#tablewraphbondslong}
=============================
----------------- --------- --------- ----------- ---------------
*D*---H···*A* *D*---H H···*A* *D*···*A* *D*---H···*A*
O2---H2···N2 0.82 1.89 2.614 (3) 147
O1---H1···O3^i^ 0.82 2.41 3.048 (4) 135
O1---H1···O2^i^ 0.82 2.13 2.897 (4) 155
----------------- --------- --------- ----------- ---------------
Symmetry codes: (i) −*x*+1, −*y*+2, −*z*.
###### Hydrogen-bond geometry (Å, °)
*D*---H⋯*A* *D*---H H⋯*A* *D*⋯*A* *D*---H⋯*A*
--------------- --------- ------- ----------- -------------
O2---H2⋯N2 0.82 1.89 2.614 (3) 147
O1---H1⋯O3^i^ 0.82 2.41 3.048 (4) 135
O1---H1⋯O2^i^ 0.82 2.13 2.897 (4) 155
Symmetry code: (i) .
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#s1}
============
The brain is regarded as the master regulator of homeostasis and metabolism. It has been suggested that bone and energy metabolism require tightly coordinated regulation so that longitudinal growth, or bone remodeling, are in accordance with energy supply and demand [@pone.0065505-Karsenty1]. Numerous studies have investigated the mechanisms involved in the central regulation of bone metabolism and possible connections between bone metabolism and energy metabolism [@pone.0065505-Iwaniec1]--[@pone.0065505-Williams1]. However, the data are contradictory regarding bone metabolism regulation. In addition, the factors that co-regulate bone metabolism and energy metabolism are not yet clear.
Leptin, a major adipokine that regulates appetite, has been widely investigated as the major factor co-regulating bone metabolism and energy metabolism [@pone.0065505-Ducy1]. Ducy et al. demonstrated that intracerebroventricular (ICV) injection of leptin induced bone loss in the spine, suggesting a role for leptin in bone metabolism regulation through a central mechanism [@pone.0065505-Ducy1]. Data from subsequent reports supported the bone catabolic effect of leptin and have shown that leptin inhibits bone formation via the sympathetic nervous system [@pone.0065505-Elefteriou1], [@pone.0065505-Takeda1]. However, several recent studies have reported contradicting data [@pone.0065505-Iwaniec1], [@pone.0065505-Bartell1], [@pone.0065505-Williams1]. Leptin administration via gene therapy into the hypothalamus did not have a significant effect on bone metabolism [@pone.0065505-Bartell1]. Leptin receptor deficient mice exhibit decreased bone mass, demonstrating leptin\'s anabolic effect on bone [@pone.0065505-Williams1]. Furthermore, ICV injection of leptin increases bone mineral density (BMD) and mineral apposition rates [@pone.0065505-Elefteriou1], a finding which is the exact opposite of the initial studies reporting a catabolic effect of central leptin [@pone.0065505-Ducy1]. Given this conflicting data, further studies are required to clarify the mechanisms underlying the central regulation of bone metabolism.
Ghrelin is a stomach hormone that acts centrally to increase appetite [@pone.0065505-Nakazato1]. Several studies have investigated the effect of chronic ICV ghrelin on energy metabolism. Chronic ICV ghrelin infusion increased food intake and weight gain [@pone.0065505-Nakazato1], reversed the effect of leptin on food intake [@pone.0065505-Kim1], increased the glucose utilization rate of adipose tissue [@pone.0065505-TheanderCarrillo1], and increased lipogenic enzymatic activity in adipose tissue and liver [@pone.0065505-SangiaoAlvarellos1]. A recent study investigated the central effects of ghrelin and leptin on body and bone marrow adiposity, as well as adipose tissue and bone marrow gene expression, and reported central ghrelin had no effect on bone marrow adiposity [@pone.0065505-Ambati1]. However, to date, no study has investigated the effect of chronic ICV ghrelin on bone metabolism.
Ghrelin and leptin have opposite effects on energy metabolism, but also on the sympathetic nervous system and other pathways [@pone.0065505-Dunbar1]--[@pone.0065505-Shan1]. Central leptin stimulates sympathetic outflow [@pone.0065505-Dunbar1] whereas central ghrelin suppresses sympathetic activity [@pone.0065505-Matsumura1]. Since the sympathetic nervous system has been suggested as the major pathway governing the effect of central leptin on bone metabolism [@pone.0065505-Ducy1], it is possible that sympathetic suppression through central ghrelin can affect bone metabolism. Therefore, the present study was designed to investigate the chronic effects of centrally administered ghrelin on bone metabolism.
Materials and Methods {#s2}
=====================
Animals and peptide {#s2a}
-------------------
Male Sprague-Dawley rats (6 weeks old) weighing 180--230 g were used. Body weight and food intake were monitored daily. All animal experiments were performed with approval from the Institutional Animal Care and Use Committee (IACUC) of the Clinical Research Institute at Seoul National University Hospital (an AAALAC accredited facility). National Research Council (NRC) guidelines for the care and use of laboratory animals were also observed (1996 revision). The standard rodent chow (Purina Rodent Chow; Biopia, Korea) was used. Ghrelin (rat) was purchased from Bachem Inc. (Bubendorf, Switzerland).The ghrelin peptide was prepared with concentration of of 0.25 µg/µl which corresponds to 1.5 µg/day (6 µg/day; 7.14 µg/Kg body weight/day).
Surgery {#s2b}
-------
Rats were anesthetized by intraperitoneal injection of 50 mg/kg zoletil and 10 mg/kg xylazine and surgically implanted with a 22-gauge stainless-steel cannula (Plastics One Inc. Roanoke, VA, USA) into the third cerebral ventricle. Osmotic mini-pumps (Model 2004, 0.25 μL/h; Alzet Corp., Cupertino, CA, USA) filled with saline or rat ghrelin peptide were implanted under the dorsal chest skin. The mini-pumps were connected to the ICV cannula via a catheter. Alzet Brain infusion kit 2 with infusion cannula of ID = 0.18 mm and OD = 0.36 mm was used. The cannula was stereotactically placed 0.72 mm posterior to the bregma on the midline and implanted 7 mm below the outer surface of the skull surgery via a stereotaxic apparatus. For the verification for correct cannula placement, cresyl violet staining and brain dissection was performed.
Study design {#s2c}
------------
Fifteen rats were divided into 3 groups: the control group (n = 5) received ICV infusions of saline for 21 days; the ghrelin *ad libitum*-fed (ghrelin *ad lib*-fed) group (n = 4) received ICV infusions of ghrelin (1.5 μg/day) for 21 days; the ghrelin pair-fed group (n = 6) received ICV infusions of ghrelin (1.5 μg/day) for 21 days and were allowed to eat only as much chow as consumed by the control group on the previous day.
Radiological analyses and bone histomorphometry {#s2d}
-----------------------------------------------
Nondestructive, three-dimensional evaluation of bone mass and architecture were performed using a microCT scanner (Skyscan 1076 for tibia and lumbar spine and Skyscan 1172 for femur; Skyscan, Aartselaar, Belgium). Lumbar spine and femur were dissected from soft tissue, fixed in 70% ethanol, and analyzed. Image acquisition of tibia and lumbar spine L3 was performed with a source voltage of 100 kV, current of 100 μA, a 0.5-mm aluminum filter, and an isotropic voxel size of 8.8 μm. Image acquisition of femur was performed with a source voltage of 70 kV, current of 141 μA, a 0.5-mm aluminum filter, and an isotropic voxel size of 11.55 μm. For the tibia metaphysis trabecular bone, 251-slice-thick volume of interest starting 150 slices distal to the proximal growth plate were analyzed. For the tibia diaphysis cortex, the mid-diaphysis cortical bone volume of interest was analyzed. For the femoral metaphysis, 101-slice-thick volumes of interest starting 150 slices proximal to the distal growth plate were analyzed. For the lumbar spine, trabecular bone volumes of interest of whole trans-axial spine body images were analyzed. For trabecular volumetric bone mineral density (BMD) analyses using microCT, phantoms with predefined densities and CTA software were used for BMD measurements, as described in the manufacturer\'s manual (Skyscan, Aartselaar, Belgium). The trabecular and cortical bone volumes of interest were outlined by interpolation of operator-drawn regions exclusively representing the trabecular and cortical bone, respectively.
BMD (g/cm^2^) of the *ex vivo* tibia, femur, and lumbar spine (L2 and L3) were measured using the dual energy X-ray absorptiometry (DXA) instrument PIXIMUS (GE Lunar, Madison, WI, USA). A phantom was scanned daily for quality control.
For histomorphometric analyses of dynamic parameters, rats were injected with calcein (20 mg/kg body weight) 2 and 6 days prior to sacrifice. Dynamic histomorphometry analyses were conducted on the lumbar spine L5 using the Bioquant program (Bio-Quant. Inc., San Diego, CA, USA) [@pone.0065505-Parfitt1].
Serum marker analyses {#s2e}
---------------------
The serum concentrations of procollagen type 1 amino-terminal propeptide (P1NP), cross-linked C-telopeptide (CTX), and tartrate-resistant acid phosphatase 5 b (TRAP-5b) were determined by ELISA, following the manufacturer\'s protocol (Immunodiagnostic Systems Inc, Scottsdale, AZ, USA). The serum concentrations of ghrelin and leptin were determined by ELISA, per the manufacturers\' protocols (Linco Research, St Charles, MO, USA and Millipore, Billerica, MA, USA, respectively). Serum insulin-like growth factor 1 (IGF-1) was determined by immunoenzymometric assay (IEMA) following the manufacturer\'s protocol (GroPep-IDS, Fountain Hills, CA, USA).
Statistical Analyses {#s2f}
--------------------
Data are presented as the mean ± SEM. Statistical analyses were performed using analysis of variance (ANOVA) with least significant difference (LSD) as a *post hoc* comparison. Significance was defined as P\<0.05. Statistical analyses were performed with SPSS for Windows (version 17.0, SPSS Inc., Chicago, IL, USA).
Results {#s3}
=======
Chronic ICV ghrelin infusion increases body weight and food intake {#s3a}
------------------------------------------------------------------
Chronic ICV ghrelin infusion (1.5 μg/day for 21 days) significantly increased body weight in ghrelin *ad lib*-fed compared with ICV saline-infused control rats (331±4 g vs. 312±9 g, P\<0.05). However, there was no difference in body weight between the control rats and ICV ghrelin-infused rats that were pair-fed the food intake of the control rats (ghrelin pair-fed) (316±11 g) (ANOVA F value = 4.21) ([Fig. 1A](#pone-0065505-g001){ref-type="fig"}). Although there was no significant difference in cumulative food intake between the ghrelin pair-fed and control groups, ghrelin *ad lib*-fed rats had significantly higher cumulative food intake compared with ghrelin pair-fed rats (P\<0.05) (ANOVA F value = 2.86) ([Fig. 1B](#pone-0065505-g001){ref-type="fig"}).
{#pone-0065505-g001}
Chronic ICV ghrelin infusion increases bone mass {#s3b}
------------------------------------------------
Chronic ICV ghrelin infusion significantly increased bone mass in ghrelin pair-fed group compared with control rats. This was indicated by increased bone volume percentage (BV/TV, 32.4±1.5% vs. 24.3±1.3%, P = 0.002) (ANOVA F value = 8.43), trabecular thickness (Tb.Th, 75.3±1.6 μm vs. 66.0±1.6 μm, P = 0.001) (ANOVA F value = 10.40), trabecular number (Tb.N, 4.3±0.2 mm^−1^ vs. 3.7±0.2 mm^−1^, P = 0.023) (ANOVA F value = 4.80), and volumetric BMD (0.164±0.004 g/cm3 vs. 0.123±0.005 g/cm3, P = 0.00005) (ANOVA F value = 20.18) of the tibia trabecular bone, as measured by microCT ([Fig. 2](#pone-0065505-g002){ref-type="fig"}). The trabecular pattern factor (Tb.Pf), a parameter inversely correlated to the connectivity, was significantly decreased in the ghrelin pair-fed group, indicating that the trabecular structure was more connected in the ghrelin pair-fed group (P\<0.05, [Fig 2](#pone-0065505-g002){ref-type="fig"}) (ANOVA F value = 6.46). This result is consistent with the decreased structure model index (SMI) observed in the ghrelin pair-fed group, which suggests that the trabecular structure was more plate-like in the ghrelin pair-fed group, as compared with the rod-like structures in the control group (P\<0.05, [Fig. 2](#pone-0065505-g002){ref-type="fig"}) (ANOVA F value = 4.45). There were no significant differences in bone mass observed between the control group and ghrelin ad-lib fed group. Similar trends were observed in femur and lumbar spine, with lower statistical significance ([Figure S1](#pone.0065505.s001){ref-type="supplementary-material"} and [S2](#pone.0065505.s002){ref-type="supplementary-material"}). Cortical area (Ct.Ar) was significantly increased in the ghrelin *ad lib*-fed group compared to the control group (P\<0.05, [Figure S3](#pone.0065505.s003){ref-type="supplementary-material"}) (ANOVA F value = 3.57). There was a tendency of increased cortical area fraction and cortical thickness in the ghrelin *ad lib*-fed group compared to the control group (P\<0.1, [Figure S3](#pone.0065505.s003){ref-type="supplementary-material"}) (ANOVA F value = 1.73 and 2.49, respectively). There were no significant differences in *ex vivo* BMD measurements between the three groups ([Table 1](#pone-0065505-t001){ref-type="table"}).
{#pone-0065505-g002}
10.1371/journal.pone.0065505.t001
###### Effect of chronic ICV ghrelin infusion on ex vivo BMD measurement of tibia, femur and spine by DXA.
{#pone-0065505-t001-1}
Control Ghrelin Ad lib-fed Ghrelin Pair-fed
----------------------------- ------------- -------------------- ------------------
Tibia ex vivo BMD (g/cm^2^) 0.119±0.002 0.125±0.004 0.127±0.003
Femur ex vivo BMD (g/cm^2^) 0.148±0.004 0.161±0.001 † 0.155±0.005
Spine ex vivo BMD (g/cm^2^) 0.154±0.002 0.158±0.003 0.157±0.003
P\<0.05. †P\<0.1. Mean±SEM.
Chronic ICV ghrelin infusion increases the mineral apposition rate {#s3c}
------------------------------------------------------------------
Chronic ICV ghrelin infusion significantly increased the mineral apposition rate (MAR) in the ghrelin pair-fed group compared with the control group (5.0±0.2 μm/d vs. 4.0±0.1 μm/d, P = 0.014, [Fig. 3](#pone-0065505-g003){ref-type="fig"}) (ANOVA F value = 5.36). The ghrelin *ad lib*-fed group tended to have a higher MAR compared with the control group (P = 0.073, [Fig 3](#pone-0065505-g003){ref-type="fig"}). No significant differences in bone formation rate were observed among the three groups.
{#pone-0065505-g003}
Effect of chronic ICV ghrelin infusion on serum markers {#s3d}
-------------------------------------------------------
Chronic ICV ghrelin infusion significantly increased serum leptin in both the ghrelin *ad lib*-fed group and ghrelin pair-fed group compared with the control group ([Fig. 4](#pone-0065505-g004){ref-type="fig"}). Chronic ICV ghrelin infusion significantly decreased serum ghrelin in the ghrelin *ad lib*-fed group compared with the control group (P\<0.05) (ANOVA F value = 3.55 and 6.31, respectively). However, there were no significant differences in IGF-1, CTX, TRAP-5b, and P1NP among the three groups.
{#pone-0065505-g004}
Discussion {#s4}
==========
Using a rat ICV administration model, we demonstrate that intracerebral infusion of ghrelin results in increased bone mass, connectivity, and mineral apposition rates in ghrelin pair-fed group. Chronic ICV administration of ghrelin increased weight gain in the ghrelin *ad lib*-fed group; however, this effect was abolished by pair feeding. These findings indicate that chronic central administration of ghrelin increased bone mass through mechanisms independent of changes in body weight. The duration of ICV ghrelin treatment (21 days) in this study could be considered relatively short to induce profound changes in bone mass. However, it should be noted that establishment of an animal model of chronic ICV treatment is technically challenging and the present study is the longest ICV ghrelin treatment attempted (21 days), with previous studies infusing ghrelin for only 6--12 days [@pone.0065505-Nakazato1]--[@pone.0065505-TheanderCarrillo1], [@pone.0065505-Ambati1]. Future study using Alzet osmotic pump model 2ML4 could extend chronic infusion up to 28 days.
Although we clearly demonstrated anabolic effects of ghrelin on bone (independent from ghrelin-induced effects on feeding), the mechanism for this effect is not known. One possible mechanism may be suppression of the sympathetic nervous system. Sympathetic activity has been reported to be suppressed by ICV administration of ghrelin [@pone.0065505-Matsumura1] and stimulated by ICV administration of leptin [@pone.0065505-Dunbar1]. Hypothalamic administration of leptin was reported to decrease bone formation due to the ability of leptin to increase sympathetic tone [@pone.0065505-Takeda1], [@pone.0065505-Ducy1], [@pone.0065505-Dunbar2]. Numerous animal [@pone.0065505-Takeda1], [@pone.0065505-Baek1] and human [@pone.0065505-Schlienger1], [@pone.0065505-Yang1] studies demonstrate that beta-blockers have protective effects on bone metabolism. Therefore, ICV ghrelin may increase bone mass through sympathetic suppression. Future studies using pharmacologic blockade of the sympathetic nervous system or local sympathectomy surgery will clarify this mechanism.
Other possible mechanisms include the secretagogue effect of ICV ghrelin on growth hormone [@pone.0065505-Kojima1]. However, the control group and ghrelin pair-fed group did not differ in serum IGF-1 levels so this does not support a role for growth hormone. These data are supported by a previous study that also reported no change in plasma IGF-1 after 7 days of ICV ghrelin treatment [@pone.0065505-Kim1]. It has also been reported that ICV administration of ghrelin increased the plasma concentration of growth hormone on day 6 but this was not sustained on day 12 [@pone.0065505-Date1]. Therefore, it is unlikely that the effect of ICV ghrelin is mediated through a growth hormone-IGF-1 axis.
In contrast to the ghrelin pair-fed group, the ghrelin *ad lib*-fed group had a slight increase in cortical bone area and no significant difference in trabecular phenotypes compared with controls. However, the ghrelin *ad lib*-fed group did exhibit increased body weight gain. Therefore, the increased cortical bone mass could be the result of increased body weight and a consequent increase in mechanical loading on bone. Recently, we and others have reported that increased fat mass could have adverse effects on bone mass because of deleterious metabolic effects [@pone.0065505-Kim2]--[@pone.0065505-Katzmarzyk1]. In addition, many researchers have suggested that weight gain and obesity could increase cortical bone mass through mechanical loading with a resultant decrease in trabecular bone due to metabolism or systemic effects [@pone.0065505-Katzmarzyk1]--[@pone.0065505-Migliaccio1]. Fatty acid lipotoxicity, numerous adipokines, inflammatory cytokines, aromatase, and insulin resistance have been suggested to mediate the deleterious effects of fat on bone metabolism [@pone.0065505-Katzmarzyk1]--[@pone.0065505-Elbaz1]. Therefore, the absence of an increase in trabecular bone mass observed in the ghrelin *ad lib*-fed group could in part result from the deleterious effects of increased fat mass. The ghrelin *ad lib*-fed group had a significantly decreased serum ghrelin concentration, likely due to a compensatory response to increased body weight. And this decrease in serum ghrelin could result in decreased bone formation, which would then neutralize the anabolic effect of ICV ghrelin on bone.
Central ghrelin is reported to have an effect on adipose tissue independent from its effects on food intake [@pone.0065505-TheanderCarrillo1], [@pone.0065505-SangiaoAlvarellos1], [@pone.0065505-PerezTilve1]. Therefore, the ghrelin pair-fed rats could have some changes in adipose tissue deposition, which could be independent of food intake and weight change. These changes in adipose tissue deposition could have some direct and indirect effect on the bone phenotype of these ghrelin pair-fed rats. Furthermore, the ghrelin effect on adipose tissue deposition may have contributed to the lack of weight gain in the ghrelin pair-fed rats. However, we did not investigate the adipose tissue phenotypes in the present study to determine the role of adipose tissue deposition in the central regulation of bone metabolism.
In contrast to the increased bone mass determined by microCT measurement, there were no significant changes in *ex vivo* BMD measured by DXA in tibia, femur, and lumbar spine among the three groups. Several possible explanations may account for the discrepancy. DXA measurement reflects both trabecular and cortical bone mass, whereas microCT gives separate estimates of BMD for trabecular and cortical bone as well as reports volumetric mineral density in g/cm^3^. Therefore, DXA measurements cannot identify the difference in trabecular bone observed by microCT. It has been reported that the measurement precision of the excised femur BMD is generally not as precise as that for the intact femur BMD *in vivo,* although the general efficacy of *ex vivo* DXA measurements were acceptable [@pone.0065505-Nagy1].
Although similar trends were observed, the effects of ICV ghrelin on femur and spine were not statistically significant. There could be a skeletal site-specific effect of ICV ghrelin. Interestingly, a previous study reported a skeletal site-specific effect of leptin with lower femur BMD and higher spine BMD observed in leptin deficient mice [@pone.0065505-Hamrick2]. However, no skeletal site-specific effect was observed in leptin receptor-deficient mice [@pone.0065505-Williams1].
We measured serum ghrelin to investigate possible leakage of ghrelin into systemic circulation from the ICV injection. However, we found a significant decrease in serum ghrelin in the ghrelin *ad lib*-fed group as compared to the control group. This result indicated that leakage to systemic circulation was unlikely. This decrease in serum ghrelin and increase in serum leptin could be a compensatory response to the weight gain of the ghrelin *ad lib*-fed group. However, the increase in serum leptin in ghrelin pair-fed group is not due to the weight gain, since there was no weight gain in the ghrelin pair-fed group. Another possible explanation is that the increase in serum leptin observed in the ghrelin *ad lib*-fed group and ghrelin pair-fed group is likely a compensatory response related to ICV ghrelin injection. Previous studies have reported a similar trend of increased blood leptin levels in ICV ghrelin injected animals [@pone.0065505-Kim1], [@pone.0065505-TheanderCarrillo1]. Circulating ghrelin and leptin may contribute to bone metabolism via a direct effect on bone cells. Generally, ghrelin increases both osteoblast and osteoclast function [@pone.0065505-Kim3]--[@pone.0065505-Fukushima1] and leptin increases osteoblast function but decreases osteoclast function [@pone.0065505-Holloway1], [@pone.0065505-Thomas1]. However, these direct effects of ghrelin and leptin were inconsistent depending on the concentrations, assays, and cell types used [@pone.0065505-Costa1], [@pone.0065505-Belloni1], [@pone.0065505-Kim4]. Therefore, the role of serum ghrelin on bone metabolism is complex, and future study specifically aimed to investigate the role of serum ghrelin on bone metabolism is required to determine this issue.
We observed no significant differences in CTX, TRAP-5b, or P1NP, which are serum biochemical markers of bone turnover. Since ghrelin was placed in a subcutaneously transplanted osmotic pump, it is possible that after 21 days of incubation at body temperature, the effect of ghrelin was diminished due to peptide degradation. Therefore, the serum biochemical markers measured after 21 days of treatment may not have captured the dynamic bone metabolism state during the treatment period. However, other possibilities could be limitations from the ELISA assay or serum preparation procedure.
The findings of the current study have important clinical implications since ghrelin or ghrelin mimetics could be utilized as potential therapeutic modalities for osteoporosis. A cross-sectional study showed that serum ghrelin positively correlated with BMD [@pone.0065505-Napoli1]. In addition, there are a number of clinical trials investigating the effects of ghrelin or ghrelin mimetics on various conditions, including sarcopenia, cancer-related cachexia, anorexia nervosa, cystic fibrosis, postoperative gastric ileus, and gastroparesis [@pone.0065505-Nass1]--[@pone.0065505-Akamizu1]. These studies have already validated the safety and efficacy of ghrelin or ghrelin mimetics.
First limitation in the present study is the dose of ICV ghrelin treatment (1.5 µg/day). This dose is slightly higher than some studies (1.0 µg/day and 1.2 µg/day) [@pone.0065505-Kim1], [@pone.0065505-Stevanovic1]. However, other studies have used higher doses (5--20 µg/day) [@pone.0065505-SangiaoAlvarellos1], [@pone.0065505-Stevanovic2], [@pone.0065505-PerezTilve1]. The dose of present study (1.5 µg/day) could be insufficient or subthreshold to induce adequate central ghrelin effect. Second limitation was the statistical power of the present study. Most of the findings are based on LSD post hoc comparisons, which is very liberal. Additional post hoc comparisons with Scheffe and Tukey HSD test were performed and some of the findings (trabecular number, cortical bone area, serum ghrelin, body weight and food intake) did not reach statistical significance. And, it should also be noted that some of the findings with tendency of P\<0.1 may be false positives. Third limitation was the conflicting fact that long duration of treatment was required to induce a sufficient bone metabolism change, whereas long duration of incubation in body temperature result in risk of ghrelin peptide degradation.
In conclusion, chronic central administration of ghrelin increases bone mass through a mechanism that is independent of body weight, suggesting that ghrelin may have a bone anabolic effect through the central nervous system.
Supporting Information {#s5}
======================
######
**Effect of chronic ICV ghrelin infusion on the femur trabecular bone phenotype.** (A) Trabecular bone volume expressed as percentage of total tissue volume (BV/TV). (B) Trabecular thickness (Tb.Th). (C) Trabecular number (Tb.N). (D) Trabecular separation (Tb.Sp). (E) Trabecular pattern factor (Tb.Pf). (F) Structure model index (SMI). (G) Trabecular volumetric BMD. (H) Representative microCT images of the distal femur. Three groups of rats (4--6 per group) were infused for 21 days with saline or ghrelin (1.5 μg/day). Rats infused with ghrelin were *ad lib*-fed or pair-fed (to saline-infused rats). \* P\<0.05; † P\<0.1. Data are presented as the mean ± SEM.
(TIF)
######
Click here for additional data file.
######
**Effect of chronic ICV ghrelin infusion on the spine trabecular bone phenotype.** (A) Trabecular bone volume expressed as percentage of total tissue volume (BV/TV). (B) Trabecular thickness (Tb.Th). (C) Trabecular number (Tb.N). (D) Trabecular separation (Tb.Sp). (E) Trabecular pattern factor (Tb.Pf). (F) Structure model index (SMI). (G) Trabecular volumetric BMD. (H) Representative microCT images of the lumbar spine. Three groups of rats (4--6 per group) were infused for 21 days with saline or ghrelin (1.5 μg/day). Rats infused with ghrelin were *ad lib*-fed or pair-fed (to saline-infused rats). \* P\<0.05; † P\<0.1. Data are presented as the mean ± SEM.
(TIF)
######
Click here for additional data file.
######
**Effect of chronic ICV ghrelin infusion on the tibia cortical bone phenotype.** (A) Total cross-sectional area inside the periosteal envelope (Tt.Ar). (B) Cortical bone area (Ct.Ar). (C) Cortical area fraction (Ct.Ar/Tt.Ar). (D) Medullary area (Ma.Ar). (E) Cortical thickness (Ct.Th). (F) Cortical porosity (Ct.Po). (G) Periosteal perimeter (Ps.Pm). (H) Endocortical perimeter (Ec.Pm). (I) Representative microCT images of the mid-diaphysis tibia. Three groups of rats (4--6 per group) were infused for 21 days with saline or ghrelin (1.5 μg/day). Rats infused with ghrelin were *ad lib*-fed or pair-fed (to saline-infused rats). \* P\<0.05; † P\<0.1. Data are presented as the mean ± SEM.
(TIF)
######
Click here for additional data file.
[^1]: **Competing Interests:**The authors have declared that no competing interests exist.
[^2]: Conceived and designed the experiments: HJC JHK JHA SWK SYK CSS. Performed the experiments: HJC KHK JYY BYJ JAS. Analyzed the data: HJC WYB JEK. Contributed reagents/materials/analysis tools: HJC KHK WYB JEK. Wrote the paper: HJC CSS.
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#S0001}
============
In February 2017, the World Health Organization (WHO) published a list of "critical" resistant bacterial species, including carbapenem-resistant *Pseudomonas aeruginosa, Acinetobacter baumannii*, and *Escherichia coli*. The same document urged the development of new, alternative treatments and/or new antimicrobials to reduce the number of deaths caused by multi-drug-resistant (MDR) bacteria[@CIT0001]
The antimicrobial peptide colistin was introduced in 1949 but in the 1980s it was withdrawn because of its nephrotoxicity and neurotoxicity. However, in the 2000s, clinicians again turned to colistin in an effort to treat infections caused by MDR Gram-negative bacteria[@CIT0002] Colistin is polycationic and interacts with negatively charged membrane components, including lipopolysaccharide (LPS), to disrupt cell membranes. Other activities contributing to colistin's bactericidal effect are ribosome binding, alterations of bacterial respiration, disturbances of the bacterial division, the induction of structural injuries, and the production of reactive oxygen species.[@CIT0003],[@CIT0004] Nevertheless, the precise mechanisms of action of colistin and other polymyxins are not yet fully understood.
Because the toxicity of colistin is dose-dependent,[@CIT0005] there is considerable interest in strategies that allow dose reductions while maintaining the therapeutic efficacy of the drug[@CIT0006] An additional consideration is the increasing frequency of colistin-resistant isolates. Among the resistance mechanisms identified thus far are modifications of LPS (*Pho*PQ and *Pmr*AB, a two-component regulatory system) and a transferable mechanism of resistance (*mcr*-1)[@CIT0007] The heterogeneous resistance (heteroresistance) of bacterial populations to colistin has also been described. However, a previous study suggested that the negative impact of colistin heteroresistance can be minimized by co-treatment with other antimicrobial agents[@CIT0008]
Synergism between colistin and other peptide molecules with a wide variety of antimicrobial activities has been explored. While colistin-carbapenem (β-lactams) is probably the best-studied combination,[@CIT0009]--[@CIT0011] synergies between colistin and quinolones, fosfomycin, and aminoglycosides have been reported as well[@CIT0012] Moreover, even though Gram-negatives are intrinsically resistant to linezolid and glycopeptides, their use in combination with colistin may be therapeutically promising. In fact, in vitro synergisms between colistin and linezolid against *A. baumannii* clinical strains[@CIT0013] and between colistin and glycopeptides in a *Galleria mellonella* model of infection[@CIT0014] were recently demonstrated. By disrupting the outer membrane to alter cytoplasmic membrane permeability, colistin allows both linezolid and glycopeptides to penetrate Gram-negative bacteria, which are otherwise impervious to these agents. A disturbance of the bacterial membrane may also underlie the observed synergies between colistin or colistin-like peptides and other, chemically unrelated antimicrobial agents.[@CIT0015] Thus, in this study we examined the role of colistin synergism with other antimicrobials with respect to: i) a reduction of the colistin dose and its impact on the drug's toxicity and ii) strategies allowing the avoidance of resistance development. Since the chloramphenicol resistance of MDR bacteria is mediated by the cellular efflux machinery, the focus of our study was the effect of colistin on cellular efflux pumps. Synergy between colistin and chloramphenicol was reported by Abdul Rahim et al[@CIT0016] Our results provide evidence of critical alterations in the efflux machinery of cells exposed to low concentrations of colistin and of the efficacy of other antimicrobials when administered in combination with colistin.
Materials and methods {#S0002}
=====================
Bacterial strains, culture conditions, and media {#S0002-S2001}
------------------------------------------------
The 24 clinical isolates used in this study differed in their clinical origins and included 7 strains of MDR *Pseudomonas aeruginosa*, 9 strains of MDR *Escherichia coli*, and 7 of MDR *Acinetobacter baumannii*. None of the strains were epidemiologically related. *Pseudomonas aeruginosa* ATCC 27853 and PAO1, *Escherichia coli* ATCC 25922, and *Acinetobacter baumannii* ATCC 17978 served as control strains. In addition, four strains of *Serratia marcescens* (two environmental and two clinical isolates) were used in the accumulation experiments.
The tryptone soy agar (TSA), tryptone soy broth (TSB), and Mueller-Hinton broth (MHB) used to culture the strains were purchased from Sharlau (Sentmenat, Barcelona, Spain).
Chemicals {#S0002-S2002}
---------
Colistin was kindly supplied by Zhejiang Shenghua Biok Biology Co., Ltd., (Shanghai, China), Chloramphenicol, amikacin, tetracycline, linezolid, acridine orange, the efflux pump inhibitor phenyl-arginyl-β-naphthylamide (PaβN), used in the accumulation assays, and the fluorescent probes diphenylhexatriene (DPH) and 1-\[4-trimethylamino-phenyl\]-6-phenyl-1,3,5-hexatriene (TMA-DPH) were purchased from Sigma-Aldrich Chemicals (Madrid, Spain). The LIVE/DEAD^®^ BacLight^™^ bacterial viability kit was from Life Technologies (Oregon, USA).
Intracellular accumulation of acridine orange {#S0002-S2003}
---------------------------------------------
Acridine orange (AO: N,N,N',N'-tetramethylacridine-3,6-diamine) produces fluorescence after it penetrates the bacterium and binds to its DNA. It is pumped out from bacterial cells by at least one efflux pump, as shown in *E. coli*[@CIT0017] and other Gram-negative bacteria[@CIT0018] In this study, AO uptake was determined: i) under standard culture conditions, ii) in the presence of the efflux pump inhibitor PaβN[@CIT0019] (20 µg/ml), and iii) in the presence of sub-inhibitory concentrations of colistin (0.25 and 0.125 µg/ml).
The assays were performed in flat-bottomed microtiter plates, with the same plate used for the three different conditions. Each well contained 100 µl of the bacterial inoculun and 100 µl of medium. Overnight cultures were diluted in Ringer's solution to an optical density (OD) of 1.5 at 520 nm. The inoculated plates were incubated for 1 h, after which fluorescence was measured in a FLUOstar OPTIMA Biogen fluorescent microplate reader. The results were expressed as the percentage increase in fluorescence vs that in wells containing bacteria and AO (=100%).
Susceptibility test {#S0002-S2004}
-------------------
The microdilution method was used to determine the minimum inhibitory concentrations (MICs) of colistin, chloramphenicol, tetracycline, vancomycin, and linezolid in each of the strains, according to EUCAST recommendations. Susceptibility was defined following EUCAST definition of antimicrobial breakpoints[@CIT0020] The MIC was also determined in the presence of the efflux pump inhibitor PaβN. Bacterial susceptibility to several different antimicrobials (chloramphenicol, tetracycline, vancomycin, and linezolid) in the presence of sub-inhibitory concentrations of colistin was evaluated as well.
Fluorescent live/dead testing {#S0002-S2005}
-----------------------------
To rule out bacterial death during the accumulation experiment and to show that AO fluorescence was due solely to an increase in the uptake of the dye, the viability of the bacteria in the presence of different concentrations of colistin was determined in a live-dead analysis.
Briefly, 1 ml of bacterial suspension (prepared as described above for the AO accumulation experiment) was incubated with colistin and centrifuged for 5 min at 18,000×g and then stained with 100 µl of the Live/Dead kit solution (LIVE/DEAD^®^ BacLight^™^). The cells were subsequently examined by confocal laser scanning microscopy (CLSM) using a ZEISS LSM 880 microscope equipped with a 488 nm argon laser, 561 nm and 633 nm diode lasers, and a 63× magnification oil immersion objective. All experiments were performed in duplicate. CLSM images were analyzed using the ZEN 2.3 software. The proportions of living vs dead bacteria were calculated.
Membrane polarization and anisotropy {#S0002-S2006}
------------------------------------
Bacterial cultures with or without colistin were washed twice in buffer and then suspended in saline buffer to a final OD~500~ of 0.4. After the addition of DPH or TMA-DPH to a final concentration of 2 μM and 1 μM, respectively, the cultures were incubated for 15 min at 37 ºC to enhance incorporation of the probes into the bacterial membranes. Anisotropy measurements were performed in cells incubated for 1 h in the presence of colistin.
Fluorescence was measured using a SLM-Aminco 8100 spectrofluorometer with a thermal jacketed cuvette holder. Fluorescence anisotropy (*r*) was calculated as: $$\documentclass[12pt]{minimal}
\usepackage{wasysym}
\usepackage[substack]{amsmath}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage[mathscr]{eucal}
\usepackage{mathrsfs}
\DeclareFontFamily{T1}{linotext}{}
\DeclareFontShape{T1}{linotext}{m}{n} {linotext }{}
\DeclareSymbolFont{linotext}{T1}{linotext}{m}{n}
\DeclareSymbolFontAlphabet{\mathLINOTEXT}{linotext}
\begin{document}
$$r = {{{I_{VV}} - G\cdot{I_{VH}}} \over {{I_{VV}} + 2G\cdot{I_{VH}}}}; G= {{{I_{VH}}} \over {{I_{HH}}}}$$
\end{document}$$
where *I*~ij~ is the fluorescence intensity measured when the excitation (*i*) and emission (*j*) polarizers are fixed in the vertical (*V*) or horizontal (*H*) position and *G* is the ratio of the sensitivities of the detection system for vertically and horizontally polarized light. Samples fluorescence emission was measured at 37 °C using excitation and emission wavelengths of 381 nm and 426 nm, respectively[@CIT0021]
Results {#S0003}
=======
Susceptibility testing {#S0003-S2001}
----------------------
The MICs for the antimicrobials investigated in this study are shown in [Table 1](#T0001){ref-type="table"}. All tested strains were susceptible to colistin according to the breakpoint defined by EUCAST. [Table 1](#T0001){ref-type="table"} also shows the effect of sub-inhibitory concentrations of colistin on the MICs of other antimicrobials. The activities of all antimicrobials tested increased when used in combination with colistin at ½ of its MIC. The greatest reduction in the MICs of these drugs was obtained in *A. baumannii*, with moderate reductions determined in *P. aeruginosa* and *E. coli*. No significant deviations were measured in the different series of experiments.Table 1Bacterial susceptibility to antimicrobials administered either alone or in combination with colistin at ½ its MIC*Acinetobacter baumannii*COLCHLCHL+COLTETTET+COLLZDLZD+COLVANVAN+COL**ABAU 137**164≤0.125\>1281\>2564\>2561**ABAU 80**1640.5\>1281\>25682562**ABAU 226**11280.541\>25616\>2568**ATCC 17978**1640.520.25\>256162564**ABAU 8**164≤0.125\>1280.5\>256162561**ABAU RUH-875**1\>1280.25320.5\>25681284**ABAU 176**11280.5\>1281\>256162568**ABAU 34**0.5640.5\>1282\>25616\>2562***Pseudomonas aeruginosa*PA 362**0.2564442\>256642562**PA 350**0.58484256321282**PA 056**0.5642164\>25632\>2568**PA 846**0.258440.5128161288**PA O1**0.516284\>2563212816**PA 023**0.580.25162\>25664\>2568**PA 666**11282162\>2563212816**PA 086**0.580.581\>2566425616**ATCC 27853**0.56441281\>256642568***Escherichia coli.*MDR 211453**0.54411256322568**MDR 39255**0.54112864256322568**MDR 208691**0.5\>1282\>12832256322568**MDR 205119**0.5641\>128161281612816**MDR 246415**1\>12816\>1282256321288**MDR 239910**116212816256162568**MDR 238192**0.512832\>128322561625616**MDR 215482**1\>12864\>12816\>2568\>2564**MDR 213873**14412816128646416**ATCC 25922**0.51110,251281612816[^1][^2]
Acridine orange accumulation {#S0003-S2002}
----------------------------
Acridine orange penetrates Gram-negative bacteria easily, although it is normally expelled by efflux pumps located in the bacterial membrane[@CIT0022] However, as shown in [Figure 1A](#F0001){ref-type="fig"}--[C](#F0001){ref-type="fig"}, in the presence of PaβN, an efflux pump inhibitor, the uptake of AO increased by nearly 200% in *E. coli* and by \~170% in *P. aeruginosa* and *A. baumannii*. Similar effects were achieved with sub-inhibitory concentrations of colistin (0.25 µg/ml), which increased AO uptake in *E. coli* by \~160%, in *P. aeruginosa* by 140%, and in *A. baumannii* by \>125%. The accumulation of AO was concentration-dependent, at least at the colistin concentrations tested (0.125 and 0.25 µg/ml). Moreover, low concentrations of colistin also increased AO accumulation by S*erratia marcescens*, a bacterium intrinsically resistant to colistin ([Figure 1D](#F0001){ref-type="fig"}). Accumulation values obtained were highly repetitive. In [Figure 1](#F0001){ref-type="fig"} standard deviations are represented.Figure 1Average acridine orange (AO) accumulation in the presence of the efflux inhibitor PaβN and sub-inhibitory concentrations of colistin. (**A**--**C**) colistin-sensitive strains; (**D**) a species intrinsically resistant to colistin.
To rule out that these results reflected differences in bacterial viability, a live/dead assay was conducted and the results were evaluated by CLSM ([Figure 2](#F0002){ref-type="fig"}). Bacterial death in the presence of colistin was \~6%, the same as in the control (without colistin). Thus, at the tested sub-inhibitory concentrations, the membrane effects of colistin did not cause a reduction in cell viability. In addition, propidium iodide was unable to penetrate the colistin-treated bacteria.Figure 2Confocal light microscopy imaging after Live/Dead staining of *Pseudomonas aeruginosa*. (**A**) Untreated cells; (**B**) cells treated with colistin at 0.25 μg/mL and (**C**) cells treated with colistin at bactericidal concentration. Similar images were obtained for *Acinetobacter baumanii* and *E. coli* (not shown). TEM of thin sections of *S. marcescens* after treatment with colistin. Blebs formed by the outer membrane are easily visualized.
Membrane polarization {#S0003-S2003}
---------------------
Anisotropy is the inverse of microviscosity. It can be measured in membranes using the hydrophobic probe DPH, which enters the inner domain of the lipid bilayer and becomes lodged parallel to the hydrocarbon chain axis of the phospholipids but also in the centre of the bilayer, parallel to the surface. Trimethylammonium (TMA)-DPH, the cationic derivative of DPH, localizes closer to the external surface of the lipid membrane, as the cationic TMA substituent acts as a surface anchor. [Table 2](#T0002){ref-type="table"} shows the DPH- and TMA-DPH-based anisotropy values in the three tested species of bacteria incubated with 2 and 0.25 µg colistin/ml. In *E. coli* MDR208691, a 1 h incubation with colistin reduced membrane anisotropy, while in *P.aeruginosa* and *A. baumannii* the reductions were lower (see DPH results in [Table 2](#T0002){ref-type="table"}). Using the TMA_DPH probe, subtle effects of colistin on the membrane of *E. coli* and *P. aeruginosa* were seen, while in *A. baumannii* colistin slightly rigidified the surface of the lipid membrane.Table 2Changes in membrane fluidity (anisotropy) in the presence of colistinColistin (μg/ml)Anisotropy (*r*)DPHTMA-DPH***E. coli***00.2787±0.02190.267±0.0390.250.26986±0.0200.265±0.0282.000.2488±0.0090.2681±0.0212***P. aeruginosa***00.3139±0.0340.284±0.02470.250.3105±0.03590.274±0.01992.000.2877±0.02130.273±0.0212***A. baumannii***00.314±0.0270.255±0.0050.250.306±0.0290.260±0.00812.000.307±0.00250.260±0.006
Discussion {#S0004}
==========
Although the toxicity of colistin resulted in the eventual discontinuation of its use as an antibiotic, with the emergence of bacterial multi-resistance the drug has been resurrected for the treatment of infections by MDR bacteria. However, the re-introduction of colistin has been accompanied by the appearance of resistance to colistin and other polymyxins, most commonly via LPS modification. A further challenge is the heteroresistance of bacterial populations; that is, the presence of resistant subpopulations within an isolate considered susceptible.[@CIT0006],[@CIT0007]
To avoid the development of colistin resistance, the colistin is used in combination with a second antimicrobial agent, based on the advantage conferred by the main mechanism of colistin action: membrane permeabilization. Ritcher et al[@CIT0023] described the increased accumulation of several antimicrobials in the presence of high concentrations of colistin (6 mM =7 µg/ml). In our study, sub-MIC concentrations of colistin were used to explore their ability to enhance the action of other antibiotics while minimizing the risk posed by colistin toxicity. The results showed that sub-MIC concentrations of colistin altered the efflux pump function of the bacterial cells, as evidenced by the increases in AO accumulation in isolates of *E. coli, A. baumannii*, and *P. aeruginosa*. All the three species are naturally susceptible to colistin, which at concentrations at or above its MIC causes catastrophic effects on the outer membrane. Moreover, similar results were obtained in *S. marcescens*, a Gram-negative species that is naturally colistin-resistant. Specifically, we were able to demonstrate severe alterations in AO efflux in *S. marcescens* cells treated with sub-inhibitory concentrations of colistin. This result suggests a mechanism of outer membrane injury in Gram-negative bacteria in which correct formation of the AcrAB-tolC complex and other efflux-pump-based complexes is prevented and efflux is therefore interrupted.
Diffusion and active transport by efflux systems are strongly dependent on membrane fluidity[@CIT0024] Vincent et al[@CIT0025] described changes in membrane polarization in bacterial cells exposed to tetracycline. As the main mechanism of action of colistin is membrane disruption, we hypothesized that concentrations of the drug below the MIC would be sufficient to alter membrane fluidity. Our results showed that low concentrations of colistin significantly increased the fluidity of the inner regions of the lipid bilayers without significantly modifying the more polar regions of the membrane, as indicated by anisotropy measurements. These colistin-induced membrane alterations would presumably i) facilitate the penetration of other antimicrobials and thus increase their biological activity and ii) decrease the lateral lipid pressure inside lipid membranes to reduce local rigidity and modify the correct packing of the efflux pumps, thus altering the efflux of drugs reaching the cell, as observed in [Figure 1](#F0001){ref-type="fig"}.
Using a live/dead assay, we showed that, at 1/2 its MIC, colistin did not cause bacterial death or injury to the cytoplasmic membrane, as the amount of propidium iodide and thus the proportion of living bacteria were identical in colistin-treated and untreated (control) cells. Therefore, the above-described results cannot be explained by differences in bacterial viability.
The cellular accumulation of AO increases in the presence of the efflux pump inhibitor PaβN since the dye is normally pumped out by efflux systems. Sub-inhibitory concentrations of colistin also led to increases in the amount of AO retained by the bacteria. Given the results of the live/dead test, the observed accumulation of AO can be attributed to the disruption of membrane permeability by colistin and to the indirect inhibition of the cellular efflux machinery.
We then demonstrated that sub-inhibitory concentrations of colistin increase the activity of antimicrobials whose mechanisms of resistance are associated with efflux systems and/or decreased permeability ([Table 1](#T0001){ref-type="table"}). The synergy between clinically relevant concentrations of colistin and chloramphenicol in *Klebsiella pneumoniae*[@CIT0016] as well as between colistin and colistin-like peptides and carbapenems in imipenem-resistant *P. aeruginosa*[@CIT0015] has been reported. An enhancement of the activity of glycopeptides targeting *A. baumannii* was also shown in vivo, in the moth *Galleria mellonella*, in which colistin increased the antimicrobial activity of vancomycin by up to 90%.[@CIT0014],[@CIT0026] Moreover, Liu et al[@CIT0013] reported the efficacy of anti-Gram-positive antimicrobials in the treatment of *A. baumannii* infections when these drugs were administered together with colistin. During the writing of the present article, Brennan-Krohn et al[@CIT0027] published a study demonstrating the synergism of colistin and a wide range of antimicrobials in both colistin-susceptible and colistin-resistant Enterobacteriaceae. The authors attributed these results to the direct effect of colistin on the bacterial outer membrane. Almost simultaneously, Fang et al[@CIT0028] reported similar data in tests of *E. coli* exposed to combinations of colistin and either azithromycin or rifampin, both in vitro and in vivo. Our data provide additional, biophysical evidence supporting those observations. Furthermore, these results together suggest lines of further research aimed at the development of compounds that, while unable to kill bacteria themselves, significantly alter bacterial membrane fluidity and thus potentiate the activity of traditional antimicrobials.
In summary, our study showed that, by altering the permeability of the outer membrane of Gram-negative bacteria, colistin acts synergistically with a broad spectrum of antimicrobials. The mechanism of action involves modification of membrane fluidity as well as the direct partial inhibition of efflux pump function.
The authors thank Dr. Benjamin Torrejón, from the Scientific and Technological Centers of the University of Barcelona (Bellvitge Campus), for his support in CLSM imaging. This work was financed by funding from the University of Barcelona. MV is a member of the ENABLE consortium. The research of this team is supported by the TV3-Marato Foundation (2018).
Author contributions {#S0005}
====================
All authors contributed to data analysis, drafting or revising the article, gave final approval of the version to be published, and agree to be accountable for all aspects of the work.
Disclosure {#S0006}
==========
The authors have no conflicts of interest to declare in this work.
[^1]: **Note:** MICs are reported in µg/ml.
[^2]: **Abbreviations:** COL, colistin; CHL, chloramphenicol; TET, tetracycline; LZD, linezolid; VAN, vancomycin.
| {
"pile_set_name": "PubMed Central"
} |
{#sp1 .757}
| {
"pile_set_name": "PubMed Central"
} |
Introduction
============
Antidepressant medications such as selective serotonin reuptake inhibitors and serotonin-norepinephrine reuptake inhibitors (SNRIs) are the first-line treatments for patients with moderate to severe major depressive disorder (MDD).[@b1-ndt-14-1261]--[@b3-ndt-14-1261] In 1993, venlafaxine was approved by the US Food and Drug Administration as the first SNRI for the treatment of MDD in adults,[@b4-ndt-14-1261] and is now marketed in more than 90 countries. In 2015, venlafaxine extended release (ER) became the third SNRI to receive regulatory approval for the treatment of MDD in Japan, thereby driving interest in data specifically addressing Japanese patients treated with venlafaxine ER.
Higuchi et al[@b5-ndt-14-1261] reported the primary results of an 8-week, double-blind, placebo-controlled, randomized, Phase III study of 538 patients conducted in Japan using fixed-dose (75 mg/day) and flexible doses (75--225 mg/day) of venlafaxine ER. The study findings showed a statistically significant difference in the change from baseline in Hamilton Rating Scale for Depression--17 items (HAM-D~17~) total score[@b6-ndt-14-1261] in the fixed-dose group (−10.76; *P*=0.031), but not in the flexible-dose group (−10.37; *P*=0.106), compared with the placebo group (−9.25). However, the flexible-dose group showed statistically significant treatment benefits compared with the placebo group (*P*\<0.05) in several secondary endpoints, such as the Montgomery--Asberg Depression Rating Scale (MADRS)[@b7-ndt-14-1261] and the Hamilton Rating Scale for Depression--6 items (HAM-D~6~).[@b8-ndt-14-1261],[@b9-ndt-14-1261] An examination of the HAM-D~17~ revealed poor improvement in sleep disturbance scores (items 4--6)[@b10-ndt-14-1261] for the flexible-dose group after Week 4, possibly due to the norepinephrine effect of venlafaxine at high doses.
Given these results, we performed exploratory post hoc subgroup analyses in order to explore the potential factors that may impact the efficacy of venlafaxine ER and treatment differences between the fixed-dose (75 mg/day) and the flexible-dose (75--225 mg/day) in patients with MDD.
Materials and methods
=====================
Study design
------------
The original study was a multicenter, randomized, double-blind, placebo-controlled, parallel-group, Phase III study to evaluate the efficacy and safety of venlafaxine ER 75 mg/day (fixed-dose) and venlafaxine ER 75--225 mg/day (flexible-dose), compared with placebo ([ClinicalTrials.gov](http://ClinicalTrials.gov): NCT01441440).[@b5-ndt-14-1261] After a 2-week screening period, eligible patients were randomized in a 1:1:1 ratio to each treatment group for 8 weeks, followed by a 2-week tapering period.
Subjects
--------
Eligible patients included outpatients aged 20 years or older who had a primary diagnosis of MDD based on criteria of the *Diagnostic and Statistical Manual of Mental Disorders*, fourth edition, and had experienced single (≥90 days) or recurrent episodes (≥28 days) of depression without psychotic features. At both screening and baseline, patients were required to have scored at least 26 on the MADRS questionnaire and the change in MADRS total score from screening to baseline was required not to have exceeded 25%. Finally, at both screening and baseline, patients were required to have a 16-item Quick Inventory of Depressive Symptomatology self-report version (QIDS~16~-SR-J)[@b11-ndt-14-1261] total score and a Clinical Global Impressions--Severity scale (CGI-S) score of at least 16 and 4, respectively.
Settings
--------
The original study was conducted at 62 investigational sites from November 2011 to March 2014 in Japan. All applicable documentation including the study protocol was approved by the Institutional Review Board and Independent Ethics Committee at each site ([Box S1](#SD1-ndt-14-1261){ref-type="supplementary-material"}). The study was conducted in agreement with all legal and regulatory requirements and the International Ethical Guidelines for Biomedical Research Involving Human Subjects (Council for International Organizations of Medical Sciences 2002), Guidelines for Good Clinical Practice, and the Declaration of Helsinki. Written informed consent was provided by all patients before study participation.
Treatment
---------
The details of drug administration and dose titration have been reported elsewhere.[@b5-ndt-14-1261] Initial dose of venlafaxine was 37.5 mg/day. In Week 1, the dose could be increased to 75 mg/day, and in Week 2, based on tolerability, it could be increased to 150 mg/day in the flexible-dose group. Further, in Week 3, the dose was force titrated to 225 mg/day in the flexible-dose group, in spite of an acceptable response at a lower dose. Dose reduction was allowed in case of intolerance to higher doses. We simply compared the point estimates among the subgroups to explore the potential factors that may impact the efficacy of venlafaxine ER and treatment differences between the fixed-dose (75 mg/day) and the flexible-dose (75--225 mg/day) groups.
Assessments
-----------
The results of all efficacy endpoints in this study have been previously published.[@b5-ndt-14-1261] In this report, HAM-D~17~, HAM-D~6~, and MADRS were used for subgroup analyses. The formula, Σ (items 1, 2, 7, 8, 10, and 13), was used to measure HAM-D~6~ score.
Statistical analysis
--------------------
As the current investigation was exploratory in nature, no statistical hypothesis testing was used. The following efficacy endpoints were analyzed: change from baseline at Week 8 in the HAM-D~17~ total score, HAM-D~6~ score, and MADRS total score. The following factors were considered for the subgroup analyses: sex (male/female); age (≤37/\>37 years \[median\]); HAM-D~17~ total score at baseline (total score ≤22/total score \>22 \[median\]); duration of MDD, defined as the duration of MDD since the occurrence of the first episode (≤22 months/\>22 months \[median\], \<12 months/≥12 months, and \<48 months/≥48 months); duration of the current depressive episode (≤6.6 months/\>6.6 months \[median\]); history of previous depressive episodes (0 \[single\]/≥1 \[recurrence\]); history of previous medications for MDD (0 \[no medication\]/≥1 \[medicated\]); and CYP2D6 phenotypes (ultra-rapid/extensive and intermediate/poor metabolizers). For each subgroup, an ANCOVA model with the treatment group as a factor and the baseline value of the respective efficacy endpoint as a covariate was fitted. Based on the ANCOVA model, the adjusted mean and its corresponding 95% CI of the treatment effect (defined as a difference from placebo) was computed for each active treatment group.
The full analysis set (FAS) was used throughout the analyses. The FAS was defined as all patients who received at least one dose of the study drug and had baseline measurement and at least one post-baseline measurement of the primary efficacy endpoint was used for the analysis. Missing values in the endpoints at Week 8 were imputed using the last observation carried forward (LOCF) algorithm. Assessments prior to the first dose of the study medication were not eligible to be carried forward.
Results
=======
Patient disposition
-------------------
Of 538 randomized patients, 537 patients received the study drug (fixed-dose group, 174; flexible-dose group, 179; placebo, 184). Overall, 475 patients completed the study period (fixed-dose, 151; flexible-dose, 158; placebo, 166). Patients in all groups had comparable demographic and baseline MDD characteristics. The mean (SD) age in the three treatment groups ranged from 38.3 (10.2) to 38.6 (11.1) years, and the mean (SD) duration of MDD ranged from 40.3 (50.0) to 52.6 (62.9) months. The mean (SD) baseline HAM-D~17~ total score and MADRS total score ranged from 22.4 (4.1) to 22.6 (4.1) and from 32.6 (4.4) to 33.2 (5.1), respectively. The distribution of the last dose during the 8-week treatment period was 4.5%, 4.5%, 10.6%, and 80.4% for 37.5 mg/day, 75 mg/day, 150 mg/day, and 225 mg/day, respectively, in the flexible-dose group.
Efficacy
--------
[Figures 1](#f1-ndt-14-1261){ref-type="fig"}[](#f2-ndt-14-1261){ref-type="fig"}--[3](#f3-ndt-14-1261){ref-type="fig"} show forest plots of the differences between each treatment group and the placebo group in the adjusted mean change from baseline at Week 8 (LOCF) in the HAM-D~17~ total score ([Figure 1](#f1-ndt-14-1261){ref-type="fig"}), HAM-D~6~ score ([Figure 2](#f2-ndt-14-1261){ref-type="fig"}), and MADRS total score ([Figure 3](#f3-ndt-14-1261){ref-type="fig"}). Both venlafaxine ER groups showed greater treatment benefits compared with the placebo group for all efficacy measures in males than in females; in patients with a low HAM-D~17~ total score at baseline (≤22) than in those with a high HAM-D~17~ total score at baseline (\>22); in patients with a long duration of MDD (\>22 months) than a short duration (≤22 months); in patients with a short duration of current episode (≤6.6 months) than a long duration (\>6.6 months); and in patients with one or more previous episodes than none ([Figures 1](#f1-ndt-14-1261){ref-type="fig"}[](#f2-ndt-14-1261){ref-type="fig"}--[3](#f3-ndt-14-1261){ref-type="fig"}). There were no consistent trends in the subgroup of patients with or without a history of previous medications for MDD, in the subgroup of patients in two age categories, and in the subgroup of patients with the two CYP2D6 phenotypes.
Although the flexible-dose group showed a favorable treatment effect compared with placebo in most subgroups, the flexible-dose group showed a relatively smaller treatment effect than the fixed-dose group in most subgroups. However, a greater treatment effect was seen in the flexible-dose group versus the fixed-dose group for all efficacy measures in subgroups of older patients (\>37 years), and in patients with a long duration of MDD (\>22 months), and a short duration of current depressive episode (≤6.6 months) ([Figures 1](#f1-ndt-14-1261){ref-type="fig"}[](#f2-ndt-14-1261){ref-type="fig"}--[3](#f3-ndt-14-1261){ref-type="fig"}). Similar trends were observed in duration of MDD in an analysis that divided the groups into durations of 12 months and 48 months ([Figures 4](#f4-ndt-14-1261){ref-type="fig"} and [5](#f5-ndt-14-1261){ref-type="fig"}, respectively).
Demographic and disease characteristics stratified by duration of MDD (≤22 months/\>22 months) are presented in [Table 1](#t1-ndt-14-1261){ref-type="table"}. Overall, demographic characteristics and baseline HAM-D~17~ and MADRS total scores were comparable between the subgroups in each treatment group. Patients with a long duration of MDD (\>22 months) had more frequent previous depressive episodes and a longer duration of the current depressive episode than patients with a short duration of MDD (≤22 months).
Discussion
==========
Although these were exploratory, post hoc subgroup analyses of a placebo-controlled clinical study, our results highlight several potential factors that may impact the efficacy of venlafaxine ER in Japanese populations.
As in comparison between subgroups regardless of the treatment groups, there was no meaningful difference in the effect of venlafaxine ER compared with placebo for different degrees of severity of depression at baseline. As shown in [Figure 6](#f6-ndt-14-1261){ref-type="fig"}, patients with more severe depression (HAM-D~17~ total score at baseline \>22) showed a greater decrease in the mean change from baseline in the HAM-D~17~ total score among all groups, including the placebo group. While some studies have shown no relationship between severity of depression and treatment response to antidepressants in MDD,[@b12-ndt-14-1261],[@b13-ndt-14-1261] others have shown greater antidepressant--placebo differences in patients with more severe depression.[@b14-ndt-14-1261]--[@b17-ndt-14-1261] This discrepancy is thought to be due to the extent of baseline score inflation.[@b18-ndt-14-1261],[@b19-ndt-14-1261] In our study, patients were first screened with MADRS, QIDS~16~-SR-J, and CGI-S to mitigate baseline score inflation, and a different score (HAM-D~17~) was used as the primary efficacy measure, although previous studies have used the same score both as an inclusion criterion and the primary efficacy measure. Patients with low baseline HAM-D~17~ scores (≤22) showed greater treatment benefits than patients with high scores (\>22) in both treatment groups compared with the placebo group, because placebo response was much higher in patients with a high baseline HAM-D~17~ score (\>22), particularly in changes from baseline in HAM-D~17~ and MADRS total scores, which contributed to smaller differences between each treatment group and the placebo group in those patients.
Patients with a long duration of MDD (\>22 months) showed greater treatment benefits than those with a short duration of MDD (≤22 months) in both venlafaxine ER treatment groups compared with the placebo group. Placebo response was much smaller in patients with a long duration of MDD ([Figure 4](#f4-ndt-14-1261){ref-type="fig"}); these results were consistent with a meta-analysis of ten clinical trials[@b13-ndt-14-1261] as well as other prior studies.[@b20-ndt-14-1261],[@b21-ndt-14-1261] A small placebo response in patients with a long duration of MDD may have contributed to the larger differences observed between each treatment group and the placebo group. Similar trends were seen in an analysis that divided the groups at 12 months and 48 months ([Figures 4](#f4-ndt-14-1261){ref-type="fig"} and [5](#f5-ndt-14-1261){ref-type="fig"}, respectively).
Comparing the effect between the treatment groups in each subgroup, greater treatment effect was seen in the flexible-dose group versus the fixed dose group in the subgroup with a long duration of MDD. These results may prove the hypothesis that a greater treatment response to venlafaxine ER (up to 225 mg/day) can be seen in patients with a longer duration of MDD. A similar observation was seen in patients in age subgroups. In the older patient (\>37 years) subgroup, the effect in the flexible-dose group was greater than that in the fixed-dose group. This result may be considered to be associated with the fact that the mean duration of MDD in the \>37 years subgroup was longer than that in the ≤37 years subgroup (data not shown).
Furthermore, in both venlafaxine ER treatment groups, patients with recurrent depressive episodes showed greater treatment benefits than those with a single depressive episode compared with the placebo group. The placebo response was much smaller in patients with recurrent episodes ([Figure 7](#f7-ndt-14-1261){ref-type="fig"}), which corresponds with the results of a meta-analysis of seven clinical trials.[@b22-ndt-14-1261] A small placebo response may have contributed to the larger differences between the venlafaxine ER groups and the placebo group in patients with recurrent depressive episodes. These results were similar to those for duration of MDD reported previously ([Figure 4](#f4-ndt-14-1261){ref-type="fig"}), as patients with a long duration of MDD (\>22 months) experienced more depressive episodes ([Table 1](#t1-ndt-14-1261){ref-type="table"}).
Interestingly, patients with a long duration of the current depressive episode (\>6.6 months) versus a short duration (≤6.6 months) showed smaller treatment benefits in both venlafaxine ER treatment groups compared with the placebo group. Treatment benefit of venlafaxine 75--225 mg/day was nearly absent in patients with a long duration of the current depressive episode ([Figure 8](#f8-ndt-14-1261){ref-type="fig"}). Since the inclusion criteria for the original study required a current depressive episode (≥90 days, single episode; ≥28 days, recurrent episode) before the screening visit, patients with more severe or treatment-resistant depression with a single episode may have been included in the subgroup. However, patients with a short duration of the current depressive episode consistently showed greater treatment benefits in the flexible-dose group than the fixed-dose group compared with the placebo group, which was also consistent with the findings observed in patients with a long duration of MDD (\>22 months) ([Figure 4](#f4-ndt-14-1261){ref-type="fig"}).
Finally, in recent years, the response to placebo observed in studies of antidepressants for MDD has increased.[@b23-ndt-14-1261] Placebo response also increases as the number of active medication arms,[@b24-ndt-14-1261] investigational sites,[@b25-ndt-14-1261] and study visits[@b26-ndt-14-1261] increase, and as the number of academic sites[@b27-ndt-14-1261] decreases. The number of study visits influences dropout rate rather than treatment response;[@b28-ndt-14-1261] that is, an increasing number of study visits increases the dropout rate. Additionally, site ratings, not centralized ratings, increase placebo response.[@b29-ndt-14-1261] The present study had two active medication arms, seven visits during the 8-week treatment phase, and was conducted with the site rating method at 62 investigational sites. These factors might have contributed considerably to the high placebo response in this study. In addition, it is important to remember that drug adherence is often low among MDD outpatients. In this study, drug adherence was determined through capsule-counting procedures and patient--physician interviews at follow-up visits.
In conclusion, despite a high placebo response, venlafaxine ER improved symptoms of MDD compared with placebo among most subgroups. It is hypothesized that patients with a longer duration of MDD may have a greater treatment response at a higher dose of venlafaxine ER (up to 225 mg/day). Further investigation in real-world settings in Japan is needed to evaluate patient groups that require higher doses of venlafaxine ER.
Data sharing statement
======================
Anonymized patient-level data will be available through the Pfizer Inc. data access request site: <http://www.pfizer.com/research/clinical_trials/trial_data_and_results/data_request>.
Supplementary material
======================
Box S1List of approving institutional review boards and ethics committeesShinagawa East One Medical Clinic institutional review board (IRB)Yokohama Minoru Clinic IRBOmuta Memorial Hospital IRBNippon Medical School, Chiba Hokusoh Hospital IRBHayakawa Clinic IRBSuzuki Hospital IRBImazato Gastroenteric Hospital IRBRiverside Clinic IRBTokyo-Eki Center Building Clinic IRBYuge Hospital IRBShinagawa Clinic IRBNational Hospital Organization central review boardHaradoi Hospital IRBSakayori Clinic IRBMiki Mental Clinic IRBTokyo Women's Medical University Hospital IRBHimorogi Psychiatric Institute IRBTokyo Kosei Nenkin Hospital IRBAino Clinic IRBEda Memorial Hospital IRBWarakukai Akasaka Clinic IRBHimeno Tomomi Clinic IRBMizuo Clinic IRBSuzuki Internal & Circulatory Medical Clinic IRBYutaka Clinic IRBKayaba Dermatology Clinic IRB
This study was sponsored by Pfizer Inc. Editorial support was provided by Pearl Gomes of Cactus Communications and was funded by Pfizer Inc. We are grateful for the contributions of trial participants, principal investigators, and all other medical personnel to this study. Full control of the manuscript content was retained by all authors.
**Disclosure**
YA, YH, RI, and TI are employees of Pfizer Japan Inc. KK was an employee of Pfizer Japan Inc. at the time of the study. YW has received speaker's honoraria from Pfizer Japan Inc., GlaxoSmithKline K.K., Otsuka Pharmaceutical Co., Ltd., Janssen Pharmaceutical K.K., Meiji Seika Pharma Co., Ltd, Eli Lilly Japan K.K., MSD K.K. a subsidiary of Merck & Co., Inc., Takeda Pharmaceutical Co., Ltd., and Mochida Pharmaceutical Co., Ltd., Mitsubishi Tanabe Pharma Corp., and Sumitomo Dainippon Pharma Co., Ltd., within the past 5 years. The authors report no other conflicts of interest in this work.
{#f1-ndt-14-1261}
{#f2-ndt-14-1261}
{#f3-ndt-14-1261}
{#f4-ndt-14-1261}
{#f5-ndt-14-1261}
{#f6-ndt-14-1261}
![Adjusted mean change from baseline at Week 8 with 95% CIs by history of previous depressive episodes (0 \[single episode\]/≥1 \[recurrent episodes\]) in each efficacy endpoint (FAS, LOCF, ANCOVA model).\
**Abbreviations:** FAS, full analysis set; HAM-D~6~, Hamilton Rating Scale for Depression--6 items; HAM-D~17~, Hamilton Rating Scale for Depression--17 items; LOCF, last observation carried forward; MADRS, Montgomery--Asberg Depression Rating Scale.](ndt-14-1261Fig7){#f7-ndt-14-1261}
{#f8-ndt-14-1261}
######
Demographic and disease characteristics by duration of MDD (≤22 months/\>22 months) (FAS)
Duration of MDD (months)
---------------------------------------------------- -------------------------- ------------- ----------- -----------
Sex Male Female Male Female
Placebo 43 (48.9) 45 (51.1) 48 (50.0) 48 (50.0)
Venlafaxine ER 75 mg/day 48 (50.0) 48 (50.0) 36 (46.2) 42 (53.8)
Venlafaxine ER 75--225 mg/day 41 (49.4) 42 (50.6) 51 (54.3) 43 (45.7)
Age, years
Placebo 39.2 (12.0) 38.0 (10.2)
Venlafaxine ER 75 mg/day 37.8 (12.6) 39.2 (10.8)
Venlafaxine ER 75--225 mg/day 38.0 (10.3) 38.4 (10.3)
Weight, kg
Placebo 62.6 (17.2) 60.8 (14.8)
Venlafaxine ER 75 mg/day 61.4 (14.1) 62.2 (14.7)
Venlafaxine ER 75--225 mg/day 60.8 (12.9) 63.7 (15.5)
BMI, kg/m^2^
Placebo 22.8 (4.5) 22.2 (4.0)
Venlafaxine ER 75 mg/day 22.3 (4.1) 23.1 (4.6)
Venlafaxine ER 75--225 mg/day 22.6 (4.0) 23.5 (4.7)
Duration of MDD, months
Placebo 9.8 (5.2) 75.7 (50.1)
Venlafaxine ER 75 mg/day 9.3 (5.2) 78.5 (53.9)
Venlafaxine ER 75--225 mg/day 7.9 (4.0) 89.7 (64.3)
Duration of the current depressive episode, months
Placebo 8.1 (4.8) 14.1 (18.7)
Venlafaxine ER 75 mg/day 7.9 (4.8) 17.3 (24.2)
Venlafaxine ER 75--225 mg/day 7.3 (4.0) 16.3 (22.3)
Number of previous depressive episodes
Placebo 0.2 (0.4) 1.3 (1.0)
Venlafaxine ER 75 mg/day 0.1 (0.3) 1.6 (1.7)
Venlafaxine ER 75--225 mg/day 0.1 (0.2) 1.4 (1.4)
Baseline HAM-D~17~ total score
Placebo 22.0 (3.5) 22.8 (4.6)
Venlafaxine ER 75 mg/day 22.8 (3.8) 22.5 (4.3)
Venlafaxine ER 75--225 mg/day 22.4 (3.9) 22.3 (4.1)
Baseline MADRS total score
Placebo 32.8 (4.9) 33.5 (5.3)
Venlafaxine ER 75 mg/day 32.6 (4.2) 32.7 (4.7)
Venlafaxine ER 75--225 mg/day 32.5 (4.7) 33.0 (4.8)
**Notes:** Sex is described as number of patients (%); all other parameters are described as mean (SD).
**Abbreviations:** BMI, body mass index; FAS, full analysis set; HAM-D~17~, Hamilton Rating Scale for Depression--17 items; MADRS, Montgomery--Asberg Depression Rating Scale; MDD, major depressive disorder.
| {
"pile_set_name": "PubMed Central"
} |
Introduction
============
A well aligned and flexible spine is important for an active and healthy life. Chronic low back pain (CLBP) is defined as back pain lasting more than 12 weeks \[[@B1]\]. This pain is a common musculoskeletal disorder affecting 80% of people some time in their lives \[[@B2]\]. The majority of lower back pain stems from benign musculoskeletal problems and is referred to as non-specific low back pain; it may be due to muscle or soft tissues sprain or strain \[[@B3]\], particularly in instances where pain arises suddenly during physical loading of the back, with the pain lateral to the spine. Over 99% of back pain instances fall within this category \[[@B4]\]. About 60%--80% of the population in industrialized countries like India, United States, Europe, Finland, and the Netherlands suffering from back pain; it is the second most common health problem after headache \[[@B5]\]. Stability of the spine is provided through ligaments and muscles of the back, lower back and abdomen.
Mechanical low back pain (MLBP) is a mechanically-derived, musculoskeletal back pain not involving nerve root compression or serious spinal diseases \[[@B6]\]. Prevalence is higher in young and active adults \[[@B7]\]. Causes of MLBP typically are attributed to an acute traumatic event, but they may also include cumulative trauma as an etiology \[[@B8]\]. The severity of an acute traumatic event varies widely, from twisting of the back to being involved in a motor vehicle collision. MLBP due to cumulative trauma tends to occur more commonly in the workplace. A systematic study review implicated a sedentary lifestyle, defined by the authors as including sitting for prolonged periods at work and during leisure time, as a risk factor for MLBP \[[@B9]\].
Low back pain (LBP) is usually self-limiting, with almost 90% of cases resolving within 6--12 weeks \[[@B10]\]. However, recurrence is high (84%) \[[@B11]\]. Risk factors for recurrence include weakness \[[@B11]\], excessive fatigability \[[@B12]\], lack of multifidus muscle recovery \[[@B13]\] and atrophy \[[@B13][@B14]\], which eliminate segmental stability.
Weakness of the superficial trunk and abdominal muscles is an important risk factor for LBP \[[@B15]\]. Strengthening these muscles markedly improves CLBP and decreases functional disability \[[@B16]\]. Another independent risk factor for CLBP is weakness and lack of motor control of deep trunk muscles, such as the lumbar multifidus (LM) and transversus abdominis (TrA) muscles \[[@B13][@B17]\]. Dysfunction of the LM crucially influences the etiology and recurrence of LBP \[[@B13][@B14][@B18]\]. Therefore, exercises to restore optimal LM function are a common aspect of current rehabilitation strategies \[[@B19][@B20]\]. More recently, attention has focused on the deepest fibers of the LM \[[@B17][@B20][@B21]\].
Understanding the anatomical structure that is painful differs from the disorder itself, and is important in order to give proper treatment. Thinking in the terms of integrated function and dysfunction might be more appropriate in diagnosis and treatment of back pain.
Therefore, in this study, we compared the efficacy of the multifidus retraining program (MRP) with conventional strengthening of abdominal and trunk muscles on pain and functional capacity in CLBP. Our hypothesis was that the MRP is more efficient than the muscle strength in the improvement of CLBP.
Materials and Methods
=====================
This study was approved by the ethical committee and research department of the Maharashtra Institute of Physiotherapy, Latur, India.
Thirty subjects (18 males and 12 females) who were computer professionals and who had at least a 2-year history of CLBP were selected and divided into two groups by purposive random sampling with 15 subjects (9 males and 6 females) in each group. One group (group A) was treated with traditional back exercise (TBE) and the other (group B) with MRP. All subjects completed the Handler 10-minute screening test for chronic back pain to rule out psychological pain. The inclusion criteria for the study were back pain felt between T12 and the gluteal folds that had lasted at least 3 months, age between 20 and 35 years, willingness to participate, ability to participate in an exercise program safely and no cognitive impairments that would limit their participation. The exclusion criteria were previous spinal surgery, trauma, rheumatologic disorders, spine infections, spine exercise training in the 3 months before the onset of the study, vertebral fracture, spinal abnormalities (scoliosis, kyphosis), inter vertebral disk prolapse, spondylolisthesis, pregnancy, malignancy, congenital abnormalities, ankylosing spondylitis, hernia, visceral problems, fibromyalgia and myofascial pain. Participants received a clear explanation of the study and provided their written informed consent.
1. Procedures
-------------
Demographic data including age, sex, height and weight were documented for descriptive statistical analysis ([Table 1](#T1){ref-type="table"}). The subjects were familiarized with the Oswestry disability questionnaire and visual analogue scale (VAS) rating of pain. These instruments were designed to give information about the back pain, which affects their ability to manage in everyday life. The patients were asked to answer all questions by placing a mark in the one box that most closely described their present condition. The pre-test scoring of both the assessment tools were done and documented. Functional outcome and pain perception were assessed with Oswestry disability questionnaire and VAS, respectively. Subjects were then randomized to the two groups as describe above. Homogeneity of variance between the groups were identical in terms of age, height, weight, pre-VAS and Oswestry disability questionnaire (MODQ) score ([Table 1](#T1){ref-type="table"}).
Both groups received exercise for 6 weeks. Based on the previous research \[[@B20][@B22]\], the exercises for MRP were designed to retrain the multifidus muscle. TBE consisted of strengthening and stretching of superficial abdominal and back muscles ([Table 2](#T2){ref-type="table"}). The exercise program was supervised by a physiotherapist in the outpatient department and the exercise register for each subject was maintained. Each exercise was meant to be repeated 20 times with a 5--8 seconds hold of each exercise. Exercise was on alternate days for 6 weeks. Subjects were allowed to rest for 2--4 minutes in between each set of exercise. The study was a pre-test and post-test experimental two group study. The subjects underwent pre-test at the starting of the study and the post-test was recorded after completion of the exercise program for 6 weeks.
2. Assessments
--------------
Assessment of severity of pain and functional disability was done at baseline and at the end of the treatment.
3. Pain
-------
Pain was assessed using a VAS consisting of a 10 cm line, with the left extremity indicating \"no pain\" and the right extremity indicating \"unbearable pain.\" Participants were asked to use the scale to indicate their current level of pain. Higher values suggested more intense pain. The pre- and post-exercise values of both MRP and TBE were documented and the mean and standard deviation (SD) was calculated for statistics.
4. Functional Disability
------------------------
Functional disability was estimated by the Oswestry disability questionnaire, a functional scale assessing the impact of LBP on daily activities. There are many functional questionnaires available for the measurement and evaluation of LBP; we felt the Oswestry questionnaire was the most appropriate. The scoring was done by adding the values circled by the subject for each of the 10 individual questions and the disability is commanded as mild or no disability (0%--20%), moderate disability (21%--40%), severe disability (41%--60%), incapacity (61%--80%) and restricted to bed (81%--100%). The pre- and post-exercise values of the TBE and MRP groups were documented and the mean and SD was calculated and recorded for statistical analysis.
Results
=======
Statistical analysis was done using the SPSS ver. 16 (SPSS Inc., Chicago, IL, USA) software to find the average and the standard deviation of age, duration, height, weight, VAS and disability score in both groups. \"F\" test was done identify the equality of the variance between the group and showed identical in all the above factors between the groups ([Table 1](#T1){ref-type="table"}). Analysis using paired *t*-test of pre and post value for both the groups showed significant improvement in both groups. Pain after exercise in group A (4.467) and group B (2.867) were analyzed using unpaired T-test value (6.32) at *p*≤0.001 levels. Similarly the functional disability after exercise in group A (17.067) and group B (9.467) were analyzed using unpaired T-test value (6.62) at *p*≤0.001 level showed that MRP produced greater improvement than TBE ([Tables 3](#T3){ref-type="table"}, [4](#T4){ref-type="table"}, [Figs. 1](#F1){ref-type="fig"}, [2](#F2){ref-type="fig"}).
Discussion
==========
This study aimed to compare the efficacy of MRP and TBE in the relief the pain and improving the functional disability of CLBP among computer professionals with an extended history of LBP. Both exercise regimens were effective in relieving pain and in decreasing functional impairment, but MRP provided better relief of pain and improved functional outcome in the CLBP patients.
Bergmark \[[@B13]\] categorized the trunk muscles in to local and global muscle system based on their main mechanical roles in stabilisation. The local muscles are capable of controlling the stiffness and intervertebral relationship of spinal segments and posture of lumbar spine \[[@B6][@B23]\]. The current focus of back pain rehabilitation has evolved from global to local muscle systems with the recognition that local muscles are important in control of segmental spinal stability. Evidence from biomechanical, morphological, histochemical, electromyography and muscle fiber studies has implicated the multificus muscles in 2/3 of the segmental stability of the spine.
Waddell et al. \[[@B24]\] concluded that LBP patients often avoid using their back in everyday life situations because of fear of pain and its consequences. We suspect that prolonged sitting posture at a computer workstation as well as lack of exercise that is the norm for many computer professionals may lead to the weakness of the multifidus muscle. This speculation is supported by previous findings findings that disuse leads to the atrophy of the back muscles especially the LM muscle, and an exercise program produces functional improvement \[[@B25][@B26]\]. Segmental instability may cause functional disorders and strain as well as pain \[[@B27]\]. Uni-segmental muscles of the lumbar spine, such as the multifidus muscle, may provide segmental control and are the primary segmental stabilizers of the spine in lumbar region \[[@B28]\]. One study identified selective atrophy of the LM after the first episode of back pain; the atrophy was unlikely to revert without specific training, and the lower muscular stability predisposed an individual to further episodes of LBP \[[@B17]\]. Our findings also support this view; the MRP group had better outcome in comparison with the generalized back exercise program. Both multifidus and transverse abdominis muscles have been suggested as primary stabilizers of the lumbar segment, minimizing compressive forces on spinal structures \[[@B29]\]. The inter-segmental LM is the most important muscle in the stabilization of spine \[[@B30]\]. Presently, poor endurance of the multifidus muscle was linked with increased recurrence of LBP.
Although many aspects of treatment for the LBP remain controversial, the superiority of active exercise to inactivity is uncontested \[[@B29]\]. Presently, both the active traditional and multifidus exercises were beneficial for LBP. Specific exercise treatment is more effective than other types of other conservative management of LBP \[[@B31]\]. We hypothesize that the pain and the disability in these patients may be due to the segmental instability caused by the weakness of LM muscles. A specific exercise program to retrain these weak muscle fibers improves the function.
MRP showed significant improvements in relieving pain and functional disability in this study and previous research \[[@B31]\] when compared to TBE. The better results of the MRP group may reflect the fact that this training program concentrates on the deep back segmental stabilizer muscle, the LM, which is week in professionals who sit for prolonged periods. The MRP regimen produced better pain reduction and functional improvement compared to TBE, which is consistent with the fact that the deep muscles provide segmental stability to the back. Subjects in the TBE group performed exercises to strengthen the superficial muscles of the abdomen and trunk. The resulting reduction of pain and improvement in the function ability of CLBP subjects is consistent with prior findings \[[@B16]\].
Even after pain remission in patients with LBP, proper deep muscle reestablishment often does not occur, with specific physical therapy focusing on those muscles being necessary \[[@B13]\]. Our findings suggest that both protocols are of clinical utility in the improvement of CLBP.
Conclusions
===========
Both the MRP and TBE regimens reduced the pain and functional disability effectively in individuals with CLBP. The improvements were superior for those receiving MRP in computer professionals with CLBP.
**Conflict of Interest:** No potential conflict of interest relevant to this article was reported.
This study was approved by the ethical committee and research department of the Maharashtra Institute of Physiotherapy, Latur, India.
{#F1}
{#F2}
###### Demographic data and homogeneity of variance
Values are presented as mean±standard deviation.
LBA, low back pain; VAS, visual analogue scale; MODQ, Oswestry disability questionnaire.
###### Exercise program
###### Data analysis and result of the VAS score between group A and B
Values are presented as mean±standard deviation.
VAS, visual analogue scale.
###### Data analysis and result of the ODQ disability score between group A and B
Values are presented as mean±standard deviation.
ODQ, Oswestry disability questionnaire.
| {
"pile_set_name": "PubMed Central"
} |
INTRODUCTION {#s1}
============
Medulloblastoma is the most frequently diagnosed malignant brain tumor in children, with a peak incidence at the age of 5 years \[[@RRU120C1]\]. Despite being rare, it can also be found in adults, in whom pediatric treatment regimes appear feasible and effective, based upon similar presentation and course \[[@RRU120C2]\]. Pathologically, medulloblastomas comprise at least seven subtypes of infratentorial central nervous system (CNS) primitive neuroectodermal tumor, which originate from the cerebellum or the roof of the fourth ventricle \[[@RRU120C3]\].
Treatment usually involves neurosurgical resection to the most reachable extent, followed by concomitant chemoradiation and additional maintenance chemotherapy \[[@RRU120C1]\]. Due to the characteristic propensity of medulloblastoma to spread throughout the CNS along cerebrospinal fluid (CSF) pathways, radiotherapy is usually administered to the complete craniospinal axis \[[@RRU120C4]\].
The unique phenomenon of early, rapid and extensive CNS dissemination is based on the high capability of medulloblastoma cells to migrate along leptomeningeal surfaces \[[@RRU120C5]--[@RRU120C7]\]. To initiate tumor cell motility, medulloblastoma cells interact with typical components of the leptomeningeal extracellular matrix (ECM), such as fibronectin, collagen I and collagen IV \[[@RRU120C7], [@RRU120C8]\]. Migration is then promoted by exposure to gradients of serous chemoattractants, which can abundantly be found within CSF or blood plasma \[[@RRU120C9], [@RRU120C10]\]. However, in addition to intrinsic promoters of cell migration, treatment-related factors have been reported as increasing tumor cell motility. Photon irradiation (IR) has been repeatedly claimed to enhance cell migration in sarcoma \[[@RRU120C11]\], colorectal cancer \[[@RRU120C12]\], and glioma \[[@RRU120C13]\]. In medulloblastoma, preclinical studies have demonstrated increased tumor cell invasiveness through upregulation of urokinase plasminogen activator receptor (uPAR) signaling \[[@RRU120C6], [@RRU120C14]\]. In D283 and UW228 medulloblastoma cells, single photon doses of 6 Gy promoted uPAR-dependent Wnt/β-catenin-signaling and consequently increased the neurosphere-forming ability \[[@RRU120C14]\]. Moreover, single photon doses of 7 Gy induced expression of α~3~, α~5~ and β~1~ integrins, leading to enhanced wound healing migration in D283 and DAOY cells \[[@RRU120C6]\]. To date, little data has focused on dose-dependent alterations in medulloblastoma cell motility, and none has addressed the potential impact of carbon ion IR.
IR with carbon ions has recently been identified as exerting inhibitory influence on tumor cell migration. Correspondingly, expression of motility-promoting factors has been found to be suppressed even at sublethal doses \[[@RRU120C11], [@RRU120C15]\]. However, controversial and ambiguous findings have also reported distinct tumor cell lines within one tumor entity as not succumbing to the decelerating effect of carbon ion IR \[[@RRU120C16]\], but as possibly increasing their invasive potential instead \[[@RRU120C17]\].
As for medulloblastoma, clinicians have repeatedly postulated the use of particle radiotherapy due to its beneficial dose distribution \[[@RRU120C3]\] and increased biological effectiveness \[[@RRU120C18]\]. Previous works have mainly focused on proton particle IR. This manuscript addresses the role of carbon ion IR in motility changes in two medulloblastoma cell lines.
MATERIALS AND METHODS {#s2}
=====================
Cell culture {#s2a}
------------
D425 medulloblastoma cells were purchased from ATCC and kept at 37°C and 5% CO~2~ in improved MEM (Biochrom AG) supplemented with 1% Penicillin/Streptomycin and 10% fetal calf serum (FCS). Med8A medulloblastoma cells were a kind gift from Michael D. Taylor (Toronto, Ontario, Canada) and were kept at 37°C and 5% CO~2~ in DMEM (FG0415 Biochrom AG) supplemented with 1% Penicillin/Streptomycin and 10% FCS. At 24 h before adhesion and migration experiments, cells were serum starved in DMEM containing 1% Penicillin/Streptomycin and 0.5% FCS. Passaging of cells was performed every week.
Migration assays using membrane coated with extracellular matrix proteins {#s2b}
-------------------------------------------------------------------------
For migration assays, polycarbonate membranes with 8-μm pores, and for adhesion assays, 96-well cell culture dishes, were coated with 50 ng/cm^2^ fibronectin, 0.5 μg/cm^2^ collagen I and 0.5 μg/cm^2^ collagen IV, kept overnight at 4°C and washed in twice distillated and salt-free water prior to the experiments. Next, 5 × 10^3^ cells were loaded into the upper chamber of a 48-well modified microchemotaxis chamber (Multiwell Chemotaxis Chamber, Neuro Probe). The lower well contained cell culture medium 10% FCS, as indicated. Lower and upper chambers were separated by an 8-µm pore size polycarbonate membrane that had been coated with fibronectin (50 ng/cm^2^), collagen I (0.5 μg/cm^2^) and collagen IV (0.5 μg/cm^2^) 24 h before the start of migration. Radiation treatments were performed 24 h before assessment of migration. Transmigrated cells were stained with DiffQuick® and counted with a Leica DC300F microscope. All assays were done in (at least) triplicates, and the wells were counted by an investigator blinded to the experimental set-up. Cell numbers are expressed as multiples of controls or as proportion of inputs.
Adhesion assay {#s2c}
--------------
At 24 h after IR, 5 × 10^2^ cells were loaded into 96-well cell culture dishes coated with fibronectin (FN), collagen I and IV and allowed to adhere for 1 h before assessment of adhesion. Adherent cells were stained and counted with a Leica DC300F microscope by an investigator blinded to the experimental set-up. All assays were done in triplicates. Cell numbers are expressed as absolute numbers.
Enzyme-linked immunosorbent assay {#s2d}
---------------------------------
For this assay, 1 × 10^5^ cells were incubated for 24 h following IR. The cell culture supernatants were then harvested and stored at −18°C. The enzyme-linked immunosorbent assay (ELISA) protocol followed the R&D Systems® guidelines for MMP2/9 assays (human MMP-9 Immoassay DMP900 quantitative determination of active and pro MMP9, mouse monoclonal anti-MMP9) (human MMP-2 Immunoassay DMP2F0 quantitative determination of MMP2, polyclonal anti-MMP2). Analysis and evaluation were performed with the Infinite® F50/Robotic ELISA plate reader (absorbance at 450 nm, correction wavelength at 570 nm) and Magellan for F50 software (Tecan Group Ltd).
Fluorescence-activated cell sorting {#s2e}
-----------------------------------
At 24 h after IR, cells were fixed with 70% ethanol and stained with labeled anibodies directed against α~ν~β~3~ (FITC mouse anti-human-CD51/61, 555505, BD), α~5~ (PE mouse anti-human CD49e, 555617, BD), β~1~ (mouse anti-human-CD29, 556049, BD) and α~ν~β~5~ (PE mouse anti-human Integrin α~ν~β~5~ FAB2528P, R&D). Med8A and D425 cells were analyzed with a three-colour fluorescence-activated cell sorting (FACS) scan flow cytometer and CellQuestPro software (BD Biosciences).
Radiation treatment {#s2f}
-------------------
Photon IR was performed using a biological X-ray irradiator with 320 kV and 12.50 mA (Precision X-Ray Inc., Model No.: XRAD 320, North Branford, USA). We applied single doses of 1, 2 and 10 Gy 24 h prior to migration experiments at a dose rate of 1.2 Gy/min. Carbon ion IR was applied with an extended Bragg peak (E = (128 ± 7) MeV/u, LET = (91.5 ± 1.5) keV/μm) at single carbon ion doses of 0.5, 1 and 3 Gy at the Heidelberg Ion Therapy Center. Carbon ion doses, assuming a relative biological effectiveness (RBE) of 3--4, were derived from comparable glioma studies \[[@RRU120C15]\], since no RBE calculations for medulloblastoma cells have been published thus far.
Statistics {#s2g}
----------
All experiments were performed at least three times using triplicates in each individual experimental set-up. Data are displayed as means ± standard deviations. Comparisons between two groups were performed with Student\'s *t*-test. Statistical significance was noted for two-tailed *P*-values of ≤0.05.
RESULTS {#s3}
=======
Radiogenic alterations of medulloblastoma cell transmigration {#s3a}
-------------------------------------------------------------
We investigated transmigration of medulloblastoma cell lines Med8A and D425 through polycarbonate membranes coated with ECM proteins FN, collagen I, and collagen IV following single doses of both photon (1, 2 and 10 Gy) and carbon ion (0.5, 1 and 3 Gy) IR. Medulloblastoma cells were motile and demonstrated robust serum-induced transmigration through collagen-I- (Med8A: *n* = 720 (14%); D425: *n* = 480 (9.6%)), collagen-IV- (Med8A: *n* = 1520 (30.4%); D425: 640 (12.8%)) and FN-coated membranes (Med8A: *n* = 1680 (33.6%); D425: *n* = 1260 (25.6%)). Following IR, we observed a significant reduction of migration in most experiments for both D425 and Med8A cells (Fig. [1](#RRU120F1){ref-type="fig"}). Impairment of migration was strongest for carbon ion--irradiated Med8A cells on collagen IV--coated surfaces (0.5 Gy: −84%, 1 Gy: −85.5%, 3 Gy: −87%). In general, carbon ion IR was more effective than photon IR in impairing medulloblastoma cell transmigration and caused significantly reduced transmigration through FN- and collagen I/IV--coated membranes in both cell lines, even at sublethal doses of 0.5 GyE. Photon IR, too, inhibited medulloblastoma transmigration in most set-ups; however, Med8A cell transmigration remained unaltered in experiments using collagen-I--coated membranes. Quantitative analyses of both photon- and carbon ion-altered transmigration are summarized in Table [1](#RRU120TB1){ref-type="table"}. Table 1.Quantitative analysis of medulloblastoma cell transmigrationTransmigration of medulloblastoma cells**Med8A (%)D425 (%)Fibronectin**control1001001 Gy photon IR74 ± 1.4 (\*)103 ± 23.4 (ns)2 Gy photon IR61 ± 17 (\*)71 ± 9.3 (\*)10 Gy photon IR51 ± 4.2 (\*)53 ± 2.9 (\*)0.5 GyE ^12^C IR57 ± 18.4 (\*)79 ± 1.4 (ns)1 GyE ^12^C IR30 ± 1.4 (\*)66 ± 14 (ns)3 GyE ^12^C IR25 ± 4.2 (\*)51 ± 4.2 (\*)**Collagen I**control1001001 Gy photon IR102 ± 21.2 (ns)89 ± 16.5 (ns)2 Gy photon IR91.5 ± 17.7 (ns)69.3 ± 9.1 (\*)10 Gy photon IR97 ± 23 (ns)55 ± 21.2 (\*)0.5 GyE ^12^C IR49 ± 9.2 (\*)81 ± 10.6 (\*)1 GyE ^12^C IR38 ± 4.2 (\*)59 ± 5.7 (\*)3 GyE ^12^C IR33 ± 4.9 (\*)24 ± 25 (\*)**Collagen IV**control1001001 Gy photon IR62 ± 7.1 (\*)109 ± 12.7 (ns)2 Gy photon IR57 ± 7.1 (\*)96 ± 44.5 (ns)10 Gy photon IR43 ± 26.9 (\*)74 ± 1.4 (\*)0.5 GyE ^12^C IR16 ± 4.2 (\*)54 ± 7.1 (\*)1 GyE ^12^C IR15 ± 2.1 (\*)49 ± 9.2 (\*)3 GyE ^12^C IR13 ± 4.2 (\*)46 ± 4.2 (\*)[^1] Fig. 1.Radiogenic inhibition of Med8A medulloblastoma cell migration. Graphical analysis of transmigration through DiffQuik®-stained 8-µm pore size polycarbonate membranes coated with Fn, collagen I and IV following IR with single photon doses \[[@RRU120C31]\] of 1, 2 and 10 Gy, and single carbon ion doses (striped) of 0.5, 1 and 3 GyE. Statistical analysis performed using Student\'s *t*-test (asterisk indicates statistical significance *P* \< 0.05).
Matrix metalloproteinase release following carbon ion and photon IR {#s3b}
-------------------------------------------------------------------
We determined secretion of matrix metalloproteinases (MMPs) into the cell supernatants following IR using ELISA. In Med8A cells, MMP9 secretion was significantly reduced after IR with carbon ions at doses as low as 1 Gy. Photon IR, too, decreased MMP9 concentration in Med8A supernatants; however, statistical significance was only reached for single doses of 10 Gy (Fig. [2](#RRU120F2){ref-type="fig"}). Supernatant concentrations of MMP9 in experiments using D425 cells remained unchanged by photon IR at all doses tested, whereas carbon ion IR supressed MMP9 secretion at single doses of both 1 and 3 Gy. In both Med8A and D425 cells, MMP2 concentration remained unaltered following both photon and carbon ion IR at all doses tested (data not shown). Fig. 2.Radiation-altered secretion of MMP9 in Med8A medulloblastoma cells. Graphical ELISA analysis of MMP9 secretion following IR with single photon doses (left group) of 1, 2 and 10 Gy, and with single carbon ion doses (right group) of 0.5, 1 and 3 GyE. Statistical analysis performed using Student\'s *t*-test (asterisk indicates statistical significance *P* \< 0.05, red bar references basal unirradiated control).
Integrin expression following carbon ion and photon IR {#s3c}
------------------------------------------------------
FACS analyses were performed in order to investigate IR-altered integrin expression. We did not observe any significant modification of α~ν~β~3~, α~ν~β~5~ or β~1~ integrin expression in Med8A and D425 cells following either photon or carbon ion IR. On the contrary, both cell lines stained significantly positive for α~5~ integrins, and following photon IR, Med8A cell surface expression of α~5~ was increased in a dose-dependent manner, yielding significant enhancements at both 2 and 10 Gy. In carbon ion--irradiated Med8A cells, expression of α~5~ was also continuously increased, with significant enhancements at both 1 and 3 Gy (Fig. [3](#RRU120F3){ref-type="fig"}). Expression of α~5~ following IR with 1 Gy photons or 0.5 Gy carbon ions was mildly increased, but did not reach statistical significance (Fig. [3](#RRU120F3){ref-type="fig"}). In D425 cells, photon IR increased α~5~ expression in a non-significant manner, while carbon ion IR significantly upregulated α~5~ expression at all doses tested (data not shown). Fig. 3.Radiation-altered expression of integrin α~5~ in Med8A medulloblastoma cells. Graphical FACS analysis of isotype-controlled integrin α~5~ surface expression following IR with single photon doses (left group) of 1, 2 and 10 Gy, and with single carbon ion doses (right group) of 0.5, 1 and 3 GyE. Statistical analysis performed using Student\'s *t*-test (asterisk indicates statistical significance *P* \< 0.05, red bar references basal unirradiated control).
Radiogenic alterations of medulloblastoma adhesion {#s3d}
--------------------------------------------------
We next assessed radiogenic modifiability of medulloblastoma cell adhesion to FN- and collagen-I/IV--coated surfaces. We observed a significant dose-dependent increase in adhesion following carbon ion IR in Med8A cells (Fig. [4](#RRU120F4){ref-type="fig"}). Photon IR also caused increased adherence to all coated surfaces; however, the effect was less pronounced than in carbon ion--irradiated cells (Fig. [4](#RRU120F4){ref-type="fig"}). The same phenomenon of both photon- and carbon ion--increased adhesion to coated surfaces was observed in D425 cells (Table [2](#RRU120TB2){ref-type="table"}). Also, radiogenic adhesion was more evident in carbon ion--irradiated cells than in photon-based experiments. Table 2.Quantitative analysis of medulloblastoma cell adhesionAdhesion of medulloblastoma cells(cell input 500 cells)**Med8A (n)D425 (n)Fibronectin**control52 ± 3.761 ± 11.31 Gy photon IR61 ± 6.3 (\*)76 ± 7.1 (ns)2 Gy photon IR71 ± 8.5 (\*)109 ± 2.2 (\*)10 Gy photon IR82 ± 10.1 (\*)118 ± 4.6 (\*)0.5 GyE ^12^C IR55 ± 9.6 (\*)71 ± 0.5 (ns)1 GyE ^12^C IR95 ± 12.7 (\*)113 ± 8.3 (\*)3 GyE ^12^C IR90 ± 11.7 (\*)150 ± 2.6 (\*)**Collagen I**control81 ± 12.030 ± 5.71 Gy photon IR95 ± 7.9 (ns)43 ± 11.7 (ns)2 Gy photon IR110 ± 15.9 (\*)54 ± 9.9 (\*)10 Gy photon IR125 ± 12.8 (\*)74 ± 10.1 (\*)0.5 GyE ^12^C IR89 ± 7.4 (ns)46 ± 5.1 (\*)1 GyE ^12^C IR118 ± 9.4 (\*)54 ± 8.0 (\*)3 GyE ^12^C IR157 ± 9.2 (\*)92 ± 14.6 (\*)**Collagen IV**control88 ± 7.124 ± 5.11 Gy photon IR104 ± 6.4 (\*)35 ± 5.4 (\*)2 Gy photon IR114 ± 5.8 (\*)46 ± 6.6 (\*)10 Gy photon IR131 ± 5.1 (\*)70 ± 10.2 (\*)0.5 GyE ^12^C IR110 ± 9.9 (\*)36 ± 6.3 (\*)1 GyE ^12^C IR147 ± 19.8 (\*)50 ± 9.6 (\*)3 GyE ^12^C IR192 ± 34.2 (\*)105 ± 12.0 (\*)[^2] Fig. 4.Radiogenic stimulation of Med8A medulloblastoma cell adhesion. Graphical analysis of cell adherence to polycarbonate surfaces coated with collagen I (left group), collagen IV (middle group), and FN (right group) following IR with single photon doses (left group) of 1, 2 and 10 Gy, and with single carbon ion doses (right group) of 0.5, 1 and 3 GyE. Statistical analysis performed using Student\'s *t*-test (asterisk indicates statistical significance *P* \< 0.05, red bar references basal unirradiated control).
DISCUSSION {#s4}
==========
In the present manuscript we investigated the modifiability of medulloblastoma cell motility through photon and carbon ion IR. In opposition to observations on various non-medulloblastoma tumor cells \[[@RRU120C11]--[@RRU120C13], [@RRU120C19]\], IR did not promote tumor cell migration but significantly impaired ECM-based transmigration, while simultaneously causing increased adherence to ECM-coated surfaces.
We have recently studied radiation-altered migration in malignant glioma cells and found migration to be significantly increased following sublethal photon IR through upregulation of FN-binding integrins \[[@RRU120C19]\]. This phenomenon has previously been observed for several other tumor cell lines, including sarcoma \[[@RRU120C11]\], colorectal \[[@RRU120C12]\] and pulmonary cancer cells \[[@RRU120C20]\]. Clinically, the finding of radiation-induced migration concerns radiation oncologists, because it imposes a theoretical risk of radiation-induced locoregional tumor infiltration and distant dissemination and, therefore, challenges current concepts of target volume delineation with limited safety margins for improved sparing of organs at risk.
Medulloblastomas are characterized by their tendency for early and rapid spread throughout the CNS \[[@RRU120C21]\]. Therefore, photon-induced migration would represent yet another stimulation of increased medulloblastoma cell motility.
In opposition to findings in other medulloblastoma cell lines, we detected migration to be decreased following both photon and carbon ion IR. Nalla *et al.* described DAOY medulloblastoma cell migration to be enhanced by 27% following single photon doses of 7 Gy \[[@RRU120C6]\]. This difference from our results may be explained by differences in experimental design. They investigated migration of cells towards one another in scratch assays, instead of analysis of transmigration away from the initial cell localization. Tabatabei *et al.* observed irradiated cells to secrete promigratory ligands \[[@RRU120C22]\] that may stimulate the reported migration in scratch assays. Therefore, the supposedly opposing notion of both promotion and inhibition of medulloblastoma cell migration through photon IR may be reconciled in terms of which migratory phenomena are being addressed: emigration from or immigration to previously irradiated tumor cell clusters.
Asuthkar *et al.* found single photon doses of 6 Gy to increase uPAR expression, suggestive of increased invasion. Therefore, they proposed targeting of uPAR in combination with photon IR as a promising option \[[@RRU120C14]\]. In contrast to these previously published works, we used two cell lines that are known to carry genetic amplifications of the *myc* oncogene. This aberration has been associated with early dissemination and poor outcome \[[@RRU120C23], [@RRU120C24]\], but it was also reported to sensitize medulloblastoma cells to IR \[[@RRU120C25]\]. These findings help to explain the opposing findings in different medulloblastoma cells, and support our results of IR-suppressed cell migration.
Inhibition of migration was strongest in FN- and collagen-IV--based experiments, which confirms previous works that have identified both FN and collagen IV to be key ECM proteins in radiation-induced motility alterations \[[@RRU120C19], [@RRU120C26], [@RRU120C27]\]. Carbon ion IR had significantly higher inhibitory effects on radiation than did photon IR. Again, this is in line with prior works that described particle IR as effectively impairing cell migration, even at sublethal doses, as were used in this project \[[@RRU120C11], [@RRU120C15]\].
Following IR, several tumor cell lines have been observed to upregulate their expression and secretion of MMPs \[[@RRU120C13], [@RRU120C20], [@RRU120C28]\]. This finding suggests MMPs might be involved in the phenomon of radiation-induced migration and infiltration, and MMP2 and MMP9 have been reported as characteristic in the context of photon-associated motility changes \[[@RRU120C13]\]. We determined concentrations of soluble MMP2 and MMP9 following both photon and carbon ion IR, and found significantly reduced MMP9 concentrations, even doses as low as 1 GyE. Following photon IR, MMP9 secretion was also suppressed, but with statistical significance only being reached for doses of 10 Gy. This finding may explain the difference in inhibitory effect on migration between photon and carbon ion IR. Measuring MMP2 after IR yielded no significant changes, suggesting an inferior role of MMP2 in the radiogenic changes in medulloblastoma cell motility. Ogata *et al.* also investigated IR-induced changes in MMP2 activity and found no influence of photon IR in sarcoma cells at various doses \[[@RRU120C11]\].
Since radiation-increased migration has repeatedly been correlated with increased integrin expression \[[@RRU120C6], [@RRU120C15], [@RRU120C19], [@RRU120C26], [@RRU120C29]\], we performed FACS analyses of various cell surface integrins. We did not detect any relevant radiogenic alteration in the expression of those integrins that have previously been identified as mediators of radiation-induced migration in other tumor entities (α~ν~β~3~, α~ν~β~5~ and β~1~). However, the expression of α~5~ was significantly enhanced in a dose-dependent manner after both photon and carbon ion IR. Our results are supported by Meineke *et al.* who found α~5~ to be significantly enhanced in a strictly time-dependent manner, with peak values being reached after 24 h following single photon doses of 5 Gy in colorectal cancer cells \[[@RRU120C26]\]. Expression of integrin α~5~ was iteratively confirmed to be a key integrin in mediating adherence to FN and collagen I/IV \[[@RRU120C30]\].
Besides its migratory impact, irradiation is also known to affect cell adherence to ECM proteins. Therefore, we finally performed adhesion experiments in order to determine radiogenic alterations that might establish a concept of radiation-decreased migration, while simultaneously increasing cell adherence. We found adhesion to FN, collagen I and IV to be significantly enhanced following both photon and carbon ion IR. A finding of increased cell adhesion to collagen I remains somewhat surprising because, in prior experiments, no corresponding decrease in cell chemotaxis was observed---as was detected for other matrix proteins; However, similar findings were reported by Tsutsumi *et al.*, who found increased adhesion to collagen I to be connected to enhanced cell invasion in non--small cell lung cancer \[[@RRU120C20]\].
The observed increase in adhesion was strictly dose-dependent, reaching statistical significance after clinically relevant doses of 2 and 10 Gy, and 1 and 3 Gy, respectively. This phenotype corresponds with the radiogenic increase in ECM protein--binding α~5~ integrins, and is supported by several other works addressing radiogenic motility alterations: Nalla *et al.* reported enhancements of adhesion by 38% and 50% in collagen I-/FN-exposed medulloblastoma cells following single photon doses of 7 Gy \[[@RRU120C6] \]; Meineke *et al.* demonstrated photon-induced upregulation of integrin expression to significantly enhance adhesion capacity in colorectal cancer cells \[[@RRU120C26]\].
Whether these adherent tumor cells are less aggressive resident cells or cells predestined to infiltrate local structures remains controversial. We are aware that our present data cannot resolve this conflict. However, we want to state that, in D425 and Med8A medulloblastoma cells, neither photon nor carbon ion IR promote migration, and that radiogenic integrin-based adhesion does not go along with an increase in invasion-initiating MMPs.
In opposition to other CNS tumors, in medulloblastoma cells, both photon and carbon ion IR effectively prevent migration through impaired MMP9-secretion and upregulation of pro-adhesive FN- and collagen-binding integrin α~5~. Therefore, radiogenic stimulation of tumor cell infiltration and dissemination must not be feared when administering radiotherapy in medulloblastoma patients. Photon-associated findings have been consistently confirmed and are even more pronounced in carbon ion--irradiation experiments, suggesting particle irradiation to be a promising radiation modality in the treatment of medulloblastoma patients.
FUNDING {#s5}
=======
This work was supported by institutional funding. Funding to pay the Open Access publication charges for this article was provided by the Department of Radiation Oncology, University Hospital of Heidelberg, Heidelberg, Germany.
[^1]: Asterisk = statistical significance, *P* \< 0.05, ns = absence of statistical significance, *P* \> 0.05.
[^2]: Asterisk = statistical significance, *P* \< 0.05, ns = absence of statistical significance, *P* \> 0.05.
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#Sec1}
============
The cerebellum is a well-known important subcortical motor structure, ensuring coordination, precision, and accurate timing of movement, and learning motor skills \[[@CR1]--[@CR4]\]. The cerebellar neuronal circuitry, organized elaborately and modularly, receives two major types of afferent inputs, mossy fibers and climbing fibers \[[@CR4], [@CR5]\]. The former originates from nuclei in the spinal cord and brainstem and carries sensory information from the periphery as well as information from the cerebral cortex, while the latter originates from the inferior olivary nucleus and sends error signals sensed from the motor performance of periphery musculatures to the cerebellum. In addition to obtaining specific and discrete information from the mossy and climbing fiber afferent systems, the cerebellum also receives nonspecific signals from the so-called third type of afferents, typically beaded fibers \[[@CR6]\], which contain various amines or neuropeptides. Although more than 20 different types of amines and neuropeptides, such as serotonin \[[@CR7]\], norepinephrine \[[@CR8]\], histamine \[[@CR9]\], orexin \[[@CR10]\], and CRF \[[@CR11], [@CR12]\], have been found in the cerebellum, their functional significance is largely unknown. In general, beaded fibers form varicose contact with Purkinje cells and interneurons in the cerebellar cortex, as well as neurons in the cerebellar nuclei, fastigial (FN), interpositus (IN) and dentate (DN) nuclei, and exert a widespread modulatory role in the cerebellar circuitry \[[@CR2], [@CR6], [@CR13]\].
Monoamines are firstly identified neurotransmitters used in the third type of afferents in the cerebellum. Among them, histamine is a newly found one in the cerebellar afferents. Although histamine was isolated from peripheral tissues as a biologically active amine more than a century ago, histamine acting as a neurotransmitter in the brain and the central histaminergic system gained general acceptance only in recent 30 years \[[@CR14]\]. Peripheral histamine is well known to be stored primarily in the tissue mast cells and enterochromaffin-like cells, and holds a pivotal position in allergic reaction, gastric acid secretion and contraction of smooth muscle tissues of the lungs, whereas central histamine tends to be considered as a "modulator for whole brain activity" \[[@CR14]--[@CR17]\]. In the cerebellum, different from serotoninergic and norepinephrinergic afferents arising from the brainstem \[[@CR6], [@CR13]\], histaminergic fibers originate from the hypothalamus, a higher center for nonsomatic visceral and autonomic regulation \[[@CR15], [@CR16]\]. In 1984, the direct hypothalamocerebellar projections were first definitively presented by Dietrichs \[[@CR18]\] in his pioneering study on cats. A subsequent series of neuroanatomical investigations from Haines, Dietrichs, and other colleagues \[[@CR19], [@CR20]\] on various mammals and nonmammalian vertebrates further substantiated the direct bidirectional connections between the cerebellum and the hypothalamus, the cerebellar-hypothalamic circuits. Since the cerebellar-hypothalamic circuits extensively exist and appear to be stronger in species ascending the phylogenetic scale, the connections may be phylogenetically old pathways \[[@CR19]\]. The neurotransmitters in the hypothalamocerebellar projections have not been well known so far, however, a growing body of data has provided strong evidence that histamine is a potential candidate and plays an important functional role in modulating activity of the cerebellar circuitry. In this review, the structure and function of hypothalamic histaminergic projections in the cerebellum are summarized and discussed.
Review {#Sec2}
======
Origination of histaminergic afferents in the cerebellum {#Sec3}
--------------------------------------------------------
In the cerebellar-hypothalamic circuits, the direct hypothalamocerebellar projections arise from widespread nuclei/regions in the hypothalamus, including the lateral, posterior, and dorsal hypothalamic areas, the dorsomedial and ventromedial nuclei, the periventricular zone/nucleus, the lateral mammillary and supramammillary nuclei, as well as the tuberomammillary (TMN) nucleus \[[@CR17], [@CR19]\]. Using an immunofluorescence technique, Ericson et al. \[[@CR21]\] demonstrated Fast Blue-labeled [l]{.smallcaps}-histidine containing neurons in the TMN after cerebellar injections. In fact, series of studies have ascertained that the TMN is not only the origination of hypothalamocerebellar histaminergic afferents (Figure [1](#Fig1){ref-type="fig"}), but also the specific sole region of origin for the whole central histaminergic system in the brain \[[@CR14], [@CR16]\].Figure 1**Hypothalamic histaminergic afferents in the cerebellum.** Cerebellar histaminergic afferent fibers originate from the tuberomammillary nucleus in the hypothalamus and project to both of the cerebellar cortex and nuclei. They parallelly modulate the Purkinje cells, granule cells and nuclear neurons via H2 and/or H1 receptors and sequentially influence the outputs of the cerebellum. CF, climbing fiber; CN, cerebellar nuclei; GC, granule cell; H1, histamine H1 receptor; H2, histamine H2 receptor; MF, mossy fiber; PC, Purkinje cell; PF, parallel fiber; TMN, tuberomammillary nucleus.
The TMN is a small nucleus located in the posterior hypothalamus. The histaminergic neurons in the TMN mostly have large somata (20--30 μm diameters) with resting potential of about -50 mV. These neurons are spontaneously active with slow regular firing rate at 1--4 Hz and mean mid-amplitude duration of action potential at 1--3 ms \[[@CR22]\]. Although hyperpolarization-activated cyclic nucleotide-gated (HCN) channels are expressed in histaminergic neurons, they are not responsible for maintaining the neuronal spontaneous activity as a pacemaker. A complex mechanism involving Na^+^, K^+^ and Ca^2+^ conductances contributes to the pacemaker properties \[[@CR14]--[@CR16]\]. Importantly, the firing rate and pattern of histaminergic neurons varies in different behavioral states, with a ratio of 1.5 between firing rates of histaminergic neurons in active and quiet waking in the cat \[[@CR23], [@CR24]\], suggesting the central histaminergic system is closely related to not only wakefulness but also movement.
Innervation of histaminergic afferents in the cerebellum {#Sec4}
--------------------------------------------------------
By means of immunocytochemistry using anti-histidine decarboxylase (HDC, the enzyme catalyzing the reaction that produces histamine) antibody or antiserum against histamine, the detailed distribution of histaminergic fibers in the cerebellum has been successively examined in the guinea pig, rat, tree shrew, and human \[[@CR9], [@CR25]--[@CR27]\]. In the rat cerebellum, HDC-immunoreactive fibers are scattered in all three cerebellar cortical layers, the molecular, Purkinje, and granular layers, rather than concentrated in any specific region \[[@CR25]\]. However, other studies did not find any histaminergic afferents in rat cerebellum \[[@CR28]\] or very low density in the cerebellar cortex \[[@CR29]\]. Similar to those in the rat, the histaminergic fibers are sparsely distributed in all cortical layers in the guinea pig cerebellum, with more denser fiber networks in the vermis and flocculus and less fiber density in the cerebellar nuclei \[[@CR26]\]. However, more histamine-immunoreactive fibers innervate cerebellar nuclei in the tree shrew \[[@CR27]\]. In human cerebellar samples, a moderate density of histaminergic afferents has also been observed in the molecular layer, and more fibers have been seen in the granular cell layer. Additionally, these fibers run parallelly to the Purkinje cell layer after traversing it perpendicularly \[[@CR9]\].
The histaminergic fibers share many morphological similarities, including distribution, orientation, branching patterns, and ending sites, with the serotoninergic, noradrenergic and neuropeptidergic axons in the cerebellar cortex (Figure [1](#Fig1){ref-type="fig"}). On the basis of these structural properties, the histaminergic afferent fibers in the cerebellar cortex are considered to be classified as multilayered fibers. Furthermore, the most endings of histaminergic fibers do not make typical synaptic specializations but form varicosities. The varicose rather than synaptic contact pattern, together with the dispersive innervation of hypothalamic histaminergic afferents in the cerebellar cortex and nuclei, indicates an extensively modulatory role of histamine in the cerebellar circuitry.
In the TMN neurons, histamine is synthesized from [l]{.smallcaps}-histidine through oxidative decarboxylation by HDC. Then, histamine is stored in neuronal somata and especially in axon varicosities, where it is carried into vesicles through the vesicular monoamine transporter VMAT-2 and released in a calcium-dependent manner upon arrival of action potentials \[[@CR14]\]. In the targets, histamine is inactivated through transfer of the methyl group from *S*-adenosylmethionine by histamine *N*-methyltransferase (HMT) or via oxidative deamination by diamine oxidase (DAO). However, HMT rather than DAO terminates histaminergic transmission in the cerebellum, since only HMT is expressed in the cerebellum \[[@CR30]\]. Inhibition of histamine methyltransferase enhance phosphoinositide turnover in the cerebellum \[[@CR30]\], which is mediated by histamine H1 receptors.
Expression and distribution of histamine receptors in the cerebellum {#Sec5}
--------------------------------------------------------------------
Up to date, four histamine receptors, H1-H4 receptors, have been cloned and identified, in which H1, H2 and H3 receptors are richly expressed in the central nervous system \[[@CR14]--[@CR16]\]. Although histamine H4 receptors are detected predominantly in the periphery, recent studies have also reported a functional expression of H4 receptors in human and rodent brain \[[@CR31]--[@CR33]\]. All histamine receptors are metabotropic and belong to the rhodopsin-like family of G protein-coupled receptors \[[@CR14]--[@CR16]\]. Among them, histamine H1, H2 and H4 receptors are postsynaptic and mediate mostly excitatory responses, whereas H3 receptors mediate presynaptic inhibitory actions as auto- or hetero-receptors \[[@CR14]--[@CR16]\]. Owing to autoradiographic mapping, in situ hybridization and immunohistochemistry techniques, expression and distribution of histamine receptors in the cerebellum have been revealed. Accumulating evidence demonstrates that all histamine H1, H2, H3 and H4 receptors exist in the cerebellum with various species difference.
### H1 receptor {#Sec6}
In situ hybridization studies have revealed that histamine H1 receptor mRNAs are expressed in granular layer and Purkinje cells of the guinea pig cerebellum \[[@CR34], [@CR35]\]. Using \[^3^H\]mepyramine or \[^125^I\]iodobolpyramine as sensitive probe, autoradiographic mapping results have showed a high density of H1 receptors in the molecular layer of the guinea pig cerebellum \[[@CR36], [@CR37]\]. Substantial levels of H1 receptors have also been observed in the cerebellum of cats and mice \[[@CR38], [@CR39]\]. However, compared with the guinea pig, mouse and cat cerebellum, much lower level of H1 receptors are expressed in rat cerebellum \[[@CR40]\].
### H2 receptor {#Sec7}
Using \[^125^I\]iodoaminopotentidine for radioligand binding and a ^33^P-labelled complementary RNA probe for in situ hybridization, an autoradiographic study have demonstrated that histamine H2 receptor and its mRNAs distribute in the guinea pig cerebellum, especially in Purkinje cell and granular layers \[[@CR41]\]. Nevertheless, in the rat brain, only low level of H2 receptor mRNA expression has been detected in the cerebellum by northern blot hybridization \[[@CR42]\]. Interestingly, in the mouse cerebellum, from developmental point of view, H2 receptor mRNA levels present an increased tendency with age \[[@CR43]\]. The expression and location of H2 receptors have also been observed in the dentate nucleus of human and monkey cerebellum \[[@CR44]\].
### H3 receptor {#Sec8}
Histamine H3 receptor, located on the somata and axon terminals of histaminergic neurons, was identified as a presynaptic autoreceptor in the rat brain by Arrang et al. in 1983 \[[@CR45]\]. Besides acting as a presynaptic autoreceptor to modulate histamine synthesis and release, H3 receptor can also exert as a presynaptic heteroreceptor to inhibit the release of various other neurotransmitters \[[@CR46]\], such as noradrenaline, acetylcholine, glutamate and GABA. The expression and distribution of H3 receptors in the cerebellum were observed in rodents, pigs and humans \[[@CR47]--[@CR49]\]. In rats, using a ^33^P-labelled riboprobe for in situ hybridization, a strong mRNA expression of H3 receptor, probably the shorter isoform \[[@CR50]\], was found in most Purkinje cells as well as in the cerebellar nuclei, including the FNs and INs \[[@CR48]\]. But there was scarce or very low detectable binding of H3 receptors in the Purkinje cells indicated by [r]{.smallcaps}-\[^3^H\]α-methylhistamine or \[^125^I\]iodoproxyfan for autoradiography \[[@CR48], [@CR49]\], suggesting H3 receptors are expressed on efferent projections rather than somata or dendrites of the Purkinje cells in rats. Furthermore, immunohistochemical analysis using affinity-enhanced anti-H3 (349--358) antibodies demonstrated that high levels of H3 receptors were detected in Purkinje cell layer but low levels in granule layer of the mouse cerebellum \[[@CR47]\], whereas high mRNA expression of the receptors was observed in the guinea pig \[[@CR51]\]. By PET, low binding of H3 receptors with \[^11^C\]GSK189254 radioligand was also detected in human and pig cerebellum \[[@CR52], [@CR53]\]. These observations indicate that H3 receptor expression in the cerebellum varies among species.
### H4 receptor {#Sec9}
H4 receptor, the newly identified histamine receptor, is expressed predominantly in peripheral tissues and cells, such as blood, lung, gut and liver \[[@CR14], [@CR54]\]. However, the expression and localization of H4 receptor in the brain remain controversial in different reports \[[@CR32], [@CR55], [@CR56]\]. By using RT-PCR technique, Nakamura et al. reported that expression of H4 receptor mRNAs was not detected in the brain \[[@CR55]\]. While, RT-PCR results from other laboratories demonstrated an expression and distribution of H4 receptor mRNAs in various brain regions, including high level expression in rat cerebellum \[[@CR32]\] and mouse cerebellar granule layer \[[@CR31]\], and low level in human cerebellum \[[@CR56]\]. The exact expression and distribution of H4 receptors in the cerebellum still needs to be further studied.
Histaminergic modulation on cerebellar neuronal activities {#Sec10}
----------------------------------------------------------
Innervation of hypothalamic histaminergic afferents on cerebellar cortex and nuclei and expression of histamine receptors in cerebellar neurons strongly suggest that histaminergic afferents may hold a key functional position in the cerebellar neuronal circuitry. In fact, a growing body of data has provided substantial evidence that histamine excites cerebellar neurons \[[@CR57]--[@CR63]\]. Although the distribution of histaminergic afferents in the rat cerebellar cortex seem to be scattered or low, electrophysiological studies show substantial evidence that histamine increases neuronal activities in cerebellar cortical circuit in rats. In 1999, Li et al. first reported that histamine induced an excitation on rat cerebellar granule cells \[[@CR57]\], the interneurons relaying mossy fiber inputs via parallel fibers to Purkinje cells. In addition, histamine was found to excite Purkinje cells \[[@CR63]\], the principle neurons in cerebellar cortical circuit, as well as neurons in the cerebellar nuclei, including the FN \[[@CR58], [@CR60]\], IN \[[@CR59]\] and DN \[[@CR62]\]. Interestingly, the effects of histamine on these cerebellar neurons are uniform postsynaptic excitation with various underlying receptor mechanisms (Figure [1](#Fig1){ref-type="fig"}). H2 receptors mediate the histamine-induced excitation on Purkinje cells and cerebellar nuclear neurons in rats \[[@CR58]--[@CR60], [@CR62], [@CR63]\], whereas H1 and H2 receptors co-mediate the excitatory effect of histamine on granule cells with a predominant contribution of H1 receptors \[[@CR57]\]. Activation of H1 receptors in guinea pig cerebellum was also found to increase intracellular Ca^2+^ concentration in Purkinje cells \[[@CR64]\]. Although H3 and H4 receptors are expressed in the cerebellum, role of them in histaminergic modulation on cerebellar neurons remains largely unclear up to date. It is only reported that H3 receptors inhibit and H2 receptors facilitate noradrenaline release in the cerebellum in guinea pigs \[[@CR65]\].
It has been well known that histamine H1 receptor is coupled to G~q/11~ protein and phospholipase C (PLC), whereas G~s~ and protein kinase A underlies H2 receptor \[[@CR14]--[@CR16]\]. Following H1 receptor activation in neurons in other brain areas, leak potassium channels are blocked, or Ca^2+^-activated cation channels and/or Na^+^-Ca^2+^ exchangers are activated \[[@CR14]--[@CR16]\]. On the other hand, activation of H2 receptors in dorsal lateral geniculate relay neurons and hippocampal pyramidal cells enhances the hyperpolarization-activated cation current (I~h~) and/or inhibits a calcium-activated potassium conductance \[[@CR14]--[@CR16]\]. The whole downstream signal transduction pathways of histamine receptors in different cerebellar cortical and nuclear neurons and the underlying ionic mechanisms have not yet been revealed.
On the other hand, histamine may influence cerebellar neuronal activity through its actions on the cerebellar glial cells. It is reported that H1, H2 and H3 receptors are all expressed in the cerebellar astrocytes \[[@CR66], [@CR67]\], including Bergmann glial cells \[[@CR68]\]. And histamine elevates several biochemical properties of astrocytes in the cerebellum, such as the activities of ornithine decarboxylase and glutamine synthetase, and incorporation of \[^3^H\]thymidine into DNA, and thus regulates growth and development of astrocytes \[[@CR69]\]. Moreover, by using fura-2-based Ca^2+^ imaging, histamine was found to induce calcium entry in rat cerebellar astrocytes \[[@CR70]\].
Intriguingly, besides cerebellar neurons, histamine also excites cerebellar target structures, in which vestibular nuclear complex in the brainstem plays a critical role in control of muscle tone and posture \[[@CR71], [@CR72]\]. The vestibular nuclear complex comprises four main nuclei, lateral (LVN), medial (MVN), inferior (IVN), and superior (SVN) vestibular nucleus. All of these four nuclei receive direct hypothalamic histaminergic innervations \[[@CR73]--[@CR75]\] and express histamine receptors \[[@CR41], [@CR48], [@CR76], [@CR77]\]. In consistent with the effect of histamine on cerebellar neurons, histamine induces an excitatory response of the neurons in vestibular nuclei. Extracellular recordings and whole-cell patch-clamp recordings in vitro showed that histamine directly excited MVN, SVN, and IVN neurons via postsynaptic H1 and H2 receptors \[[@CR78]--[@CR80]\] and depolarized LVN neurons through H2 receptors \[[@CR81]\]. Na^+^-Ca^2+^ exchangers coupled to H1 receptors and HCN channels linked to H2 receptors contribute to the histamine-induced depolarization on MVN neurons \[[@CR81], [@CR82]\]. Presynaptic H3 receptor also holds a key position in vestibular nuclear circuit \[[@CR83], [@CR84]\] and even in vestibular compensation \[[@CR83]--[@CR85]\], however, its role in modulation of vestibular nuclear neuronal activity has not been reported.
It is noteworthy that the actions of histamine on cerebellar cortical and nuclear neurons as well as vestibular nuclear neurons are homogeneous excitation. Thus, the hypothalamic histaminergic afferent system acts to uniformly and parallelly excite components in the cerebellar circuitry as well as the cerebellar target structure, vestibular nuclear complex. Due to histaminergic varicose endings and histamine metabotropic receptors, the hypothalamic histaminergic afferent system may not transmit fast signals, but act as a biasing force to influence electrophysiological properties of cerebellar and vestibular neurons and hold their excitability and sensitivity at an appropriate level for responding to inputs coding changes in internal and external environments. In this way, the histaminergic afferent inputs may extensively modulate the sensorimotor integration in the cerebellar and vestibular circuits and sequentially influence cerebellar-related motor behaviors.
Physiological function of histaminergic afferents in the cerebellar-related behaviors {#Sec11}
-------------------------------------------------------------------------------------
The central histaminergic nervous system has been implicated in many nonsomatic basic physiological functions, such as sleep-waking cycle, energy and endocrine homeostasis, synaptic plasticity, and learning \[[@CR14]--[@CR16]\]. Recently, role of histamine and histaminergic system in somatic motor control receives increasing attention. Intraventricular administration of histamine produced a biphasic effect in spontaneous locomotor activity with an initial transient hypoactivity followed by hyperactivity \[[@CR86], [@CR87]\]. Depletion of brain histamine or knockout of histamine receptors influenced motor behaviors \[[@CR88]--[@CR90]\]. The activity levels, such as wheel-running and spontaneous locomotion, in the HDC knock-out mice were lower than those in the wild types \[[@CR91]\]. Knockout of H1 receptors in mice altered ambulatory activity and reduced exploratory behavior \[[@CR89]\]. The H3 receptor-deficient mice showed a decrease in overall locomotion, wheel-running behavior, and stereotypic responses \[[@CR90]\]. Interestingly, bilateral microinjection of histamine into the cerebellar FNs or INs, two final output nuclei of the spinocerebellum, does not influence overground locomotion in rats in an open field \[[@CR58], [@CR61]\]. However, microinjection of histamine into the FNs and INs significantly lengthens the endurance time of rats on an accelerating rota-rod (Figure [2](#Fig2){ref-type="fig"}) and shortens the time that rats spend traversing a balance beam, which is mediated by H2 receptors \[[@CR58], [@CR61]\], indicating a promotion of histamine on cerebellum-mediated motor balance and motor coordination. Furthermore, microinjection of histamine into bilateral FNs narrowed stride width of footprint but did not influence wire suspension, whereas microinjection of histamine into bilateral INs increased stride length and promoted suspension \[[@CR58]\] (Figure [3](#Fig3){ref-type="fig"}), suggesting that cerebellar histaminergic afferent system may precisely modulate trunk, proximal and distal muscles via biasing the FN and IN.Figure 2**Histamine promotes motor balance and motor coordination in accelerating rota-rod via H2 receptors in the cerebellar interpositus nuclei. (A)** Motor performances of rats microinjected with normal saline, GABA, histamine, ranitidine (antagonist for H2 receptor) and triprolidine (antagonist for H1 receptor) in accelerating rota-rod. **(B)** Reversal effect of histamine on ranitidine-injected rats. \**P* \< 0.05; \*\**P* \< 0.01. Modified from Song et al., Neuroscience, 140:33--43, 2006.Figure 3**Histamine precisely modulates trunk, proximal and distal muscles through the cerebellar fastigial and interpositus nuclei, respectively. (A)** Walking track of hindfeet of a normal rat. **(B)** Microinjection of histamine into the fastigial rather than interpositus nuclei induced a narrower stride width. **(C)** Microinjection of histamine into the interpositus but not fastigial nuclei induced a longer stride length. **(D)** Microinjection of histamine into the interpositus but not fastigial nuclei increased the endurance time of suspension. \*\**P* \< 0.01. SL, stride length; SW, stride width. Modified from He et al., Behav Brain Res., 228:44--52, 2012.
Besides somatic motor control, cerebellum also actively participates in many basic nonsomatic regulations and even high cognitive functions \[[@CR17], [@CR92]\]. Interestingly, recently, histamine has been found to be involved in the cerebellar-mediated emotional memory consolidation. Microinjection of histamine into the cerebellar vermis impairs emotional memory consolidation in mice in the elevated plus-maze \[[@CR93]\]. The impairment is mediated by H1 rather than H2 receptors \[[@CR94]\]. However, via H2 receptors in the cerebellum, histamine enhances memory consolidation of inhibitory avoidance learning in mice \[[@CR95]\]. These results indicate that cerebellar histaminergic afferent system may be extensively involved in cerebellar physiological functions.
Histamine and cerebellar ataxia {#Sec12}
-------------------------------
Cerebellar ataxia, a form of ataxia associated with lesions to the cerebellum, is a complex motor disturbance that involves the planning and execution of movements and reduces movement accuracy and coordination \[[@CR96]\]. Cerebellar ataxia presents with symptoms of an inability to coordinate balance, gait, extremity, and eye movements \[[@CR97]\]. Since histaminergic afferent system plays an important role in cerebellar functions, histaminergic reagents may become potential drugs for treatment of cerebellar ataxia. Betacerc (betahistidine, an antagonist for H3 receptor and a weak agonist for H1 receptor) ameliorates symptoms of static ataxy in patients with cerebellar ataxia \[[@CR98]\]. Ciproxifan, a potent H3 receptor antagonist, enhances MK-801 (dizocilpine, a non-competitive antagonist for NMDA receptor) produced ataxia and motor impairment \[[@CR99]\]. Cetirizine, selective H1 receptor antagonist, decreases the falling off latency from the rota-rod and potentiates the effects of ethanol-induced ataxia \[[@CR100]\]. The reasons why betacerc and ciproxifan exert opposite effects on ataxias still needs further investigation, Betacerc is also a weak agonist for H1 receptors and different causes of ataxias may be account for it. Although these clinical and experimental results are very preliminary, they provide a new insight and indicate a possibility of using histaminergic reagents to ameliorate symptoms of cerebellar ataxia.
Conclusion {#Sec13}
==========
Histaminergic afferent system in the cerebellum, despite being a third type of cerebellar afferents, plays an important modulatory role in the cerebellar circuitry and actively participates in the cerebellar somatic motor and nonsomatic functions. Different from the serotoninergic and noradrenergic fibers originating from lower brainstem, histaminergic afferents in the cerebellum arise from the hypothalamus, a higher center for visceral and autonomic regulation. Thus, the hypothalamocerebellar histaminergic projections bridge nonsomatic center, the hypothalamus, to somatic structure, the cerebellum. These connections and especially the histaminergic modulations may not only endow the cerebellar circuitry with an appropriate functional state, but also form a vital part of the somatic-nonsomatic integration, which is critical for generating an integrated and coordinated behavioral response to changes in internal and external environment.
Although clinical use of histaminergic reagents in the therapy for cerebellar ataxia is still in exploration, intensive studies on function and receptor and ionic mechanisms of the histaminergic modulation on cerebellar circuitry may provide a new target for clinical treatment of cerebellar ataxia.
**Competing interest**
The authors declare that they have no competing interests.
**Authors' contributions**
BL drafted the manuscript and figures. JNZ designed the review, conceived the figures, and drafted the manuscript. JJW gave comments and suggestions based on the area reviewed and helped to draft the manuscript. JNZ and JJW developed conclusions. All authors read and approved the final manuscript.
Researches from our group were supported by grants 31070959, 31071021, 31171050, 31330033, and 91332124 from the National Natural Science Foundation of China; RFDP grant 20100091110016, SRFDP/RGC ERG Joint Research Scheme grant 20130091140003 and NCET Program from the State Educational Ministry of China; grant BK2011014 from the Natural Science Foundation of Jiangsu Province, China; and grant 2013 T60520 from China Postdoctoral Science Foundation.
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#s0001}
============
Psychosocial ill health in society is growing rapidly. The transition to parenthood is a joyful event for the majority of new mothers and fathers; however, it can include depressive symptoms and parental stress ([@C1]--[@C3]), which can negatively affect themselves, their partner, and not least their child ([@C4]--[@C6]). Depressive symptoms and stress in parents are associated with an increased risk of separation ([@C7]), which makes it even more important to focus on depressive symptoms and parental stress.
The prevalence of depressive symptoms in early parenthood is reported to be 5%--20% in mothers ([@C5],[@C7],[@C8]) and 3%--10% in fathers ([@C6]) and is associated with problems in children, such as stress and internalizing problems in both boys and girls ([@C10]). There is a link between depressive symptoms and parental stress ([@C3],[@C11]), and they influence one another over time ([@C12]). Mothers often report higher parental stress than fathers ([@C10],[@C12]--[@C16]); parents with poor health have higher stress levels than parents with good health ([@C17]). Predictors of parental stress are: low education, not living with the partner in early pregnancy ([@C16]), child's behaviour problem, low self-esteem, lack of social support, breast-feeding problems, lack of time with the child, and depression ([@C3]).
Parental stress affects parents' health negatively, especially mothers ([@C17]). Parents with lower sense of coherence (SOC) report more parental stress than parents with higher SOC ([@C15]). A new source of support is websites where parents can get information, advice, and new friends ([@C18]). Some parents lack social support and role models ([@C19]). Even though governments in many countries support gender equality, studies in early parenthood often focus on mothers. Further, studies considering mothers and fathers in the same family are still rare. The aim of the present study was to determine whether there is an association between depressive symptoms and parental stress among mothers and fathers in early parenthood. Our hypothesis was that parents with depressive symptoms had higher parental stress compared with parents without depressive symptoms.
Materials and methods {#s0002}
=====================
The BiT study (*Barnhälsovård i Tiden*, Child Health Care Today) was a longitudinal cohort study in the county of Västmanland, Sweden, which has a population of about 250,000. The participants were Swedish-speaking parents of children born in the years 2004--2006 from eight Child Health Centres CHCs. During this period, 521 children were born in the study region. The CHC nurses gave all Swedish-speaking new parents oral and written information about the study and asked if they would like to participate. The parents were assured that their participation was voluntary and that they could withdraw from the study at any time, with no effects on their child's health care. The Central Ethics Committee in Stockholm approved the study. A baseline questionnaire was distributed to the parents by the nurse during their first visit to the CHC and was then returned by the parents in prepaid envelopes. After 3 weeks, a reminder was sent by post, and a second reminder was given by telephone after 5 weeks. Three and 18 months after their children's births, new questionnaires and return envelopes were sent to the parents' homes, with reminders as at baseline. The parents were informed that it was important to complete the questionnaires individually. The parents were recruited consecutively, regardless of whether or not the child was their first. [Figure 1](#F1){ref-type="fig"} provides an overview of the participants and instruments using a timeline from baseline to 18 months after childbirth. These parents have been used for a previous study, but now other extended measures have been carried out ([@C20]).
{#F1}
Baseline questionnaire {#s0003}
----------------------
The baseline questionnaire contained demographic questions about the parent's age, the child's sex, whether or not it was the first child, and the parent's education level and socio-economic status (SES) (divided into three categories: manual workers, non-manual employees, and self-employed). The parents were also asked whether they had a parental role model with the question 'Have you had any role model for your role as a mother/father?'; the response options were: 'Yes' or 'No, I have no role model'. This was followed by 'If yes, who?'
Three-month questionnaire {#s0004}
-------------------------
*Edinburgh Postnatal Depressive Scale (EPDS).* The EPDS is a screening instrument to identify depressive symptoms ([@C21]), validated among mothers and fathers ([@C22]) and translated into several languages including Swedish. The scale includes 10 items with a total score of 0--30, where a higher score indicates more depressive symptoms. To estimate a person's level of depressive symptoms, all 10 questions must be answered. A cut-off point of ≥10 is recommended to identify risk of postnatal depression. In the statistical analyses, EPDS scores were treated as continuous data, and a cut-off point of ≥10 was used.
*Sense of Coherence (SOC).* To measure SOC, we used the three-question instrument SOC-3 suggested by Lundberg and Nyström Peck ([@C23]). The three questions address: Manageability ('Do you usually see a solution to problems that others find hopeless?); Meaningfulness ('Do you usually feel that your daily life is a source of personal satisfaction?)'; and Comprehensibility ('Do you usually feel that things that happen in your daily life are hard to understand?'). The response alternatives were 'No' (2 points), 'Yes, sometimes' (1 point), or 'Yes, usually' (0 points); with reversed scoring for comprehensibility. The total score is 0--6, where a score of ≤2 indicating a strong SOC and ≥3 indicating a poor SOC ([@C24]).
Eighteen-month questionnaire {#s0005}
----------------------------
*Swedish Parenthood Stress Questionnaire (SPSQ).* The SPSQ measures parental stress and consists of 34 items within five sub-areas. *Incompetence* focuses on general experiences of care-giving, feelings of incompetence in the parental role, and the difficulty of parenthood. *Role restriction* is about interests and activities restricted by parental responsibilities. *Social isolation* includes social contacts outside the family. The sub-area *Spouse relationship problems* describes the social experiences within the family. *Health problems* deals with the parents' physical health such as physical fitness, fatigue, and infections. The response options range from 'Strongly disagree' to 'Strongly agree' on a scale from 1 to 5, with a total possible score of 170 ([@C25]). Higher scores indicate higher stress.
Statistical analyses {#s0006}
--------------------
Descriptive statistics for the variables are presented as a percentage (number) except for age, which is presented as mean and standard deviation (SD). Comparisons between mothers and fathers were analysed using statistical methods for paired data, because in the context data from the mother and the father who are a couple are dependent. The association between EPDS scores and SPSQ scores in mothers and fathers was assessed using the Spearman's rank correlation (*ρ*), and the occurrence of depressive symptoms and SPSQ scores was assessed with the Wilcoxon signed-rank test. Linear regression analysis was performed to calculate the regression coefficients (*β*) with corresponding 95% confidence intervals for the outcome SPSQ scores of the mothers and fathers. We first performed univariable analyses of all relevant independent variables. Variables with *P* \< 0.2 were then considered in the multivariable analyses. These were performed with forward and backward stepwise selection (5% for inclusion and 5.1% for exclusion) to protect against potential collinearity that could disturb the analyses. Adjusted *R*^2^ was used to identify the proportion of the total variability explained by the model. A residual goodness-of-fit test was used to test the overall fit of the logistic regression model. SPSS (version 20) was used for the statistical analyses, and statistical significance was set at *P* \< 0.05 (two-sided).
Results {#s0007}
=======
The participants' demographic data, SES, parental role model status, and SOC are presented in [Table 1](#TB1){ref-type="table"}.
######
Parents' age, whether this was their first child, education, socio-economic status (SES), whether parental role model, and sense of coherence (SOC).
Mothers Fathers
-------------------------------- -------------- --------------
Parent's age, mean (SD) (*n*) 30 (5) (401) 33 (6) (396)
First child, % (*n*)
Yes 40 (161) 44 (175)
No 60 (240) 56 (221)
Education, % (*n*)
Comprehensive school, 9 years 6 (26) 6 (25)
High school, ≤12 years 60 (241) 74 (290)
University, \>12 years 34 (133) 20 (77)
SES, % (*n*)
Manual workers 59 (235) 70 (277)
Non-manual workers 38 (151) 22 (86)
Self-employed 4 (15) 8 (30)
Parental role model, % (*n*)
Yes 53 (210) 71 (276)
No 48 (188) 29 (111)
SOC at 3 months, % (*n*)
Strong 88 (296) 84 (278)
Poor 12 (40) 16 (55)
SOC at 18 months, % (*n*)
Strong 86 (274) 84 (259)
Poor 14 (44) 17 (52)
Depressive symptoms 3 months after childbirth {#s0008}
---------------------------------------------
The EPDS score was significantly higher for the mothers (mean 5.5, SD 4.3) than for the fathers (mean 4.0, SD 3.8) (*P* \< 0.001). Overall, 18% (*n* = 59) of the mothers and 9% (*n* = 28) of the fathers had an EPDS score ≥10. The results indicate that nearly a quarter (23%) of the children had at least one parent with depressive symptoms (not shown in the table). The correlation between the mothers' and fathers' EPDS scores was *ρ* = 0.296 (*P* \< 0.001).
Parental stress 18 months after childbirth {#s0009}
------------------------------------------
The mothers estimated higher total parental stress than the fathers (*P* \< 0.001) ([Table 2](#TB2){ref-type="table"}). Both mothers and fathers had their highest level of stress in the sub-area 'Role restriction'. The mothers had the lowest level of stress in the sub-area 'Social isolation', and the fathers in the sub-area 'Incompetence'. The mothers perceived higher levels of stress than the fathers did in all sub-areas except for 'Social isolation', where the fathers perceived higher stress. The correlation between the mothers' and fathers' SPSQ scores was *ρ* = 0.387 (*P* \< 0.001).
######
Swedish Parenthood Stress Questionnaire: five sub-area scores in mothers and fathers 18 months after childbirth.
Sub-area Number of mothers/fathers Mothers (mean ± SD) Fathers (mean ± SD) *P* value[^a^](#TF1){ref-type="table-fn"}
--------------------------------------- --------------------------- --------------------- --------------------- -------------------------------------------
Incompetence 314/311 2.1 ± 0.6 1.9 ± 0.6 \<0.001
Role restriction 312/308 3.3 ± 0.8 3.1 ± 0.8 \<0.001
Social isolation 316/310 2.0 ± 0.6 2.2 ± 0.6 \<0.001
Spouse relationship problems 314/310 2.2 ± 0.9 2.1 ± 0.7 0.004
Health problems 316/311 2.6 ± 0.8 2.5 ± 0.7 0.027
Total[^b^](#TF2){ref-type="table-fn"} 307/301 2.4 ± 0.5 2.3 ± 0.5 \<0.001
Wilcoxon signed-rank test.
Mothers and fathers who answered all sub-areas of parental stress.
Depressive symptoms 3 months and parental stress 18 months after childbirth {#s0010}
---------------------------------------------------------------------------
The correlation between the mother's EPDS score and SPSQ score was *ρ* = 0.467 (*P* \< 0.001), and between the mother's EPDS score and father's SPSQ score it was *ρ* = 0.312 (*P* \< 0.001). The correlation between the father's EPDS score and mother's SPSQ score was *ρ* = 0.170 (*P* = 0.005), and between the father's EPDS score and father's SPSQ score it was *ρ* = 0.489 (*P* \< 0.001) (not shown in the table).
Linear regression analyses {#s0011}
--------------------------
Significant predictors in the multivariable regression analyses of mother's SPSQ score at 18 months were mother's EPDS score at 3 months, mother's SOC at 18 months, and father's SPSQ score at 18 months ([Table 3](#TB3){ref-type="table"}). Significant predictors of father's SPSQ score at 18 months were father's EPDS score at 3 months, father's SOC at 18 months, and mother's SPSQ score at 18 months ([Table 3](#TB3){ref-type="table"}). The results from the univariable regression analyses are presented in Supplementary Tables I and II (available online).
######
Multivariable regression analysis for mothers' and fathers' Swedish Parenthood Stress Questionnaire (SPSQ) scores 18 months after childbirth.
SPSQ score
----------------------------------------------------- ------------------- ------ ---------
Mothers (*n* = 264)[^a^](#TF3){ref-type="table-fn"}
Mother's EPDS score at 3 months 0.04 (0.02--0.05) 5.75 \<0.001
Mother's SOC at 18 months 0.52 (0.37--0.67) 6.70 \<0.001
Father's SPSQ score at 18 months 0.21 (0.11--0.32) 3.98 \<0.001
Fathers (*n* = 252)[^b^](#TF4){ref-type="table-fn"}
Father's EPDS score at 3 months 0.05 (0.04--0.07) 6.98 \<0.001
Father's SOC at 18 months 0.20 (0.06--0.39) 2.87 0.004
Mother's SPSQ score at 18 months 0.29 (0.19--0.36) 5.05 \<0.001
Adjusted *R*^2^ 35.9%.
Adjusted *R*^2^ 30.9%.
Discussion {#s0012}
==========
The main finding for both mothers and fathers was the association between the parents' depressive symptoms and parental stress. The mother's parental stress after 18 months was positively associated with her own depressive symptoms and/or SOC and/or the partner's parental stress. This was similar for the fathers. As is well known, Sweden is one of the most gender-equal countries in the world in terms of extensive and egalitarian parental leave policies ([@C26]). Although many studies confirm differences between mothers and fathers during early parenthood, we found no significant difference between genders concerning parental stress in the present study. Own depressive symptoms and poor SOC affect parental stress for both mothers and fathers. This is in accordance with Edhborg et al., who showed that Swedish fathers with depressive symptoms experience the first year after childbirth similarly to mothers, with turbulent everyday life involving parental stress and lack of social support ([@C13]). Furthermore, we state that a partner's stress affects the other parent's stress. We speculate that a parent's stress can affect the partner's stress because in Sweden both parents share the duties in the family more than previously, as found by Strandh and Nordenmark ([@C27]). Perhaps a feeling of losing control could occur if a parent who was used to taking decisions alone is now required to involve the partner. Dissatisfaction with partner support and parental stress are stronger predictors of poor health in fathers than in mothers ([@C17]), which may partly explain why the partner's stress affects the other parent's stress as described in our results.
The association we found between depressive symptoms and parental stress is in accordance with Saisto et al. ([@C3]). A possible explanation may be that a person who experiences depressive symptoms might be more susceptible to parental stress. We also found that parents with poor SOC experienced higher parental stress than did parents with strong SOC. A possible explanation may be that some new parents do not find their situation manageable, with the child taking a lot of their time, and therefore experience more stress. The highest level of parental stress was in the sub-area of 'Role restriction' for both mothers and fathers, which is in accordance with other studies ([@C11],[@C15]). Parents with poor SOC have a strained economic situation and come from lower social classes ([@C28]), which can also contribute to parental stress.
Social support (contact with friends and relatives) has a positive effect on SOC ([@C28]). Therefore, maintaining contact with friends and relatives could be important for new parents. Today, parents seek friends through social media and various websites, but this seems to create more parental stress, especially for mothers ([@C18]). Perhaps they require more practical support than advice about parenthood. This is in accordance with Saisto et al. ([@C3]), who describe parents' lack of support in early pregnancy. In future studies, it could be interesting to investigate how parental stress and child-care are affected by social media, which might put pressure on parents as they attempt to live up to all the demands of society and their friends.
The mothers experienced more depressive symptoms and higher parental stress than the fathers, which is consistent with previous studies ([@C15],[@C29]). The mothers perceived higher total parental stress than the fathers in all sub-areas except for 'Social isolation', where the fathers perceived higher parental stress than mothers did. A possible explanation may be that mothers experience greater conflicts between work and household demands than fathers ([@C27]). Even though Sweden has one of the most comprehensive and egalitarian parental leave policies in the world ([@C26]), mothers still use three-quarters of the parental leave taken ([@C30]). Another possible explanation for women experiencing more parental stress may be that they can express their feelings better than men, as they are more used to expressing emotions.
A Swedish study found that one partner's depressive symptoms affect the other partner's bonding with the child ([@C29]). In the present study, we found that a parent's parental stress affected the other parent's stress; however, a parent's depressive symptoms did not affect the partner's stress. We must emphasize how important it is that parental stress affects not only the individual but also the partner and ultimately the child's well-being. The parents' well-being obviously affects the child's well-being, and the child's well-being affects the well-being of the parents, which can lead to a vicious circle that affects family health adversely.
Strengths and limitations {#s0013}
-------------------------
Studies often focus on mothers' experiences of parenthood, but studies focusing on the fathers' experiences are rare. A strength of the present study is that a large number of mothers and fathers of the same child participated. Another strength is that parents from eight different CHCs participated, and this reduced the risk of influences of individual nurses.
The first limitation is that the parents who answered the questionnaires might have fewer depressive symptoms and parental stress than the non-responding parents, as motivation and strength are required to complete the questionnaires. However, two advantages of using questionnaires are that they only take a short time to complete and they have simple scoring systems. A second limitation is that only Swedish-speaking parents were included. Finally, the result might not be generalizable to contexts other than that described in the present study.
Conclusions {#s0014}
===========
In the present study, we found an association between mothers' and fathers' depressive symptoms and parental stress during early parenthood. However, a parent's stress obviously affects the partner, and this knowledge is important for health professionals as well as parents. This knowledge can be useful for health professionals in preventing and detecting depressive symptoms and parental stress, and for developing interventions. Both mothers and fathers should receive support and guidance during pregnancy and while bringing up their children to optimize the conditions for raising the child.
We thank all the mothers, fathers, and CHC nurses who participated in the study.
Birgitta Kerstis, Eva Nohlert, John Öhrvik, and Margareta Widarsson participated in the data analysis, the drafting and critical review of the manuscript. John Öhrvik provided statistical expertise. All authors have seen and approved the final version of the manuscript.
Disclosure statement {#s0015}
====================
The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.
Funding information {#s0016}
===================
This study was funded by the Centre for Clinical Research Västmanland and the Uppsala-Örebro Regional Research Council (Regionala forskningsrådet I Uppsala- och Örebroregionen RFR 151751). The study sponsors had no role in the data analysis, data interpretation, or writing of the report.
[^1]: Supplemental data for this article can be accessed [here]{.ul}.
| {
"pile_set_name": "PubMed Central"
} |
Introduction
============
Breast cancer is a leading cause of morbidity and mortality among female cancer patients worldwide. The disease is categorized according to at least three distinct molecular and clinical subtypes with dramatically different outcome and response to therapeutic agents [@b1],[@b2]. The luminal subtype (estrogen receptor \[ER\]-positive) is the most common and heterogeneous group and is treated with endocrine therapy. The human epidermal growth factor receptor 2 (HER2 or ERBB2)-amplified subtype has recently been successfully treated with anti-HER2 targeted approaches. The triple-negative breast cancer (TNBC) subgroup, lacking expression of hormonal receptors and HER2, has an unfavorable prognosis and is currently treated after surgery with standard chemotherapy due to its significant chemosensitivity. The incidence of TNBCs, also known as basal-like breast tumors, is increased in patients with germline Breast Cancer 1, Early Onset (BRCA1) mutations [@b2] and in premenopausal women of African ancestry [@b3]. Heterogeneity and complexity of breast cancer have been extensively detailed by large genomic and transcriptomic studies [@b4],[@b5] that must be translated into a personalized management of the disease based on genetic and molecular traits of each breast tumor, but also accounting for social and environmental conditions of each patient. In this context, it remains controversial whether racial/ethnic information can help to explain molecular features, etiological diversity and progression of breast cancer. Although there is abundant epidemiological evidence that race/ethnicity is associated with disparities in cancer incidence and mortality [@b6], and some studies have shown that such differences are affected by social--economic conditions leading to inequality in access to effective medical care or exposure to risk factors [@b6],[@b7], it is still unclear whether race/ethnic variability in cancer also reflects differences in cancer biology.
Compared with women in Western countries, Asian women still have lower incidence rates of breast cancer and lower median age at the diagnosis (48--50 years) [@b8]; however, the disease has increased dramatically over the past 30 years in urban areas of China and other Asian developing countries. Fan et al. [@b9] highlighted how socioeconomic development, accelerated urbanization, higher lifetime expectancy, and the aging of the Shanghai population have greatly contributed to the rapid change in risk factors and the increase in breast cancer incidence and mortality. Since this increase is likely to continue in the next decades across all of China [@b8], there is an urgent need for molecular and clinical detailing of breast cancer in this population. Epidemiological and genomic differences between Asian and Caucasian breast cancer have been reported based on data from independent analyses of Asian and Caucasian series [@b10],[@b11] and from heterogeneous cohorts of Asian/Pacific Islander women [@b12],[@b13], but no studies have been conducted to compare Chinese and Caucasian samples simultaneously, even though studies of breast cancer aggressiveness in Afro-American women have underscored the great advantage of simultaneous analysis in large cohorts of women of African or Caucasian ancestry in the US [@b3],[@b7]. Herein, we investigated biological diversity among breast cancers from Chinese and Caucasian women by comparing comprehensive gene and microRNA (miRNA) profiles of samples analyzed simultaneously and derived from a unique transethnic cohort of Chinese Han breast cancer patients surgically resected at Fudan Cancer Center of Shanghai (China) and of Caucasian Italian patients surgically resected at Fondazione IRCCS Istituto Nazionale dei Tumori of Milan (INT, Italy). Chinese and Caucasian comprehensive transcriptomic data were examined for patterns of similarity and dissimilarities using unsupervised and supervised analysis, and breast cancer molecular taxonomy was explored and compared in both Chinese and Caucasian gene profiles.
Materials and Methods
=====================
Samples and clinical data
-------------------------
Tissue samples from 78 Chinese Han and 97 Caucasian Italian consecutive primary breast tumors were collected for gene and miRNA profiling at Breast Surgery Units of Cancer Hospital-Fudan University of Shanghai and INT of Milan in 2007. Specimens were frozen in liquid nitrogen within 1 h in the surgical room at Shanghai and within 4 h in the Pathological Department at Milan, after which all samples were kept at −80°C in the Italian Center. Bilateral synchronous breast tumors, that were 2% and 3% in Chinese and Caucasian series, respectively, were excluded from the profiled cohort. Clinical--pathological data on consecutive breast cancer from 1057 Chinese and 1047 Italian patients were collected in years 2005 and 2006 in the same Institutions. The study was approved by the Medical Ethics Committee of both Institutions. All data and tissue samples were obtained upon informed consent and according to respective institutional rules.
Immunohistochemical analysis
----------------------------
ER status was determined by immunohistochemistry using mouse monoclonal anti-human ER antibody (Dako, Carpinteria, CA, clone 1D5) on Chinese tumors or using ER-ICA (estrogen receptor immunocytochemical assay) assay on Italian tumors (Abbott, Abbott Park, IL) according to the manufacturer's recommendations. Tumors were considered receptor-positive if more than 1% of malignant cells showed nuclear staining. To assess agreement of ER status determination between the two hospitals, independent pathologists in Milan and Shanghai reviewed a random sample of 100 breast cancer cases and found a Cohen's κ statistic of 0.94, indicating substantial agreement beyond chance and an overall concordance of 98%.
Subtyping of breast cancer samples according to immunohistochemical (IHC) determination of their hormonal and HER2 receptors was based on Carey classification [@b3]: Luminal HER2− (ER+ and/or PR+, HER2−); Luminal HER2+ (ER+ and/or PR+, HER2+); HER2+/ER− (ER−, PR−, HER2+) and ER−/PR−HER2−. HER2/neu status was determined using polyclonal rabbit anti-Human c-erbB-2 antibody (Dako) in both Chinese and Italian tumors. Staining was quantified using a score of 0, 1, 2 and 3; tumors with a score of 3 were considered HER2 positive.
RNA extraction
--------------
Total RNA was extracted using the miRNeasy Mini Kit (Qiagen, Valencia, CA) at INT from 50 to 100 mg of each Chinese and Italian tissue sample homogenized using a bench-top homogenizer (MM200, Retsch, Germany) in 1 mL of TRIzol reagent (Invitrogen, Life Technologies, Grand Island, NY). An aliquot of 1--2 *μ*g of total RNA was retained for miRNA analysis, while the remaining material was cleaned-up and treated with RNAse-free DNAse to remove genomic DNA traces, using the RNeasy Mini Kit (Qiagen) according to the manufacturer's instructions. Qualitative analysis of RNA was performed using Agilent RNA 6000 Nano kit and Agilent 2100 Bioanalyzer.
Gene and miRNA profiling
------------------------
RNA samples were processed for gene and miRNA profiling on the Illumina platform at the Functional Genomics Facility of INT. For gene profiling analysis, 800 ng of total RNA was reverse transcribed, biotin-labeled, and amplified using Illumina RNA TotalPrep Amplification kit (Ambion, Life Technologies, Grand Island, NY) as per the manufacturer's instructions. One *μ*g of each cRNA amplified sample was added to Hyb E1 hybridization buffer containing 37.5% (w/w) formamide and hybridized to array HumanHT-12-v3 expression Bead Chip (Illumina, Inc., San Diego, CA) at 58°C for 18 h. Array chips were washed and stained using 1 *μ*g/mL of Cy3-streptavidin (Amersham Biosciences). For miRNA profiling, miRNAs were amplified with the Illumina human_v2 MicroRNA expression profiling kit, based on the DASL (cDNA-mediated Annealing, Selection, Extension, and Ligation) assay according to the manufacturer's instructions. Briefly, 400 ng/sample of total RNA was converted to cDNA. After annealing with miRNA-specific oligo pools, PCR amplification and fluorescent labeling, probes were hybridized on Illumina miRNA BeadChips, allowing analysis of 1146 sequences covering 97% of miRNAs present in the miRBase v12.0. After hybridization, fluorescent signals were detected by the Illumina BeadArray^TM^ reader.
Microarray data preprocessing
-----------------------------
Raw expression data were collected from scanned images using Illumina BeadStudio v3.3.8 (Illumina, Inc.) and processed using the *lumi* package [@b14] from Bioconductor v2.11 [@b15]. Probes from gene expression data were reannotated using the *illuminaHumanv3.db* package v1.16.0. Quality control was performed on raw and processed data by evaluation of array intensity distributions, distances between arrays, and principal component analysis (PCA) for the identification of outliers. All samples passed quality-control procedures. Gene and miRNA raw data were log2-transformed, normalized with Robust Spline Normalization and filtered, keeping only the probes with a detection *P* \< 0.01 in at least one sample. Multiple probes representing the same gene were collapsed and the probe with the highest detection rate, that is, the percentage of samples in which the probe had a detection *P* \< 0.01, was selected. In the case of equal detection rates, the most variant probe according to interquartile range (IQR) was selected. Final data included 15,929 unique genes and 848 unique miRNAs. Expression data were deposited in the Gene Expression Omnibus data repository (GEO) with accession number GSE59595. Gene expression and microRNA raw data from Buffa et al. [@b16] were used as independent datasets in the comparison of Chinese and Caucasian Italian profiles and were downloaded from GEO (accession numbers GSE22219 and GSE22216, respectively) and subjected to the above preprocessing procedure. Common features in GSE22219 (Illumina human Ref-8 v1.0) and Chinese-Italian datasets were selected by gene symbol.
Gene expression data from 81 Taiwanese Chinese Han patients (GSE48390 [@b17]) and from 100 women of National University Hospital of Singapore (GSE36772) were downloaded from GEO and used to validate the lower prevalence of luminal A tumors in Chinese women. Both cohorts were profiled on Affymetrix platform. Raw CEL files were processed using the frozen robust multiarray analysis method [@b18]. Multiple probes representing the same gene were collapsed selecting the probe with the highest average expression.
Unsupervised analysis
---------------------
Groups of correlated genes and miRNAs were identified in Chinese and Caucasian Italian datasets separately, using the approach described by Callari et al. [@b19]. Briefly, genes with IQR \> 0.4 from the Chinese dataset were partitioned using hierarchical clustering with agglomerative average linkage as linkage criterion and 1- the Pearson's correlation coefficient as a distance measure. The resulting dendrogram was cut at a correlation value of 0.6 and clusters containing at least 10 genes were selected. Such thresholds were selected in order to consider sufficiently large groups of genes with biologically meaningful correlations, ensuring representation of pivotal breast cancer gene-clusters such as "ER" and "proliferation-related." A dendrogram was then generated for each cluster using Caucasian Italian gene expression data and the same correlation thresholds. A cluster was considered validated if at least three genes had a correlation \>0.6 in the Italian dataset. The same approach was applied to Chinese miRNA data, but in view of the different number of features and data characteristics, the correlation threshold was set at 0.4; clusters containing at least five miRNAs were selected and validated on the Caucasian Italian dataset if at least three miRNAs passed the correlation threshold. The same procedure was repeated in the reverse order, starting from the Caucasian Italian gene and miRNA datasets and validating the clusters in the corresponding Chinese datasets or starting from the GSE22219 and GSE22216 datasets and validating selected clusters in both Chinese and Caucasian Italian data. Overrepresentation of gene lists relevant to breast cancer [@b1],[@b20]--[@b22] was analyzed using the hypergeometric test, in which *P* \< 0.01 was considered significant.
Sample classification according to breast cancer molecular signatures
---------------------------------------------------------------------
Samples were classified as ER-positive and ERBB2-positive by setting the expression value thresholds corresponding to the local minima of the bimodal density distribution of the each gene, calculated using the *turnpoints* function of the *pastecs* package [@b23]. Hierarchical clustering with average linkage and Euclidean distance was applied to classify samples into breast intrinsic molecular subtypes, using the PAM50 gene signature [@b24] or to assign samples to the claudin-low subtype using the gene list of Prat et al. [@b25]. Five PAM50 genes (*NUF2*, *CXXC5*, *MIA*, *ORC6L*, and *MYBL2*) were not present on the platform or were filtered out as never detected.
Identification and stability analysis of extracellular matrix (ECM) clusters was performed using the Large Average Submatricies (LAS) biclustering algorithm and the statistical methods described in Triulzi et al. [@b26]. A list of genes including *CTSS*, *GZMK*, *MMP7*, *MMP9*, *SELL*, *SPOCK2*, and *VCAM1* [@b21] and the proposed ECM3 signature [@b26] were used to identify ECM1 and ECM3 clusters, respectively. Similarity between the intrinsic and ECM subtypes identified in the two datasets was assessed by subclass mapping using the SubMap module of GenePattern software v3.8 [@b27]. In addition, PCA was applied on global miRNA data using the first two principal components to compare the distributions of tumors in different subtypes.
Statistical analysis of gene and miRNA expression data
------------------------------------------------------
Differentially expressed genes in the Chinese and Italian samples or among biological subgroups were identified by linear modeling as implemented in the *limma* package [@b28], using ethnicity and molecular subtype as covariates. Multiple-testing correction was performed using the Benjamini--Hochberg false discovery rate (FDR). Genes and miRNAs with FDR \< 0.05 and absolute fold-change ≥2 between Chinese and Italian samples or in at least one of the possible pairwise comparisons between subtypes were considered significantly differentially expressed. Association between expression data and storage time was assessed through linear modeling using the *lm* function in R. Significance threshold was set at nominal *P* \< 0.001. Breast cancer signatures were downloaded from the GeneSigDB database [@b29] using the keywords "human" and "breast."
Statistical analysis of clinic-pathological variables
-----------------------------------------------------
Association between categorical variables and ethnicity was assessed using the chi-squared test or Fisher's exact test. Partial correlation was used to estimate the degree of association between ER status and age adjusted for the effect of a set of controlling variables (tumor size, grade, nodal status, ER−/PR−/HER2− vs. HER2+/ER− vs. Luminal HER2− vs. Luminal HER2+). Significance threshold was set at *P* \< 0.05.
Results
=======
Clinico-pathological features of Chinese/Caucasian cohort
---------------------------------------------------------
Table[1](#tbl1){ref-type="table"} lists the clinico-pathological characteristics of the profiled breast tumors from the consecutive cohort of Fudan Cancer Center of Shanghai and INT of Milan. Age of disease onset differed significantly in the two groups, with median ages of 50 and 60 years in the Chinese and Caucasian series, respectively, and with 47% of Chinese versus 68% of Caucasian women older than 50 years. Although not statistically significant, the percentage of ER-positive patients as assessed immunohistochemically was lower in the Chinese group (67% vs. 77%), according to Fan et al. [@b9], with a similar trend evidenced when ER status was assessed based on gene expression. Other clinico-pathological features such as histological grade, size and lymph nodal status were not significantly associated with race/ethnicity. Comparison of 1057 Chinese and 1047 Caucasian consecutive breast cancer patients from the same Institutions (Table[2](#tbl2){ref-type="table"}) to verify the prevalence of ER-positive disease in a larger cohort revealed a significant disparity between Chinese and Caucasian women in both age of incidence and ER status; 54% of women in the Chinese series were older than 50 years versus 69% of Caucasians (median age 51 and 59, respectively) and 63% of Chinese versus 85% of Caucasian tumors were ER-positive.
######
Clinical, pathological, and molecular characteristics of 78 Chinese and 97 Caucasian primary consecutive breast cancer patients surgically treated at Cancer Hospital-Fudan University of Shanghai and INT of Milan, respectively, and analyzed for gene and miRNA expression
Chinese series (%) Caucasian series (%) *P*-value
-------------------- -------------------- ---------------------- -----------
Number of patients 78 (100) 97 (100)
Age (years)
Median 50 60 0.009
Range 32--78 35--85
≤50 41 (53) 31 (32)
\>50 37 (47) 66 (68)
Size (cm)
Median 2.5 2.0 0.124
Range 1.0--7.0 0.4--11.0
≤2.0 47 (60) 45 (46)
\>2.0 31 (40) 50 (52)
NA 0 (0) 2 (2)
Histological grade
I + II 40 (51) 46 (47) 0.195
III 26 (33) 48 (49)
NA 12 (16) 3 (4)
Lymph node
Positive 35 (45) 46 (47) 0.498
Negative 36 (46) 36 (37)
NA 7 (9) 15 (16)
ER status (IHC)
Positive 52 (67) 75 (77) 0.162
Negative 26 (33) 22 (23)
ER status (gene)
Positive 45 (58) 69 (71) 0.090
Negative 33 (42) 28 (29)
HER2 status (IHC)
Positive 10 (13) 10 (10) 0.780
Negative 68 (87) 87 (90)
HER2 status (gene)
Positive 12 (15) 11 (11) 0.574
Negative 66 (85) 86 (89)
IHC subtypes
ER−/PG−/HER2− 17 (22) 13 (13) 0.218
HER2+/ER− 6 (8) 4 (4)
Luminal HER2− 51 (65) 70 (72)
Luminal HER2+ 4 (5) 10 (10)
ER, estrogen receptor; NA, not available; IHC, immunohistochemistry; HER2, human epidermal growth factor receptor 2.
######
Age and ER status of 1057 and 1047 primary consecutive breast cancer patients surgically treated in the years 2005--2006 at Cancer Hospital-Fudan University of Shanghai and INT of Milan, respectively
Chinese series (%) Caucasian series (%) *P*-value
-------------------- -------------------- ---------------------- -----------
Number of patients 1057 (100) 1047 (100)
Age (years)
Median 51 59 \<0.001
Range 23--85 25--92
≤50 487 (46) 324 (31)
\>50 570 (54) 723 (69)
ER status
Positive 663 (63) 885 (85) \<0.001
Negative 394 (37) 162 (15)
ER (estrogen receptor) status was assessed by immunohistochemistry.
Age at onset was significantly associated with ER status in both Chinese (adjusted mean age of 49.4 years in ER-negative and 53.3 years in ER-positive, *P* = 0.0049) and Caucasian series (adjusted mean age of 51.4 years in ER-negative and 57.1 years in ER-positive, *P* = 0.046).
Comparison of global gene and miRNA expression profiles of Chinese and Caucasian samples
----------------------------------------------------------------------------------------
Variability in specimen collection, storage, processing and analysis represents a major source of bias in translational research, producing artifacts and misinterpretation of high-throughput results. Accordingly, all specimens initially collected in Shanghai and Milan were subsequently stored, randomly processed, and analyzed in identical experimental conditions in the Italian center to minimize preanalytical, instrumental, and computational variability between the two series, enabling direct comparison of the gene and miRNA profiles. Agglomerative hierarchical clustering applied to the Chinese--Caucasian dataset validated homogeneity of Chinese and Caucasian expression profiles, since all samples clustered independently of their origin and tended to group according to hormonal receptor status (Fig.[1A](#fig01){ref-type="fig"}). Comparable results were observed in clustered miRNA expression profiles of Chinese and Caucasian samples (Fig.[1B](#fig01){ref-type="fig"}) and on cluster analysis of selected genes and miRNAs with high variation across samples (IQR \> 1) (data not shown). The association between storage time before freezing and gene and microRNA expression was assessed in the Italian dataset to verify whether this technical source of variability could affect the molecular comparison of the 2 groups. Forty-seven (0.3%) of 15,929 genes analyzed and 1 (0.1%) microRNA of 848 showed a significant dependence to storage time (*P* \< 0.001, [Table S1](#sd2){ref-type="supplementary-material"}). No overlap was observed between these 47 genes and 484 of 554 human breast signatures retrieved from GeneSigDB collection; the remaining 70 signatures included from 1 to 5 of the storage-associated genes ([Table S2](#sd2){ref-type="supplementary-material"}). Therefore, the storage time of samples before freezing impacted the expression of a very limited number of genes and miRNAs and the genes were not involved in pathways or biological processes relevant for breast cancer.
{#fig01}
To compare the molecular architecture of Chinese and Caucasian transcriptomes, we tested whether similar clusters of correlated genes, presumably cotranscribed and involved in the same biological processes, were identifiable in both datasets. We identified 59 clusters including at least 10 genes with a Pearson correlation greater than or equal to 0.6 in the Chinese profile and confirmed 54 of the clusters (92%) in the Caucasian dataset ([Table S3](#sd2){ref-type="supplementary-material"} and Fig.[1C](#fig01){ref-type="fig"}). The same approach was applied to miRNA data, where 31 (97%) of 32 clusters identified in the Chinese dataset were validated in the Caucasian dataset (Fig.[1D](#fig01){ref-type="fig"}). Because a major pitfall of this approach rests in the higher probability of validation in larger clusters, the size of clusters identified in Chinese and validated in Caucasian sets was examined to verify the conservation of gene number in each cluster pair; a similar number of genes (Pearson's coefficient = 0.94) and miRNAs (Pearson's coefficient = 0.95) was included in Chinese and Caucasian clusters. The entire procedure was repeated using Caucasian gene and miRNA datasets to identify clusters validated in the corresponding Chinese datasets, i.e., 43/54 (80%) gene clusters and 24/24 (100%) miRNA clusters ([Fig. S1A](#sd1){ref-type="supplementary-material"} and [B](#sd1){ref-type="supplementary-material"} and [Table S4](#sd2){ref-type="supplementary-material"}), confirming that expression of pivotal transcriptional features in breast cancer is basically alike in Chinese and Caucasian samples. Finally, Chinese and Caucasian profiles were both used as validation datasets for gene and miRNA clusters identified in an independent dataset which integrated both gene (GSE22219) and miRNA (GSE22216) expression data from 207 samples of primary breast tumor samples from Caucasian English patients profiled on the Illumina platform. A quota of variability derived by different annotation and global gene number of the independent dataset affected the comparison at the gene level, and 53% of gene and 82% of miRNA clusters, identified in the independent dataset were validated both in Chinese and Caucasian Italian datasets ([Table S5](#sd2){ref-type="supplementary-material"} and [Fig. S2A](#sd1){ref-type="supplementary-material"} and [B](#sd1){ref-type="supplementary-material"}). Functional characterization of Chinese and Caucasian gene-clusters, by testing whether gene lists relevant to breast cancer [@b1],[@b20]--[@b22] were over-represented, revealed a number of clusters of correlated genes in both Chinese and Caucasian datasets that were associated with: (1) intrinsic classification [@b1],[@b20] defining luminal epithelial/ER and ERBB2 clusters and clusters of genes involved in proliferation and epithelial differentiation; (2) ECM components, including structural and adhesion molecules [@b21]; and (3) immunological mechanisms according to Rody et al. [@b22], represented by hemopoietic cell kinase (HCK), lymphocyte-specific kinase (LCK), major histocompatibility complex (MHC) and interferon clusters (Table S6).
Subtyping Chinese and Caucasian samples by intrinsic genes
----------------------------------------------------------
Chinese and Caucasian samples were assigned to breast cancer intrinsic subtypes by unsupervised hierarchical clustering using the list of 50 intrinsic genes from Parker et al. [@b24] (Fig.[2A](#fig02){ref-type="fig"} and [B](#fig02){ref-type="fig"}). None of the 47 genes affected by the storage time before freezing was included in the PAM50 signature ([Table S2](#sd2){ref-type="supplementary-material"}). Six Caucasian and seven Chinese samples remained unclassified by PAM50 analysis (Fig.[2A](#fig02){ref-type="fig"} and [B](#fig02){ref-type="fig"}), but analysis of expression profiles of genes characterizing the claudin-low subgroup of basal-like tumors [@b25] suggested that nine unclassified samples were indeed claudin-low tumors ([Fig. S3](#sd1){ref-type="supplementary-material"}). Subclass mapping, a method that statistically measures the similarity of predetermined subtypes in independent datasets, was used to test the correspondence of intrinsic molecular subtypes observed in Chinese and Caucasian gene expression data. The heat map of the subclass association matrix from Chinese and Caucasian intrinsic subtypes revealed near-identity of the subgroups identified in both datasets (Fig.[2C](#fig02){ref-type="fig"}). Finally, PCA of global miRNA transcriptomes showed that like gene expression, miRNA profiles tended to group Chinese and Caucasian samples according to molecular subtypes, with a more significant separation of both Chinese and Caucasian basal-like samples independent of ethnic origin (Fig.[2D](#fig02){ref-type="fig"}).
{#fig02}
The prevalence of intrinsic subtypes in Chinese and Caucasian series differed significantly (*P* = 0.006, Table[3](#tbl3){ref-type="table"} and [Fig. S4](#sd1){ref-type="supplementary-material"}), due mainly to an imbalance in luminal A and luminal B samples, which are characterized by lower and higher expression of proliferation genes, respectively. In the Caucasian Italian dataset, 67% of samples were assigned to the luminal group (51% luminal A and 16% luminal B), while 15% and 12% were classified as basal-like and ERBB2 subtypes, respectively. These proportions were consistent with the intrinsic classification of the Caucasian English tumors from dataset GSE22219 ([Fig. S4](#sd1){ref-type="supplementary-material"}, *P* = 0.525). In the Shanghai series, 13% of tumors were classified as basal-like and 22% as ERBB2, while luminal tumors represented 56% (27% luminal A and 29% luminal B). To validate the lower prevalence of luminal A subtype observed in our Chinese series, samples profiled in two public datasets of Chinese consecutive primary breast cancer patients were assigned to PAM50 intrinsic subtypes (Table[3](#tbl3){ref-type="table"}). GSE48390 dataset included 81 samples from Taiwanese Chinese Han patients and GSE36772 included 100 samples from Singapore, where the population is constituted by 75% of Chinese. Accordingly to Shanghai series, breast tumors from Taiwan and Singapore showed a significant difference in the frequency of intrinsic subtypes in comparison to Caucasian cohorts, with higher disparity in luminal subtypes ([Fig. S4](#sd1){ref-type="supplementary-material"}, *P* range: 0.00001--0.056). The ratio between luminal A and luminal B was 0.93, 0.44 and 0.66 for Shanghai, Taiwan and Singapore patients, respectively, whereas it was 3.19 for Caucasian Italian and 2.16 for Caucasian English patients (GSE22219, Table[3](#tbl3){ref-type="table"}). Comparison of intrinsic subtypes frequency among the three Chinese series revealed no significant differences ([Fig. S4](#sd1){ref-type="supplementary-material"}, *P* range: 0.284--0.714).
######
Frequency of PAM50 intrinsic subtypes in independent series of primary breast cancer from Caucasian and Chinese patients
Caucasian series Chinese series
------------------------------------ ------------------ ---------------- ---------- ---------- -----------
Number of samples 97 207 78 81 100
Sample origin Milan Oxford Shanghai Taiwan Singapore
GEO identifier GSE59590 GSE22219 GSE59590 GSE48390 GSE36722
PAM50 intrinsic subtype [@b24] (%)
Basal-like 15 (15) 34 (16) 10 (13) 8 (10) 8 (8)
ERBB2 12 (12) 32 (15) 17 (22) 15 (19) 17 (17)
Luminal A 49 (51) 88 (41) 21 (27) 15 (19) 25 (25)
Luminal B 15 (16) 42 (19) 23 (29) 35 (43) 38 (38)
Undetermined 6 (6) 20 (9) 7 (9) 8 (10) 12 (12)
Luminal A/Luminal B ratio 3.19 2.16 0.93 0.44 0.66
GEO, Gene Expression Omnibus.
Subtyping Chinese and Caucasian samples by ECM genes
----------------------------------------------------
Differential expression of ECM-related genes identified distinct subgroups with different clinical outcome in breast carcinoma [@b21]. LAS biclustering of 738 ECM-related genes identified Chinese and Caucasian ECM3 clusters (Fig.[3](#fig03){ref-type="fig"}A and B), whose gene content and stability parameters ([Tables S7](#sd2){ref-type="supplementary-material"} and [S8](#sd2){ref-type="supplementary-material"}) were comparable to those of ECM3 clusters described in 6 independent datasets [@b26], confirming ECM3 as the most robust ECM cluster in Caucasian and Chinese breast cancers. Likewise, LAS biclustering identified ECM1 clusters with a similar gene pattern and stability in both Chinese and Italian datasets (Fig.[3A](#fig03){ref-type="fig"} and [B](#fig03){ref-type="fig"}, [Tables S7](#sd2){ref-type="supplementary-material"} and [S9](#sd2){ref-type="supplementary-material"}). Consistent with our previous studies [@b21],[@b26], the most significant features in determining ECM3 clusters in both Chinese and Caucasian samples were *SPARC*, *BGN*, *CDH11*, *FN1*, *LAMA4*, *MMP2* and several collagens (*COL1A1*, *COL1A2*, *COL5A1*, *COL5A2*, *COL5A3*, *COL6A1*, *COL6A2*, *COL6A3*), whereas ECM1 clusters were characterized by genes encoding cell adhesion molecules (*SELL*, *ITGA4*, *ITGAL*, *ITGAX*, *ITGB2*, *ITGB7*) and enzymes such as granzymes (*GZMA*, *GZMB*, *GZMK*, *GZMH*, *GZMM*), metallopeptidases (*ADAM7* and *ADAM28*) and cathepsins (*CTSC*, *CTSH*, *CTSS*, *CTSW*, *CTSZ*). Subclass mapping revealed extensive similarity in the global gene profiles of ECM clusters identified in the Chinese and Caucasian datasets (Fig.[3C](#fig03){ref-type="fig"}), and PCA of miRNA profiles (Fig.[3d](#fig03){ref-type="fig"}) supported the transcriptional similarity of ECM3 and ECM1 Chinese and Caucasian tumors.
{#fig03}
Prevalence in Chinese and Caucasian series of ECM1 (22% and 21%, respectively) and ECM3 (18% and 32%, respectively) subtypes was not significantly different (*P* = 0.25). Chinese and Caucasian ECM3 tumors were mainly ER-positive (78% and 83%, respectively) while Chinese and Caucasian ECM1 tumors were principally ER-negative (70% and 75%, respectively).
Differential expression of genes and miRNAs among Chinese and Caucasian samples
-------------------------------------------------------------------------------
To quantify differences between Chinese and Caucasian samples, differentially expressed genes and miRNAs were identified through direct comparison between all samples followed by comparison among intrinsic subtypes. The number of differentially expressed genes and miRNAs (absolute fold-change ≥2 and FDR \< 0.05) according to the molecular subtype significantly outnumbered the differentially expressed features between Chinese and Caucasian samples (Fig.[4](#fig04){ref-type="fig"}), suggesting that most of the variability in Chinese and Caucasian expression data reflects biological differences associated with molecular subtype rather than ethnic origin. Moreover, one of the 11 genes and the only miRNA differentially expressed between Chinese and Caucasian samples were significantly affected by storage time before freezing, indicating that a fraction of the observed differences derived from technical factors.
{#fig04}
Discussion
==========
We find here that comprehensive molecular portraits of breast cancer transcriptomes originating from gene and miRNA expression of Chinese Han and Caucasian Italian subjects are remarkably similar. We therefore provide evidence that intrinsic classification of Chinese breast cancer, as carried out previously in Asian cohorts [@b30],[@b31], is indeed equivalent to Caucasian breast tumors, since equal entities were identified in Chinese and Caucasian cohorts classified by PAM50.
We took advantage of a unique collection of breast cancer frozen specimens from Shanghai Fudan Cancer Center profiled simultaneously with an analogous Caucasian Italian series that yielded transcriptomic data devoid of batch effects. Gene and miRNA profiles of Chinese and Caucasian samples were compared using an unsupervised approach that highlighted similar clusters of correlated features in Chinese and Caucasian samples of our study and in an independent dataset of Caucasian patients, indicating that the taxonomy of transcriptional elements regulating breast cancer biology was the same in the Chinese and Caucasian groups. Interestingly, the transcriptomic diversity between the Caucasian Italian and English groups profiled in different hospitals seemed greater than that between the Chinese and Caucasian groups profiled simultaneously on the same platform.
Partition of gene expression data using predetermined gene lists including "intrinsic" or "ECM" genes identified Chinese and Caucasian subgroups with equivalent gene and miRNA profiles. Thus, breast tumors in the Chinese and Caucasian groups appear to be subjected to the same molecular and differentiation processes. Similar transcriptional profiles and cluster characteristics of Chinese and Caucasian ECM3 and ECM1 tumors suggest similarity in the tumor-surrounding stroma characteristics, including tumor infiltration, since the expression of genes related to immunological processes [@b22] was also consistent between Chinese and Caucasian samples. Finally, although the miRNA breast cancer transcriptome is less well-defined, we found an overall transcriptional similarity of Chinese and Caucasian miRNA patterns, indicating that the main control circuitries regulating gene expression are conserved in Chinese breast cancer. Nevertheless, we cannot exclude the possibility that some genes or miRNAs differentially expressed among Chinese and Caucasian samples have a slight functional effect or that functional traits not associated with transcriptional changes, such as protein modifications or some metabolic pathways, might differ.
Although intrinsic subtypes of the Chinese and Caucasian series were biologically similar, their prevalence differed significantly, with a reduced fraction of luminal A tumors in Chinese samples from Shanghai. This finding, that was validated in 2 independent series of Chinese women from Taiwan and Singapore, might indicate an actual difference in the prevalence of this luminal tumor subgroup in the Chinese series, as also suggested by our findings of a significant disparity in prevalence of ER-positive disease in a large cohort of Chinese and Caucasian patients from Fudan and INT, respectively. Consistent with our data, a study of breast cancer patients in Shanghai [@b32] and a multicenter analysis of breast cancer patients from 7 distinct hospitals across China [@b33] reported 65% and 57% ER-positive tumors, respectively. The younger age of tumor onset in Chinese series might explain the reduced prevalence of luminal tumors compared to Caucasians, considering that age and ER status were associated with Chinese women in our study. This assumption is supported by the recent continuous and simultaneous increase of age of tumor onset and prevalence of ER positivity of breast cancer in China [@b8]. Rapid lifestyle changes that Chinese women experienced in the last decades, including dietary patterns and reproductive factors that are well known to influence the hormonal profile, and the biological similarity of breast cancer from Caucasian and Chinese women, point to the likelihood that the incidence of low-grade ER-positive tumors in postmenopausal Chinese women will increase to levels currently seen in Caucasian breast cancer [@b34]. Further large comparative studies of Asian and Caucasian cohorts homogeneous for geographic origin and socioeconomic status are needed to fully understand the significance of age or other risk factors in the prevalence of ER- positive disease in China.
The substantial similarity of breast cancer transcriptomes from our different ethnic groups is in agreement with findings in comparative analyses of Afro-American and Caucasian women with breast cancer, in whom similarity at the transcriptional and protein levels [@b35] and similar DNA methylation status of 7/8 frequently hypermethylated genes in breast tumor tissues [@b36] were found. Nevertheless, epidemiological differences in breast cancer incidence/mortality and strikingly distinct somatic mutation spectra between subjects and groups have been reported [@b6], indicating that cancer diversity across ethnic populations is complex and not well understood. Complexity in collecting homogeneous and comparable data in multiethnic cohorts and small sample size, study design and socioeconomic conditions, in addition to heterogeneity of breast cancer disease, all affect transethnic studies. A paradigmatic example is the reportedly increased mortality of Afro-American women partially due to socioeconomic disparity but also associated with the prevalence of the most aggressive TNBC subtype in premenopausal women [@b37]; upon subtype-specific analysis, differences in mortality risk among Afro-American and Caucasian women with TNBC disease vanished [@b38].
We lacked information on the mutational landscape of our Chinese cohort, and several reports on breast cancer in Asian women indicated both mutational similarity [@b39] and variations in comparison to Caucasian women [@b40]--[@b43]. It is well-recognized that a complex pattern of somatic mutations involving thousands of genetic entities in cancer of single individuals [@b4],[@b5] indeed merges in a Darwinian way to a restricted number of functional alterations common to cancer of different tissues [@b44]. In this context, our data suggest that although the somatic mutational landscape of breast cancer in Chinese and Caucasian patients might differ, the resulting transcriptomes of both groups retain a similar global architecture distinctive of breast cancer, leading to equivalent functional relationships. This finding has ramifications for translational medicine and personalized approaches since similarity of breast cancer biology across ethnic groups will influence clinical management of patients worldwide and facilitate development of anti-cancer drugs.
This work was supported by the Ministero degli Affari Esteri (Significant bilateral projects between Italy and China) and the Shanghai Science and Technology Project of International Cooperation (no. 09410704800 and 12410707700). We thank Edoardo Marchesi and Lorenzo Bertola for mRNA extraction, the Tissue Bank of Fondazione IRCCS Istituto Nazionale dei Tumori for Italian specimens, and Laura Mameli for secretarial support.
Conflict of Interest
====================
None declared.
Supporting Information
======================
Additional Supporting Information may be found in the online version of this article:
**Figure S1.** Number of genes (A) and miRNAs (B) included in each cluster identified in the Caucasian Italian dataset and validated in the Chinese dataset. Chinese and Caucasian cluster size showed a Pearson correlation of 0.96 and 0.87 for genes and miRNAs, respectively.
**Figure S2.** Number of genes (A) and miRNAs (B) included in each cluster identified in GSE22219 and validated in both Caucasian Italian and Chinese datasets. For genes, size of cluster pairs showed a Pearson correlation of 0.77 and 0.79 for GSE22219 versus the Chinese dataset and versus the Italian dataset, respectively; for miRNAs, size of cluster pairs showed a Pearson correlation of 0.71 and 0.76 for GSE22216 versus the Chinese dataset and versus the Italian dataset, respectively.
**Figure S3.** Unsupervised hierarchical clustering of merged Chinese and Caucasian Italian samples using the claudin-low gene list. Nine of the thirteen Chinese and Caucasian samples designated as unclassified by the PAM50 signature were identified as claudin-low tumors.
**Figure S4.** Square-matrix reporting the *P*-values obtained from the comparison of PAM50 intrinsic subtypes frequencies between each pair of Chinese and Caucasian group. Colors represent the strength of the significance: red, significant; green, not significant.
######
**Table S1.** List of genes significantly associated with the storage time before freezing.
**Table S2.** Overlap between genes associated with the storage time before freezing and genes included in the breast cancer signatures retrieved from GeneSigDB database.
**Table S3.** Genes included in clusters of correlated genes identified in the Chinese dataset and validated in the Caucasian Italian dataset. The value 0 indicates that the gene was not included in the gene cluster in the validation phase.
**Table S4.** Genes included in clusters of correlated genes identified in the Caucasian Italian dataset and validated in the Chinese dataset. The value 0 indicates that the gene was not included in the gene cluster in the validation phase.
**Table S5.** Genes included in clusters of correlated genes identified in the public Caucasian dataset GSE22219 and validated in our Chinese and Caucasian Italian datasets. The value 0 indicates that the gene was not included in the gene cluster in the validation phase.
**Table S6.** Enrichment of relevant breast cancer biological signatures in gene-clusters identified in Chinese, Caucasian Italian, and Caucasian English datasets.
**Table S7.** List of genes defining ECM1 and ECM3 subtypes in Chinese and Caucasian Italian datasets.
**Table S8.** Stability analysis of ECM3 clusters. Statistical methods used to compare Chinese and Caucasian ECM clusters are detailed in the supplementary material of Triulzi et al. [@b26].
**Table S9.** Stability analysis of and ECM1 clusters. Statistical methods used to compare Chinese and Caucasian ECM clusters are detailed in the supplementary material of Triulzi et al. [@b26].
[^1]: These authors contributed equally to this study.
[^2]: **Funding Information** This work was supported by the Ministero degli Affari Esteri (significant bilateral projects between Italy and China) and the Shanghai Science and Technology Project of International Cooperation (no. 09410704800 and 12410707700).
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#s1}
============
Ubiquitylation describes the covalent attachment of ubiquitin, a 76 amino acid polypeptide, to proteins which occurs as a multi-step process (reviewed in [@pgen.1002261-Welchman1], [@pgen.1002261-Hicke1]). E1-activating enzymes activate ubiquitin and transfer it to E2-conjugating enzymes. E3-ubiquitin ligases mediate the conjugation of ubiquitin from the E2 to the target protein. Repeated ubiquitylation cycles lead to the formation of polyubiquitin chains attached on target proteins. Polyubiquitylated proteins are delivered to the 26S proteasome for degradation. However, non-proteolytic roles of ubiquitylation have also been described (reviewed in [@pgen.1002261-Chen1], [@pgen.1002261-Mukhopadhyay1]). From E1 to E3, there is increasing complexity. For example, the *Drosophila* genome encodes only one E1 enzyme, termed UBA1, which is required for all ubiquitin-dependent reactions in the cell [@pgen.1002261-Lee1]. In contrast, there are hundreds of E3-ubiquitin ligases which are needed to confer substrate specificity.
Programmed cell death or apoptosis is an essential physiological process for normal development and maintenance of tissue homeostasis in both vertebrates and invertebrates (reviewed in [@pgen.1002261-Degterev1]). A highly specialized class of proteases, termed caspases, are central components of the apoptotic pathway (reviewed in [@pgen.1002261-Kumar1]). The full-length form (zymogen) of caspases is catalytically inactive and consists of a prodomain, a large and a small subunit. Activation of caspases occurs through dimerization and proteolytic cleavage, separating the large and small subunits. Based on the length of the prodomain, caspases are divided into initiator (also known as apical or upstream) and effector (also known as executioner or downstream) caspases [@pgen.1002261-Kumar1]. The long prodomains of initiator caspases harbor regulatory motifs such as the caspase activation and recruitment domain (CARD) in CASPASE-9. Through homotypic CARD/CARD interactions with the adapter protein APAF-1, CASPASE-9 is recruited into the apoptosome, a large multi-subunit complex, where it dimerizes and auto-processes leading to its activation [@pgen.1002261-Bao1], [@pgen.1002261-Riedl1]. Activated CASPASE-9 cleaves and activates effector caspases (CASPASE-3, -6, and --7), which are characterized by short prodomains. Effector caspases execute the cell death process by cleaving a large number of cellular proteins [@pgen.1002261-Timmer1].
In *Drosophila*, the initiator caspase DRONC and the effector caspases DrICE and DCP-1 are essential for apoptosis [@pgen.1002261-Xu1]--[@pgen.1002261-Muro1]. Like human CASPASE-9, DRONC carries a CARD motif in its prodomain [@pgen.1002261-Dorstyn1]. Consistently, DRONC interacts with ARK, the APAF-1 ortholog in *Drosophila* (also known as DARK, HAC-1 or D-APAF-1) [@pgen.1002261-Kanuka1]--[@pgen.1002261-Zhou1] for recruitment into an apoptosome-like complex which is required for DRONC activation [@pgen.1002261-Kanuka1], [@pgen.1002261-Quinn1]--[@pgen.1002261-Yuan1]. After recruitment into the ARK apoptosome, DRONC cleaves and activates the effector caspases DrICE and DCP-1 [@pgen.1002261-Dorstyn3], [@pgen.1002261-Yuan1]--[@pgen.1002261-Snipas1].
Caspases are subject to negative regulation by inhibitor of apoptosis proteins (IAPs) (reviewed in [@pgen.1002261-ORiordan1], [@pgen.1002261-Vaux1]). For example, DRONC is negatively regulated by *Drosophila* IAP1 (DIAP1) [@pgen.1002261-Meier1], [@pgen.1002261-Chai1]. *diap1* mutations cause a dramatic cell death phenotype, in which nearly every mutant cell is apoptotic, suggesting an essential genetic role of *diap1* for cellular survival [@pgen.1002261-Wang1]--[@pgen.1002261-Lisi1]. DIAP1 is characterized by two tandem repeats known as the Baculovirus IAP Repeat (BIR), and one C-terminally located RING domain [@pgen.1002261-Hay1]. The BIR domains are required for binding to caspases [@pgen.1002261-Meier1], [@pgen.1002261-Chai1], [@pgen.1002261-Zachariou1]. The RING domain provides DIAP1 with E3-ubiquitin ligase activity, required for ubiquitylation of target proteins [@pgen.1002261-ORiordan1], [@pgen.1002261-Vaux1]. Importantly, the BIR domains can bind to caspases independently of the RING domain [@pgen.1002261-Meier1], [@pgen.1002261-Zachariou1].
Usually, IAPs bind to and inhibit activated, i.e. processed caspases, including CASPASE-3, CASPASE-7 and CASPASE-9 as well as the *Drosophila* caspases DrICE and DCP-1 (reviewed in [@pgen.1002261-ORiordan1], [@pgen.1002261-Vaux1]). However, a notable exception to this rule is DRONC. DIAP1 binds to the prodomain of full-length DRONC [@pgen.1002261-Meier1], [@pgen.1002261-Chai1], [@pgen.1002261-Zachariou1]. This unusual behavior suggests an important mechanism for the control of DRONC activation. Indeed, it has been shown that the RING domain of DIAP1 ubiquitylates full-length DRONC *in vitro* [@pgen.1002261-Chai1], [@pgen.1002261-Wilson1]. It has also been proposed that DIAP1 ubiquitylates auto-processed DRONC [@pgen.1002261-Muro2]. These ubiquitylation events are critical for the control of apoptosis, as homozygous *diap1* mutants which lack a functional RING domain (*diap1^ΔRING^*) are highly apoptotic [@pgen.1002261-Lisi1]. Because the BIR domains are intact in *diap1^ΔRING^* mutants, binding of DIAP1 to DRONC is not sufficient for inhibition of DRONC under physiological conditions, and ubiquitylation is the critical event for DRONC inhibition.
Although the importance of DIAP1-mediated ubiquitylation of DRONC is well established, it is still unclear how this ubiquitylation event leads to inactivation of DRONC and of caspases in general. Because DRONC protein accumulates in *diap1* mutant cells that are kept alive by expression of the effector caspase inhibitor P35, generating so-called 'undead' cells, it has been proposed that DIAP1-mediated ubiquitylation triggers proteasomal degradation of full-length DRONC in living cells, thus protecting them from apoptosis [@pgen.1002261-Muro2], [@pgen.1002261-Chai1], [@pgen.1002261-Ryoo1], [@pgen.1002261-Shapiro1]. However, degradation of full-length DRONC in living cells has never been observed and non-degradative models have also been proposed [@pgen.1002261-Wilson1]. Furthermore, ubiquitylation of mammalian CASPASE-3 and CASPASE-7 has been demonstrated *in vitro* [@pgen.1002261-Suzuki1]--[@pgen.1002261-Schile1]. However, evidence for proteasome-dependent degradation of these caspases *in vivo*, i.e. in the context of a living animal, is lacking. In fact, a non-degradative mechanism has been demonstrated for the effector caspase DrICE in *Drosophila* [@pgen.1002261-Ditzel1].
Here, we further characterize the role of ubiquitylation for the control of DRONC activation. Consistent with a previous report [@pgen.1002261-Wilson1], we find that ubiquitylation of DRONC by DIAP1 is critical for inhibition of DRONC\'s pro-apoptotic activity. Using loss and gain of *diap1* function, we provide genetic evidence that DIAP1-mediated ubiquitylation of full-length DRONC regulates this initiator caspase through a non-degradative mechanism. We find that the conjugation of ubiquitin suppresses DRONC processing and activation. Interestingly, once DRONC is processed and activated, it has reduced protein stability. Finally, we show that *dronc* transcripts accumulate in P35-expressing 'undead' cells, accounting for increased DRONC protein levels in these cells. These data refine the current model of caspase regulation by IAPs.
Results {#s2}
=======
Overexpression of DIAP1 fails to suppress apoptosis of *Uba1* mutant cells {#s2a}
--------------------------------------------------------------------------
It has previously been shown that complete loss of ubiquitylation due to mutations of the E1 enzyme *Uba1* causes apoptosis in eye imaginal discs as detected by an antibody that recognizes cleaved, i.e. activated, CASPASE-3 (CAS3\*) [@pgen.1002261-Lee1], [@pgen.1002261-Fan1], [@pgen.1002261-Pfleger1] (see also [Figure 1A](#pgen-1002261-g001){ref-type="fig"}). Because ubiquitylation of DRONC does not occur in *Uba1* mutants, we hypothesized that inappropriate activation of DRONC accounts for the apoptotic phenotype of *Uba1* mutants. To test this possibility, we targeted *dronc* by RNA interference (RNAi) in *Uba1* mutant cells in eye imaginal discs using the MARCM system and labeled for apoptosis using CAS3\* antibody. In this system, *Uba1* mutant cells expressing *dronc* RNAi are positively marked by GFP. Consistent with our hypothesis, knock-down of *dronc* strongly reduces apoptosis in *Uba1* mutant clones ([Figure 1B](#pgen-1002261-g001){ref-type="fig"}). Furthermore, we tested clones doubly mutant for *Uba1* and *ark*, the *Drosophila* ortholog of APAF-1 that is required for DRONC activation (see [Introduction](#s1){ref-type="sec"}). Apoptosis induced in *Uba1* mutant clones is strongly suppressed if *ark* function is removed ([Figure S1](#pgen.1002261.s001){ref-type="supplementary-material"}). These observations suggest that the apoptotic phenotype in *Uba1* clones is caused by inappropriate activation of DRONC, presumably due to lack of ubiquitylation.
{ref-type="sec"}). Presence of GFP marks the location of *Uba1* clones (see arrow). (B) TUNEL labeling of *Uba1* mosaic eye-antennal imaginal discs expressing an RNAi transgene targeting *dronc* (*UAS-droncIR* (inverted repeat)) using the MARCM technique (see [Material and Methods](#s4){ref-type="sec"}). Clones are positively marked by GFP. TUNEL-positive cell death is largely blocked by *dronc* knockdown (B′ and B″). (C) Strong overexpression of *diap1* in *Uba1* clones (magenta in C′″) fails to rescue the apoptotic phenotype, as visualized by CAS3\* labeling (red in C′). *Uba1* clones are marked by GFP due to the MARCM technique. Please note that *diap1* is so strongly overexpressed in the clones that we had to adjust the settings in such a way that endogenous DIAP1 in wild-type tissue is below the detection limit (C′″). Genotypes: (A) *hs-FLP UAS-GFP*; *FRT42D Uba1^D6^*/*FRT42D tub*-*Gal80*; *tub*-*GAL4*. (B) *hs-FLP UAS-GFP*; *FRT42D Uba1^D6^*/*FRT42D tub*-*Gal80*; *tub*-*GAL4*/*UAS-droncIR*. (C) *hs-FLP UAS-GFP/UAS-diap1*; *FRT42D Uba1^D6^*/*FRT42D tub*-*Gal80*; *tub*-*GAL4*.](pgen.1002261.g001){#pgen-1002261-g001}
However, the protein levels of DIAP1 are increased in *Uba1* mutant clones [@pgen.1002261-Lee1], [@pgen.1002261-Pfleger1]. There are two possibilities to explain the apoptotic phenotype in *Uba1* mutants despite increased DIAP1 levels. First, the DIAP1 levels may not be sufficiently increased to inhibit DRONC. Alternatively, binding of DIAP1 to DRONC alone may not be sufficient for inhibition of DRONC; instead, ubiquitylation by DIAP1 is required to block DRONC activation, as previously suggested [@pgen.1002261-Wilson1]. To distinguish between these two possibilities, we strongly overexpressed *diap1* in *Uba1* mutant clones in eye discs using the MARCM system and imaged for apoptosis by CAS3\* labeling. Surprisingly, despite massive expression of *diap1* (\>20 fold over wild-type levels; [Figure 1C′″](#pgen-1002261-g001){ref-type="fig"}), apoptosis still proceeds in *Uba1* mutant clones ([Figure 1C′](#pgen-1002261-g001){ref-type="fig"}), even though expression of the same transgene can block strong apoptotic phenotypes in several apoptotic paradigms ([Figure S2](#pgen.1002261.s002){ref-type="supplementary-material"}). Apparently, overexpression of DIAP1 is not sufficient to inhibit DRONC and to protect *Uba1* mutant cells from apoptosis. Because DIAP1 is the key regulator of DRONC and because DRONC is required for the apoptotic phenotype of *Uba1* mutant cells, as evidenced by knock-down of *dronc* ([Figure 1B](#pgen-1002261-g001){ref-type="fig"}), our data provide genetic evidence that binding of DIAP1 is not sufficient for DRONC inhibition in *Uba1* mutant cells.
Consistent with this view, it has previously been shown that DIAP1 does ubiquitylate full-length DRONC *in vitro* [@pgen.1002261-Muro2], [@pgen.1002261-Chai1], [@pgen.1002261-Wilson1]. We tested whether DIAP1 can also ubiquitylate DRONC *in vivo*. Because the available DRONC antibodies failed to immunoprecipitate endogenous DRONC, we transfected DRONC-V5 along with DIAP1^+^ or DIAP1^ΔRING^ mutants (CΔ6, lacking the last six C-terminal residues, and F437A changing a critical Phe residue in the RING domain to Ala [@pgen.1002261-Silke1]) and His-tagged Ubiquitin into *Drosophila* S2 cells. Ubiquitylated proteins were affinity purified under denaturing conditions using Ni columns. The eluates were subsequently examined by immunoblotting with anti-V5 antibodies to detect ubiquitylated forms of DRONC. Under these conditions, DIAP1^+^ readily ubiquitylates full-length DRONC in S2 cells ([Figure 2](#pgen-1002261-g002){ref-type="fig"}), whereas the RING mutants DIAP1^CΔ6^ and DIAP1^F437A^ were significantly impaired in their ability to ubiquitylate DRONC ([Figure 2](#pgen-1002261-g002){ref-type="fig"}). These results indicate that DIAP1 ubiquitylates full-length DRONC in a RING-dependent manner in cultured cells.
{#pgen-1002261-g002}
Overexpression of DIAP1 does not induce degradation of DRONC {#s2b}
------------------------------------------------------------
Because DIAP1 readily ubiquitylates DRONC, it has been postulated that DIAP1-mediated ubiquitylation leads to proteasomal degradation of DRONC [@pgen.1002261-Muro2], [@pgen.1002261-Chai1], [@pgen.1002261-Ryoo1]. However, this has never been rigorously tested *in vivo*. Therefore, we examined, whether overexpression of *diap1* in wild-type animals can influence DRONC protein levels *in vivo*. To this end, we generated clones overexpressing *diap1* (marked by absence of GFP) in eye discs, and analyzed the protein abundance of DRONC. Interestingly, despite high expression of *diap1* ([Figure 3A′″](#pgen-1002261-g003){ref-type="fig"}), the levels of DRONC remained unchanged and were not influenced by DIAP1 ([Figure 3A′](#pgen-1002261-g003){ref-type="fig"}). The anti-DRONC antibody used in this assay is specific for DRONC ([Figure S3](#pgen.1002261.s003){ref-type="supplementary-material"}). Importantly, the *diap1* transgene used produces a functional DIAP1 protein that is able to inhibit apoptosis in several paradigms ([Figure S2](#pgen.1002261.s002){ref-type="supplementary-material"}). Therefore, these data suggest that overexpressed DIAP1 does not target DRONC for degradation in living cells.
{#pgen-1002261-g003}
REAPER-induced loss of DIAP1 does not increase DRONC protein levels {#s2c}
-------------------------------------------------------------------
Because of the surprising observation that overexpressed DIAP1 does not cause degradation of DRONC, we tested whether removal of DIAP1 changes DRONC protein levels. Expression of the IAP antagonist *reaper* (*rpr*) induces DIAP1 degradation and apoptosis [@pgen.1002261-Ryoo2]--[@pgen.1002261-Yoo1]. Therefore, we examined whether RPR-induced degradation of DIAP1 changes DRONC protein levels. If DIAP1 targets DRONC for degradation, we would expect that DRONC protein levels would accumulate in response to *rpr* expression. Expression of *rpr* in eye imaginal discs posterior to the morphogenetic furrow (MF) using the *GMR* promoter (*GMR*-*rpr*) is well suited for this analysis. The MF is a dynamic structure that initiates at the posterior edge of the eye disc and moves towards the anterior during 3^rd^ instar larval stage [@pgen.1002261-Wolff1], [@pgen.1002261-Cagan1] ([Figure 4A](#pgen-1002261-g004){ref-type="fig"}, arrow). Expression of *rpr* by *GMR* is induced in all cells posterior to the MF [@pgen.1002261-Ellis1] (red in [Figure 4A](#pgen-1002261-g004){ref-type="fig"}). Therefore, *GMR*-*rpr* eye discs provide a continuum of all developmental stages in which cells close to the MF have only recently induced *rpr* expression, while cells towards the posterior edge of the disc have been exposed to *rpr* progressively longer. Therefore, if DRONC accumulates during any of these stages, we should be able to detect it. In wild-type eye discs, DRONC protein is homogenously distributed throughout the disc. Only in the MF, higher levels of DRONC are detectable (arrowhead in [Figure 4B″](#pgen-1002261-g004){ref-type="fig"}). This high expression of DRONC in the MF serves as an orientation mark. DIAP1 protein levels are low anterior to the MF, but increase in the MF (arrowhead) and posterior to it in wild-type discs ([Figure 4B′](#pgen-1002261-g004){ref-type="fig"}). In *GMR*-*rpr* eye discs, overall DIAP1 levels are reduced in the *rpr*-expressing domain posterior to the MF ([Figure 4C′](#pgen-1002261-g004){ref-type="fig"}), but particularly strongly reduced in the CAS3\*-positive area ([Figure 4C′, D′](#pgen-1002261-g004){ref-type="fig"}, arrow) consistent with previous reports [@pgen.1002261-Ryoo2]--[@pgen.1002261-Yoo1]. However, accumulation of DRONC is not observed ([Figure 4C″, D″](#pgen-1002261-g004){ref-type="fig"}). In contrast, it appears that DRONC levels are also reduced. They are still high in the MF ([Figure 4C″](#pgen-1002261-g004){ref-type="fig"}, arrowhead), but drop immediately thereafter.
{#pgen-1002261-g004}
We also related DRONC levels to caspase activation. In the MF, where CAS3\* activity is not detectable, DRONC is still high ([Figure 4D′, D″](#pgen-1002261-g004){ref-type="fig"}; arrowhead), but in the CAS3\*-positive area, DRONC levels are reduced ([Figure 4D′, D″](#pgen-1002261-g004){ref-type="fig"}; arrow). These data indicate that loss of DIAP1 does not cause accumulation of DRONC protein implying that DIAP1 does not induce degradation of DRONC. In contrast, it appears that DIAP1 stabilizes DRONC at least under these conditions (see [Discussion](#s3){ref-type="sec"}).
"Undead" *diap1* mutant cells induce transcription of *dronc* {#s2d}
-------------------------------------------------------------
Finally, we analyzed DRONC protein levels in *diap1^ΔRING^* mutants which cannot ubiquitylate DRONC [@pgen.1002261-Wilson1]. It has previously been shown that clones of the RING mutant *diap1^22^* ^-*8s*^ accumulate DRONC protein [@pgen.1002261-Ryoo1], [@pgen.1002261-Shapiro1] implying that ubiquitylation by the RING domain of DIAP1 causes degradation of DRONC. We repeated these experiments and indeed confirmed that DRONC levels are increased in *diap1^22^* ^-8s^ mutant clones ([Figure S4](#pgen.1002261.s004){ref-type="supplementary-material"}). Thus, these results appear inconsistent with the data presented in [Figure 3](#pgen-1002261-g003){ref-type="fig"} and [Figure 4](#pgen-1002261-g004){ref-type="fig"} in which manipulating DIAP1 levels did not provide evidence for DIAP1-mediated degradataion of DRONC. However, one caveat with the *diap1^22^* ^-8s^ experiment was the use of the caspase inhibitor P35 which kept *diap1^22^* ^-8s^ mutant cells in an 'undead' condition [@pgen.1002261-Ryoo1]. It has been pointed out that the 'undead' state may change the properties of the affected cells (reviewed by [@pgen.1002261-Martin1]) and in fact abnormal induction of transcription in 'undead' cells has been reported [@pgen.1002261-Ryoo1], [@pgen.1002261-Huh1]--[@pgen.1002261-Wells1]. Thus, to explain the conflicting results between the *diap1^22^* ^-8s^ data [@pgen.1002261-Ryoo1] and our data shown here, we hypothesized that *p35*-expressing 'undead' *diap1^22^* ^-*8s*^ clones induce *dronc* transcription, leading to accumulation of DRONC protein. To test this hypothesis, we used a transcriptional *lacZ* reporter containing 1.33 kb of regulatory genomic sequences upstream of the transcriptional start site of the *dronc* gene fused to *lacZ* (*dronc1.33-lacZ*) [@pgen.1002261-Daish2], [@pgen.1002261-Cakouros1]. Compared to controls ([Figure 5A, 5A′](#pgen-1002261-g005){ref-type="fig"}) and consistent with the hypothesis, *dronc1.33-lacZ* reporter activity is increased in *p35*-expressing 'undead' *diap1^22^* ^-*8s*^ cells in wing imaginal discs and matches the increased DRONC protein pattern ([Figure 5B′-5B′″](#pgen-1002261-g005){ref-type="fig"}). We also produced 'undead' cells in eye imaginal discs by co-expression of the IAP-antagonist *hid* and the caspase inhibitor *p35* in the dorsal half of the eye disc using a *dorsal eye-* (*DE-*) *GAL4* driver ([Figure 5C](#pgen-1002261-g005){ref-type="fig"}). Similar to wing discs, *dronc* reporter activity is increased in 'undead' cells in the dorsal half of the eye ([Figure 5D](#pgen-1002261-g005){ref-type="fig"}). Expression of *p35* alone does not trigger transcription of *dronc* ([Figure 5E](#pgen-1002261-g005){ref-type="fig"}) suggesting it is not the mere presence of P35 which causes *dronc* transcription, but the 'undead' nature of the affected cells.
!["Undead" *diap1* mutant cells trigger transcription of *dronc*.\
Shown are 3^rd^ instar larval wing (A,B,F) and eye imaginal discs (C,D,E) labeled for DRONC protein levels (blue) and *dronc* transcriptional activity (red) using the *dronc1.33-lacZ* reporter (ß-GAL labeling). (A,A′) Co-labeling for DRONC protein (A) and *dronc* reporter activity (A′) of a wild-type wing disc expressing the *dronc1.33-lacZ* transgene. (B-B′″) A *diap1^22-8s^* mosaic wing disc expressing *p35* under *nub-GAL4* control in a *dronc1.33-lacZ* background. A mutant clone in the wing pouch is highlighted by an arrow in the GFP-only channel (B). DRONC protein (B′) and ß-GAL immunoreactivity as readout of *dronc1.33-lacZ* activity (B″) are increased in the same cells and overlap (B′″). Please note that the *dronc1.33-lacZ* reporter produces nuclear ß-GAL, while DRONC protein appears cytoplasmic. (C) GFP expression in the eye imaginal disc indicates the dorsal expression domain (arrow) of the *dorsal eye* (*DE*)-*GAL4* driver [@pgen.1002261-Morrison1]. (D) Increased *dronc* reporter activity in the dorsal half of the eye imaginal disc (arrow) in undead cells obtained by co-expression of *hid* and *p35* using *DE-GAL4*. (E) Expression of *p35* alone by *DE-GAL4* does not induce *dronc* reporter activity. (F-F″) A *diap1^22-8s^* mosaic wing disc in a *dronc1.33-lacZ* background which does not express *p35*. *diap1^22-8s^* mutant clones are marked by the absence of GFP (F). An arrow points to a representative *diap1^22-8s^* clone in the wing pouch. In the same position, neither DRONC protein (F′) nor *dronc* reporter activity (F″) are increased. Note, that this clone is present in the wing pouch which has the capacity to upregulate DRONC and *dronc* transcription in the 'undead', *p35*-expressing condition (see panel B″). Genotypes: (A) *dronc1.33-lacZ*/+. (B) *ubx-FLP*; *nub-GAL4 UAS-p35*/*dronc1.33-lacZ*; *diap1^22-8s^ FRT80*/*ubi-GFP FRT80*. (C) *DE-GAL4 UAS-GFP*/+. (D) *UAS-p35 UAS-hid/dronc1.33-lacZ*; *DE*-*GAL4*. (E) *UAS-p35/dronc1.33-lacZ*; *DE-GAL4*. (F) *ubx-FLP*; *nub*-*GAL4*/*dronc1.33-lacZ*; *diap1^22-8s^ FRT80*/*ubi-GFP FRT80*.](pgen.1002261.g005){#pgen-1002261-g005}
These observations may explain why DRONC protein accumulates in 'undead' *diap1^22^* ^-*8s*^ mutant cells, but they still do not rule out the possibility that DRONC protein accumulates in *diap1^22^* ^-*8s*^ mutants due to lack of ubiquitylation and thus degradation. To clarify this issue we examined *dronc1.33-lacZ* and DRONC levels in *diap1^22^* ^-*8s*^ mutant clones without simultaneous *p35* expression. Without the inhibition of apoptosis by P35, *diap1^22^* ^-*8s*^ clones die rapidly. Nevertheless, we were able to recover wing discs which contained small *diap1^22-8s^* mutant clones. In these clones, neither *dronc1.33-lacZ* nor DRONC levels are detectably increased ([Figure 5F](#pgen-1002261-g005){ref-type="fig"}). Notably, these clones are located in the wing pouch in which we observed accumulation of *dronc* reporter activity and DRONC protein under 'undead' conditions ([Figure 5B″](#pgen-1002261-g005){ref-type="fig"}). Thus, the 'undead' condition of *p35*-expressing *diap1^22^* ^-*8s*^ mutant cells causes accumulation of DRONC protein due to induction of *dronc* transcription, explaining the observations of Ryoo et al. (2004) [@pgen.1002261-Ryoo1]. In the absence of *p35* expression, transcription of *dronc* and accumulation of DRONC protein are not observed, providing additional evidence that ubiquitylation of DRONC by the RING domain of DIAP1 does not trigger degradation of DRONC.
Ubiquitylation controls processing and thus activation of DRONC *in vivo* {#s2e}
-------------------------------------------------------------------------
Our *in vivo* analysis implies that DIAP1-mediated ubiquitylation does not trigger proteasomal degradation of DRONC. To identify the role of ubiquitylation for regulation of DRONC activity, we analyzed the fate of DRONC protein in RING mutants of *diap1*. Of note, these mutants retain their ability to bind to DRONC, because DRONC binding is not mediated by the RING domain, but by the BIR2 domain [@pgen.1002261-Meier1], [@pgen.1002261-Chai1], [@pgen.1002261-Zachariou1]. The RING mutant *diap1^33-1s^* is especially suitable for this analysis because a premature stop codon results in deletion of the entire RING domain but leaves the BIR domains intact [@pgen.1002261-Wilson1] ([Figure 6A](#pgen-1002261-g006){ref-type="fig"}), thus abrogating its E3 activity, but not caspase binding. Importantly, *diap1^33-1s^* is characterized by a strong apoptotic phenotype, suggesting inappropriate caspase activation [@pgen.1002261-Lisi1], [@pgen.1002261-Ryoo1]. Indeed, we showed previously that *diap1^ΔRING^* mutant phenotypes are dependent upon DRONC, because *dronc* mutants suppress *diap1^ΔRING^* phenotypes [@pgen.1002261-Xu1]. Therefore, ubiquitylation of DRONC by DIAP1 is critical to maintain cell survival.
![Ubiquitylation controls processing of DRONC.\
(A) Top: schematic outline of the domain structure of DIAP1^+^ (wild-type) and RING-deleted DIAP1^33-1s^. Not drawn to scale. Immunoblots of embryonic extracts of stage 6--9 wild-type (wt) and heterozygous *diap1^33-1s^* mutants were probed with anti-DIAP1 (upper panel) and anti-DRONC antibodies (middle panel). The RING-depleted *diap1^33-1s^* allele produces a stable protein (DIAP1^33-1s^) that is detectable by its faster electrophoretic mobility (upper panel). In RING-depleted *diap1^33-1s^* embryos a significant portion of processed DRONC is present (middle panel) which likely accounts for the apoptotic phenotype of *diap1^33-1s^* embryos [@pgen.1002261-Lisi1]. These extracts were obtained from a cross of heterozygous males and females. Thus, only one quarter of the embryos is homozygous mutant for *diap1^33-1s^*; yet, processed DRONC is easily detectable. The anti-DRONC antibody is specific for the large subunit of DRONC. Lower panel: loading control. (B) Extracts of imaginal discs from wild-type (wt) and mosaic *Uba1* imaginal discs were analyzed by immunoblotting using an antibody raised against the small subunit of DRONC. Clones of the temperature sensitive allele *Uba1^D6^* were induced at the permissive temperature in first larval instar and then shifted to the non-permissive temperature (30°C) during third larval instar 12 hours before dissection (see [Material and Methods](#s4){ref-type="sec"}). This treatment ensures that approximately 50% of the mosaic disc is mutant for *Uba1*. Although only 50% of the disc tissue is mutant for *Uba1*, processed DRONC is easily detectable. Lower panel: loading control.](pgen.1002261.g006){#pgen-1002261-g006}
We examined the cause of the *diap1^33^* ^-*1s*^ apoptotic phenotype. First, as a control, we determined whether the *diap1^33^* ^-*1s*^ gene produces a stable protein *in vivo*. We chose to analyze stage 6--9 embryos, because normal developmental cell death starts at stage 11 [@pgen.1002261-Abrams1]; therefore, stage 6--9 *diap1^33^* ^-*1s*^ mutant embryos allow analysis of DIAP1 in the absence of upstream apoptotic signals. In immunoblots of embryonic extracts obtained from a cross of heterozygous *diap1^33^* ^-*1s*^ males and females, the DIAP1^33-1s^ protein is easily distinguished from wild-type DIAP1^+^ protein due to its faster electrophoretic mobility ([Figure 6A](#pgen-1002261-g006){ref-type="fig"}, top panel). The presence of the DIAP1^33-1s^ protein suggests that the apoptotic phenotype in *diap1^33^* ^-*1s*^ mutant embryos is not caused by instability of the mutant protein. Interestingly, the protein levels of DIAP1^+^ and RING-deleted DIAP1^33-1s^ are similar ([Figure 6A](#pgen-1002261-g006){ref-type="fig"}, top panel) suggesting that loss of the RING domain does not influence the protein stability of DIAP1 in the absence of pro-apoptotic signals.
Next, we analyzed DRONC protein in extracts from *diap1^33^* ^-*1s*^ mutant embryos. Consistent with the data in [Figure 4](#pgen-1002261-g004){ref-type="fig"} and [Figure 5](#pgen-1002261-g005){ref-type="fig"}, we do not detect a significant increase in the protein levels of DRONC in these embryos ([Figure 6A](#pgen-1002261-g006){ref-type="fig"}, middle panel). However, a significant amount of DRONC is present in a processed form in extracts of stage 6--9 *diap1^33-1s^* mutant embryos which is absent in control extracts from wild-type embryos ([Figure 6A](#pgen-1002261-g006){ref-type="fig"}, middle panel). Therefore, DRONC processing, and thus activation, occurs in RING-depleted *diap1^33^* ^-*1s*^ mutant embryos despite the fact that the BIR domains of DIAP1 are unaffected. The processed form of DRONC in this mutant of MW ∼36 kDa corresponds to the major apoptotic form of DRONC composed of the large subunit and the prodomain of DRONC [@pgen.1002261-Muro3]. This finding, and the one presented in [Figure 1](#pgen-1002261-g001){ref-type="fig"}, confirms that binding of DIAP1 to DRONC is not sufficient to prevent processing and activation of DRONC. Instead, the RING domain is required to control DRONC processing. Because the RING domain contains an E3-ubiquitin ligase activity [@pgen.1002261-Ryoo1], [@pgen.1002261-Hays1]--[@pgen.1002261-Yoo1] and because ubiquitylation of full-length DRONC does not trigger proteasomal degradation ([Figure 3](#pgen-1002261-g003){ref-type="fig"}, [Figure 4](#pgen-1002261-g004){ref-type="fig"}, and [Figure 5](#pgen-1002261-g005){ref-type="fig"}), we conclude that ubiquitylation of DRONC by the RING domain of DIAP1 is necessary to inhibit DRONC processing and thus activation.
To further characterize the role of ubiquitylation in the regulation of DRONC processing, we performed an immunoblot analysis with extracts from wild-type and *Uba1* mosaic imaginal discs, which, under our experimental conditions, are about half mutant for *Uba1* and half wild-type. Immunoblot analysis demonstrated that a significant amount of DRONC is processed in *Uba1* mosaic discs ([Figure 6B](#pgen-1002261-g006){ref-type="fig"}). Thus, these data further support the notion that ubiquitylation of full-length DRONC is necessary for inhibition of DRONC processing.
Discussion {#s3}
==========
In this paper, we provide three take-home messages. First, we provide genetic evidence that binding of DIAP1 to DRONC is not sufficient for inhibition of DRONC. Instead, ubiquitylation of DRONC controls its apoptotic activity, consistent with the apoptotic phenotype of *diap1^ΔRING^* mutants, that retain caspase binding abilities. Second, DIAP1-mediated ubiquitylation of full-length DRONC does not lead to its proteasomal degradation; rather, ubiquitylation directly controls processing and activation of DRONC. Interestingly, processed and active DRONC shows reduced protein stability. Third, 'undead' cells accumulate *dronc* transcripts.
Binding of DIAP1 is not sufficient for Dronc inhibition {#s3a}
-------------------------------------------------------
Based on biochemical studies *in vitro* and overexpression studies in cultured cells, many of cancerous origin, it was initially proposed that binding of IAPs to caspases through their BIR domains is sufficient to inhibit caspases [@pgen.1002261-Deveraux1]--[@pgen.1002261-Silke2]. However, when ubiquitylation of caspases by IAPs was described [@pgen.1002261-Chai1], [@pgen.1002261-Wilson1], [@pgen.1002261-Suzuki1], [@pgen.1002261-Huang1], it was unclear what role ubiquitylation would play for control of caspase activity, especially since for none of them, ubiquitin-mediated degradation has been observed (see below). Because the RING domain is also required for auto-ubiquitylation of DIAP1 [@pgen.1002261-Ryoo2]--[@pgen.1002261-Yoo1], mutations of the RING domain would be expected to increase DIAP1 protein levels and protect cells from apoptosis. However, the opposite phenotype, massive apoptosis, was observed [@pgen.1002261-Lisi1]. Nevertheless, despite the biochemical studies showing that the BIR domains of DIAP1 are sufficient for interaction with DRONC [@pgen.1002261-Meier1], [@pgen.1002261-Chai1], [@pgen.1002261-Zachariou1], one could argue that DIAP1^ΔRING^ mutants have lost the ability to interact with DRONC *in vivo*. While we cannot exclude this possibility due to the inability of our antibodies to immunoprecipitate endogenous proteins, another experiment supports the notion that ubiquitylation is necessary for DRONC inhibition: when wild-type *diap1* was strongly overexpressed in an ubiquitylation-deficient *Uba1* mutant background, DRONC-dependent apoptosis was not inhibited ([Figure 1C](#pgen-1002261-g001){ref-type="fig"}), suggesting that binding of DIAP1 is not sufficient for inhibition of DRONC. Instead, ubiquitylation is critical for inhibition of DRONC activity.
DIAP1 does not control protein levels of full-length DRONC {#s3b}
----------------------------------------------------------
The current model holds that DIAP1-mediated ubiquitylation leads to proteasomal degradation of full-length DRONC in living cells [@pgen.1002261-Muro2], [@pgen.1002261-Chai1], [@pgen.1002261-Ryoo1]. However, our data do not support this model *in vivo*. This model is based on biochemical ubiquitylation studies without *in vivo* validation and does not take into account that ubiquitylation often serves non-proteolytic functions [@pgen.1002261-Welchman1], [@pgen.1002261-Chen1], [@pgen.1002261-Mukhopadhyay1]. Both overexpression and loss of *diap1* does not cause a detectable alteration of the protein levels of DRONC ([Figure 3](#pgen-1002261-g003){ref-type="fig"}, [Figure 4](#pgen-1002261-g004){ref-type="fig"}, [Figure 5](#pgen-1002261-g005){ref-type="fig"}), arguing against a degradation model. The only example where DRONC accumulation has been observed is in P35-expressing 'undead' *diap1^ΔRING^* mutant cells [@pgen.1002261-Ryoo1], [@pgen.1002261-Shapiro1], and we showed here that the 'undead' nature of these cells causes transcriptional induction of *dronc* ([Figure 5](#pgen-1002261-g005){ref-type="fig"}). Together, these observations argue against a degradation model of full-length DRONC and favor a non-traditional (non-proteolytic) role of ubiquitylation regarding control of DRONC activity. Similarly, DIAP1-mediated ubiquitylation of the effector caspase DrICE inactivates this effector caspase through a non-degradative mechanism [@pgen.1002261-Ditzel1].
Interestingly, in *GMR-rpr* eye imaginal discs, DRONC protein levels appear to be reduced in apoptotic cells compared to living cells ([Figure 4C″, 4D″](#pgen-1002261-g004){ref-type="fig"}). Due to the apoptotic activity of REAPER, DRONC is present in its processed and activated form. Reduced protein stability of DRONC has also been reported after incorporation into the ARK apoptosome where it is processed and activated [@pgen.1002261-Shapiro1]. Combined, these observations suggest that while DIAP1-mediated ubiquitylation of full-length DRONC does not trigger its degradation, processed and activated DRONC has reduced protein stability and may indeed be degraded. It is unclear whether degradation of activated DRONC is mediated by DIAP1, as proposed previously [@pgen.1002261-Muro2]. In *GMR-rpr* eye imaginal discs, reduced DRONC levels correlate with a reduction of DIAP1 protein ([Figure 4C′, 4D′](#pgen-1002261-g004){ref-type="fig"}). This correlation indicates that DIAP1 may actually stabilize DRONC protein, at least in part. Alternatively, because DRONC and DIAP1 form a complex [@pgen.1002261-Meier1], REAPER-induced degradation of DIAP1 may result in co-degradation of complexed DRONC in the process. Further studies are needed to determine the cause of decreased DRONC stability in apoptotic cells.
In addition to *Drosophila* DRONC, mammalian CASPASE-3 and CASPASE-7 have been reported to be ubiquitylated *in vitro* [@pgen.1002261-Suzuki1], [@pgen.1002261-Huang1]. However, proteasome-mediated degradation of these caspases *in vivo* has not been reported. Although a decrease of CASPASE-3 levels has been noted upon overexpression of XIAP, this was shown for an artificial CASPASE-3 mutant, in which the order of the subunits was reversed and the Cys residue in the active site changed to Ser [@pgen.1002261-Suzuki1]. This catalytically inactive CASPASE-3 mutant is not proteolytically processed [@pgen.1002261-Suzuki1]. Therefore, physiological *in vivo* evidence for IAP-mediated degradation of mammalian caspases is lacking.
Moreover, the phenotype of a RING-deleted *XIAP* mutant mouse is consistent with our data [@pgen.1002261-Schile1]. The *XIAP^ΔRING^* mutant, which was generated by a knock-in approach in the endogenous *XIAP* gene, is characterized by increased caspase activity [@pgen.1002261-Schile1]. Intriguingly, the protein levels of CASPASE-3, CASPASE-7 and CASPASE-9 did not significantly change in the *XIAP^ΔRING^* mutant despite the fact that ubiquitylation of CASPASE-3 was reduced. However, processing of these caspases was easily detectable in *XIAP^ΔRING^* mutants [@pgen.1002261-Schile1]. These data are very similar to the ones presented here for *diap1^33^* ^-1s^ ([Figure 6](#pgen-1002261-g006){ref-type="fig"}), and together strongly suggest that non-proteolytic ubiquitylation controls caspase processing and activity in both vertebrates and invertebrates.
Non-proteolytic roles of ubiquitylation have been described in recent years and are involved in many processes including signal transduction, endocytosis, DNA repair, and histone activity (reviewed in [@pgen.1002261-Welchman1], [@pgen.1002261-Chen1], [@pgen.1002261-Mukhopadhyay1]). Two types of ubiquitylation lead to non-proteolytic functions. Monoubiquitylation is involved in endocytosis, DNA repair and histone activity. In fact, mammalian CASPASE-3 and CASPASE-7 have been found to be monoubiquitylated *in vitro* [@pgen.1002261-Huang1]. However, it is unclear whether DRONC is monoubiquitylated by DIAP1. The presence of high molecular-weight ubiquitin conjugates *in vitro* ([Figure 2](#pgen-1002261-g002){ref-type="fig"}) suggests that DRONC may be polyubiquitylated, at least under the experimental conditions [@pgen.1002261-Chai1], [@pgen.1002261-Wilson1].
Polyubiquitylation serves both proteolytic and non-proteolytic functions depending on the Lysine (K) residue used for polyubiquitin chain formation. In general, if polyubiquitylation occurs via K48, the target protein is subject to proteasome-mediated degradation. If it occurs on a different Lys residue, such as K63, non-proteolytic functions are induced [@pgen.1002261-Welchman1], [@pgen.1002261-Chen1], [@pgen.1002261-Mukhopadhyay1]. The best studied examples of both K48- and K63-linked polyubiquitylation are in the NF-κB pathway (reviewed in [@pgen.1002261-Chen1], [@pgen.1002261-Wertz1]). While K48-polyubiquitylation leads to proteasomal degradation, K63-linked polyubiquitin chains act as scaffolds to assemble protein complexes required for NF-κB activation [@pgen.1002261-Chen1], [@pgen.1002261-Wertz1]. It is unknown what type of polyubiquitin chain is attached to DRONC, but it is possible that it is not K48-linked. Interestingly, in this context it has been shown that auto-ubiquitylation of DIAP1 preferentially involves K63-linked polyubiquitin chains [@pgen.1002261-HermanBachinsky1]. By analogy, DIAP1 may ubiquitylate DRONC through formation of K63-linked polyubiquitin chains. This will be an interesting avenue to explore in future experiments.
Conjugated monoubiquitin and polyubiquitin chains can serve as docking sites for factors containing ubiquitin-binding domains (UBD) [@pgen.1002261-Hicke1], [@pgen.1002261-Mukhopadhyay1], [@pgen.1002261-Adhikari1]. The UBD-containing factors interpret the ubiquitylation status of the target protein, and trigger the appropriate response. For example, K48-linked polyubiquitin chains are recognized by Rad23 and Drk2 which deliver them to the proteasome [@pgen.1002261-Hicke1]. Other forms of ubiquitin conjugates are recognized by different UBD-containing factors which control the activity of the ubiquitylated protein. Therefore, it is possible that an as yet unknown UBD-containing protein binds to ubiquitylated DRONC and controls its activity. For example, such an interaction could block the recruitment of ubiquitylated DRONC into the ARK apoptosome. Another possibility is that ubiquitylation could block dimerization of DRONC, which is required for activation of DRONC [@pgen.1002261-Snipas1].
"Undead" cells trigger *dronc* transcription {#s3c}
--------------------------------------------
'Undead' cells can be obtained by expression of the effector caspase inhibitor P35 [@pgen.1002261-Hay2]. In these cells, apoptosis has been induced, but cannot be executed due to effector caspase inhibition. Nevertheless, the initiator caspase DRONC is active in 'undead' cells and can promote non-apoptotic processes [@pgen.1002261-Fan1]. Recent work has suggested that 'undead' cells may alter their cellular behavior. For example, 'undead' cells change their size and shape, and have some migratory abilities to invade neighboring tissue [@pgen.1002261-Martin1]. They are also able to promote proliferation of neighboring cells causing hyperplastic overgrowth [@pgen.1002261-Kondo1], [@pgen.1002261-Ryoo1], [@pgen.1002261-Huh1]--[@pgen.1002261-Wells1] (reviewed by [@pgen.1002261-Bergmann1], [@pgen.1002261-Fan2]). Altered transcription of the TGF-ß/BMP member *decapentaplegic* (*dpp*), the Wnt-homolog *wingless* (*wg*), and the p53 ortholog *dp53* has also been reported in 'undead' cells [@pgen.1002261-Ryoo1], [@pgen.1002261-PerezGarijo1]--[@pgen.1002261-Wells1]. Intriguingly, while *dpp* and *wg* are usually not expressed in the same cells [@pgen.1002261-Tabata1], 'undead' cells co-express them ectopically, clearly indicating an altered transcriptional program.
As part of this altered transcriptional program, we showed that 'undead' cells also stimulate transcription of the initiator caspase *dronc* ([Figure 5](#pgen-1002261-g005){ref-type="fig"}). Interestingly, *p35* expression in normal cells does not induce *dronc* transcription suggesting that it is not the mere presence of P35 that induces *dronc* transcription, but instead the 'undead' condition of the affected cells. This transcriptional induction of *dronc* provides an explanation why DRONC protein levels are increased in 'undead' cells. It may also help to explain another observation regarding 'undead' cells. It has been demonstrated that although these cells are unable to die, they maintain the apoptotic machinery indefinitely [@pgen.1002261-Martin1], [@pgen.1002261-Yu2]. Therefore, as part of this maintenance program, 'undead' cells stimulate *dronc* transcription. This is likely not restricted for *dronc*. Martin et al. (2009) [@pgen.1002261-Martin1] also showed that DrICE protein levels remain high in 'undead' cells which may also be caused by increased *drICE* transcription. It is also possible that the induction of *dp53* by 'undead' cells [@pgen.1002261-Wells1] is part of this maintenance program, because we have shown that Dp53 induces expression of *hid* and *rpr* [@pgen.1002261-Fan3] and a positive feedback loop between *dp53*, *hid* and *dronc* exists in 'undead' cells [@pgen.1002261-Wells1]. This may all occur at a transcriptional level. The mechanism by which 'undead' cells stimulate expression of *dpp*, *wg*, *dp53*, *dronc* and potentially *drICE* are currently unknown and are avenues for future research.
Material and Methods {#s4}
====================
*Drosophila* genetics {#s4a}
---------------------
Fly crosses were conducted using standard procedures at 25°C. The following mutants and transgenes were used: *Uba1^D6^* [@pgen.1002261-Lee1]; *ark^G8^* [@pgen.1002261-Srivastava1]; *diap1^22-8s^* and *diap1^33^* ^-1s^ [@pgen.1002261-Wilson1]; *vps25^N55^* [@pgen.1002261-Herz1]; *dronc^I29^* [@pgen.1002261-Xu1]; *UAS-droncIR* (*dronc* inverted repeats) [@pgen.1002261-Leulier1]; *GMR-rpr* [@pgen.1002261-White1]; *dronc1.33-lacZ* [@pgen.1002261-Daish2], [@pgen.1002261-Cakouros1], *ubx-FLP* [@pgen.1002261-Emery1], *nub-GAL4* [@pgen.1002261-Brand1], *DE-* (*dorsal eye-*) *GAL4* [@pgen.1002261-Morrison1], and *UAS-hid* [@pgen.1002261-Zhou2]. *nub-FLP* is *nub-GAL4 UAS-FLP*. *UAS-p35* and *UAS-FLP* were obtained from Bloomington. *Uba1^D6^* is a temperature sensitive allele which at 25°C is a hypomorphic allele, but at 30°C is a null allele [@pgen.1002261-Lee1]. In the experiments in [Figure 1](#pgen-1002261-g001){ref-type="fig"}, [Figure 6B](#pgen-1002261-g006){ref-type="fig"}, and [Figure S1](#pgen.1002261.s001){ref-type="supplementary-material"}, *Uba1^D6^* and *Uba1^D6^ ark^G8^* mosaic larvae were incubated at 25°C; 12 hours before dissection the temperature was shifted to 30°C. This treatment allows recovery of *Uba1^D6^* null mutant clones, which otherwise are cell lethal.
Generation of mutant clones and expression of transgenes {#s4b}
--------------------------------------------------------
Mutant clones were induced in eye-antennal imaginal discs using the *FLP*/*FRT* mitotic recombination system [@pgen.1002261-Xu3] using *ey*-*FLP* [@pgen.1002261-Newsome1]. In this case, clones are marked by loss of GFP. Expression of *diap1* and *dronc* RNAi in *Uba1^D6^* clones ([Figure 1](#pgen-1002261-g001){ref-type="fig"}) was induced from *UAS-diap1* or *UAS-droncIR* transgenes using the MARCM system [@pgen.1002261-Lee2]. Here, clones are positively marked by GFP. For induction of *diap1*-expressing clones in [Figure 3](#pgen-1002261-g003){ref-type="fig"}, the *UAS-diap1* transgene was crossed to *hs*-*FLP*; *tub*\<*GFP*\<*GAL4* (\< = FRT). Clones are marked by the absence of GFP. MARCM clones and *diap1*-overexpressing clones were induced in first instar larvae by heat-shock for one hour in a 37°C water bath. Expression of *UAS-p35* in *diap1^22-8s^* mosaic discs was accomplished by *nub*-*GAL4*.
Immunohistochemistry {#s4c}
--------------------
Eye-antennal imaginal discs from third instar larvae were dissected using standard protocols and labeled with antibodies raised against the following antigens: DIAP1 (a kind gift of Hermann Steller and Hyung Don Ryoo); cleaved CASPASE-3 (CAS3\*) (Cell Signaling Technology) and anti-ß-GAL (Promega). The DRONC antibody used for immunofluorescence was raised against the C-terminus of DRONC in guinea pigs [@pgen.1002261-Wilson1]. This antibody is specific for DRONC ([Figure S3](#pgen.1002261.s003){ref-type="supplementary-material"}). Cy3- and Cy-5 fluorescently-conjugated secondary antibodies were obtained from Jackson ImmunoResearch. In each experiment, multiple clones in 10--20 eye and wing imaginal discs were analyzed, unless otherwise noted. Images were captured using an Olympus Optical FV500 confocal microscope.
Ubiquitylation assays {#s4d}
---------------------
Schneider S2 cells were co-transfected with pMT-DRONC C\>A V5, pAcDIAP1 (wt or CΔ6, F437A, respectively, described in [@pgen.1002261-Ditzel1]) and pAc His-HA-Ub at equal ratios. Expression of DRONC was induced overnight with 350 µM CuSO~4~. Cells were lysed under denaturing conditions and ubiquitylated proteins were purified using Ni^2+^-NTA agarose beads (QIAGEN). Immunoblot analysis was performed with α-V5 (Serotec) and α-DIAP1 antibodies [@pgen.1002261-Meier1], [@pgen.1002261-Zachariou1].
Immunoblot analysis {#s4e}
-------------------
For the immunoblots in [Figure 6A](#pgen-1002261-g006){ref-type="fig"}, embryos were collected, decorionated and snap frozen in liquid nitrogen. Embryos were sonicated in Laemmli SDS loading buffer while being frozen. The equivalent of 30 lysed embryos was loaded per lane. Immunoblots were done using standard procedures. The anti-DRONC antibody used in [Figure 6A](#pgen-1002261-g006){ref-type="fig"} is a peptide antibody raised against the large subunit of DRONC. The anti-DRONC antibody used in [Figure 6B](#pgen-1002261-g006){ref-type="fig"} was raised against the C-terminus of DRONC in guinea pigs.
Supporting Information {#s5}
======================
######
Loss of *ark* suppresses apoptosis in *Uba1* clones. *Uba1 ark* mosaic eye-antennal disc labeled for cleaved CASPASE-3 (CAS3\*) antibody (red). These discs were incubated at 30°C 12 hours before dissection (see [Material and Methods](#s4){ref-type="sec"}). Absence of GFP marks the location of *Uba1 ark* clones (see arrows). There is scattered apoptosis detectable. However, this occurs throughout the disc and does not correlate with the positions of the *Uba1 ark* double mutant clones. Genotype: *ey-FLP*; *FRT42D Uba1^D6^ ark^G8^*/*FRT42D ubi-GFP*.
(TIF)
######
Click here for additional data file.
######
*UAS-diap1* rescues *GMR-hid* and apoptosis induced in *vps25* mutants. Because the *UAS-diap1* transgene failed to suppress apoptosis in *Uba1* clones ([Figure 1C](#pgen-1002261-g001){ref-type="fig"}), we tested its ability to inhibit the strong apoptotic phenotype in two other paradigms. (A) Overexpression of the IAP-antagonist *hid* specifically in the fly eye under *GMR* promoter control gives rise to a strong eye ablation phenotype due to massive induction of apoptosis [@pgen.1002261-Grether1]. (B) Coexpression of *UAS-diap1* partially suppresses the *GMR-hid* eye ablation phenotype [@pgen.1002261-Hay1]. (C) *vps25* mutant clones induce a strong apoptotic phenotype. *vps25* encodes an component involved in endosomal protein sorting [@pgen.1002261-Herz1]. The apoptotic phenotype of *vps25* and *Uba1* as well as other phenotypes caused by inactivation of these genes are very similar, and both mutants were obtained in the same genetic screen [@pgen.1002261-Lee1], [@pgen.1002261-Herz1]. The left panel is the merge of GFP and anti-cleaved CASPASE-3 (CAS3\*) labeling, the right panel (C′) displays only the CAS3\* channel. White arrows mark a few clones as examples. (D) Overexpression of *diap1* completely suppresses the strong apoptotic phenotype of *vps25* mutant clones. The experimental conditions applied here are identical to the *Uba1* experiment in [Figure 1C](#pgen-1002261-g001){ref-type="fig"}. The left panel is the merge of GFP and anti-cleaved CASPASE-3 (CAS3\*) labeling, the right panel (D′) displays only the CAS3\* channel. Genotype: *hs-FLP UAS-GFP/UAS-diap1*; *FRT42D vps25^N55^*/*FRT42D tub*-*Gal80*; *tub*-*GAL4*. Genotypes: (A) *GMR-hid GMR-GAL4*. (B) *UAS-diap1; GMR-hid GMR-GAL4*. (C) *ey-FLP*; *FRT42D vps25^N55^*/*FRT42D* P\[*ubi-GFP*\]. (D) *ey-FLP*; *FRT42D vps25^N55^*/*FRT42D* P\[*ubi-GFP*\].
(TIF)
######
Click here for additional data file.
######
Specificity of the anti-DRONC antibody. The specificity of the anti-DRONC antibody used for immunofluorescence in [Figure 3](#pgen-1002261-g003){ref-type="fig"}, [Figure 4](#pgen-1002261-g004){ref-type="fig"}, and [Figure 5](#pgen-1002261-g005){ref-type="fig"} was verified in *dronc^I29^* mosaic eye (A) and wing (B) imaginal discs. The *dronc^I29^* allele contains a premature STOP codon at position 53 [@pgen.1002261-Xu1]. *dronc^I29^* clones were induced using the MARCM system, hence they are positively marked by GFP (arrows). The anti-DRONC antibody does not produce labeling signals in the mutant clones (arrows in A′ and B′, and the merge in A″ and B″), demonstrating that it is specific for DRONC. Genotype: *hs*-*FLP*; *dronc^I29^ FRT80*/*ubi*-*GFP FRT80*.
(TIF)
######
Click here for additional data file.
######
"Undead" *diap1^22-8s^* cells accumulate DRONC protein autonomously. (A, A′) Using MARCM, *p35*-expressing, 'undead' *diap1^22-8s^* mutant clones (green) were induced in eye discs and labeled for DRONC protein (red). DRONC protein autonomously accumulates in P35-expressing *diap1^22-8s^* clones (arrows). Similar results were obtained in wing discs (data not shown). Genotype: *hs-FLP tub-GAL4 UAS-GFP*/+; *UAS-p35*/+; *diap1^22-8s^ FRT80*/*tub-GAL80 FRT80.*
(TIF)
######
Click here for additional data file.
We would like to thank Hans-Martin Herz for the *vps25* data in [Figure S2](#pgen.1002261.s002){ref-type="supplementary-material"}; Audrey Christiansen for the outline of eye imaginal discs in [Figure 4A](#pgen-1002261-g004){ref-type="fig"}; Hugo Bellen, Georg Halder, Sharad Kumar, Hyung Don Ryoo, Hermann Steller, and Kristin White for antibodies and fly stocks; and the Bloomington Stock Center for fly stocks.
The authors have declared that no competing interests exist.
This work is supported by NIH grants (R01s GM068016, GM081543) and by an anonymous donor to AB. MB is supported by a fellowship of the Deutsche Forschungsgemeinschaft (DFG). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
[^1]: **¤:** Current address: Department of Cancer Biology, University of Massachusetts Medical School, Worcester, Massachusetts, United States of America
[^2]: Conceived and designed the experiments: AB PM TVL YF. Performed the experiments: TVL YF SW MS MB. Analyzed the data: AB PM TVL YF SW MS MB. Contributed reagents/materials/analysis tools: AB PM TVL YF SW MS MB. Wrote the paper: TVL AB.
| {
"pile_set_name": "PubMed Central"
} |
Background
==========
As obligate blood-feeding arthropods, ticks feed on terrestrial vertebrates and are one of the most important disease vectors worldwide. Ticks are involved in carrying and transmitting a diverse group of pathogenic bacteria, which exert severely negative impacts on human health, livestock production as well as wildlife \[[@B1]-[@B3]\]. Hard ticks (Ixodidae) usually have three life stages (larva, nymph, adult) in their lifespan and feed on distinct host species at different developmental stages. Thus, they are prone to acquiring various pathogens and spreading among vertebrate species during the whole life cycle \[[@B4]\].
In addition to pathogens, vertically-transmitted symbionts and foreign nonpathogenic bacteria also colonize the ticks \[[@B5]-[@B7]\]. Symbionts are closely associated with the host fitness \[[@B8],[@B9]\] and transmission and virulence of pathogens \[[@B10],[@B11]\]. To date, it is well-known that ticks harbour various symbionts, including *Coxiella*\[[@B12]\], *Francisella*\[[@B13]\], *Wolbachia*\[[@B14]\], *Rickettsia*\[[@B15]\], *Arsenophonus*\[[@B16]\], *Candidatus* Midichloria mitochondrii \[[@B17]\]. Moreover, some of them have crucial effects on the biological characteristics of ticks. For example, the obligate *Coxiella*-like symbiont manipulates the reproduction of the host *Amblyomma americanum*\[[@B18]-[@B20]\]. *Ca.* Midichloria mitochondrii, the symbiont of *Ixodes ricinus*, has the unique ability to enter and destroy mitochondria within ovarian cells of host \[[@B21],[@B22]\], which is coupled with the process of engorgement and molt \[[@B23],[@B24]\]. In addition, colonization of *Dermacentor variabilis* by *R. peacockii* can prevent secondary infection with other *Rickettsia* bacteria, including pathogenic species \[[@B25]\]*.* Clay \[[@B5]\] found that the infection frequency of *Arsenophonus* was negatively correlated to *Rickettsia* in *A. americanum*, indicating that infection with symbionts may affect the microbial community structure and their interaction.
Ticks could also obtain allochthonous nonpathogenic bacteria by sucking host blood and contact with the natural environment \[[@B26]-[@B28]\]. Although allochthonous bacteria are not directly related to the pathogenicity of hosts, they can reflect the habitats of hosts and even have the potential to impact on the density and transmission of host-borne agents \[[@B29]\]. Overall, diverse microbial communities can exert great impact on the fitness of ticks and the colonization and transmission of tick-borne pathogens.
*Haemaphysalis longicornis* (Acari: Ixodidae), a three-host tick, extensively distributed in China \[[@B30]\], Korea, Japan, New Zealand, and Australia \[[@B31],[@B32]\], has inflicted serious damage on public health and economics. It is reported that this tick species transmitted a variety of pathogens \[[@B33],[@B34]\], including *Babesia microti*\[[@B35]\], *Ehrlichia chaffeensis*\[[@B36]\], *Anaplasma bovis*\[[@B37]\], *A. phagocytophilum*\[[@B38]\], *C. burnetii*\[[@B39]\], Spotted fever group rickettsiae \[[@B40]\], and *Borrelia burgdorferi*\[[@B41]\]. With regard to the human health and economic significance, pathogens carried by *H. longicornis* have been well investigated in China \[[@B42]-[@B47]\]. However, little is known about the broader array of the microbial community in *H. longicornis* under natural conditions.
In the present study, we investigated the composition of bacterial communities associated with *H. longicornis* using PCR-RFLP and DNA sequencing approaches, and evaluated the putative symbionts*.*
Methods
=======
Sample collection and storage
-----------------------------
*H. longicornis* samples were collected in Xiaowutai National Natural Reserve Area in China by flag dragging. A total of 500 ticks were collected and stored at -80°C. We also collected 1,000 ticks, which were reared on rabbits as described by Liu *et al*. \[[@B48]\]. Colonies of these ticks were reared on the ears of rabbits. Rabbits were maintained in a room with 50-55% relative humidity (RH) at 25--27°C and exposed to daylight. After detachment, ticks were collected and incubated in cotton-plugged glass tubes filled with folded filter paper in an incubator with 75 ± 5% RH and 6/18 h of L/D cycle at 26 ± 1°C. The protocol of all animal experiments was approved by the Institutional Animal Care and Use Committee of Hebei Normal University.
DNA extraction
--------------
The genomic DNA for constructing eubacterial 16S rRNA gene libraries were extracted from a group of adults (10 females and 10 males). To assess the prevalence, vertical transmission and infection sites, DNA was isolated from individual field-collected adults, and pooled ticks at different developmental stages (500 eggs, 200 larvae and 50 nymphs) and different tissues (ovaries, salivary glands, Malpighian tubules and midguts), respectively. Before DNA extraction, the whole tick samples were sterilized as described previously \[[@B5]\], and dissected tissues were washed in sterile phosphate-buffered saline (PBS) (137 mM NaCl, 2.7 mM KCl, 4.3 mM Na~2~HPO~4~ · 7H~2~O, 1.4 mM KH~2~PO~4~, pH 7.4) for three times. Subsequently, all samples were snap-frozen in liquid nitrogen and purified using the DNeasy Tissue Kit (Qiagen, Germany) according to the manufacturer's instructions.
Eubacterial 16S rRNA gene library construction, RFLP analysis and sequencing
----------------------------------------------------------------------------
The eubacterial 16S rRNA gene clone library was constructed by amplifying about 1500 bp fragment of 16S rRNA gene using eubacterial universal primers 27 F/1492R \[[@B49]\]. Purified PCR products were cloned into pEASY-T1 vector (TransGen, China) and transformed into *Escherichia coli* TOP10 competent Cell (TransGen, China). The positive clones were subjected to restriction fragment length polymorphism (RFLP) analyses using both HaeIII and RsaI restriction endonucleases. Sequencing of the positive clones with different restriction fragment patterns were performed by Sangon Biotech Company (China).
Phylogenetic analyses
---------------------
The 16S rRNA gene sequences of the most closely related species were retrieved from GenBank and then aligned. The similarity analyses were performed using the CLUSTAL_W program \[[@B50]\]. The phylogenetic tree was constructed using neighbour-joining method implemented in MEGA4.0 \[[@B51]\]. The stability of the tree clustering was evaluated by means of a bootstrap analysis of 1000 datasets \[[@B52]\].
Diagnostic PCR
--------------
To assess their prevalence, vertical transmission and infection site of symbionts, diagnostic PCR assay was performed with four sets of specific primers (Table [1](#T1){ref-type="table"}). The PCR reaction contained 20 mM Tris--HCl (pH 8.4), 50 mM KCl, 1.5 mM MgCl~2~, 200 μM each dNTP, 2.5 U Platinum *Taq* DNA Polymerase (Invitrogen, America), and 0.5 mM each primer. PCR cycling conditions were as follows: 1 cycle of 94°C for 2 min; 30 cycles of 94°C for 30 s, 55°C for 30 s, and 72°C for 15 s, and 72°C for 10 min.
######
Oligonucleotide primers used for PCR amplification and sequencing
**Primer name** **Genera or Species** **Target gene** **Nucleotide sequence (5′-3′)** **Annealing temperature (°C)** **Approx product size (bp)** **Reference**
-------------------- ----------------------- ----------------- --------------------------------- -------------------------------- ------------------------------ ---------------
CLS F *Coxiella* 16S rRNA CACGTAGGAATCTACCTTGTAG 55 90 55
CLS R CGTTTTGTTCCGAAGAAATTAT
ALS 82 F *Arsenophonu* 16S rRNA AGGGAGCTTGCTTCCTGGCCGG 59 130 55
ALS 198R CGAAGGTGTGAGGCCTAATGG
*Rickettsia* 354 F *Rickettsia* 16S rRNA CAGCAATACCGAGTGAGTGATGAAG 56 350 69
*Rickettsia* 647R AGCGTCAGTTGTAGCCCAGATG
NCLS F NCLS 16S rRNA TCCCTGGCGGCGAGTGG 55 110 This study
NCLS R CGTATTAGAGATTAGAGAAACC
*Rr*190.70p *Rickettsia* *rompA* ATGGCGAATATTTCTCCAAAA 48 540 57
*Rr*190.602n GTTCCGTTAATGGCAGCATCT
Results
=======
Diversity of bacterial communities from field-collected *H. longicornis*
------------------------------------------------------------------------
Our results revealed diverse microbial communities associated with field-collected *H. longicornis* (Table [2](#T2){ref-type="table"}). In the 16S rRNA gene clone libraries, the most abundant sequences (belonged to *Coxiella* and designated as CLS-Hl) in both males and females share 99.5% similarity to the symbiont of *H. longicornis* (GenBank: AB001519) \[[@B53]\]*.* In females, the second abundant sequence (designated as NCLS-Hl) belonged to *Coxiella*, sharing 95.3% similarity to *Coxiella*-like symbionts of *Rhipicephalus sanguine* (GenBank: D84559)*.* Following NCLS-Hl, sequences belonged to *Rickettsia* (designated as RLS-Hl) and *E. coli. Rickettsia* bacteria shared the highest sequence similarity (99.8%) with symbiotic *Rickettsia* of *D. varibilis* (GenBank: U55820) \[[@B54]\]. In males, the second abundant microbes (designated as ALS-Hl) belonged to *Arsenophonus*, sharing over 99.5% and 97.9% similarity to symbionts of *D. silvarum* (GenBank: JN866582) \[[@B55]\] and *D. variabilis* (GenBank: AY265342) \[[@B56]\], respectively*.* Following ALS-Hl, there were RLS-Hl and NCLS-Hl. Interestingly, the top four abundant bacteria in females were consistent with those in males.
######
**Eubacterial 16S rRNA gene clone libraries from females and males of*Haemaphysalis longicornis***
**Classification** **Closest taxon (GenBank No.)** **Identity %** **Sequences**^**\#**^**(F/M)\*** **Accession No. (GenBank)**
----------------------------------------------------------------------- ---------------------------------------- ---------------- ---------------------------------- -----------------------------
**Proteobacteria**
α-proteobacteria *Rhizobium* spp. (EF363715) 99% 4/0 JN866565
*Rickettsia* spp. (CP003319)^§^ 99% 9/12 JN866571
*Devosia psychrophila* (GU441678) 98% 0/1 JN866597
*Asticcacaulis excentricus* (CP002395) 99% 2/0 JN866569
β-proteobacteria *Janthinobacterium lividum* (EU275366) 99% 0/5 JN866578
γ-proteobacteria *Escherichia coli* (CP001673) 100% 8/7 JN866566
*Moraxella osloensis* (AB643592) 99% 4/0 KF421818
*Pseudomonas* spp*.* (AB334768) 99% 0/7 JN866577
*Coxiella*-like symbionts of *H. longicornis* (AB001519)^§^ 99% 36/19 JN866564
*Coxiella*-like symbionts of *R. sanguine* (D84559)^§^ 95% 21/11 JN866567
*Arsenophonus*-like symbionts of *Dermacentor silvarum* (JN866582)^§^ 99% 5/16 JN866572
**Bacteroidetes**
Sphingobacteria *Mucilaginibacter* sp. (GU139695) 98% 1/0 JN866568
*Sphingomonas asaccharolytica* (AY509241) 98% 0/1 JN866579
**Actinobacteria**
Actinobacteria *Williamsia maris* (NR024671) 99% 0/2 JN866598
*Conexibacter woesei* (CP001854) 98% 0/1 JN866581
^\*^Source of the sequences from female ticks (F) or male ticks (M).
^\#^Sequences indicate the number of 16S rRNA gene sequences obtained for each 16S rRNA gene clone libraries.
^§^Sequences are related phylogenetically to previously reported tick-associated bacteria.
Symbionts
---------
*OmpA* gene sequence analyses revealed that RLS-Hl belonged to *Ca.* Rickettsia hebeiii, with 99.1% sequence similarity \[[@B57]\]. To confirm vertical transmission of symbionts, eggs and the first generation of laboratory-reared ticks (larvae, nymphs and adults) was tested by diagnostic PCR assay. The putative symbionts were found in all samples detected, with the exception of NCLS-Hl. These results indicated that CLS-Hl, ALS-Hl and RLS-Hl were inherited from progeny to the next generation, while NCLS-Hl was not (Figure [1](#F1){ref-type="fig"}). To determine the stability of vertical transmission, we investigated the presence of each bacterium in the 7th generation of laboratory-reared ticks (F7). The results showed only CLS-Hl was detected in F7 ticks, while ALS-Hl and RLS-Hl were not, suggesting that CLS-Hl can be maintained stably in laboratory-reared ticks.
{#F1}
In natural field populations, CLS-Hl showed 100% infection rate; ALS-Hl occurred in 33.3% (14/36) females and 83.3% (10/12) males; the prevalence of RLS-Hl was 88.9% (32/36) and 100% (12/12) in females and males, respectively. In the laboratory population, the prevalence of CLS-Hl was 100%.
Analysis of the infection site showed that three types of symbionts infected all tissues detected (accessory glands, testes, salivary glands, midguts) and there is no significant difference among all tissues in males. In females, CLS-Hl colonized specifically in the ovaries and Malpighian tubules, while ALS-Hl and RLS-Hl infected all tissues tested (ovaries, salivary glands, Malpighian tubules and midguts) (Figure [2](#F2){ref-type="fig"}).
{#F2}
Discussion
==========
In this study, we showed that the natural *H. longicornis* harboured a diverse group of bacterial species. Previous studies have shown that some tick species were colonized with complex microbial communities, including *A. americanum*, *I. ricinus*, *I. scapularis* and *Rh. microplus*\[[@B5]-[@B7],[@B26]-[@B28],[@B58],[@B59]\]. Investigation of microbiota in the ticks revealed the complexity of ecological interactions between host and microbe and provided insight for the biological control of ticks.
Symbionts
---------
Interestingly, some sequences were phylogenetically related to previously reported tick-associated bacteria, including (1) two types of *Coxiella*-like bacteria, CLS-Hl and NCLS-Hl, (2) *Arsenophonus*-like symbiont, ALS-Hl, (3) *Rickettsia*-like symbiont, RLS-Hl. It is now widely accepted that the co-infection with multiple symbionts occurs in many arthropod hosts \[[@B60]\]. However, co-infection with symbionts has been demonstrated only in a few tick species, such as *A. americanum*\[[@B5]\], *Ornithodoros moubata*\[[@B61]\], *D. silvarum*\[[@B55]\] and *Rh. sanguineus*\[[@B62]\]. We reported here for the first time that *Coxiella*-like, *Arsenophonus*-like and *Rickettsia*-like symbionts were detected simultaneously in wild *H. longicornis*. Interestingly, we identified two different types of *Coxiella*. CLS-Hl, with 95.1% similarity to the *C. burnetii*, shared 99.5% sequence identify with the symbionts of *H. longicornis* previously reported \[[@B39],[@B53]\]. However, NCLS-Hl, with 95.1% and 95.3% similarity to the *C. burnetii* and symbionts of *Rh. sanguine*, respectively, was regarded as a novel *Coxiella*-like agent and has never been described before. Phylogenetic analyses revealed (Figure [3](#F3){ref-type="fig"}) that CLS-Hl and NCLS-Hl form an individual clade, although both of them belong to the cluster of hard tick species (Ixodidae). CLS-Hl was clustered with symbionts of *Haemphysalis* (*H. longicornis* and *H. shimoga*), while NCLS-Hl was grouped together with the symbionts of *Rhipicephalus* (*Rh. sanguineus* and *Rh. turanicus*). However, the determination of vertical transmission does not support that NCLS-Hl was a symbiont. Further studies are needed to confirm the pathogenicity of this novel *Coxiella* agent to vertebrate hosts.
{#F3}
In ticks, *Coxiella* bacteria are the most common agent, exhibiting diverse genotypes in different tick species \[[@B5],[@B12],[@B63]-[@B68]\]. By comparing the phylogeny between the tick mitochondrial 16S rRNA gene and the *Coxiella* agents carried by ticks, Almeida \[[@B12]\] showed that ticks were phylogenetically associated with their *Coxiella* agents, suggesting that *Coxiella* symbionts and ticks shared a long co-evolution process. In *H. longicornis*, *Coxiella* symbionts from different areas, such as Japan \[[@B53]\] and Korea \[[@B39]\], shared identical genotypes. This wide geographical distribution indicates that *Coxiella* symbionts form ancient symbiotic associations with their host before spreading.
Previous studies suggested that *Coxiella* symbionts in *A. americanum*, with 100% infection, vertical transmission, reduced-genome, regulation of the reproduction of the host, was closely related to the fitness of the ticks. Besides, in *O. rostratus*\[[@B12]\], *D. silvarum*\[[@B55]\] and *Rh. sanguineus*\[[@B62]\], *Coxiella*-like symbionts exhibited nearly perfect prevalence and were presumed as obligate symbionts. In this study, 100% infection rate, perfect vertical transmission, colonization in specialized tissues, CLS-Hl was regarded as obligate and contributed to the fitness of ticks.
*Arsenophonus* is one of the four major inherited symbionts of arthropods with an infection rate of 5% \[[@B60]\]. *Arsenophonus* has been reported to infect *A. americanum*\[[@B5]\], *D. variablilis*\[[@B16]\], *D. andersoni*\[[@B16]\] and *D. silvarum*\[[@B55]\]. Interestingly, the sequence of ALS-Hl was identical to that of *D. silvarum* collected from the same place as the *H. longicornis*, possibly suggesting horizontal transmission of *Arsenphonus*-like symbionts between different genera of ticks. Previous studies also provided evidence of horizontal transmission of symbionts between *D. variabilis* and *D. andersoni*\[[@B16]\]. Therefore, *Arsenophonus* possessed not only intragenus, but also intergenus horizontal transmission.
*Rickettsia*, with both symbiotic and pathogenic lifestyle, inhabits multiple tick species. Pathogenic rickettsiae have been extensively investigated, while only few studies on symbiotic rickettsiae have been reported. To date, *Rickettsia*-like symbionts have been identified in *D. variabilis*\[[@B54]\]*, D. andersoni*\[[@B69]\]*, D. silvarum*\[[@B55]\]*I. scapularis*\[[@B15]\] and *A. americanum*\[[@B5]\]. In this study, we showed molecular evidence for the vertical transmission of the *Rickettsia*-like symbiont (*Ca.* Rickettsia hebeiii) in *H. longicornis*. Moreover, high prevalence in the natural environment suggested interaction between RLS-Hl and their hosts.
However, RLS-Hl and ALS-Hl were not detected in F7 laboratory-reared ticks, suggesting that they cannot be maintained stably from one generation to the progeny in the laboratory. This phenomenon was also observed in *I. ricinus*, whose symbionts IricES1 were lost from female adults to progeny in the laboratory \[[@B70]\]. The transmission efficiency of symbionts decreased significantly in the laboratory, which was partially due to the competition among symbionts and bottlenecks during vertical transmission \[[@B71]\]. Alternatively, transmission reduction may be due to temperature differences between natural and laboratory environments \[[@B72]\].
Conclusions
===========
A diverse array of microbial communities were identified from field-collected *H. longicornis*. Three types of symbionts were detected in a single host simultaneously. Moreover, analysis on the prevalence, vertical transmission and infection sites supported obligate symbiotic association between *Coxiella* symbionts and its host. The role of *Coxiella* symbionts in the host fitness and the interaction among microbial communities remains to be elucidated. Our investigation of microbial communities in the ticks revealed the complexity of ecological interactions between host and microbe and provided insight for the biological control of ticks.
Abbreviations
=============
F7: The 7th generation of laboratory-reared ticks; CLS-Hl: *Coxiella*-like symbiont of *H. longicornis*; NCLS-Hl: Novel *Coxiella*-like symbiont of *H. longicornis*; RLS-Hl: *Rickettsia*-like symbiont of *H. longicornis*; ALS-Hl: *Arsenophonus*-like symbiont of *H. longicornis.*
Competing interests
===================
The authors declare that they have no competing interests.
Authors' contributions
======================
Liu L-M and Liu J-Z conceived and designed the study, constructed 16S rRNA gene libraries and performed RFLP analyses, drafted the manuscript, and critically revised the manuscript. Liu J-N carried out RFLP analyses and diagnostic PCR. Liu Z, Yu Z-J, Xv S-Q and Yang X-H, participated in sample collection, study implementation and data collection and helped to revise the manuscript. Li T, Li S-S and Guo L-D participated in sample collection and tick feeding. All authors read and approved the final manuscript.
Acknowledgements
================
This work was supported by grants from the National Natural Science Foundation of China (Grant No. 31272372) and the Specialized Research Fund for the Doctoral Program of Higher Education of China (Grant No. 20131303130001).
| {
"pile_set_name": "PubMed Central"
} |
Introduction
============
Several parameters of quality have been defined for optimal colonoscopy, such as cecal intubation, adverse event rates, duration of withdrawal, and bowel preparation [@JR059-1] [@JR059-2] [@JR059-3] [@JR059-4] [@BR059-5]. High-quality bowel preparation is noted to be of primary importance for an effective colonoscopic examination.
Among bowel preparation agents, osmotic laxatives (i. e., mainly water-soluble polymers that are neither digested nor absorbed in the small intestine) have been widely used for more than three decades [@JR059-6]. The osmotic action of these agents increases the colonic fluid volume by maintaining water in the colonic lumen, thereby distending the bowel and indirectly increasing bowel peristalsis.
More recently, osmotic laxatives containing sodium phosphate have been proposed, in order to decrease the volume of fluid intake and improve patient acceptability [@JR059-7] [@JR059-8] [@JR059-9] [@JR059-10]. However, recent studies have reported that these agents, in solution or in a tablet formulation, may induce aphthous or erosive injuries of the stomach and colon in some patients [@JR059-11] [@JR059-12] [@JR059-13] [@JR059-14] [@JR059-15]. In addition, after the introduction of sodium phosphate tablets (NaPT) in France, several individual case safety reports of gastritis or stomach ulcerations were registered. To date, no risk factors, such as preexisting lesions, history, co-morbidities, and concomitant drug use, have been identified. Although these lesions are frequently asymptomatic and can be viewed as "false-positive" findings, it is important to explain their pathophysiology and evolution more clearly. For instance, although the specific mechanisms inducing such lesions remain unknown, it has been suggested that an increase in the production of free radicals may be involved [@JR059-16]. However, defining these osmotic laxative -- induced lesions in humans remains difficult. Therefore, it was thought that an experimental model would be helpful by allowing time course studies to be done. Such an experimental model should fulfill several criteria, enabling (1) prolonged and stable exposure to laxative or control substances, (2) in vivo and possibly repeated endoscopic examinations, and (3) histologic and functional analysis of the exposed mucosa.
In an attempt to reproduce the gastric lesions observed in humans, we developed a pig model to determine the effects of the prolonged and direct application of NaPT on the gastric mucosa by using a combination of endoscopic, probe-based confocal laser endomicroscopy (pCLE) and histologic and functional methods.
Materials and Methods
=====================
Animals and tablets
-------------------
A total of 14 pigs (weighing between 30 and 40 kg) were used in this study. Experiments were done in a dedicated surgical facility of the Laboratoire des Grands Animaux, INSERM U 643, in accordance with French Veterinary Regulations and Ethics Committee standards (agreement E.44010). During all procedures, the animals were under general anesthesia, induced with 5 % isoflurane and 60 % nitrous oxide and subsequently maintained with 2 % isoflurane.
Each NaPT (Colokit^®^; Mayoly Spindler, Chatou, France[)]{.smallcaps} weighed 1500 mg and was composed of 1102 mg of monobasic sodium phosphate monohydrate and 398 mg of dibasic anhydrous sodium phosphate. Each placebo tablet (PlaT) weighed 1704 mg and was composed of 170.4 mg of polyethylene glycol, 8.52 mg of magnesium stearate, 1354.68 mg of lactose (Tablettose^®^; Meggle, Wasserburg, Germany), and 170.4 mg of microcrystalline cellulose (Avicel^®^; FMC BioPolymer, Philadelphia, Pennsylvania, USA).
Study design
------------
The study design is summarized in [Fig. 1](#FI059-1){ref-type="fig"}. An initial study (study 1) was conducted with five pigs to determine the feasibility of the procedure and to characterize early lesions observed after NaPT had been applied for 1.5 hours. Two consecutives studies were then performed to determine the time course of the lesional aspects at 1.5 hours, 24 hours (study 2, n = 5), and 72 hours (study 3, n = 4) after the application of NaPT and PlaT. In order to consider potential mechanical effects of the tablets, a normal mucosal area, hereafter referred as the control site, was also analyzed. Thus, three different sites of analysis were considered: control, PlaT, and NaPT.
{#FI059-1}
At each considered time (0, 1.5, 24, and 72 hours), both the macroscopic and microscopic (pCLE) aspects of each site were analyzed in vivo during upper gastrointestinal endoscopy. At the end of the study, an endoscopic mucosal resection (EMR) was performed on each site, and specimens were sent for further ex vivo morphologic and functional analysis.
Endoscopic procedure
--------------------
During all procedures, performed by a single experienced endoscopist (E.C.), the animals were in prone position and under general anesthesia. The endoscopic appearance of the gastric mucosa was assessed meticulously with a standard resolution endoscope (EG450; Fujifilm, Tokyo, Japan). First, particular attention was paid to washing the gastric cavity with syringes of water. The best area on which to apply both tablets (PlaT and NaPT) was then defined as a stable area in the fundus with no visible contractions. One NaPT tablet was captured with an endoscopic snare, stabilized within a cap placed at the distal tip of the endoscope, and placed into the stomach at the level of the fundus. The same procedure was immediately repeated with a PlaT, which was placed 2 to 3 cm to the right of the previous tablet. Finally, one endoscopic clip was positioned between the two tablets in order to allow re-identification of these areas at 24 and 72 hours ([Fig. 2](#FI059-2){ref-type="fig"}). The macroscopic aspect of each site of analysis was described and graded as follows: grade 0, no lesion; grade 1, mild lesion defined by the presence of exanthema; grade 2, moderate lesion defined by the presence of aphthous ulcer and/or black clots; grade 3, severe lesion defined by the presence of ulcer and/or active bleeding.
{#FI059-2}
Probe-based confocal laser endomicroscopy examination
-----------------------------------------------------
The pCLE procedure was performed by a single experienced endoscopist (E.C.), as follows. The confocal miniprobe (ColoFlexUHD, Cellvizio; Mauna Kea Technologies, Paris, France) was gently positioned onto the gastric mucosa in contact with either the drug tablet or the placebo tablet, or onto the control site. A volume of 5 mL of 10 % fluorescein sodium (Novartis Pharma, France) was injected intravenously as a contrast agent. The pCLE video sequences were recorded from the time of injection to 5 minutes after injection for further "off-line" analysis.
Following the procedure, each video clip was reviewed by one endoscopist (M.P.), who was blinded to the procedure. The six best images from each pCLE video sequence were extracted and transferred to an external drive in PNG (portable network graphics) format to allow further semiquantitative analysis. The final data set comprised 678 images from 14 pigs. The following parameters were assessed: epithelial lining irregularity, architectural disorganization, and fluorescein intensity, which was measured in both the crypt lumen and the intercryptic area. ImageJ 1.42q software was used to measure fluorescein intensity (National Institutes of Health, Bethesda, Maryland, USA). Epithelial irregularity and architectural disorganization were assessed with a semiquantitative score ([Table 1](#TB059-1){ref-type="table"}) adapted from Li et al. [@JR059-17] *.*This score was assessed independently by two procedure-blinded investigators (M.D. and N.M.).
###### Classification of endomicroscopic lesions.
0 (absent) 1 (moderate) 2 (severe)
--------------------------------- ---------------------------------------- ---------------------------------------------------------------- ----------------------
Surface epithelium architecture Regular thickness Irregular thickness Destroyed epithelium
Crypts architecture Regular arrangement and size of crypts Irregular arrangement of crypts, enlargedspaces between crypts Crypt destruction
Ex vivo assessment of paracellular permeability
-----------------------------------------------
Fundic specimens obtained by EMR were immediately placed into Krebs solution (0.187 g/L NaH~2~PO~4~∙2H~2~O, 6.84 g/L NaCl, 0.35 g/L KCl, 2.10 g/L NaHCO~3~, 1.98 g/L glucose, 0.368 g/L CaCl~2~∙2H~2~O, 0.244 g/L MgCl~2~∙6H~2~O) at 4 °C. For each site, the EMR specimen was microdissected to separate the mucosa from the submucosal layer (in the plane of the submucosal blood vessels). Specimens of isolated fundic mucosa were then mounted in dedicated Ussing chambers (Physiologic Instruments, San Diego, California, USA) with a chamber surface of 0.0314 cm^2^. Each chamber contained 2 mL of Ham's F12 Nutrient Mixture (Life Technologies, Grand Island, New York, USA) containing 0.1 % fetal calf serum. The media were continuously oxygenated by a gas flow of 95 % O~2~ and 5 % CO~2~ and maintained at 37 °C. After an equilibration period of 30 minutes, 200 µL of apical medium was replaced with 200 µL of fluorescein -- 5.6 sulfonic acid (1 mg/mL). Every 30 minutes, the fluorescence level of basolateral aliquots of 150 µL was measured over a period of 180 minutes with a fluorometer (Thermo Scientific). Paracellular permeability was determined by calculating the slope of the changes in fluorescence intensity over time with a linear regression fit model (GraphPad Software, La Jolla, California, USA).
Histologic evaluation
---------------------
For each site, a portion of the EMR specimen was fixed in 4 % paraformaldehyde in phosphate-buffered saline. After tissue washing in phosphate-buffered saline, the specimen was dehydrated and embedded in paraffin. Sequential sections of mucosa (5 µm) were obtained and, after staining in hematoxylin and eosin, were analyzed with an Olympus IX50 microscope (Olympus America, Center Valley, Pennsylvania, USA). Pictures were captured with a DP71 digital camera (Olympus) connected to a computer through a frame grabber card (Cell\^B software, Olympus). To assess the mucosal morphology, we analyzed mucosal thickness, determined by measuring the distance between the surface epithelium and the underlying muscularis mucosae (only in study 1).
Statistical analysis
--------------------
Because of the limited sample size, nonparametric tests were used for statistical analysis. Two groups were compared with either the Mann -- Whitney test or the Wilcoxon test. More than two groups were compared with either the Kruskal -- Wallis test or the Dunn test. Comparisons between groups, and over time, were performed with a two-way analysis of variance (ANOVA) followed by a Bonferroni test. All statistical analyses were performed with GraphPad Prism 5.00 for Windows (GraphPad Software).
Results
=======
Induction of specific, early gastric mucosal lesions with sodium phosphate tablets
----------------------------------------------------------------------------------
All tablets were successfully applied onto the fundic mucosa under endoscopic guidance ([Fig. 2](#FI059-2){ref-type="fig"}). At 1.5 hours, all tablets remained in position at the original site and were partially dissolved.
First, we used white light endoscopy to grade the mucosal changes induced by NaPT. All sites in all animals were macroscopically normal (grade 0) at the time of NaPT application. At 1.5 hours, only grade 1 lesions were observed in 8 of the 14 animals (57 %) at the NaPT sites. No grade 2 or 3 lesions were observed. No lesions were observed at control sites.
Second, we used pCLE to analyze and quantify microscopic mucosal changes in vivo. With semiquantitative score, we indicated that all sites initially had normal epithelial lining and glandular architecture (grade 0). At 1.5 hours, no epithelial irregularity or architectural disorganization was seen at control sites. In contrast, a significant increase in epithelial irregularity and architectural disorganization scores was noted at NaPT sites ([Fig. 3 a, b](#FI059-3){ref-type="fig"}). Blinded interobserver analysis showed an acceptable match between two independent investigators (kappa = 0.52). Further quantitative pCLE analysis showed that changes in pit intensity over 1.5 hours were significantly greater at NaPT sites than at control sites. In contrast, changes in intercryptic intensity did not differ between control and NaPT sites during this time interval ([Fig. 3 c, d](#FI059-3){ref-type="fig"}).
![ Short-term effects of sodium phosphate tablets (NaPT) 1.5 hours after application onto the fundic mucosa of pigs, assessed with probe-based confocal laser endomicroscopy (studies 2 and 3) and graded. **a** Semiquantitative lesion score of surface epithelium architecture. **b** Semiquantitative lesion score of crypts architecture. **c** Variations of fluorescein intensity in the crypt lumen \[(H1.5 -- H0) × 100 /H0\]. **d** Variations of intercryptic intensity. Wilcoxon signed-rank test: \* *P* ≤ 0.05, \*\*\**P* \< 0.0001; Mann -- Whitney test: ^\#^ *P* \< 0.0025. PlaT, placebo tablet; Ct, control; AU, arbitrary units.](10-1055-s-0034-1377934-i059ei3){#FI059-3}
Third, in addition to in vivo endoscopic analysis, we performed both histologic and functional analysis on EMR specimens collected at the end of this part of the study. We showed that mucosal thickness at NaPT sites was similar to that at control sites (0.33 ± 0.05 mm vs. 0.27 ± 0.04 mm; n = 5; *P* \> 0.05). Consistently, fundic paracellular permeability was not changed at NaPT sites in comparison with control sites (0.28 ± 0.04 vs. 0.33 ± 0.05; n = 5; *P* \> 0.05). Furthermore, there was no correlation between paracellular permeability and pit or crypt intensity.
Finally, in order to consider potential mechanical effects of NaPT on mucosal lesions, we performed a comparative study with PlaT. Initially, all sites were macroscopically normal (grade 0) and remained normal 1.5 hours after application of the PlaT (n = 9). PlaT induced a slight but significant increase in epithelial irregularity and architectural scores ([Fig. 3 a, b](#FI059-3){ref-type="fig"}). However, epithelial irregularity and architectural disorganization scores were significantly higher at NaPT than at PlaT sites ([Fig. 3 a, b](#FI059-3){ref-type="fig"}). PlaT did not induce a change in pit intensity at 1.5 hours after application, in contrast to NaPT ([Fig. 3 c](#FI059-3){ref-type="fig"}). Similarly, PlaT did not induce a change in intercryptic intensity at 1.5 hours after application, in contrast to NaPT ([Fig. 3 d](#FI059-3){ref-type="fig"}).
Evolution of gastric mucosal lesions induced by sodium phosphate tablets
------------------------------------------------------------------------
This part of the study attempted to monitor the evolution of the mucosal lesions induced by NaPT at 24 and 72 hours after tablet application ([Fig. 4](#FI059-4){ref-type="fig"}). At 24 and 72 hours, with white light endoscopy, all application sites could be easily re-identified by the presence of the endoscopic clips. There were no residual tablets at these sites.
{#FI059-4}
First, macroscopic examination 24 hours after NaPT application revealed that 2 of 9 pigs had grade 1 lesions (22 %), 1 had grade 2 lesions (11 %), and the remaining 6 had no lesion (67 %). In 2 of 9 pigs (22 %), the PlaT site showed grade 1 lesions, and the remaining pigs had no lesion at the PlaT site. Finally, no lesion was visible at the control site in any pig. At 72 hours, no lesion was identifiable in any animal for any conditions (NaPT, PlaT, and control).
Second, pCLE at PlaT and control sites showed that crypt pit intensity and intercryptic intensity remained unchanged over the different time points studied ([Fig. 5](#FI059-5){ref-type="fig"}). In contrast, at NaPT sites, the increased semiquantitative scores and crypt pit intensity observed at 1.5 hours remained unchanged at 24 hours. However, at 72 hours, the semiquantitative scores and crypt pit intensity were significantly reduced in comparison with the values at 24 hours and were similar to those measured before the application of NaPT ([Figs. 4, 5](#FI059-4){ref-type="fig"}).
![ Time course of endomicroscopic aspects at 1.5, 24, and 72 hours after the application of sodium phosphate tablets (NaPT) or placebo tablets (PlaT) in pigs. Control (Ct): unexposed mucosal area. **a** Semiquantitative lesion score of surface epithelium architecture. **b** Semiquantitative lesion score of crypts architecture. **c** Variations of fluorescein intensity in the crypt lumen \[(H1.5 -- H0) × 100 /H0\]. **d** Variations of intercryptic intensity. Wilcoxon signed-rank test: \* *P* ≤ 0.05 vs. value at H0. Mann -- Whitney test: ^\#^ *P* \< 0.05 vs. value of Ct at identical time and ^‡^ *P* \< 0.05 vs. value of PlaT at identical time. AU, arbitrary units.](10-1055-s-0034-1377934-i059ei5){#FI059-5}
Finally, we performed functional analysis on EMR specimens collected at 24 and 48 hours after tablet application. Fundic paracellular permeabilities were similar at each site and at all time points (0.19 ± 0.07 vs. 0.22 ± 0.07 vs. 0.18 ± 0.07 and 0.10 ± 0.02 vs. 0.09 ± 0.04 vs. 0.10 ± 0.03, respectively, for NaPT, PlaT and Ct sites at 24 and 72 hours, respectively).
Discussion
==========
The present study demonstrated that a 90-minute application of NaPT onto fundic mucosa induces endomicroscopically detectable changes in the epithelial lining. These changes persisted for 24 hours after NaPT application but were resolved at 72 hours. In addition, standard video endoscopy revealed minor macroscopic lesions, such as local exanthema or superficial lesions, at the site of NaPT application in a subgroup of animals. These lesions disappeared by 72 hours after application.
Few studies have specifically determined the gastrotoxic effects of orally administered drugs, and this characterization remains difficult and limited in humans [@JR059-18]. In the present study, we developed an original approach for characterizing both the features of injury (macroscopic and microscopic) and the functional consequences of direct, drug-induced gastric mucosal toxicity. Our rationale for developing an experimental animal model to document drug-induced injuries in the gastric mucosa follows: (1) an animal model is ethical because serial endoscopic examinations over either a short or prolonged interval cannot be performed in humans; (2) the model allows the development of lesions to be monitored over a long period of time; and (3) large submucosal dissections can be done. The pig was chosen for its suitability as a model in terms of gastric anatomy, gastric size, and physiology, which are considered to be close to those of humans [@JR059-19]. Furthermore, in this model, tools similar to those used in humans can be used for the morphologic and functional assessment of mucosal lesions.
We were able to place the tablets at the chosen sites and, more importantly, to keep them in place without difficulty for 1.5 hours. This was likely facilitated by the deep anesthesia of the animals, which helped to inhibit gastric wall motility. In these conditions of prolonged tablet application, we anticipated the induction of maximal injury to the exposed area.
Based on the macroscopic appearances, our results show the presence of mucosal lesions after NaPT application. These observations confirm previous findings in humans, especially concerning fundic injuries and injuries in the colon [@JR059-11] [@JR059-12] [@JR059-14]. Although nonphysiologic and rather drastic experimental conditions were used, with direct and prolonged application of tablets onto a "dry" area of mucosa, we did not observe any grade 3 lesions. At 24 hours after the initial procedure, grade 1 and grade 2 lesions were observed in 3 of the 9 animals (33 %), whereas there were no lesions in the other 6 animals (66 %). In addition, 72 hours after the procedure, no lesion could be identified on the previously exposed gastric area. Thus, our design additionally documents the time course of injuries and their spontaneous reversibility. Observed injuries were reversible in less than 72 hours, which appears to be consistent with healing of the gastric mucosa following superficial lesions [@JR059-20].
The pCLE analysis showed that microscopic architectural disorganization occurred early, reflected by variation in fluorescein intensity in the crypt lumen as early as 1.5 hours after the application. The placebo tablet did not induce any significant changes in any of the parameters studied. By contrast, 24 hours after the application of PlaT and NaPT, significant changes were observed for all parameters except intercryptic intensity. However, the changes were smaller with PlaT than with NaPT. Overall, this finding suggests that early macroscopic mucosal changes could be due to active compound (see below), whereas later effects (24 hours) might involve mechanical as well as excipient components ([Fig. 5](#FI059-5){ref-type="fig"}). In fact, despite the observed pCLE changes, the functional consequences were probably minimal because (1) there was no change in mucosal permeability assessed with the Ussing chamber and (2) no morphologic change in mucosal thickness was noted.
The mechanisms and components involved in the occurrence of such lesions remain unknown. They are likely related to sodium monophosphate because similar observations have been reported with other compounds containing sodium phosphate [@JR059-11] [@JR059-12] [@JR059-14].
Despite the fact that this well-characterized animal model included a time course analysis, our study has some limitations. First, as always when an animal model is used, the data may not be considered to reflect human conditions strictly. In addition, the anesthetic conditions may represent a potential bias because the animals were maintained in one position, so that gastric motility was inhibited and changes in mucosal defense mechanisms or in gastric mucosal blood flow could not be formally excluded. However, despite its nonphysiologic character, this condition does allow the creation of maximal direct toxic effects on the mucosa. Finally, although endoscopic sessions were repeated in the animals, the relatively long sampling time period (48 hours between the 24 -- and 72-hour sessions) does not allow a precise definition of the time taken for complete mucosal healing.
Although no severe lesions were observed in this study, the methods presented here could be valuable for gaining pathophysiologic insight into the evolution of severe digestive mucosal lesions. For instance, the model could be used to test the potential additional effects of corticosteroid therapies, especially in combination with other drugs known to induce gastric injury, such as nonsteroidal anti-inflammatory drugs and aspirin or platelet inhibitors. Likewise, from a mechanistic point of view, it could be interesting to use such an experimental model to test for the effects of simultaneously administered gastroprotective agents.
In conclusion, this study documents, in particularly severe experimental conditions, the macroscopic and microscopic acute gastric injuries induced by NaPT, which are commonly used for colonoscopic bowel preparation in humans. It illustrates both the superficial nature and spontaneous reversibility of the lesions, with healing occurring in less than 72 hours. Although such experimental conditions are never observed in clinical practice, and no increased risk factor associated with any drug has been reported in these conditions, the results indicate that further studies are needed to document the potential consequences of gastric toxicity in patients with increased risk factors, such as those with a history of gastric ulcers or gastric hemorrhage.
Financial Support
=================
Declaration of funding interests
--------------------------------
This study was funded in part by Mayoly Spindler (experimental costs only) and by INSERM.
**Competing interests:** Emmanuel Coron has served as a speaker for Mayoly Spindler and Pentax and as a consultant for Mauna Kea Technologies. Stanislas Bruley des Varannes has served as a speaker for Mayoly Spindler, is a consultant for Alfa Wassermann and Given Imaging, and has received research funding from Mayoly Spindler.
| {
"pile_set_name": "PubMed Central"
} |
Introduction
============
One of the main goals of professional soccer clubs and their youth academies is to develop young, talented players into successful professional players ([@B27]). Many clubs and national associations (e.g., Germany, Belgium, Portugal, Netherlands) have created programs for talent identification and development (TID) to provide the best training environment and conditions for young players with noticeable potential; enrollment in these programs often starts during early adolescence ([@B15]; [@B28]; [@B33]). A player's success in a soccer match depends on complex multidimensional performance that is influenced by technical, tactical, physical, anthropometric, and mental factors ([@B48]; [@B19]). To ensure the highest possible efficiency in the TID process, performance tests of different soccer domains outside the context of the soccer match and notational analysis (observation and quantitative/qualitative analysis of the technical and tactical actions performed during a match) are often used in addition to coaches' expert and highly subjective assessment methods ([@B2]).
During the last two decades, technical-tactical skills and physical fitness in particular have been frequently explored and identified as key determinants of young players' game performance, serving as discriminants between elite, subelite and non-elite youth soccer players ([@B40]; [@B53]; [@B26]; [@B52]; [@B33]). In particular, technical skills such as dribbling the ball, passing, shooting and ball mastery are considered critical game rudiments ([@B47]) and have been recognized as important motor factors within TID programs ([@B55]; [@B15]). Previous research suggests that technical skills develop most rapidly during the prepubertal and pubertal (10--15 years) phases ([@B55]; [@B56]; [@B33]). Specifically, the test of speed dribbling is the best discriminator of performance levels among soccer players ([@B55]; [@B29]). Elite young players display significantly better performance in strength, speed, agility and aerobic/anaerobic endurance ([@B55]; [@B42]; [@B50]) than subelite young players. However, any connections between these findings and current game performance in youth players should be made with caution because differences in physical fitness and tactical skills are often caused by differences in the speed of biological maturation ([@B54]; [@B8]; [@B40]; [@B53]; [@B9]). Early maturing soccer players generally show higher levels of explosive performance, sprinting, agility and aerobic endurance ([@B18]; [@B56]; [@B50]). The relationship between biological maturation and the performance of technical skills is contradictory to the results of some studies confirming the influence of biological maturation status on the performance of technical skills tests ([@B46]; [@B50]) and other studies finding a lack of influence of biological maturation on the performance of technical skills ([@B18]; [@B57]).
Recently, several studies have emphasized the importance of motor coordination, i.e., non-specific motor coordination, in the process of TID in youth soccer ([@B57]; [@B12]; [@B13]; [@B14]; [@B50]). Furthermore, these studies showed that motor coordination is a significant long-term predictor of specific aerobic fitness and explosive leg power in young soccer players ([@B12]; [@B13]) and does not depend on biological maturation ([@B57]; [@B50]). However, the direct relationship between motor coordination and specific technical skills (e.g., speed dribbling) was not explored in prepubescent soccer players. In another study, [@B14] measured motor coordination performance among club players (playing in the two highest youth soccer leagues) and dropout players (those who dropped to lower soccer leagues) over the 8-year period from age 8 years to age 16 years and found that the club players performed significantly better than the dropout players on all motor coordination tasks and on aerobic endurance and speed. The authors suggested that motor coordination performance is essential for discriminating between players in a high-level training program and dropout players from the age of 9 years until late puberty. Although the direct relationship between motor coordination and specific technical skills was not investigated in this study, one could hypothesize that the dropped players had overall worse specific technical skills and worse motor coordination than club players.
In many studies focused on motor development, the term "motor coordination" has been used to denote motor competence, motor proficiency or fundamental motor skills (FMS) to describe goal-directed human movement ([@B49]). For the purpose of our study, we decided to use the term FMS to describe the level of general motor competence. In general, according to several key motor development theoretical models, FMS are frequently defined as the "elements" of more advanced complex movements required to participate in sports, games, or other context-specific physical activity ([@B7]; [@B21]). However, no clear research evidence indicates whether this theory is valid in prepubertal soccer players. Once FMS are mastered, the learning of sport-specific skills can occur more quickly and be more effective ([@B21]). FMS are traditionally divided into object control/ball/manipulative skills (e.g., throwing, catching, dribbling), locomotor skills (e.g., running, jumping, galloping), and balance/stability skills (e.g., non-locomotor skills such as body rolling, one-foot balance, stretching, twisting) ([@B21]). Although children have the developmental potential to master most FMS by the age of 6 years ([@B21]), recent research highlights that children and adolescent youth do not perform FMS to their expected developmental capabilities ([@B45]). [@B45] further demonstrated that while levels of FMS vary by country, performance levels remain consistently low, with the majority of children and adolescents failing to surpass 50% mastery in most skills.
To our knowledge, little attention has been paid to the importance of FMS in the process of acquiring technical skills (e.g., dribbling, receiving, or passing a ball) in prepubescent soccer players. Moreover, current research describes the direct relationships between technical skills and other physical, motor control, or morphological factors but have not described how those factors interact with or mediate specific soccer skills. Although there is clear evidence concerning the relationships between FMS, physical fitness and biological maturation, there is a lack of information about the influence of FMS on the performance of soccer technical skills in prepubescent-aged players. We hypothesized that FMS strengthen the influence of physical fitness and biological maturation on technical skills (e.g., speed dribbling the ball). Therefore, the aim of this study was to investigate the role of FMS in the relationships between physical fitness, biological maturation and technical skills in prepubescent soccer players.
Materials and Methods {#s1}
=====================
Methodological Approach
-----------------------
Cross-sectional measurement was performed during the competitive part of the soccer season. The participants were familiarized with the experimental protocol 1 week prior to the experiment and did not perform any exhausting activity 72 h before the experiment. After the participants' body mass (BM) was estimated, they performed the battery of FMS, speed dribbling and physical fitness tests within 1 day (two training sessions). The FMS and sit-ups (part of Unifittest 6--60) tests were performed indoors on a teraflex surface during the morning training session between 9 and 11 am. The rest of physical fitness tests were then conducted during afternoon training session on the outdoor ground with artificial grass between 3 and 5 pm.
Participants
------------
The research sample consisted of forty U12 soccer players (mean ± SD; age 11.5 ± 0.3 years; height 145 ± 7.0 cm; body mass 37.2 ± 4.1 kg). The players were members of teams from two clubs in the Prague district of the Czechia that participated in the highest Czech youth league level. These two clubs were randomly selected from a basic sample (a total of fourteen clubs in the Prague district) and then were asked to participate in the study. The weekly cycle consisted of four training sessions (7--8 h) focused primarily on technical-tactical skills during exercises and games and one competitive match. The inclusion criteria were a minimum of 6.4 years of experience with organized soccer and full attendance in ongoing habitual training cycles. Exclusion criteria were any medical problems that compromised participation or performance in the study, such as soft tissue injury, delayed muscle soreness, recent illness or recent recovery from injury. The research was approved by the Ethics Committee of the Faculty of Physical Education and Sport, Charles University, and all participants and their parents signed an informed consent form.
Fundamental Motor Skills
------------------------
The Bruininks-Oseretsky Test -- 2nd edition (BOT-2; short version) was used to assess fundamental fine and gross motor skills ([@B5]). The BOT-2 has demonstrated high inter-rater reliability (*r* ≥ 0.90), test-retest reliability (*r* ≥ 0.80) and construct validity ([@B11]). The short version contains sixteen items divided into eight dimensions (see [Table 1](#T1){ref-type="table"}). Raw scores from BOT-2 were transformed into standard scores according to age by ASSIST software (MN, United States). Standard scores were then used for the final analysis.
######
List of dimensions and items of the BOT-2 motor test.
Fine motor precision Balance
---------------------------------------------------- ----------------------------------------------------
Drawing lines through crooked paths Walking forward on a line
Folding paper Standing on one leg on a balance beam -- eyes open
**Fine motor integration** **Running speed and agility**
Copying a square One-legged stationary hop
Copying a circle **Upper limb coordination**
Copying a star Dropping and catching a ball -- both hands
Copying a pencil Dribbling a ball -- alternating hands
**Manual dexterity** **Strength**
Transferring a penny Full push-ups
**Bilateral coordination** Sit-ups -- 30 s
Jumping in place -- same side synchronized
Tapping feet and fingers -- same side synchronized
Physical Fitness Tests
----------------------
Three physical fitness parameters were measured (shuttle run 4 m × 10 m, standing broad jump, and 20-m progressive shuttle run). These three tests are included in the Unifittest 6--60 test battery, which is standardized for the Czech context ([@B38]; [@B6]) with a satisfactory level of reliability and validity ([@B38]). *Shuttle run* 4 m × 10 m, which assesses coordination and speed, was performed twice by each player, with 3--4 min of rest between the two trials. In a start position, the player stood on the starting line without moving into the space between photocells. The player sprinted to the opposite marker (10 m), turned and returned to the starting line directly adjacent to the photocell gate. This was performed twice to cover a 40-m distance. The time of the faster trial was recorded. An infrared timing gate (Alge Timing GmbH, Lustenau, Austria) placed at approximately hip height was used for the start and finish points. *Standing broad jump*, an indicator of explosive power in the lower limbs, was performed three times by each player, with 2 min of rest between trials. The player stood behind a line marked on the ground. A two-foot takeoff and landing area was used, and players were instructed to jump as far as possible while swinging their arms and bending their knees to provide forward momentum. The longest jump was recorded and used for the analysis. *Progressive shuttle run* 20 m is a measure of maximal aerobic fitness. The player continuously ran between two lines 20 m apart, keeping pace with recorded beeps, which accelerated each minute. The test was stopped when the player failed to reach the line (within two meters) after two consecutive warnings. Finally, from each test item, a standard score was obtained. The composite score of all tests on a scale from 0 to 20 was calculated as a marker of physical fitness.
Predicted Maturity Offset
-------------------------
Maturity offset was estimated according to [@B41] equations. Although these equations have been widely used in the sport environment (e.g., [@B59]; [@B22]; [@B23]; [@B39]), studies have pointed to the limits of this predictive method in both sexes ([@B36],[@B37]; [@B35]). In the current study, we considered the finding of [@B32], who stated that maturity offset predicted from the Mirwald equations matched the observed peak high velocity (PHV) in 12-year-old boys, to be important.
Y-PHV = −9.326 + (length of lower limbs ^∗^ sitting height) -- (0.001663 ^∗^ \[decimal age ^∗^ length of lower limbs\]) + (0.007216 ^∗^ \[decimal age ^∗^ sitting height\]) + 0.02292 ^∗^ \[weight/height\]
Body weight was assessed with an accuracy of 0.1 kg using medical calibrated weight (CAS DBI-C, Lesak and Zemánek s.r.o., Czechia). A portable anthropometer (A 226, Trystom, spol. s.r.o., Czechia) with a balancing point to determine the right vertical position of the anthropometer was used for the measurement of height and sitting height (0.1 cm).
Speed Dribbling Test
--------------------
The short dribbling test (SDT) ([@B3]; [Figure 1](#F1){ref-type="fig"}) is a test of dribbling the ball at a high speed with changes in direction around a defined track. Players are required to dribble as fast as possible around the cones without touching them. The test is finished by stopping the ball in a square (defined with blue cones) at the end of the track. Each player underwent one training and one competitive trial. If the player touched any cone during dribbling, the trial failed, and the player was allowed an additional trial. Time was measured by a telemetric photocell system (Alge Timing GmbH, Lustenau, Austria), and the best time was recorded. Similar agility dribbling tests have shown acceptable reliability, with an intraclass correlation coefficient between 0.78 and 0.89 and coefficient of variation of 2.4 and 3.9 ([@B51]; [@B1]; [@B10]).
![Short dribbling test ([@B3]).](fphys-10-00596-g001){#F1}
Data Analysis
-------------
Multiple regression path analysis (MRPA) was used to test the hypothesized links, with successive multiple regression equations calculated to estimate path coefficients. Mardia's, Henze-Zirkler's and Royston's multivariate normality tests were performed using the R package MVN, version 4.0.2, in R 3.4.1, with a cut-off of p greater than 0.05 for accepting the normality of multivariate data. The exogenous independent variables were physical fitness and maturity offset. The interacting endogenous variable was FMS. All of these variables were analyzed first using linear regressions and then using multiple regressions. Subsequently, the final path analysis model was selected. In the path model, there were specified direct paths from the exogenous variable to the endogenous variable and from the exogenous and endogenous variables to the SDT. Finally, for variables that had statistically significant predictive power (*p* \< 0.05) for FMS or for SDT performance, specific indirect effects via FMS were investigated. MRPA and Pearson's correlations were performed using M-plus software version 6.0 ([@B43]). All data can be found in the "[Supplementary Table S1](#SM1){ref-type="supplementary-material"}" [Supplementary File](#SM1){ref-type="supplementary-material"}.
Results
=======
The means and standard deviations of the basic descriptive statistics and correlation coefficients of FMS, physical fitness, maturity offset and speed dribbling are shown in [Tables 2](#T2){ref-type="table"}, [3](#T3){ref-type="table"}, respectively. Significant moderate associations were observed between speed dribbling and FMS, speed dribbling and physical fitness, and FMS and physical fitness. However, there was no association between speed dribbling and maturity offset.
######
Basic descriptive statistics (*n* = 40).
Mean SD Median Interquartile range 95% confidence interval
------------------------- -------- ------ -------- --------------------- -------------------------
Age (years) 11.50 0.30 11.63 0.42 ±0.09
Height (cm) 145.00 7.00 149.45 6.53 ±1.98
Body mass (kg) 37.20 4.10 37.15 7.85 ±2.05
Index BMI (kg/m^2^) 17.52 1.89 17.06 2.08 ±0.59
FMS (ss) 57.33 8.88 58.50 17.25 ±2.75
Physical fitness (ss) 21.05 3.34 21.00 4.50 ±1.03
Maturity offset (years) −2.88 0.30 −2.90 0.34 ±0.09
Speed dribbling (s) 13.68 1.53 13.73 2.46 ±0.47
FMS, fundamental motor skills; SD, standard deviation; ss, standard score.
######
Correlation matrix of the study variables.
Maturity offset Physical fitness FMS
---------------------- ----------------- ------------------ -----------
**Maturity offset** 1
**Physical fitness** −0.21^∗∗^ 1
**FMS** −0.29^∗∗^ 0.50^∗∗^ 1
**Speed dribbling** −0.03 −0.42^∗∗^ −0.60^∗∗^
FMS, fundamental motor skills;
∗∗
indicates statistical significance of p \< 0.01.
In the first step, we analyzed the predictive power of FMS, physical fitness and maturity offset (independent variables) on speed dribbling performance (dependent variable). The linear regression results ([Table 4](#T4){ref-type="table"}) showed that FMS and physical fitness are significant predictors of speed dribbling performance. Nevertheless, only the effect of FMS (*R*^2^ = 0.36; *t* = 2.97; *p* = 0.003) was significant, while the effect of physical fitness was not (*R*^2^ = 0.18; *t* = 1.64; *p* = 0.100).
######
Linear regressions of physical fitness, FMS and maturity offset on speed dribbling.
Independent variable *B* *SE B* β *t* *p*-Value
---------------------- ------- -------- ------- ------ -------------
**Physical fitness** −0.93 0.31 −0.43 3.28 0.001^∗∗^
**FMS** −3.5 0.74 −0.60 5.95 \<0.001^∗∗^
**Maturity offset** −0.01 0.03 −0.03 0.19 0.85
FMS, fundamental motor skills;
∗∗
indicates statistical significance of p \< 0.01; B, unstandardized beta; SE B, standard error for the unstandardized beta; β, the standardized beta; t, t-test statistic.
Since clear evidence of the relationship between biological age and physical fitness in pubescent soccer players has been reported in previous research (biologically advanced players achieve better performance in physical fitness), we verified whether FMS and maturity offset are significant predictors of the level of physical fitness. The multiple regression model showed that the effects of FMS and maturity offset explain 22% of physical fitness performance variability (*R*^2^ = 0.22). Furthermore, from [Table 5](#T5){ref-type="table"}, it is clear that FMS are significantly better predictors for physical fitness than maturity offset in prepubescent players.
######
Multiple regression of FMS and maturity offset on physical fitness.
Independent variable *B* *SE B* β *t* *p*-Value
---------------------- ------- -------- ------ ------ -----------
**FMS** −0.82 0.06 0.48 3.25 0.003^∗∗^
**Maturity offset** 0.18 1.64 0.08 0.50 0.62
**Adjusted *R*^2^** 0.22
FMS, fundamental motor skills;
∗∗
indicates statistical significance of p \< 0.01; B, unstandardized beta; SE B, standard error for the unstandardized beta; β, the standardized beta; t, t-test statistic.
Considering these findings, we decided to use the analysis model where FMS plays the role of mediator between physical fitness (as the independent variable) and speed dribbling (as the dependent variable representing specific soccer skills). In the 1st path analysis model, we specified both direct and indirect paths between physical fitness and speed dribbling. The 1st path analysis model ([Figure 2](#F2){ref-type="fig"}) showed that the direct effect of physical fitness on speed dribbling was non-significant (standard estimation = −0.17, *p* = 0.247), and the indirect effect through FMS was significant (standard estimation = −0.26, *p* = 0.005). To obtain better results, we decided to formulate the 2nd path analysis model without a direct effect ([Figure 3](#F3){ref-type="fig"}). The 2nd model was approved as significant and acceptable with empirical data explaining more than 25% of the model. Generally, FMS were a significant mediator between physical fitness and speed dribbling (standard estimation = −0.31, *p* = 0.001).
![First path analysis model with direct and indirect effects of physical fitness and FMS on speed dribbling \[Satorra--Bentler χ^2^ (*df* = 0) = 0; *p* = 0.00; RMSEA = 0.0; SRMR = 0.0; CFI = 0.0; TLI = 0.0\]; FMS, fundamental motor skills. ^∗∗^*p* \< 0.01; ^∗∗∗^*p* \< 0.001.](fphys-10-00596-g002){#F2}
![Second path analysis model with only the indirect effect of physical fitness on speed dribbling with FMS as the mediator variable \[Satorra--Bentler χ^2^ (*df* = 1) = 1.3; *p* = 0.254; RMSEA = 0.08; SRMR = 0.04; CFI = 0.99; TLI = 0.97\]; FMS, fundamental motor skills. ^∗∗^*p* \< 0.01; ^∗∗∗^*p* \< 0.001.](fphys-10-00596-g003){#F3}
Discussion
==========
The present study examined the possible role of FMS in the relationships between physical fitness and biological maturation and speed dribbling as a soccer-specific soccer skill in young soccer players. We found that FMS were a significant mediator of the relationship between physical fitness and speed dribbling. Notably, biological maturation did not prove to be a significant contributor to speed dribbling performance through FMS. Despite moderate correlations between physical fitness and speed dribbling, the path model did not reveal a direct influence of physical fitness or biological maturation on speed dribbling. These findings suggest the need for a certain level of FMS (fine and gross motor skills) to acquire soccer-specific motor skills. Generally, both quantitative (physical fitness) and qualitative (FMS) motor aspects were found to be significant contributors to the performance of soccer-specific motor skills, represented by speed dribbling. Our path model revealed that FMS, physical fitness and other related factors have a prior effect on specific skill, whereas biological maturation might explain only 8.7% of motor coordination ([@B20]).
Our results suggest that FMS mastery significantly increases the influence of physical fitness on the performance of soccer-specific skills in young players. These findings are in accordance with [@B7] statement that FMS are basic elements for later skillfulness in a range of sport and game domains. Our results cannot be compared with similar data measured on soccer players. However, similar conclusions have been found in research involving combat sports ([@B4]), where high correlations were found between specific karate skills and FMS (*r* = 0.74) in 5- to 7-year-old members of karate clubs. This study suggested that children with higher FMS also have better karate techniques, while others have difficulties acquiring these techniques. Unfortunately, recent research has documented very poor or insufficient FMS performance in preschool and school-aged children ([@B44]; [@B16]; [@B25]; [@B30],[@B31]) combined with generally unresolved inactivity in children ([@B17]), which may result in impaired acquisition of more complex and difficult sport skills or delays in mastering the required skills. Similarly, the players in our study showed only an average level of FMS even though they were considered to be capable of high performance. Since a higher level of FMS and soccer-specific skills were found in players selected for Belgian professional clubs ([@B14]) than in "dropout" players, FMS and soccer-specific skills seem to be crucial to the identification of gifted players and their likelihood of remaining in high-level talent development programs.
Our participants were at a specific age (U12) where physical development plateaus (in reactive strength and jumping) with a change in the mechanical properties of the lower limb (decreased relative leg stiffness), which has been previously observed ([@B34]). This might explain why biological maturation in U12 children is not strongly related to physical fitness or motor control testing ([@B20]) and why separate values of physical fitness and biological maturation are insufficient for predicting a player's ability to acquire soccer-specific skills. Our path models suggest that the best performance of soccer-specific skills will occur in players with adequate levels of both FMS and physical fitness. This finding is in agreement with the finding that FMS were a long-term predictor of explosive power in soccer players from childhood to young adulthood ([@B13]). Therefore, we believe that well-developed FMS and the simultaneous development of PF are necessary in pre-PHV boys and that well-coordinated players will improve in power and performance with age due to increased tendon stiffness in late adolescence ([@B13]).
The harmony between physical fitness and biological maturation has been highlighted by several authors (e.g., [@B40]; [@B53]; [@B57]) to discriminate elite, subelite and non-elite soccer players during talent identification. Moreover, FMS were found to be a long-term predictor of soccer-specific aerobic performance in elite pubertal soccer players ([@B12]) and children ([@B24]). Specifically, children with low FMS performed worse on all fitness tests (50-m run, standing broad jump and endurance shuttle run), where endurance shuttle test differences increased between low- and high-FMS groups over 5 years ([@B24]). Therefore, we highlight the importance of FMS not only for soccer-specific motor skills but also for separate components of physical fitness, such as explosive power and aerobic endurance, during long-term motor development.
Several possible limitations associated with this study should be noted. The present study utilized a cross-sectional design; thus, the role of FMS in the relationships between physical fitness, biological maturation and soccer-specific motor skills should be interpreted with caution. A longitudinal follow-up of young soccer players, especially during the pubertal phase (aged 12--15 years), may provide a more accurate explanation of this mediation effect. Another possible limitation is related to the non-inclusion of psychological variables such as motivation or self-confidence, which certainly influence game performance and likely also affect the development of new soccer-specific motor skills. Lastly, although several authors consider speed dribbling the most valid soccer skill test, the inclusion of additional soccer-specific skills (e.g., passing, shooting, receiving,or multifaceted tests) could elucidate the role of FMS in the acquisition of soccer-specific motor skills ([@B58]). Therefore, future research should focus on (1) performing longitudinal research to verify the role of FMS in acquiring soccer-specific skills during the pubertal phase, (2) testing more soccer-specific skills and psychological variables with respect to players' positions, and (3) using a process-oriented (assessment of movement quality) test for FMS assessment.
Conclusion
==========
This is the first study to evaluate the role of FMS in a complex theoretical model with the relationships between physical fitness, biological maturation and soccer-specific motor skill (measured using speed dribbling) in young soccer players. Our results showed that FMS significantly strengthened the influence of physical fitness on the performance of speed dribbling, a soccer-specific motor skill, and thus play an important role in the process of the acquisition of sport-specific motor skills in prepubertal elite soccer players. Conversely, physical fitness and biological maturation alone did not significantly influence speed dribbling performance. Generally, it appears that developing and improving a wide range of basic FMS as building blocks for more complex and more difficult soccer-specific motor skills is necessary during the long-term training process. Based on these findings, FMS could be included in TID programs for young elite soccer players, especially during childhood and before puberty. Thus, it is recommended that youth soccer coaches and practitioners carefully consider providing training on FMS (fine motor, locomotor, object control, balance), especially during childhood, with an emphasis on the quality of movements.
Data Availability
=================
All datasets generated for this study are included in the manuscript and/or the [Supplementary Files](#SM1){ref-type="supplementary-material"}.
Ethics Statement
================
This study was carried out in accordance with the recommendations of "name of guidelines, name of committee" with written informed consent from all subjects. All subjects gave written informed consent in accordance with the Declaration of Helsinki. The protocol was approved by the "name of committee."
Author Contributions
====================
JK, MM, and PS involved in the conceptualization of the study design and the drafting of the manuscript. JK and MM involved in data collection. JK involved in performing an overview of the previous research. MM involved in conducting the statistical analysis. PW and EM-C helped with the data assessment and interpretation.
Conflict of Interest Statement
==============================
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
**Funding.** This study was supported by the project UNCE No. 032, GA19-12150S and the Progress Q19 Social-Sciences Aspects of Human Movement Studies II program.
The authors would like to thank the participating players and their coaches (particularly Mr. Ales Vytlacil and Mr. Oldrich Smerda) from AC Sparta Prague for their effort and support during the research activities connected with this study.
Supplementary Material
======================
The Supplementary Material for this article can be found online at: <https://www.frontiersin.org/articles/10.3389/fphys.2019.00596/full#supplementary-material>
######
Click here for additional data file.
[^1]: Edited by: Filipe Manuel Clemente, Polytechnic Institute of Viana do Castelo, Portugal
[^2]: Reviewed by: Mustafa Söğüt, Kırıkkale University, Turkey; Ricardo Franco Lima, Polytechnic Institute of Viana do Castelo, Portugal
[^3]: This article was submitted to Exercise Physiology, a section of the journal Frontiers in Physiology
| {
"pile_set_name": "PubMed Central"
} |
INTRODUCTION
============
It is well known that medical students, interns, and residents experience higher rates of depression than the general population.^[@R1]^ Mata et al.^[@R2]^ recently conducted a systematic review and meta-analysis and calculated the mean prevalence of depression among residents to be 28.8% (range, 20.9--43.2%). Most of the reported studies investigated the prevalence of depression among internal medicine residents or combined medical and surgical residents in their study populations. Depression among plastic surgery residents has not been specifically investigated.
In this article, we report on a cross-sectional study investigating the prevalence of depression among plastic surgery residents enrolled in the Saudi Plastic Surgery Program.
MATERIALS AND METHODS
=====================
An institutional review board approval was obtained from King Saud University (Project \#E17-2293). This is a cross-sectional e-mail survey, and all plastic surgery residents enrolled in the Saudi Plastic Surgery Program in Riyadh (Saudi Arabia) were included (n = 39). The Beck Depression Inventory II (BDI-II) was used for data collection.^[@R3]^ The total score is 63. Mild, moderate, and severe depression is diagnosed when the score is 10--19, 20--29, and 30--63, respectively. We also added to the test questionnaire questions regarding level of residency, number of hours/week, number of night duties per month, and whether or not the resident was allowed to take the day off following the night duty. The test questionnaire was formulated using the Survey Monkey Web site ad sent by e-mail to all residents. The survey was conducted during the second half of academic year (April 2017) to avoid the stress associated with entering a new level of training.
RESULTS
=======
Of the 39 surveys, 34 were returned, giving a response rate of 87%. There were 19 men and 15 women. Only 32% of the residents were married. The majority of residents (73%) were below the age of 30 years. More than half of the residents (56%) reported working more than 60 hours per week, and 44% reported having more than 7 night duties per month. Only 11% of residents were all allowed to take the day off following the nigh duty.
The Depression Inventory showed that mild, moderate, and severe depressive symptoms were prevalent in 20.6%, 38.2%, and 11.8%, respectively. This meant that 70.6% of the residents showed depression symptoms.
We compared the rate of depression to the demographic data and residency level. In total, 56% of the depressed residents were male, and this was similar to the percentage of male participants of the survey (56% of participants were male); 34.8% of the depressed residents were married, and that was also similar to the percentage of the marital status of participants of the survey (32% of participants were married). The Saudi Plastic Surgery Program is a 6-year program (R1-R6), and Figure [1](#F1){ref-type="fig"} shows the prevalence of depression among R1-R6 residents. The chart shows a trend of lower rates of depression in senior (R5, R6) residency.
{#F1}
DISCUSSION
==========
After an extensive review of the literature, we could not find any studies that specifically investigated the prevalence of depression among plastic surgery residents. The overall prevalence of depression in our study (70.6%) is alarming and is much higher than the mean prevalence of 28.8% reported by Mata et al.^[@R2]^ in their recent systematic review. Further studies are required to compare our prevalence to other international plastic surgery residency programs.
The high prevalence rate on our study and the lack of comparative studies required another literature review on the factors associated with depression among residents, and we have summarized these causes in Table [1](#T1){ref-type="table"}. It is reasonable to believe that surgical residents are under more stress than medical residents. The systematic review of Mata et al.^[@R2]^ and our review of the literature did not find any articles comparing the prevalence of depression in surgical versus medical residents. However, 1 study^[@R4]^ showed that the prevalence of "burnout" is much higher among surgical when compared with medical residents (58% versus 27%). "Burning out" is known to be associated with insomnia, suicidal ideation, and depression.^[@R4],[@R5]^ We did not specifically investigate burnout in our study, which is measured using other inventories such as the Maslach Burnout Inventory.^[@R5]^ Long working hours (over 60--70 h/wk) are also known to be associated with high rates of depression, burning out, errors of prescribing wrong medication doses, and other medical errors during residency.^[@R6],[@R7]^ The stress of junior residency^[@R4]^ and sleep deprivation^[@R8]^ are also associated with "depressive symptoms" (Table [1](#T1){ref-type="table"}).
######
Factors Associated with Depression among Residents In-Training
A high prevalence of depression is also associated with specific personalities and characters^[@R9]^ and in residents with emotional exhaustion.^[@R10]^ We did not investigate temperament and character of our residents, and such assessment requires other inventories such as the Temperament and Character Inventory. One study showed that residents with high "harm avoidance" and low "self-directedness" are more prone for depression.^[@R9]^
An American Foundation for Suicide Prevention evaluated physician depression and suicide tendency and concluded that lack of support and rough institutional policies were important predisposing factors.^[@R11]^ Two studies^[@R6],[@R8]^ suggested that the female gender is a risk factor for depressive symptoms during residency. However, most other studies (and our study as well) did not find that gender is a factor.
Several studies^[@R1]--[@R11]^ recommended preventive methods to decrease the prevalence rate of depression among residents such as reducing the number of working hours, giving the post-call day off, and calling for a shift in institutional policies to support residents. Furthermore, treatment is indicated in residents with depressive symptoms. An American Council of Review Committee of residents and fellows conducted an inquiry on resident's wellness and depression.^[@R12]^ The group recommended to increase the awareness of stress and depression during residency. They also recommended the promotion of a supportive culture and other measures to be taken by all stakeholders of postgraduate medical education.^[@R12]^ Screening residents in-training for depressive symptoms may not be practical. However, "facilitated discussion group intervention" has been found effective to reduce the prevalence of burnout and depression among high risk junior residents.^[@R13]^ Residents with moderate-to-severe depression require formal psychiatric assessment and anti-depression medications. An alarming finding was reported by Stoesser and Cobb^[@R14]^ who investigated depression among residents at the University of Utah: only 27% of residents with moderate to severe depression were receiving medications for their depression.
Our study has 2 main limitations: the small number and the use of self-data. Similar studies in other plastic surgery residency programs are also required for comparison.
The Saudi plastic surgery training program has unique characteristics that may raise the stress level among residents. The program is the only postgraduate plastic surgery program in the country. Of the several hundred candidates applying every year, only 6--7 residents are accepted. The program is under the supervision of the "Saudi Commission for Heath Specialties." The promotion to a subsequent residency level (i.e., R1 to R2, R2 to R3, and so on) depends on 2 factors: the yearly evaluation (from the clinical rotations) and passing the yearly examination (the pass mark is 60%). The program's policies do not allow the promotion of a resident if he/she fails the examination, even if the yearly evaluation is satisfactory, or if the program director is happy with the resident's attitude, clinical performance, and technical skills. This yearly examination "phobia" may increase stress levels, especially among junior residents. The yearly examination is uniform for all residents, regardless of their level of training. Examination questions include the entire curriculum of plastic surgery. Hence, it is relatively more difficult for junior residents. The program director (1 of the co-authors) has supervised the study, and suggestions are being made to make 2 sets of examinations, 1 for junior and 1 for senior residents. Another suggestion that is being made is psychiatric assessment of the residents and treatment of depression if required. We also plan to repeat the study in a few years, after the implementation of these suggestions.
In conclusion, the high prevalence rate of depressive symptoms among plastic surgery residents is alarming, and further studies are warranted. The problem has received no attention in the plastic surgery literature.
Supported by the College of Medicine Research Center, Deanship of Scientific Research, King Saud University, Riyadh, Saudi Arabia.
**Disclosure:** The authors have no financial interest to declare in relation to the content of this article. The Article Processing Charge was paid for by the authors.
| {
"pile_set_name": "PubMed Central"
} |
Dobbelstein M, Levine AJ. Mdm2: Open questions. Cancer Sci. 2020;111:2203--2211. 10.1111/cas.14433
1. INTRODUCTION {#cas14433-sec-0001}
===============
The Mdm2 oncoprotein and its association with p53 were discovered 30 years ago,[^1^](#cas14433-bib-0001){ref-type="ref"}, [^2^](#cas14433-bib-0002){ref-type="ref"}, [^3^](#cas14433-bib-0003){ref-type="ref"} and a cornucopia of activities and regulatory pathways have been associated with it. A number of reviews summarize what we believe to know about Mdm2 and its paralogue and association partner Mdm4.[^4^](#cas14433-bib-0004){ref-type="ref"}, [^5^](#cas14433-bib-0005){ref-type="ref"}, [^6^](#cas14433-bib-0006){ref-type="ref"}, [^7^](#cas14433-bib-0007){ref-type="ref"}, [^8^](#cas14433-bib-0008){ref-type="ref"}, [^9^](#cas14433-bib-0009){ref-type="ref"}, [^10^](#cas14433-bib-0010){ref-type="ref"} However, open questions remain, similar to the white spots labeled "terra incognita", Latin for "unknown territory", on ancient geographical maps or globes. Such unexplored areas were alternatively labeled "hic sunt dracones" (here are the dragons) to indicate the risks and uncertainties for the traveler, and this metaphor could apply to the unexplored features of Mdm2 as well. In this review, we will raise *questions about Mdm2 and Mdm4* that we consider worth pursuing in future research, reaching from molecular structures and intracellular activities all the way to development, evolution, and cancer therapy. We anticipate that such research will not only close a few gaps in our knowledge but could add new dimensions to our current view.
2. HOW EXACTLY DOES MDM2 ACT ON P53? {#cas14433-sec-0002}
====================================
The current view emphasizes that Mdm2 forms a complex with p53 and mediates its ubiquitination, followed by proteasomal degradation. However, even this standard summary about Mdm2 leaves open questions such as the following.
*What is the exact structure of the complex formed between Mdm2 and a p53 tetramer?* The fact that the full‐length proteins have never been crystallized, neither alone nor in a complex, makes it difficult to answer this question. However, the advent of cryo‐electron microscopy could enable progress in this field. These types of data with the p53‐DNA complex (but without Mdm2) were reported already.[^11^](#cas14433-bib-0011){ref-type="ref"}
*Are the two aminoterminal domains of both proteins representing the only relevant interaction surface? How would the other domains of Mdm2 and p53 fold in relation to each other? What about the dynamics of the complex -- does it "breathe" to carry out the transfer to ubiquitin onto various sites on p53? And what is the structural difference between the p53‐Mdm2 complex and the TAp73‐Mdm2 complex?* -- the latter forming with high efficiency but without detectable destabilization of TAp73.[^12^](#cas14433-bib-0012){ref-type="ref"}, [^13^](#cas14433-bib-0013){ref-type="ref"}
*What is the effect of additional binding partners on the structure of Mdm2 and the complex with p53?* The complex of Mdm2 and Mdm4 is held together, at least in part, through the RING finger domains of both proteins,[^14^](#cas14433-bib-0014){ref-type="ref"} and this association can be separated from the ubiquitin ligase activity of Mdm2,[^15^](#cas14433-bib-0015){ref-type="ref"} but again, it is subject to ongoing research how the other domains are positioned within the complex of the Mdm2/Mdm4 heterodimer and the Mdm2/Mdm2 homodimer. Additional partners include but are certainly not limited to E2 ligases, p14ARF, and the ribosomal L5/L11/5S‐RNA complex. The structures of these complexes remain to be determined, including functional consequences. p53‐Bound DNA might well reshape the p53‐Mdm2 complex, and the same notion holds for chromatin‐associated binding partners of Mdm2, such as members of the polycomb repressor complexes.[^16^](#cas14433-bib-0016){ref-type="ref"}, [^17^](#cas14433-bib-0017){ref-type="ref"}, [^18^](#cas14433-bib-0018){ref-type="ref"}
Taken together, despite our knowledge on single domains within p53 and Mdm2, we are far from understanding the higher order structures of full‐length proteins and their multiple complexes resulting in alternative functions and transcriptional patterns.
3. IS THERE A MODIFICATION CODE FOR MDM2/MDM4 AND P53, DEPENDING ON THE KIND OF CELLULAR STRESS AND THE DESIRED RESPONSE? {#cas14433-sec-0003}
=========================================================================================================================
Numerous posttranslational modifications were identified on Mdm2[^19^](#cas14433-bib-0019){ref-type="ref"} and Mdm4[^20^](#cas14433-bib-0020){ref-type="ref"} as well as p53.[^21^](#cas14433-bib-0021){ref-type="ref"} Many (though not all) of these modifications enhance p53 activity and diminish the ability of Mdm2 to bind and degrade p53. The phosphorylations by AKT[^22^](#cas14433-bib-0022){ref-type="ref"}, [^23^](#cas14433-bib-0023){ref-type="ref"}, [^24^](#cas14433-bib-0024){ref-type="ref"} and by ATM[^25^](#cas14433-bib-0025){ref-type="ref"} (on different residues and with partially opposing effects) are only the most prominent examples.
The function of p53 and Mdm2/Mdm4 is to receive information (largely through the Mdm2 protein) about intrinsic and extrinsic cellular stresses and respond (through the p53 protein) by selective programs of transcriptional activation that either repair the damage produced by the stress and restore homeostasis, or kill the cell, eliminating the consequences of the damage. There are at least 10 stress signals (recognized by stress identifiers and transmitters) that act by inhibiting Mdm2 levels or activity and increase p53 levels, and at least 4 stress signals that act to increase Mdm2 levels or activity and decrease p53 transcriptional functions. Signaling to Mdm2, as well as the p53 responses, are accomplished by either protein modifications, eg phosphorylation, acetylation, methylation, ubiquitination, and sumoylation, protein‐protein interactions, or RNA‐protein interactions. P53 enhances the fidelity of cellular growth, replication, and division. It not only responds to a stress signal, but when multiple stress signals perturb cell division, the p53 protein modifications and interactions integrate the stresses that are to be responded to and program responses accordingly. As such the Mdm2‐p53 node integrates many diverse functional signal transduction pathways and as such that node is highly connected to a large amount of information that mediates cellular responses.
These considerations at least suggest that there is a code of modifications on both Mdm2/Mdm4 and p53, reflecting the stress input and the biological effects, such as cell cycle arrest, senescence, or cell death of different kinds. Such a code, if it exists, might well depend on the cell type and signaling activities. The future challenge will consist of an integrated understanding of *how combinations of modifications on Mdm2 and Mdm4 are achieved in a stress‐specific way, and how this will affect the p53‐driven response*.
4. HOW DOES COMPARTMENTALIZATION OF MDM2 AND MDM4 AFFECT THEIR FUNCTIONS? {#cas14433-sec-0004}
=========================================================================
Both Mdm2 and Mdm4 can adopt diverse intracellular localizations, and these are subject to dynamic changes. We have reported that Mdm2 shuttles between the nucleus and the cytoplasm, through specific import and export signal sequences.[^26^](#cas14433-bib-0026){ref-type="ref"} Similar findings were also reported for p53,[^27^](#cas14433-bib-0027){ref-type="ref"} and the transport of each binding partner can alter the p53 response.[^28^](#cas14433-bib-0028){ref-type="ref"} Moreover, Mdm2 undergoes relocalization when associating with binding partners. For instance, p14ARF is capable of relocalizing Mdm2 to nucleoli,[^29^](#cas14433-bib-0029){ref-type="ref"}, [^30^](#cas14433-bib-0030){ref-type="ref"} whereas the acetyl transferase KAT5/Tip60 takes it to promyelocytic leukemia protein (PML) nuclear bodies.[^31^](#cas14433-bib-0031){ref-type="ref"} More recently, it turned out that Mdm4 predominantly localizes to the cytoplasm in the absence of Mdm2 but travels to the nucleus in the presence of it.[^32^](#cas14433-bib-0032){ref-type="ref"} Mdm4 localization can further be regulated by phosphorylation through AKT and Chk1, each leading to its association with 14‐3‐3 proteins.[^33^](#cas14433-bib-0033){ref-type="ref"}, [^34^](#cas14433-bib-0034){ref-type="ref"} It should be noted, however, that most of these experiments were carried out by overexpressing Mdm2/Mdm4, making it even more important to address the precise compartmentalization of the endogenous proteins, their changes in cellular stress situations, and their impact on p53 activities and cellular responses.
5. HOW DOES MDM2 AFFECT CELL FATE, INDEPENDENT OF P53? {#cas14433-sec-0005}
======================================================
Although Mdm2 is widely known as a negative regulator of p53, a variety of additional functions have been reported. Most of them were observed by overexpressing Mdm2 from transfected plasmids, raising the uncertainty of their physiological relevance. Still, these activities point out that Mdm2 is at least capable of doing much more than counteracting p53. For instance, overexpressed Mdm2 *hinders* cell cycle progression in the majority of cell types analyzed,[^35^](#cas14433-bib-0035){ref-type="ref"} at least suggesting that it could also have tumor‐suppressive properties[^36^](#cas14433-bib-0036){ref-type="ref"} that contribute to p53 activity as an effector.[^37^](#cas14433-bib-0037){ref-type="ref"} Moreover, in a similar setting, Mdm2 binds the MRN complex (in particular Nbs1) and negatively regulates DNA repair.[^38^](#cas14433-bib-0038){ref-type="ref"} On top of this, Mdm2 binds a variety of RNA molecules,[^39^](#cas14433-bib-0039){ref-type="ref"}, [^40^](#cas14433-bib-0040){ref-type="ref"}, [^41^](#cas14433-bib-0041){ref-type="ref"} including that of p53,[^42^](#cas14433-bib-0042){ref-type="ref"} and this could affect the translation of mRNAs. It remains to be determined whether endogenous Mdm2 also carries out such functions, particularly in the context of p53‐induced expression of Mdm2, but also in a setting where p53 is absent. Of note, endogenous Mdm2 can contribute to stemness and chromatin modifications (H2A K119ub1 as well as H3K27me3) in cells that lack p53 altogether.[^16^](#cas14433-bib-0016){ref-type="ref"} Moreover, and again in the absence of p53, Mdm2 supports the progression of DNA replication forks.[^43^](#cas14433-bib-0043){ref-type="ref"} *A challenging question is whether such activities are enhanced when p53 increases the levels of Mdm2?*
6. DOES MDM2 PROMOTE OR PREVENT TUMOR DEVELOPMENT? {#cas14433-sec-0006}
==================================================
When p53 is deleted, mice are prone to cancer.[^44^](#cas14433-bib-0044){ref-type="ref"} Some p53 target genes were studied in a similar fashion, as they are at least not completely essential for the development of a mouse. In this way, it turned out that none of these p53 target genes, when knocked out, recapitulates the phenotype of p53‐null mice, ie the susceptibility to cancer. Even triple knock‐outs of cdkn1a/p21, bbc3/puma, and pmaip1/noxa in p53‐proficient mice did not induce cancer formation.[^45^](#cas14433-bib-0045){ref-type="ref"} This either means that p53 suppresses tumors by transcription‐independent mechanisms, or otherwise that a p53 target gene was missing from the reported analyses -- and that this missing gene was responsible for tumor suppression. One important gene that could not be analyzed in this way is *mdm2*. Like the other genes mentioned, it is strongly p53‐responsive. However, knocking down *mdm2* in p53‐proficient mice results in exaggerated p53 activity and embryonic lethality.[^46^](#cas14433-bib-0046){ref-type="ref"}, [^47^](#cas14433-bib-0047){ref-type="ref"} Thus, it is difficult to determine if Mdm2 (and perhaps Mdm4) are carrying out a tumor‐suppressive activity in addition to their p53‐regulating function, although this hypothesis has been raised for some time.[^36^](#cas14433-bib-0036){ref-type="ref"}
One fact that might support such a scenario is that Mdm2 overexpression, eg by gene amplification, is a rare event in comparison to p53 mutations. Only a subset of sarcomas, as such a relatively rare tumor species, contains such amplifications of Mdm2 on a regular basis ([www.cbioportal.org](http://www.cbioportal.org)). In contrast, p53 is mutant in roughly 1 out of 2 tumors, including the most common cancer species. Likewise, even when p53 is wildtype, silencing p14ARF represents a far more common way of dampening p53 activity, in comparison to Mdm2 amplifications. *Could tumor‐suppressive activities of Mdm2 represent a reason for this failure to observe Mdm2 amplifications in carcinomas? Could this be a reflection of the deleterious effects of Mdm2 overexpression in most cultivated cells?*
7. ARE THERE ANY TUMOR‐PROMOTING MUTATIONS IN MDM2? {#cas14433-sec-0007}
===================================================
If Mdm2 has tumor‐suppressing activities, it is expected that these will be difficult to separate from p53 regulation. Otherwise, we would probably find more cancers with a mutation of Mdm2 that preserves p53‐binding but silences such tumor‐suppressive activities. However, it remains possible that Mdm2 evolved to comprise p53‐regulating as well as tumor‐suppressing activities on similar domains, decreasing the likelihood of cancer. Still, if Mdm2 can be activated by posttranslational modifications, *why don\'t we find activating mutations of Mdm2 in cancer?* Wouldn\'t it seem "easy" for cancer evolution to enhance its binding to p53 and/or its ubiquitin ligase activity, its stability, or its robustness against phosphorylation and inhibition by ATM?
Another question is whether Mdm2 alterations do in fact exist in tumors, but perhaps not as often as classical missense mutations. Rather, the complex splice pattern of Mdm2[^48^](#cas14433-bib-0048){ref-type="ref"}, [^49^](#cas14433-bib-0049){ref-type="ref"} might be altered, by dysfunctional splice regulators or even by mutations in Mdm2 introns that would still need to be identified. It remains subject to future research whether some variations in Mdm2 splicing are enhanced in tumors, and if so, whether the resulting Mdm2 variant might antagonize p53 while abolishing additional, cytotoxic activities of Mdm2.
8. WHAT IS THE ROLE OF MDM2 IN DEVELOPMENT, IF ANY? {#cas14433-sec-0008}
===================================================
When p53 is deleted, this strongly increases the induction of stem cells by the Yamanaka protocol.[^50^](#cas14433-bib-0050){ref-type="ref"} Interestingly, even within this background of p53‐null cells, the removal of Mdm2 decreases the efficiency of stem cell induction.[^16^](#cas14433-bib-0016){ref-type="ref"} We have correlated this phenomenon with the ability of Mdm2 to support the activity of polycomb repressor complexes. However, the question remains *whether Mdm2 and p53 affect the pool sizes of stem cells* in vivo*, or whether they otherwise govern the development of an organism*? At first glance, it appears that the major role of Mdm2 and Mdm4 in development consists in the p53 antagonism. As soon as p53 is removed along with Mdm2 (or Mdm4), mice are born at near‐Mendelian ratios (although they are still as cancer prone as the p53 single knockouts).[^51^](#cas14433-bib-0051){ref-type="ref"} However, these are animals that are kept under very artificial conditions. In nature, even developing organisms are facing stresses that include infectious diseases, malnutrition, and predators. It remains to be determined whether Mdm2 might add robustness to stem cells or other aspects of development under such stresses, and whether this might comprise p53‐independent activities as well.
9. WHAT IS THE ROLE OF MDM2 IN AGING? HOW DOES THE P53‐MDM2 AXIS FUNCTION TO AFFECT THE RATE OF AGING? {#cas14433-sec-0009}
======================================================================================================
Cellular senescence occurs in response to intrinsic and extrinsic stresses where the cell withdraws from its cell cycle progression and loses its capacity to replicate. This is commonly irreversible and the cell may lose some of its functions and alter its morphology while it continues to metabolize. Replicative senescence (RS)[^52^](#cas14433-bib-0052){ref-type="ref"} is observed when cells duplicate for a certain number of generations, resulting in telomere shortening. This in turn is recognized as breaks in the DNA by the ATM protein (along with the MRN complex), which phosphorylates and activates checkpoint kinase 2 (Chk2), and the resultant protein modifications weaken the Mdm2 protein binding to p53. P53 is then activated (increased in concentration) as a transcription factor producing p21, PAI‐1, PML, and microRNA‐34a, which contribute to the inhibition of Cyclin E‐Cdk2, partially blocking the release of E2F from its negative regulator, the retinoblastoma protein (Rb). At the same time, increased levels of the p16 tumor suppressor protein inhibit Cyclin D‐Cdk4/6 and complete the inhibition of Rb's release of E2F from an Rb‐E2F complex blocking entry into S‐phase.
A second mediator of cellular senescence results from oncogene‐induced senescence (OIS).[^53^](#cas14433-bib-0053){ref-type="ref"} In this case, the mutational activation of *Ras*, *Myc*, or other oncogenes results in the transcription and translation of the ARF protein, which binds to Mdm2 and blocks its ability to ubiquitinate p53. P53 levels increase, resulting in a senescence program similar to the one described above. When this happens in vivo, the p53 transcriptional program includes several cytokines of the innate immune system (interleukin‐6, tumor necrosis factor, and macrophage inhibitory cytokine‐1) which attract natural killer cells, CD‐8 T cells, and monocytes/macrophages, which kill the senescent cells and clean up the debris. This is called the senescence‐associated secretory phenotype (SASP).[^54^](#cas14433-bib-0054){ref-type="ref"}
Over a lifetime, RS, DNA damage repair (DDR) senescence, and OIS produce cells that secrete inflammatory cytokines as part of the SASP. With aging, the efficiencies of the innate immune system and the adaptive immune system decline, and senescent cells remain in vivo resulting in chronic diseases like rheumatoid arthritis, inflammatory disorders, autoimmune diseases, and even cancers. It has been hypothesized that persistence of senescent cells in the body is the (or a) cause of the aging process. Kirkland and his associates have shown that adding isogenic senescent cells to mice results in reduced survival of these mice compared to untreated age‐matched mice. This effect was more pronounced in older than in younger mice. There are senolytic drugs, dasatinib and quercetin, that have been shown to preferentially kill and remove senescent cells from a mouse and reduce the levels of SASP in the animals.[^55^](#cas14433-bib-0055){ref-type="ref"} When these drugs have been used to treat older mice with high levels of senescent cells, they reduced the physical dysfunctions and extended the life span of these mice by 36% when compared with untreated controls. These drugs also reduced the number of senescent cells and the levels of the SASP in human explants tested in vitro.
These kinds of studies implicate Mdm2 and p53 in playing a central role in aging by initiating senescent cells in response to RS, DDR, and OIS. Is there any evidence to support these ideas? Scrable and her group inserted into the germ line of mice a splice variant of p53, deltaNp53 missing the amino‐terminal transcriptional activator. The resultant mice had a much reduced lifespan, aging more rapidly than normal mice.[^56^](#cas14433-bib-0056){ref-type="ref"} A mutation at the amino‐terminal residue, ser‐15 (ser‐18 in murine p53), which is phosphorylated by the ATM kinase resulting in p53 activation and senescence, produced a mouse that (unexpectedly) also had accelerated aging.[^57^](#cas14433-bib-0057){ref-type="ref"} Extra copies of the p53 gene are lethal. However, up to 4n copies of p53 can be tolerated with extra copies of p19 ARF, which binds to and regulates Mdm2 activity,[^58^](#cas14433-bib-0058){ref-type="ref"} resulting in an increased lifespan and an improved age‐related health decline. The conclusions from these observations are that physiological regulation of p53 through Mdm2 delays the aging process, whereas chronic excessive p53 prevents cancer but accelerates the aging process.[^59^](#cas14433-bib-0059){ref-type="ref"} In support of these ideas, Lessel and his colleagues[^60^](#cas14433-bib-0060){ref-type="ref"} have described a human family where a mutation in the *Mdm2* gene is linked to premature aging. Somewhat similarly, a germline mutation in Mdm4 leads to shortened telomeres in patients as well as in mice.[^61^](#cas14433-bib-0061){ref-type="ref"} *The question remains whether the balance between tumor suppression and aging can be manipulated in a favorable manner by therapeutic interventions, to postpone the aging process while still avoiding cancer*.
10. WHAT IS THE ADVANTAGE OF MDM2 IN EVOLUTION? {#cas14433-sec-0010}
===============================================
The available evidence for the role of Mdm2 in development of the mouse suggests that the only absolutely essential function of Mdm2 is the regulation of p53. This means that Mdm2 forms a genetically inseparable part of the p53 system. It raises the question *why Mdm2 evolved as a separated gene, rather than relying on p53 itself for its functions and its regulation*. Does the evolution of Mdm2 make the p53‐Mdm2 system either more efficient in tumor suppression, or more controllable to avoid unnecessary p53‐driven cell death? Rather than integrating inputs such as DNA damage signaling on Mdm2 as well as p53, it would have been more efficient to focus all input and output signaling pathways on p53 itself (eliminating Mdm2), eg through posttranslational modifications on p53 that determine its activities and stability.
Nature uses such an alternative regulatory pathway in the case of p63, a p53 paralogue that appears to be the evolutionary precursor in invertebrates of p53 in vertebrates. In vertebrates, p63 has many of the same responses to DNA damage in the germ line that p53 has in the soma. Irradiated oocytes undergo apoptosis depending on p63, not p53.[^62^](#cas14433-bib-0062){ref-type="ref"} To activate p63 in this context does not seem to require Mdm2 regulation. Rather, phosphorylation of p63 induces a conformational switch that allows the transcriptionally active isoform(s) of p63 to expose its transactivation domain.[^63^](#cas14433-bib-0063){ref-type="ref"}, [^64^](#cas14433-bib-0064){ref-type="ref"} It even seems that p63 and its regulation represent the evolutionary older mechanism, in comparison to p53 and Mdm2.[^65^](#cas14433-bib-0065){ref-type="ref"} In comparison to p63, p53 has lost much of this C‐terminal, regulatory domain.
*Why would p53 not use a similar route of regulation, and why do we instead see Mdm2 as its major regulator?* *What makes the difference in this regard between somatic cells and germ cells?* One can speculate that having a separate recipient of stress signaling (Mdm2) could increase the capacities of the system for multiple stress inputs. If signaling pathways end on Mdm2, all they need to do is to *inactivate* its ability to bind and/or ubiquitinate p53. This can be achieved on multiple domains within Mdm2, interfering either with p53 binding, or ubiquitin ligase activity, or just the conformation of Mdm2 that enables the transfer of ubiquitins specifically on p53. In contrast, if a molecule such as p53 needs to become *more active* in response to a signaling pathway, the opportunities to carry this out are more limited, eg by modifying specific intramolecular switches. Moreover, having a negative regulator between stress signaling and p53 might also lower the chances of inactivating p53 by mutation. If intramolecular switches on p53 would be obligatory for its activation, any mutation within such switchable domains could abolish its tumor‐suppressive activity. In contrast, when Mdm2 is carrying out the regulation of an otherwise constitutively active p53 molecule, relatively fewer mutations on p53 can dramatically reduce its activity, and that is what we see in only a handful of hotspot mutations, which represent more than 33% of the p53 mutations observed in cancers.
If some of the above considerations are valid, early evolved versions of Mdm2 should be useful to determine the most well‐preserved receptor functions for damage signaling. The discovery of Mdm2 in placozoans (summarized by Lane and Verma[^66^](#cas14433-bib-0066){ref-type="ref"}) and other ancient organisms should help to clarify the most conserved functions of it.
The p53 pathway has a central node containing p53 and Mdm2 proteins. Feeding into this pathway are 10‐15 different types of stress signals that turn on the p53 transcription factor or turn off the p53 transcription factor. Each stress signal has a sensor (ATM for DNA breaks), a transmitter, Chk2 for DNA breaks, and a receptor protein that is modified by the detector and/or transmitter. This receptor is Mdm2 and in turn regulates the level or activity of p53. The p53 transcription factor protein is modified to integrate the stress signals being transmitted and then chooses a transcriptional program that results in either the repair of a stress or damage or decides to kill the cell (by one of several mechanisms) so as to eliminate the damaged cells and prevent cancers from arising. Between the inputs and the outputs of the p53 signal transduction pathway and the information it integrates, ie the stress signals and responses (higher order information), the p53 pathway is the most connected pathway to other cellular functions and the most informed about those functions through one central node. However, the organization or structure of this set of pathways appears vulnerable, not robust at all, because a single mutation in the *p53* gene can disrupt this entire set of multiple pathways in the cell. The fact that the *p53* gene is the single most common mutation in human cancers supports this notion. It is the most vulnerable node in the cell to mutation for cancer development. *Why did evolution choose to construct the p53/Mdm2 pathway in this fashion?*
Almost all of the input signals coming into the central node are received and interpreted by the Mdm2 protein. A p53 mutation does not affect the reception of a stress signal. The insertion of Mdm2 into the p53 pathway, where p53 transcribes the *Mdm2* gene while the Mdm2 protein ubiqitinates p53 and thereby mediates the degradation of the p53 protein, does a number of new things. First, it sets up an autoregulatory loop so that the levels (activities) of these two proteins oscillate 180 degrees out of phase, which does not let either protein get to a very high concentration (like a thermostat) without the loop being broken. Second, a mutation in the *p53* gene has no effect on the ability of Mdm2 to detect a stress signal while a mutation in a structurally autoregulated p63 would fail to respond to both an input and an output signal. If in addition to p53, Mdm2 has other substrates that respond to stresses, then these other substrates will still function after a stress signal. Thus, it is possible that the input stress signals are more robust, or backed up, than the output signals of p53, because the Mdm2 gene was inserted into the pathway during evolutionary history. If this idea is correct it brings up the question *what other functions does Mdm2 have in response to stress signals, in addition to regulating p53 activity? If the Mdm2 pathway splits into p53 regulation and other responses to a stress, are we missing 50% of what this node does for a cell?*
11. WHAT IS THE NEED FOR MDM4? {#cas14433-sec-0011}
==============================
The above considerations apply even more urgently to the evolution of Mdm4. Here, it is even more puzzling why another paralogue of Mdm2 evolved, and what the evolutionary advantage of it might be as opposed to one Mdm2‐like molecule. Moreover, like Mdm2, Mdm4 turned out to be essential for murine development, if and only if functional p53 is present. Yet, Mdm2 and Mdm4 do have a few differences. They are regulated differently and their levels and "activity" differ. *What regulators and signaling pathways act on Mdm4 rather than Mdm2, and how do the two proteins divide up their tasks?* Our current knowledge on this was recently summarized.[^20^](#cas14433-bib-0020){ref-type="ref"}The Mdm2 KO mouse dies earlier in fetal life than the Mdm4 KO mouse but both are rescued by a p53 KO. This argues that: (i) different efficiencies of the proteins are prevalent during development; (ii) different locations or cell types "prefer" either of the p53 regulators during development; and (iii) each protein is produced at different times in development.[^67^](#cas14433-bib-0067){ref-type="ref"} *Where and when is Mdm2 vs. Mdm4 active and required for p53 regulation during development?*While *Mdm2* gene amplifications can occur mainly in sarcomas, Mdm4 amplifications occur in approximately 10% of glioblastomas and invasive breast carcinomas according to The Cancer Genome Atlas ([www.cbioportal.org](http://www.cbioportal.org)). This is even more surprising as the two proteins are supposedly most active in a heterodimeric complex. *What is the determinant and mechanism behind this tumor tissue specificity?*
One possibility is that Mdm4 could have been inserted into the p53 pathway because it receives other stress signals not received by Mdm2. *If this is correct then the question is why is there not an Mdm5, 6, 7, 8, and 9 for additional stress signals?*
Mdm2 has a large number of spliced forms.[^48^](#cas14433-bib-0048){ref-type="ref"} Perhaps the reason for this is that each spliced form receives a different stress signal than the others. This could represent a way of partitioning Mdm2 into Mdm5, 6, 7, 8, and 9, each protein responding to a diverse stress. A mutation in one Mdm2 spliced form would not necessarily impact another spliced form from signaling to p53 or the alternate Mdm2 substrates. In this way, most mutations in the Mdm2 protein would have only a small effect on cancer production and not necessarily be selected for, while p53 mutations would be much more common. In addition, any mutation in the *Mdm2* gene that inactivates its function of regulating p53 levels should be a lethal mutation, because too much WT p53 would kill a cell. The mutations in the *Mdm2* gene are most commonly gene amplifications (too much Mdm2 protein, not too little) and, as mentioned above, seem to be tissue restricted to mesenchymal tumors, especially liposarcomas. *Does this suggest that different stem or progenitor cells of different tissues may have different p53‐Mdm2 signal transduction pathways, resulting in tissue‐specific mutations contributing to cancers?*
Mdm4 is regulated by alternative splicing as well, and the emerging mechanism(s) are defined more sharply than for Mdm2. In particular, exon 6 can be skipped from Mdm4, giving rise to a shorter and less stable isoform, as reviewed recently.[^20^](#cas14433-bib-0020){ref-type="ref"} The inclusion of this exon correlates with tumor‐associated deletions of the gene encoding ribosomal protein RPL22,[^68^](#cas14433-bib-0068){ref-type="ref"} at least suggesting that such tumors control p53 by giving preference to the synthesis of Mdm4. The splicing pattern can be manipulated by clinically available drugs. For instance, inhibition of cyclin‐dependent kinase 4 (CDK4) leads to exon skipping in Mdm4, synergistically with PRMT5 inhibition.[^69^](#cas14433-bib-0069){ref-type="ref"} Taken together, *this raises the question whether tumor‐specific, full‐length Mdm4 can serve as a biomarker or even target in tumor therapy*.
12. IS THERE A FUTURE OF TARGETING MDM2 AND/OR MDM4 FOR THERAPY? {#cas14433-sec-0012}
================================================================
With the advent of Nutlin[^70^](#cas14433-bib-0070){ref-type="ref"} 20 years ago, Mdm2 became druggable, raising the hope that the most successful tumor suppressor could now be activated at will, at least in those 50% of all tumors that retain WT copies of p53. Still, however, no FDA approval has been reached for Mdm2 antagonists, despite multiple attempts to prove their efficacy in the clinics. *Why did this turn out to be so difficult?* Liposarcoma seemed like an ideal tumor entity to be cured by Mdm2 antagonizing drugs, given its 90% frequency of *Mdm2* gene amplifications and the strong in vitro response to Mdm2 antagonists of liposarcoma‐derived cells. However, insufficient cell killing, perhaps because of the pharmacokinetic and dynamics, and the occurrence of p53 mutations were hampering clinical successes so far. On top of missing efficacy, toxicities including myelosuppression but also severe nausea and diarrhea (cell death in the gut) were strongly impairing the quality of patients' lives.[^71^](#cas14433-bib-0071){ref-type="ref"} So is it time to give up?
Alternatively, *are there better ways to target Mdm2 antagonists towards tumors, and to enhance p53‐mediated tumor cell death?* Surprisingly, we are still not sure about the determinants that render cells susceptible towards death in response to Mdm2 inhibitors, aside from functional p53. Within cultivated cells with a WT p53 status, there is still a large spectrum of responses, reaching from reversible cell cycle arrest through sustainable senescence all the way to efficient cell death.[^72^](#cas14433-bib-0072){ref-type="ref"} If we knew the mechanisms underlying these differential responses, we might have a chance to target Mdm2 along with such determinants of survival. Then, however, it would still be subject to investigation whether corresponding drug combinations had acceptable toxicities.
*Could we improve the efficacy of Mdm2 antagonists by eliminating Mdm2 altogether*, rather than "just" blocking its interaction with p53? Recently, proteolysis‐targeting chimera (PROTAC)[^73^](#cas14433-bib-0073){ref-type="ref"} drugs to degrade Mdm2 were reported.[^74^](#cas14433-bib-0074){ref-type="ref"} If Mdm2 had oncogenic activities on top of its action on p53, such PROTACs should be more capable of interfering with tumor cell proliferation than the classical Mdm2 antagonists. However, as outlined above, Mdm2 also displays a number of activities that are at least capable of hindering proliferation. If such activities are lost, Mdm2‐targeting PROTACs might be less efficient. Thus, the biological effects of this exciting class of small compounds still remain to be determined when targeting Mdm2.
*What are the advantages of targeting Mdm4, alone or in addition to Mdm2?* As Mdm4 represents an essential partner of Mdm2 for many activities, it is at least conceivable to design compounds that bind and inhibit Mdm4. This could have advantages over targeting Mdm2, at least in a subset of malignancies that rely on the enhanced synthesis of full‐length Mdm4.[^75^](#cas14433-bib-0075){ref-type="ref"} Interestingly, depletion of Mdm4 interferes with the proliferation of breast cancer cells even when p53 is mutant, through activation of the CDK inhibitor p27/CDKN1B.[^76^](#cas14433-bib-0076){ref-type="ref"} This further argues that Mdm4 has p53‐independent, cancer‐promoting activities that might be druggable. Moreover, small compounds can induce the dimerization of Mdm2 and Mdm4 while antagonizing their binding to p53.[^77^](#cas14433-bib-0077){ref-type="ref"} Together, these considerations raise the question *which molecular interfaces and activities of Mdm4 would be most helpful to target by small compounds?*
*Could Mdm2/Mdm4 antagonists serve different purposes in addition to tumor cell killing?* One possibility could be the protection of nontransformed cells and tissues against the toxicities of chemotherapy, in the context of p53‐mutant tumors. The idea is to first halt the cell cycle of most normal cells, including tissues that are otherwise rapidly proliferating, eg the bone marrow or the epithelia of the gut and the skin. Mdm2‐antagonists would keep them from dividing, whereas they would not halt the proliferation of tumor cells carrying mutant p53. Subsequently, the treatment with chemotherapeutics should selectively affect the tumor, as long as such chemotherapeutics depend on cell proliferation for their efficacy. This applies to nucleoside analogues that are incorporated during S phase -- they are essentially of no effect when the cell cycle is arrested.[^78^](#cas14433-bib-0078){ref-type="ref"} But similar principles are also applicable to topoisomerase inhibitors (most effective in S phase) or drugs targeting the mitotic spindle, such as taxanes[^79^](#cas14433-bib-0079){ref-type="ref"} (most effective during mitosis). This concept, sometimes referred to as "cyclotherapy",[^80^](#cas14433-bib-0080){ref-type="ref"}, [^81^](#cas14433-bib-0081){ref-type="ref"} has been entertained for two decades but nonetheless has not been clinically established. What would be required to get it to work? Two major obstacles need to be overcome. First, Mdm2 antagonists not only prevent cell cycle progression, but they can also induce cell death, and they are therefore cytotoxic to the bone marrow and the gut on their own. Thus, if possible, new drugs, treatment schedules, and pharmacokinetics would be required to achieve Mdm2 antagonism with as little tissue damage as possible, while still inducing thorough (but reversible) cell cycle arrest. Second, the regimen applied to the (still cycling) tumor cell would need to be harsh enough to eliminate most, if not all, tumor cells within the relatively short time window during which normal cells can be kept arrested. This would probably require a higher degree of cytotoxicity than what is presently achieved by chemotherapeutics.
Medical applications of Mdm2 antagonists might not be limited to the context of cancer. *So what else could these drugs be used for?* Like some chemotherapeutics, eg methotrexate, Mdm2 antagonists might be useful in immunosuppression. The adaptive immune response is based on the clonal proliferation of few B and T cells with receptors to a given antigen. Halting this proliferation can be very effective to overcome autoimmune diseases such as systemic lupus erythematosus or scleroderma. It remains to be determined whether drugs that disrupt the p53‐Mdm2 interaction can be used for such purposes as well.
At present, we are facing a bewildering discrepancy between the unique frequency of p53 mutations in human malignancies and the lack of clinically successful therapeutic approaches that directly target the p53/Mdm2 regulatory system. *How can we change this?*
CONFLICT OF INTEREST {#cas14433-sec-0014}
====================
Arnold J. Levine is on the Board of Directors of PMV Pharma, a company that develops p53 regulators. He is also a member of the board of directors of Pharmabody, a monoclonal antibody company presently involved with hemophilia, complement disorders, and infectious diseases. Matthias Dobbelstein has no conflict of interest.
This work was initiated in preparation for the 10th Mdm2 Workshop in Tokyo, which will now be held in 2022. Our goal is to address a series of questions that both stimulate and focus the field upon the regulatory node of the p53‐Mdm2 signal transduction pathway. This pathway is central to the homeostasis, vitality and responses to stress brought about by many diverse pathologies. By publishing this set of challenging questions now, in 2020, we are hoping to hear about some of the answers to these questions at the meeting in 2022. Meetings focused on topics like those discussed in this review are perfect places to advance the field. Everyone, from many diverse fields, is welcome to join this meeting.
| {
"pile_set_name": "PubMed Central"
} |
Background {#Sec1}
==========
Cigarette and alcohol consumption are major causes of preventable deaths worldwide. Smoking is the cause of approximately six million deaths per year \[[@CR1]\] secondary to the development of different cancers as well as cardiovascular and respiratory diseases \[[@CR2]\]. Alcohol consumption, on the other hand, is associated with more than 200 diseases and accidents, causing 5.1% of the global burden of disease \[[@CR3]\].
Rural-to-urban migrants are thought to increase their cigarette \[[@CR4]--[@CR7]\] and alcohol \[[@CR5]--[@CR9]\] consumption after migration, not only because they usually migrate from low to high consumption settings, but also because they often suffer from high levels of stress and poor mental health, which are related with greater alcohol and cigarette consumption \[[@CR10], [@CR11]\]. Some studies, however, have found that certain rural-to-urban migrants do not consume more alcohol and cigarette products, and even consume less, probably due to a protective effect of rural backgrounds, or economic limitations that prevent substance purchase \[[@CR12], [@CR13]\].
Most of the published studies in rural-to-urban migrants have cross-sectional designs, and compare the prevalence of current smoking, alcohol intake, or alcohol intoxication between only two of the three population groups: migrants vs. rural \[[@CR14]\] or vs. urban groups \[[@CR7], [@CR8], [@CR15]\]. Few cross-sectional studies have contrasted smoking and alcohol consumption in rural-to-urban migrants with their rural groups of origin and their urban counterparts: one study in India \[[@CR9]\], two studies in China \[[@CR6], [@CR16]\], and one study with data from China, Ghana, India, Mexico, Russia and South Africa \[[@CR17]\]. Their results suggest mixed estimates of smoking and alcohol use between rural, migrant and urban groups.
Although cross-sectional data is important to evaluate the association of increased substance consumption between rural-to-urban migrants to their rural and urban counterparts, longitudinal data is necessary to better understand its epidemiology and identify appropriate preventive interventions. To our knowledge, only two longitudinal prospective studies in rural-to-urban migrants have assessed these behavioral risk factors. The first is a study in Tanzania that evaluated smoking rates and weekly alcohol consumption before and one-to-three months after migration \[[@CR5]\]. The second is a study in Indonesia that evaluated smoking initiation and changes in smoking quantity among recent (\<3 years) migrants \[[@CR18]\].
Smoking and alcohol consumption are behaviors whose negative health impacts are closely related to the quantity and frequency of consumption. Yet, before quantifying units of intake, it is also important to address major patterns of consumption across rural, urban, and migrant groups to identify meaningful approaches to prevention.
In addition, epidemiological studies use many different definitions to address substance consumption, which inhibits an adequate comparison. For instance, the concept of "current smoking" has various definitions, and therefore various prevalence rates \[[@CR19]\]. Therefore, it is also necessary to compare these definitions and evaluate their appropriateness of use in rural-to-urban migrant studies.
In summary, migrant populations appear to be at increased risk to consume harmful substances such as alcohol and tobacco. To guide regional policy in places of high-density rural-to-urban migration, local studies of substance use are necessary. As such, this study aimed to compare the prevalence of tobacco smoking and heavy drinking among rural, urban, and rural-to-urban migrants in Peru, as well as the incidence of new smoking and new heavy drinking cases at 5-years follow-up.
Methods {#Sec2}
=======
Study design {#Sec3}
------------
This is a secondary data analysis of the PERU MIGRANT Study. The PERU MIGRANT is an ongoing longitudinal study aimed to evaluate cardiovascular risk factors in rural, urban and rural-to-urban migrant population. The methods used have been previously described \[[@CR20]\].
Participants and procedures {#Sec4}
---------------------------
Briefly, a random sampling of three different population groups was conducted. 1) Rural population: people born and living in the village of San Jose de Secce, Ayacucho, located at 3,239 m above sea level. 2) Rural-to-urban migrants: people born in rural settings of Ayacucho who have migrated and were living in Lima at enrollment. 3) Urban population: people born and living in Lima, the capital of Peru. Both the urban and migrant populations were taken from Las Pampas de San Juan de Miraflores, a periurban coastal setting located in the south of Lima.
In San Jose de Secce, our rural setting, people usually work as farmers, with hard physical labor and few televisions. In comparison, in our urban setting, people usually work in service industries or factories, and spend their leisure time watching television. In addition, poverty rates and illiteracy are lower in our urban than our rural setting (18 vs 80%, and 2 vs 33%, respectively) \[[@CR21]\]. Additional information, including a map of the settings, is available elsewhere \[[@CR20]\].
To differentiate rural and urban settings, we relied on population size, according to thresholds proposed by the Department of Agriculture's Rural--urban Continuum Codes of the United States, that have been extensively used \[[@CR22], [@CR23]\]. This institution define rural and urban areas as places with fewer or more than 2,500 inhabitants respectively. Accordingly, in 2006, our urban setting had over 350 thousands and our rural setting less than 1000 inhabitants.
The baseline study was conducted in 2006--2007 after a general census in both study settings. A single-stage random sampling, stratified by sex and age groups, was performed in all three population groups. Trained community health workers administered the questionnaires and collected anthropometric measurements as well as laboratory samples. During 2012--2013, a follow-up visit was carried out, which included a survey and anthropometric measurements.
Variables definition {#Sec5}
--------------------
### Cross-sectional outcomes {#Sec6}
For the baseline analysis, the outcomes of interest were: lifetime smoking, current smoking, and heavy drinking.
Different definitions of smoking status are currently used in other epidemiological studies. In order to compare these definitions, we calculated prevalence rates for each of our population groups using four of the most popular definitions across the literature (Table [1](#Tab1){ref-type="table"}). For instance, definition 1 is similar to the one used in the Global Adult Tobacco Survey \[[@CR24]\] and UK Labour Force Survey \[[@CR25]\]. Definition 2 is similar to the one used in the National Survey on Drug Use and Health of United States \[[@CR19]\]. Definition 3 is similar to the one used in the CDC' National Health Interview Survey \[[@CR26]\]. Definition 4 is similar to the one used by the New Zealand Ministry of Health \[[@CR27]\].Table 1Prevalence of never, former, and current smokers according to four different definitionsDefinitionsDefinition 1\
Based on having smoked and smoking daily or occasionallyDefinition 2\
Based on having smoked and havingsmoked in the last monthDefinition 3\
Based on having smoked at least 100 cigarettes and smoking daily or occasionallyDefinition 4\
Based on having smoked at least 100 cigarettes and having smoked in the last monthNever smokersExperimental smokers or those who have never smoked cigarettesExperimental smokers or those who have never smoked cigarettesThose who have not smoked 100 cigarettes in their lifetimeThose who have not smoked 100 cigarettes in their lifetimeFormer smokersHaving smoked, but currently do not smoke daily nor occasionallyHaving smoked, but not in the last 30 daysThose who smoked ≥ 100 cigarettes in their lifetime, but currently do not smoke daily nor occasionallyThose who smoked ≥ 100 cigarettes in their lifetime, but currently do not smoke daily nor occasionallyCurrent smokersThose who currently smoke daily or occasionallyThose who have smoked in the last monthThose who smoked ≥100 cigarettes in their lifetime, AND currently smoke daily or occasionallyThose who smoked ≥100 cigarettes in their lifetime, AND have smoked in the last monthPrevalence rates In our population, per study groupN% (95% CI)N% (95% CI)N% (95% CI)N% (95% CI)Prevalence of never smokers Urban6131.0 (24.5--37.4)6131.0 (24.5--37.4)13267.3 (60.8--73.9)13267.3 (60.8--73.9) Migrant23139.9 (35.9--43.9)23139.9 (35.9--43.9)49286.2 (83.3--89.0)49286.2 (83.3--89.0) Rural9648.0 (41.1--54.9)9648.0 (41.1--54.9)17593.1 (89.5--96.7)17593.1 (89.5--96.7)Prevalence of former smokers Urban8543.1 (36.2--50.1)9447.7 (40.7--54.7)2914.8 (9.8--19.8)3417.3 (12.0--22.6) Migrant25143.4 (39.3--47.4)27547.5 (43.4--51.6)396.8 (4.8--8.9)427.4 (5.2--9.5) Rural4120.5 (14.9--26.1)7035.0 (28.4--41.6)42.1 (0.1--4.2)52.7 (0.4--5.0)Prevalence of current smokers Urban5125.9 (19.8--32.0)4221.3 (15.6--27.0)3517.9 (12.5--23.2)3015.3 (10.3--20.3) Migrant9716.8 (13.7---19.8)7312.6 (9.9---15.3)407.0 (4.9---9.1)376.5 (4.5---8.5) Rural6331.5 (25.1---37.9)3417.0 (11.8---22.2)94.8 (1.7---7.8)84.3 (1.4---7.1)
Although the 100 cigarettes threshold has not been associated with increased clinical risks, it has been extensively used to pragmatically identify lifetime smokers (current or former smokers) and differentiate them from experimental or new-onset smokers, so we decided to use it \[[@CR28]\]. In addition, since we are evaluating consumption patterns across two settings, the "occasionally smoking" term could be interpreted differently between people living in these diverse environmental settings. As such, we used the "one-month smoking" threshold to more objectively identify current smokers. Consequently, definition 4, which includes both 100-cigarettes and one-month smoking thresholds, was used to define lifetime smoking and current smoking outcomes for our research questions and statistical models.
Heavy drinking was assessed by the question "In the last year, how often have you consumed 6 or more standard alcohol drinks on one occasion?" Those who answered "monthly," "weekly," or "daily or almost daily" were classified as "heavy drinkers." One standard alcohol drink, as established by the National Institute on Alcohol Abuse and Alcoholism, is defined as a 12 oz (355 mL) glass of beer, a 5 oz (148 mL) cup of wine, or 1.5 oz (44.3 mL) of distilled spirits, \[[@CR29]\].
In our settings, people commonly smoke branded cigarettes, and drink either branded beer or artisanal *cañazo* (sugarcane brandy) \[[@CR30]\]. *Cañazo* is considered as a distilled spirit, since its alcohol concentration is approximately 40%.
### Longitudinal outcomes {#Sec7}
For the longitudinal analysis, we used two outcomes: new smokers and new heavy drinkers. New smokers were individuals classified as never smokers in the baseline survey who reported having smoked in the last month during the follow-up survey. New heavy drinkers were individuals who denied heavy drinking or did it less than monthly in the baseline, but reported heavy drinking at least monthly at follow-up.
### Exposure {#Sec8}
For both, the cross-sectional and the longitudinal analyses, the exposure of interest was the study group, categorized as rural, urban, and rural-to-urban-migrant groups.
### Other variables {#Sec9}
Other smoking-related variable was daily smoking, defined as participants who answered "I smoke at least a cigarette per day." to the question "At present, how often do you smoke cigarettes?" Average and median number of daily cigarettes smoked were also estimated among daily smokers.
Demographic variables included in the analyses as potential confounders were: age (\<50 or ≥50 years), sex, education level (none or some primary education, complete primary education, and at least some secondary education), possessions weighted assets index, and positive mental health (PMH).
Assets index was based on the number of assets available at the participant's household, divided in tertiles for each population group (lowest, middle, and highest), and then combined in one single variable. PMH, an expression of a healthy mind, was measured by an adaptation of the General Health Questionnaire (GHQ-12), and treated as a continuous variable, as detailed elsewhere \[[@CR31]\].
Statistical analysis {#Sec10}
--------------------
For the descriptive analysis, means and standard deviations (SD), medians and interquartile ranges (IQR), as well as frequencies and percentages, were utilized. We performed bivariate analyses in order to compare sex, age, education level, assets index, PMH, and daily smoking according to population groups, using Chi-squared or ANOVA tests. We also used the Kruskal-Wallis test to compare the number of daily cigarettes smoked among daily smokers according to population groups.
For cross-sectional analysis, we generated crude and adjusted Poisson regression models with robust variance and estimated prevalence ratios (PR) and 95% confidence intervals (95% CI) in order to assess the associations between exposures (population groups, sex, age, education level, asset index, and PMH) and three dichotomous outcomes: lifetime smoking, current smoking, and heavy drinking. Adjusted models included all exposures mentioned.
For longitudinal analysis, we performed Poisson regression models to report risk ratios (RR) and 95% CI for two outcomes: incidence of new smokers and incidence of new heavy drinkers. For both associations, we generated crude and adjusted models using the same aforementioned exposures and confounders as in cross-sectional models.
We also made post-hoc analyses in the migrant group, which was categorized according to the time since first migration at the baseline survey (\<15 years, 15 to 30 years, or \>30 years). In each of these categories, prevalence and incidence rates of smoking and of heavy drinking were calculated. Fisher's exact test was used to evaluate differences in these categories.
Ethical considerations {#Sec11}
----------------------
Ethical approval for the baseline study was obtained from Institutional Review Boards at Universidad Peruana Cayetano Heredia, in Lima, Peru, and the London School of Hygiene and Tropical Medicine, in London, United Kingdom. The follow-up phase was reviewed and approved by the same Peruvian institution. All enrolled participants gave written informed consent.
Results {#Sec12}
=======
Population characteristics {#Sec13}
--------------------------
We analyzed data from 988 participants: 200 rural, 589 urban-to-rural migrants, and 199 urban residents. Sex was evenly distributed across all three groups. Median ages (IQR) were 47 (37--57), 46 (39--55), and 48 (38--56) years old in the rural, migrant, and urban groups, respectively. The proportion of those having completed at least a year in secondary education was lower in the rural group, intermediate in the migrant group, and higher in the urban group (Table [2](#Tab2){ref-type="table"}).Table 2Baseline characteristics of the study population, per study groupVariablesRural (*n* = 200)\
N (%)Migrant (*n* = 589)\
N (%)Urban (*n* = 199)\
N (%)*p*Women106 (53.0)309 (52.5)107 (53.8)0.949Age ≥ 50 years84 (42.0)252 (42.8)89 (44.7)0.846Education level\<0.001 None or some primary education132 (66.0)183 (31.1)13 (6.6) Complete primary education29 (14.5)99 (16.8)23 (11.6) At least some secondary education39 (19.5)306 (52.0)162 (81.8)Assets index\<0.001 Lowest123 (61.5)242 (41.1)67 (33.7) Middle14 (7.0)156 (26.5)69 (34.7) Highest63 (31.5)191 (32.4)63 (31.7)Positive mental health score5.9 ± 1.96.5 ± 1.86.8 ± 1.8\<0.001Daily smokers1 (0.5)15 (2.6)17 (8.6)\<0.001*P* values were calculated using the chi-squared or the ANOVA test
We evaluated four smoking status definitions. The prevalence of never smoking was higher among the rural population than the migrant/urban populations, for all evaluated definitions. The prevalence of current smoking followed a trend (urban \> migrant \> rural) as per definitions 3 and 4 (which included the 100 cigarettes threshold), was higher in the rural group than in the migrant/urban groups as per definitions 1 and 2. The prevalence of current smoking in definition 1 (defined as having smoked occasionally or daily) was 21.6%, 33.3%, and 85.3% higher than the prevalence of current smokers in definition 2 (defined as having smoked in the last month), among urban, migrant, and rural subjects respectively (Table [1](#Tab1){ref-type="table"}).
The prevalence of daily smoking was higher in the urban group, intermediate in the migrant group, and lower in the rural group. Within daily smokers, the median of cigarettes smoked per day was 1.0 in the rural group, 2.0 in the migrant group, and 2.7 in the urban group (Kruskal-Wallis test *p* = 0.67).
Cross-sectional and longitudinal models for smoking {#Sec14}
---------------------------------------------------
In the adjusted model, compared with the migrant group, the prevalence of lifetime smoking was 129% higher (*p* \< 0.01) among urban dwellers, and 45% lower (*p* = 0.047) among rural dwellers. In addition, our adjusted model with the migrant group as reference revealed a 129% higher (*p* \< 0.01) prevalence of current smoking among the urban group, but showed no significant differences with the rural group (*p* = 0.214) (Table [3](#Tab3){ref-type="table"}).Table 3Factors associated with lifetime smoking, current smoking, and heavy drinkingVariablesLifetime smokingCurrent smokingHeavy drinkingCrude PR (95% CI)Adjusted^a^PR (95% CI)Crude PR (95% CI)Adjusted^a^ PR (95% CI)Crude PR (95% CI)Adjusted^a^ PR (95% CI)Study group Urban**2.36 (1.77--3.14)2.29 (1.64--3.20)2.36 (1.50--3.72)2.29 (1.26--4.16)**1.08 (0.63--1.83)0.91 (0.49--1.68) MigrantRefRefRefRefRefRef Rural**0.50 (0.28--0.88)0.55 (0.31--0.99)**0.66 (0.31--1.39)0.60 (0.27--1.33)1.41 (0.88--2.27)1.19 (0.71--2.00)Sex FemaleRefRefRefRefRefRef Male**6.04 (3.99--9.16)6.05 (3.78--9.69)6.39 (3.41--11.97)7.08 (3.35--14.95)6.92 (3.81--12.57)6.53 (3.41--12.47)**Age \< 50 yearsRefRefRefRefRefRef ≥ 50 years1.15 (0.87--1.54)1.24 (0.91--1.68)0.83 (0.53--1.29)1.00 (0.61--1.63)**0.57 (0.37--0.89)0.55 (0.35--0.88)**Education level None or some primary educationRefRefRefRefRefRef Complete primary education**2.12 (1.18--3.82)**1.33 (0.70--2.55)**2.83 (1.22--6.57)**1.38 (0.58--3.26)1.27 (0.66--2.45)0.74 (0.37--1.48)At least some secondary education**3.68 (2.34--5.79)**1.53 (0.87--2.69)**3.77 (1.88--7.53)**0.98 (0.44--2.19)1.54 (0.95--2.48)0.77 (0.44--1.33)Assets index LowestRefRefRefRefRefRef Middle**1.83 (1.24--2.70)**1.26 (0.83--1.90)1.62 (0.89--2.95)1.17 (0.60--2.29)1.09 (0.64--1.86)1.02 (0.59--1.78) Highest**2.14 (1.51--3.05)1.48 (1.02--2.16)2.23 (1.32--3.75)**1.58 (0.88--2.85)1.44 (0.91--2.27)1.16 (0.72--1.87)Positive mental health (continuous variable)**1.15 (1.03--1.28)**0.95 (0.87--1.04)**1.22 (1.02--1.46)**1.02 (0.86--1.22)1.16 (0.99--1.36)1.06 (0.91--1.25)^**a**^Adjusted by all the variables listed in the tableBold numbers indicate significant associations, *p* \< 0.05
From the longitudinal adjusted models, among those classified as never smokers during the baseline survey, the risk of smoking in the last month during the 5-year follow up was 175% higher in the urban group than the migrant group (*p* = 0.043), with no significant differences between the migrant and rural groups (*p* = 0--349) (Table [4](#Tab4){ref-type="table"}).Table 4Risk factors for smoking and heavy drinking incidenceVariablesNew smokingNew heavy drinkingIncidenceCrude RR (95% CI)Adjusted^a^ RR (95% CI)IncidenceCrude RR (95% CI)Adjusted^a^ RR (95% CI)Study group Urban15/145 = 10.3%**2.65 (1.03--6.81)2.75 (1.03--7.34)**5/155--3.2%1.48 (0.51--4.25)1.27 (0.39--4.11) Migrant15/475 = 3.2%RefRef10/463 = 2.2%RefRef Rural9/158 = 5.7%1.90 (0.76--4.74)1.57 (0.61--4.05)6/153 = 3.9%1.50 (0.55--4.05)1.14 (0.40--3.27)Sex Female9/444 = 2.0%RefRef2/446 = 0.4%RefRef Male30/334 = 9.0%**3.93 (1.66--9.30)4.43 (1.64--11.99)**19/325 = 5.8%**12.99 (3.05--55.41)12.25 (3.06--49.03)**Age \< 50 years28/448 = 6.3%RefRef14/441 = 3.2%RefRef ≥ 50 years11/330 = 3.3%0.45 (0.18--1.10)0.53 (0.19--1.45)7/330 = 2.1%0.67 (0.27--1.64)0.75 (0.27--2.03)Education level None or some primary education6/265 = 2.3%RefRef4/261 = 1.5%RefRef Complete primary education7/116 = 6.0%**4.20 (1.26--14.06)**2.72 (0.68--10.85)3/115 = 2.6%1.77 (0.40--7.79)0.87 (0.16--4.65) At least some secondary education26/396 = 6.6%2.78 (0.92--8.33)1.36 (0.33--5.57)14/393 = 3.6%2.47 (0.82--7.42)0.92 (0.25--3.36)Assets index Lowest10/340 = 2.9%RefRef5/336 = 1.5%RefRef Middle12/187 = 6.4%1.38 (0.50--3.81)0.97 (0.32--2.97)7/191 = 3.7%2.56 (0.82--7.94)2.29 (0.55--9.62) Highest17/251 = 6.8%1.67 (0.69--4.03)1.44 (0.56--3.70)9/244 = 3.7%2.51 (0.85--7.40)3.00 (0.78--11.53)Positive mental health (continuous variable)1.14 (0.90--1.43)0.98 (0.76--1.25)1.11 (0.87--1.40)0.88 (0.65--1.19)^a^Adjusted by all the variables listed in the tableBold numbers indicate significant associations, *p* \< 0.05
Cross-sectional and longitudinal models for heavy drinking {#Sec15}
----------------------------------------------------------
Heavy drinking prevalence was similar among rural and urban groups, when compared to the migrant reference group (Table [3](#Tab3){ref-type="table"}). Likewise, among those who did not report heavy drinking in the baseline survey, the risk of heavy drinking during follow-up was similar among rural and urban groups, when compared to the migrant reference group (Table [4](#Tab4){ref-type="table"}).
Cross-sectional and longitudinal models for other variables {#Sec16}
-----------------------------------------------------------
In the baseline analysis, the prevalence of lifetime smoking, current smoking, and heavy drinking, were higher in men than women (*p* \< 0.01). Heavy drinking prevalence was lower in participants older than 50 years old (*p* = 0.013). Higher prevalence of lifetime smoking was noted among those with a higher assets index (*p* = 0.041) (Additional file [1](#MOESM1){ref-type="media"}: Table S1 and Table [3](#Tab3){ref-type="table"}). In the longitudinal analysis, the incidence of new smoking and new heavy drinking were higher among men than among women (*p* \< 0.01) (Table [4](#Tab4){ref-type="table"}).
Discussion {#Sec17}
==========
Main results {#Sec18}
------------
We found that lifetime smoking prevalence was higher in urban dwellers, intermediate in migrants, and lower in rural dwellers. This indicates that following on a process of internal migration, subjects are more exposed to smoking behaviors than rural dwellers, and experience more smoking. However, smoking incidence was not different between rural and migrant groups. In comparison, prevalence and incidence of heavy drinking were similar between rural, migrant and, urban groups.
Smoking patterns {#Sec19}
----------------
When comparing four "current smoking" definitions, it is clear that prevalence found with definitions that asked for "occasionally or daily smoking" tend to be higher than the prevalence found with definitions that asked for "smoking in the last month". Moreover, these differences seem to be higher among rural dwellers, suggesting that more objective definitions of current smoking are needed, especially in low-consumption settings.
In addition, in definitions 1 and 2, which did not utilize the 100 cigarettes threshold, no trend was observed. Implying, perhaps, that secular changes are occurring, and rural dwellers have recently started smoking; thus, they endorsed smoking in the last month but denied smoked more than 100 cigarettes in their lifetime. Although not statistically significant, this seems to be reinforced by the higher, smoking incidence in the rural group.
Our population has a low smoking prevalence, consistently with previous studies conducted in Peru \[[@CR32], [@CR33]\]. Conversely, prevalence of daily smoking in high income countries reaches 13.7% in United States \[[@CR34]\] and 25% in Finland \[[@CR35]\]. Among daily smokers, the mean of cigarettes per day consumed by our population was under five cigarettes per day, lower than in other Latin American cities as Santiago (Chile), Quito (Ecuador), Bogota (Colombia), and Mexico City \[[@CR33]\], and other countries as United States \[[@CR26]\] and China \[[@CR36]\].
Smoking prevalence and incidence rates were lower in migrants than in the urban population, which suggest that the rural background have a protective effect. Lifetime smoking was significantly higher in the migrant than in the rural group. This smoking pattern is similar to that found with Chinese rural-to-urban migrants \[[@CR4]\], where behavior habit adoption has been attributed to acculturation, stress \[[@CR11]\] and poor mental health \[[@CR37]\]. Conversely, in the adjusted analysis, PMH, a mental health status evaluation, was not associated with smoking prevalence or incidence. However, since our migrant group lived in an urban environment for an average of 32 years, it is possible we are not observing an association of PMH that is given only among more recent migrants. On the other hand, current smoking was not significantly different among rural and migrant groups, possibly due to the low prevalence of current smokers and the small sample size.
Smoking prevalence rates followed a consistent trend in our population (urban \> migrant \> rural). We found four studies that also compared smoking prevalence rates between the rural-to-urban migrants, urban, and rural groups. A study in Yi migrants (China) found the same trend but only in men, while smoking prevalence among women was higher in the migrant group than in the rural and urban groups \[[@CR6]\]. A multi-country study found that migrants had a higher ever smoking prevalence than rural dwellers in Mexico, while no significant difference was found in China, Ghana, India, Russia, and South Africa \[[@CR17]\]. These mixed results may be explained by differences in time since migration, smoking patterns, or the acculturation process between population groups included in those studies and ours. In addition, studies in India \[[@CR9]\] and China \[[@CR16]\] found that male migrants had lower cigarette consumption than rural and urban males. In these last studies, the migrant group was composed of workers, so it is possible that selective migration of people with the best education and predisposition to improve their lifestyle would explain their low consumption, while migration in our study was influenced by political violence lived in Ayacucho \[[@CR38], [@CR39]\] which could have reduced this selective migration effect.
From the longitudinal point of view, only two prospective studies evaluating smoking in rural-to-urban migrants were found. A study in Tanzania \[[@CR5]\] compared current smoking rates before migration and one to three months after migration, and found a non-significant increase in smoking rates (from 16.2 to 23.5%) only in men, while no women reported smoking in either evaluation. Another study in Indonesia \[[@CR18]\] followed-up recent migrants, approximately 65% migrating less than three years ago, and found no significant increase in smoking initiation, but a clear increase in the number of daily cigarettes smoked. These studies suggest a slight increase in smoking rates after migration, but give no information about risk in long-term settled migrants.
While lifetime smoking prevalence was higher in migrants than in the rural population, smoking incidence was not significantly different between those populations. This may suggest that the risk of initiating smoking in migrants could increase during the first years post-migration, and later decrease over time. To evaluate this assumption, we made a post-hoc analysis in our migrant group, and found that prevalence rates of having smoked in the last month were 0.0, 14.9, and 11.3% among those who migrated \<15, 15 to 30, and \>30 years prior to the baseline assessment (Fisher's exact *p* = 0.244). Accordingly, incidence rates were 0.0, 3.5, and 1.3% in these sub-groups (Fisher's exact *p* = 0.334). These findings suggest a higher smoking risk at 15 to 30 years of migration.
Heavy drinking {#Sec20}
--------------
The prevalence of heavy drinking in the last year was similar between the urban, migrant, and rural groups. Accordingly, studies that evaluated alcohol intake in Guatemala \[[@CR14]\] and alcohol dependence in Canada \[[@CR40]\], found similar patterns between rural-to-urban migrant and rural groups. However, data from Tanzania \[[@CR5]\] found that weekly alcohol consumption prevalence increased after migration, and studies assessing monthly drinking in Vietnam \[[@CR7]\] and alcohol intoxication in China \[[@CR15]\] found that rural-to-urban migrants had higher rates than the urban population.
Three previous studies have compared alcohol intake rates between the rural-to-urban migrants, urban, and rural groups. One of them \[[@CR6]\] made in Yi migrants (China), found that migrants had similar prevalence rates of current alcohol use than rural and urban dwellers. The other study, also in China \[[@CR16]\], surveyed migrants recruited in workplaces and reported that they had higher alcohol intoxication rates than rural and urban dwellers. While the first study resembles our results, the second did not, possibly because their participants were workers, younger (mean age 25 years), and therefore possibly more prone to alcohol drinking, than our migrant group with mean age of 48 years. The third study found similar alcohol use between rural and migrant groups in Ghana, India, Mexico, Russia, and South Africa, while migrants had lower alcohol use than rural dwellers in China \[[@CR17]\].
We only found one longitudinal study that evaluated alcohol intake in rural-to-urban migrants in Tanzania, which found that weekly alcohol consumption has a non significant increase after migration \[[@CR5]\]. In our longitudinal analysis, incidences of heavy drinking were similar between the three study groups. As for smoking, we verified whether heavy drinking increases after the first years of migration in post-hoc analysis in our migrant group, and found that prevalence rates of heavy drinking were 18.2, 7.1, and 8.6% among those who migrated \<15, 15 to 30, and \>30 years previous to baseline assessment (Fisher's exact *p* = 0.270). However, incidence rates were 0.0, 3.9, and 0.8 in these sub-groups (Fisher's exact *p* = 0.099). These findings, along with the Tanzania study, are consistent with a non-significant increase in heavy drinking in the first years after migration.
Subjects in our rural and urban settings, although living in different socio-environmental contexts, had similar heavy drinking rates. This may reflect similar heavy drinking manners in our urban and rural populations, along with similar alcohol access besides economic differences, probably because low resource individuals in our settings can purchase low-cost artisanal alcoholic drinks \[[@CR30]\]. Some studies suggest that recent migrants may present migration-related psychological distress, which was associated to higher alcohol consumption \[[@CR41], [@CR42]\]. However, PMH, a mental health status evaluation, was not associated to heavy drinking in late-term migrants.
Classically, smoking has been associated with alcohol intake and with heavy drinking \[[@CR43], [@CR44]\]. However, in Peru, alcohol intake prevalence is much higher than smoking prevalence \[[@CR45]\], and our results show that increasing smoking rates in migrants are not accompanied by an increase in heavy drinking rates. A possible explanation is that low-resource populations do not have enough money to buy cigarettes, but they can manufacture low-cost artisanal alcohol drinks \[[@CR30]\]. Accordingly, asset index was associated with the prevalence of lifetime smoking, but not with the prevalence of heavy drinking.
Public health relevance {#Sec21}
-----------------------
Our results present smoking and heavy drinking patterns in a rural-to-urban internal migration in Peru, which may be similar to other rural-to-urban internal migrations in Peru and other developing countries.
Our results indicate that migrants are at risk to increase their smoking patterns, especially in the first years after migration. Thus, smoking interventions in migrant populations appear more beneficial if oriented to prevent smoking initiation rather than cessation. These observations are not against major tobacco control policies that remain to be sustained as beneficial public health policies at the country-level \[[@CR46]\].
Post-hoc analyses show not significant lower smoking rates and higher heavy alcohol drinking rates among those who migrated in the past 15 years. Future studies in recent migrants could identify which would be the best moment for preventive interventions in smoking and alcohol consumption.
It is also important to take into account that Peruvian rural settings could have a higher use of artisanal alcohol distilled drinks with high alcohol concentration \[[@CR30]\], which may have a higher concentration of aliphatic alcohol \[[@CR47]\], and therefore represent an additional risk of liver damage \[[@CR48]\].
Strengths and limitations {#Sec22}
-------------------------
This study has assessed smoking and heavy drinking in well-defined rural, urban, and migrant populations. This allows a better understanding of the influence of rural--urban migration in the consumption patterns, and can be used to improve health interventions targeted towards these migrants.
However, some limitations deserve consideration. First, all the variables studied were self-reported, with the inherent social desirability bias. Nevertheless, previous studies in other countries reinforce use of self-reporting as a reliable method to measure the smoking status \[[@CR49]--[@CR51]\] and alcohol consumption \[[@CR52], [@CR53]\] in the general population. Second, there are some confounders that we could not address, such as smoking/drinking status before migration, or reason of migration. Finally, some studies have reported that female migrants may be at higher risk of cigarette \[[@CR54], [@CR55]\] and alcohol consumption \[[@CR56]\] compared to males. However, since we did not have enough cases to stratify by sex, we could not explore this.
In addition, we have to highlight that the smoking status categorization we used in this study is primarily driven by frequency of consumption (having consumed cigarettes in the last month), without considering the amount component (how many cigarettes have been smoked). Thus, in our population, in which the number of cigarettes consumed is low, current smokers will have a lower smoking-related risk than current smokers in other countries with high smoking prevalence \[[@CR57]\]. Also, our results suggest an overestimation of current smoking rates among rural dwellers when using definitions based on "occasionally/daily smoking". Thus, studies using this definition may find different smoking rates and different measures of association than those found in our study.
Conclusion {#Sec23}
==========
Our results show that migrants have a higher smoking prevalence than the rural population, but lower than the urban population, suggesting a potential protective effect of being exposed to a rural background. However, smoking incidence was similar in rural and migrant groups and higher in the urban group, suggesting that despite their current exposure to urban environments, migrants do not align with their urban peers. In addition, our results suggest that different definitions of smoking status could lead to different smoking rates and potentially different measures of association. On the other hand, heavy drinking prevalence and incidence were similar across all three population groups, suggesting similar heavy drinking culture and alcohol access between our urban and rural settings.
Additional file {#Sec24}
===============
Additional file 1: Table S1.Outcomes of interest and evaluated co-variables. (DOCX 17 kb)
95% CI
: 95% confidence intervals
GHQ-12
: General Health Questionnaire
IQR
: Interquartile ranges
PMH
: Positive mental health
PR
: Prevalence ratios
RR
: Risk ratios
SD
: Standard deviations
Not applicable.
Funding {#FPar1}
=======
The PERU MIGRANT Study baseline assessment work was funded through by a Wellcome Trust Master Research Training Fellowship and a Wellcome Trust PhD Studentship to JJM (Grant number 074833/Z/04/A), and its follow up by Universidad Peruana Cayetano Heredia (Fondo Concursable No. 20205071009). AB-O is supported by a Wellcome Trust Research Training Fellowship in Public Health and Tropical Medicine (Grant number 103994/Z/14/Z).
Availability of data and materials {#FPar2}
==================================
The baseline database is available in <https://figshare.com/articles/PERU_MIGRANT_Study_Baseline_dataset/3125005>. The follow-up data is available under request.
Authors' contributions {#FPar3}
======================
ATR, ABO, GFA, and JJM designed the study and performed the statistical analysis. RHG and LM helped with the data interpretation. All authors were major contributor in writing the manuscript and gave final approval for manuscript submission.
Competing interests {#FPar4}
===================
The authors declare that they have no competing interests.
Consent for publication {#FPar5}
=======================
Not applicable.
Ethics approval and consent to participate {#FPar6}
==========================================
Ethical approval for the baseline study was obtained from Institutional Review Boards at Universidad Peruana Cayetano Heredia, in Lima, Peru and the London School of Hygiene and Tropical Medicine, in London, United Kingdom. The follow-up phase was reviewed and approved by the same Peruvian institution. All enrolled participants gave written informed consent.
| {
"pile_set_name": "PubMed Central"
} |
Introduction
============
This paper describes a set of fundamental concepts from causality theory that can be used to critically analyze summary measures of population health. It then uses these concepts to argue that health measures based on hypothetical outcome removal \[[@B1],[@B2]\] are ambiguous and potentially misleading. Thorough analyses require a multivariate framework to capture what is known about sources of morbidity and mortality. Because of the unavoidable shortcomings of single-valued summaries, multidimensional indices should be considered for summary measures of population health.
The first task is to define cause and effect in a manner precise enough for logical manipulation and quantification. Three types of models have achieved widespread usage:
• *Counterfactual or potential-outcome models*. The counterfactual conceptualization of causality was glimpsed by Hume \[[@B3]\] and was operationalized by statisticians in the 1920s and 1930s in the form of potential-outcome models \[[@B4]\]. Related counterfactual models have also received much attention in philosophy \[[@B5],[@B6]\], and potential outcome models are now common in social sciences \[[@B7],[@B8]\] and epidemiology (see citations below) as well as in statistics.
• *Structural-equations models*. These models can be traced to early work in path analysis in the 1920s and were given full expression by econometricians in the 1940s.
• *Graphical models*(causal diagrams). These models also originated in path analysis, but were not fully developed until the early 1990s \[[@B9]-[@B11]\].
In his book on causal theory, Pearl \[[@B11]\] details the above approaches and their histories, emphasizing that they are logically equivalent (isomorphic) for all practical purposes. This means that an analysis using one approach can be translated into either of the other two approaches, while maintaining logical consistency. Greenland and Poole \[[@B12]\] and Greenland and Brumback \[[@B13]\] describe the connections between potential outcome models and the sufficient-component cause models familiar to epidemiologists, and Greenland \[[@B14]\] discusses potential outcome models in relation to more traditional approaches to causal inference.
Because of the emphasis on counterfactual models in the literature on measures of cause and effect, the following development is based on them. Further details of counterfactual theory for health science research are described elsewhere \[[@B12],[@B14]-[@B26]\]. For details on its relation to missing-data models, see Rubin \[[@B27]\]. The present paper concerns only issues of defining effects and their implications for policy formulation. Equally important is quantitative accounting for study biases in estimating effects; see Greenland \[[@B28],[@B29]\], Lash and Fink \[[@B30]\], and Phillips \[[@B31]\] for recent developments in that topic.
Basic concepts of counterfactual causality
==========================================
Actions, outcomes and counterfactuals
-------------------------------------
To minimize ambiguity, a counterfactual model requires reasonably precise definitions of the following model ingredients:
• At least one *target*subject of interest in whom or which causation is to be studied (e.g. a specific person or population).
• A list of two or more possible alternative actions, *x*~0~, *x*~1~, etc., that could have been applied at or over a span of time, one of which may be the actual action taken.
• An outcome measure, *Y*, taken at a point in time or over a period of time, following completion of the action.
As an example, the subject could be Kosovo, the action *x*~1~could be revocation of its autonomy by Yugoslavia in 1988, an alternative *x*~0~could be that revocation never occurred, and the outcome measure could be the mortality rate of pre-1999 residents during 1999. Because only one of the possible actions *x*~0~, *x*~1~, etc. can take place, all but one of the actions must become *counterfactual*, or contrary to fact. In the example, the actual (or factual) action was *x*~1~= \"revocation of autonomy\"; thus, the action *x*~0~= \"no revocation\" is counterfactual. (This example also illustrates the truism that to do nothing is an action, corresponding to zero on an action scale.)
If *x*~*a*~is the actual action taken (from the list *x*~0~, *x*~1~, etc.), we may observe the outcome *Y*(*x*~*a*~) that follows that action. A *counterfactual outcome model*posits that, for any counterfactual action, *x*~*c*~(from the same list), there is also a well-defined outcome, *Y*(*x*~*c*~), that would have followed *that*action. The entire list of such outcomes *Y*(*x*~0~), *Y*(*x*~1~), *Y*(*x*~2~), \... is called the set of *potential outcomes*of the subject; it includes both the actual outcome *Y*(*x*~*a*~) as well as all the outcomes of counterfactual actions.
In the example, the actual action, *x*~*a*~, was *x*~1~and was eventually followed by a mortality rate, *Y*(*x*~1~), in 1999. *Y*(*x*~1~) is difficult to assess but nonetheless exists as a matter of fact; it is part of what occurred subsequently to the action *x*~1~. A counterfactual model for 1999 mortality posits that, had the counterfactual action *x*~0~been taken instead (that is, if no revocation had occurred), the mortality would have equaled some number, *Y*(*x*~0~). Given that *x*~0~is counterfactual, it is logically impossible for us to observe *Y*(*x*~0~). Nonetheless, the counterfactual model treats this number as a precise, though unknown, quantity.
The idea of treating the outcome, *Y*(*x*~*c*~), under a counterfactual action, *x*~*c*~, as a precise quantity has been a source of much controversy and misunderstanding (e.g. \[[@B23],[@B32],[@B33]\]). Some major misunderstandings are addressed below:
1\) The counterfactual approach does *not*require that the outcome, *Y*(*x*~*c*~), be precisely defined for every possible counterfactual action, *x*~*c*~. In the above example, if we are interested only in contrasting revocation (*x*~1~) with no action (*x*~0~), our model need not mention any other actions. That is, we can limit the action list to just those actions of interest.
2\) The counterfactual approach is *not*inherently deterministic in either the classical or quantum-mechanical sense \[[@B15],[@B34]\]. The potential outcomes, *Y*(*x*~0~), *Y*(*x*~1~), etc., may represent different values for a statistical parameter in a classical probability model. For example, they may be expected rates in Poisson models. Alternatively, they may represent different mixtures of superposed states (different wave functions) in quantum models. Indeed, some theoreticians regard counterfactuals as essential for formulating coherent explanations of quantum phenomena \[[@B35]\].
3\) The counterfactual approach extends the partial quantification of outcomes, *Y*(*x*~*c*~), under counterfactual actions embedded in ordinary discourse. In the example, some (though not all) observers of the events in Kosovo 1988--1999 speculated that the actual 1999 mortality, *Y*(*x*~1~), was probably greater than *Y*(*x*~0~), the mortality that would have occurred had autonomy never been revoked. This speculation arises from the following tentative explanation of actual events in Kosovo: revocation of autonomy, (*x*~1~), caused Albanian resistance to increased Serb authority, which in turn caused Serbian leaders to extend their \"ethnic cleansing\" policy to Kosovo. Had there been no revocation, (*x*~0~), this tragic causal sequence of events would not have occurred.
Cause and effect
----------------
The speculative explanation in the third bulleted item above is an example of an informal causal hypothesis. Consideration of such hypotheses has led to the following definition: An *effect*of taking an action, *x*~*j*~, rather than another action, *x*~*k*~, on an outcome measure, *Y*, is a numerical contrast of that measure (e.g. the difference or ratio) under the two different actions. The contrast is called an *effect measure*. In the example, the contrast *Y*(*x*~1~) - *Y*(*x*~0~) is an effect of revocation *x*~1~versus no revocation *x*~0~; this effect measure is known as the mortality difference due to x~1~versus *x*~0~. Similarly, *Y*(*x*~1~) / *Y*(*x*~0~) is the effect measure known as the mortality ratio due to x~1~versus *x*~0~.
Many common ideas and a few surprises follow from the above definitions, among them:
4\) An effect is a *relation*between the outcomes that would follow *two*different actions, *x*~*j*~and *x*~*k*~, in just *one*subject (a population or single person). It is thus meaningless to talk of (say) \"the effect of smoking a pack a day\"; one must at least imply a reference (baseline) action for \"the effect\" to have meaning. While smoking a pack a day can cause lung cancer relative to no smoking, it can also prevent lung cancer relative to smoking two packs a day.
5\) If *Y*(*x*~*j*~) = *Y*(*x*~*k*~), we say that having *x*~*j*~happen rather than *x*~*k*~had no effect on *Y*for the subject; otherwise, we say that having *x*~*j*~happen rather than *x*~*k*~caused the outcome to be *Y*(*x*~*j*~) and prevented the outcome from being *Y*(*x*~*k*~). For example, we may say that smoking prevents survival past age 70 years just as surely as it causes death by age 70. Similarly, we may say that not smoking causes survival past age 70 just as surely as it prevents death by age 70. Thus, the distinction between causation and prevention is merely a matter of whether we are talking of an action, *x*~*j*~, and *its*consequence, *Y*(*x*~*j*~) (causation of *Y*(*x*~*j*~)), or an action, *x*~*j*~, and a consequence, *Y*(*x*~*k*~), of an alternative action *x*~*k*~≠ *x*~*j*~(prevention of *Y*(*x*~*k*~)).
6\) At least one of the actions, *x*~*j*~, *x*~*k*~, in an effect measure must be counterfactual. Thus, we can never observe an effect measure separate from an outcome measure. In the example, we observed the mortality *Y*(*x*~1~) = *Y*(*x*~0~) + \[*Y*(*x*~1~) - *Y*(*x*~0~)\], so the mortality difference, *Y*(*x*~1~) - *Y*(*x*~0~), is mixed with the reference (baseline) mortality rate, *Y*(*x*~0~), in our observation. The best we can do is make an informed estimate of *Y*(*x*~0~), which is the outcome that would have happened under the counterfactual action *x*~0~, and from that estimate deduce an estimate of the effect measure (by subtraction, in this example).
Causation, confounding and association
--------------------------------------
Problem 6 is considered a fundamental problem of all causal inference. It was recognized by Hume \[[@B36]\] and is now known as the *identification problem*of cause and effect. All causal inferences (and hence all intervention plans) depend on accuracy in estimating or predicting at least one unobserved potential outcome following one counterfactual action. We ordinarily make this prediction based on observations of other subjects (controls) who experienced actual actions different from the subject of interest. For example, we might estimate that the mortality Kosovo would have experienced in 1999 had there been no revocation, *Y*(*x*~0~), would equal the mortality it experienced in 1988, before violence began growing. In making this estimate, we run the risk of error because, *even under the action*x~0~(no revocation), Kosovo mortality might have changed between 1988 and 1999. If so, we say that our estimate is *confounded*by this unknown change.
Denote by *Y*~1988~the mortality experienced by Kosovo in 1988. We can then restate the last problem as follows: we do not observe *Y*(*x*~0~), so we cannot directly compute a measure of the effect of *x*~1~versus *x*~0~. If, however, we believed the speculative explanation given in the above third bulleted item, we might also think that *Y*~1988~is not too far from *Y*(*x*~0~), and so substitute *Y*~1988~for *Y*(*x*~0~) in our measures. Thus, if we also observe *Y*(*x*~1~), the actual 1999 mortality, we would estimate the effect measure *Y*(*x*~1~) - *Y*(*x*~0~) with the observed mortality difference *Y*(*x*~1~) - *Y*~1988~.
The latter observed difference is called a *measure of association*, because it contrasts two different *subjects*(Kosovo in 1999 versus Kosovo in 1988), rather than one subject under two different *actions*in our list (Kosovo in 1999 under revocation versus Kosovo in 1999 with no revocation). Because of the identification problem (we cannot see *Y*(x~0~)), we must substitute a measure of association for the measure of effect. In this usage, the observed difference will misrepresent the effect measure by an amount equal to the difference of the two:
\[*Y*(*x*~1~) - *Y*~1988~\] - \[*Y*(*x*~1~) - *Y*(*x*~0~)\] = *Y*(*x*~0~) - *Y*~1988~
This quantity measures the amount of *confounding*in the association measure (the observed difference) when it is used as a substitute for the effect measure. Like the effect measure itself, the confounding measure contains the unobserved *Y*(*x*~0~) and so can only be estimated, not observed directly. Suppose, however, that we know of reasons why *Y*(*x*~0~) and *Y*~1988~would differ, such as changes in age structure over time. We can then attempt to adjust *Y*~1988~for these suspected differences, in the hopes of getting closer to *Y*(*x*~0~). *Standardization*is probably simplest example of such adjustment \[[@B18],[@B26]\].
The presumption underlying use of an adjusted effect measure is that it accounts for all important differences between the unobserved (counterfactual) reference outcome, *Y*(*x*~0~), and the substitute, *Y*~1988~, in the above example. The presumption is debatable in most applications; for example, some would argue that \"ethnic cleansing\" would have spread to Kosovo even without autonomy revocation and Albanian resistance. This problem of uncontrolled confounding is but one of many methodological problems in estimating effects that are discussed in textbooks (e.g. \[[@B26]\]).
The effects of outcome removal
==============================
Consider a question that asks about the health burden attributable to *y*~1~versus *y*~0~, where *y*~1~and *y*~0~are not actions in the earlier sense, but are themselves alternative *outcomes*such as AIDS death and CHD death. For example, *y*~1~could be \"subject dies of lung cancer\" and *y*~0~could be \"subject does not die of lung cancer\". As in the earlier framework, these outcomes are mutually exclusive possibilities for just one subject at any one time; hence, at least one must become counterfactual. Because they are not interventions, however, there is severe ambiguity in any definition of another outcome, *T*, as a function of the potential outcomes, *y*~1~and *y*~0~, because *T*depends in a critical fashion on *how y*~1~and *y*~0~are caused.
To see this, suppose *T*is years of life lived beyond age 50 (which is age at death minus 50). How would one have brought about *y*~0~(prevented the lung-cancer death) if the subject were a male lifelong heavy smoker who developed lung cancer at age 51 and died from it at age 54 (and so had *T*(*y*~1~) = 4 years of life after age 50)? If *y*~0~had been achieved by convincing the subject to never start smoking, *T*(*y*~0~) could be much larger than *T*(*y*~1~), because the risks of many other causes of death (e.g. CHD) would have been much lower as a consequence of never smoking. But if *y*~0~had been achieved via an unusually successful new chemotherapy for lung tumors, *T*(*y*~0~) might be little changed from *T*(*y*~1~). This would occur if, shortly after remission, the subject had a fatal myocardial infarction whose occurrence was traceable to smoking-induced coronary stenosis.
The problem just described has long been recognized in discussions of estimating the impact of \"cause removal\" or \"removal of competing risks\" when the \"causes\" or \"risks\" at issue are outcomes rather than actions or treatments \[[@B37]\]. These outcomes are not subject to direct manipulation independent of the earlier history of the subject. Therefore, any realistic evaluation of the impact of their removal must account for other effects of the *means*of removal.
A similar problem arises in the evaluation of ordinary treatments whenever noncompliance can occur. In general, only advice or prescriptions are under control of the health practitioner; what a patient actually receives is affected not only by advice or prescription, but also by the many complex social and personality factors that influence compliance. This leads to manifold problems in evaluating the effects of received treatment \[[@B38]\], for then the treatment a subject receives is only an outcome, *Y*(*x*~*j*~), of an earlier prescriptive action, *x*~*j*~. In most cases, however, this initial action is unambiguous.
Suppose we could avoid the ambiguity problem by introducing a pair of well-defined alternative actions, *x*~1~and *x*~0~such that *x*~1~causes *y*~1~and prevents *y*~0~relative to *x*~0~. That is, suppose *y*~1~will follow *x*~1~, whereas *y*~0~will follow *x*~0~, so that we have *y*~1~= *Y*(*x*~1~) and *y*~0~= *Y*(*x*~0~) with *y*~1~≠ *y*~0~. We may still face a serious confounding problem in the form of \"dependent competing risks\". Consider again the heavy smoker who develops lung cancer at age 54, with treatment, *x*~0~, being successful chemotherapy. It could be a mistake to calculate this subject\'s life expectancy, *T*(*x*~0~), from that of other heavy smokers of the same age and sex who had not developed lung cancer, because such smokers may differ in ways that render them not only less susceptible to smoking-induced lung cancer, but also less susceptible to other smoking-induced cancers (perhaps because they have better DNA-repair mechanisms).
More generally, even if the means of removal is precisely defined, feasible and has no side effects, there is rarely a basis to believe, and often good reason to doubt, that removal of a particular outcome (such as a particular cause of death) would be followed by risks similar to risks among persons who, in the absence of intervention, do not experience the removed outcome \[[@B37],[@B39]\]. Unfortunately, standard statistical procedures for projecting outcomes under cause removal (such as Kaplan-Meier/product limit methods and traditional \"cause-deleted\" life tables) are based on this similarity assumption.
In view of the problems just described, it is reasonable to conclude the following:
7\) Projections of the impact of outcome removal (e.g. removal of a particular ICD9 cause of death \[[@B1]\]), rather than an action that brings about outcome reduction, may not be useful for program planning. Except perhaps in some unusually simple cases (e.g. smallpox eradication), the effects of actions and policies do not correspond to simple cause removal.
8\) Even when we have a treatment that specifically and completely prevents an outcome, biased effect estimates are likely if one simply projects the experience of those who naturally lack the outcome onto those who avoid the outcome because of the treatment. Only ongoing follow-up of successfully treated subjects can reliably identify the impact of outcome removal.
Problem 7 implies that summary measures for policy formulation should refer to effects of operationalizable actions (e.g. anti-smoking campaigns, food-distribution programs), rather than effects of removing the outcomes targeted by those actions (e.g. smoking, cancer, malnutrition). Only rarely will the two effects coincide. Focusing on the outcome removal presents a grossly overoptimistic picture of what can actually be accomplished, since the latter is determined by what feasible interventions are available. Focusing on outcome removal has the potential of diverting resources away from where it will do the most good -- outcomes with feasible and effective preventives -- toward outcomes that, while more common and costly, have less hope of successful and economical prevention. Finally, a focus on outcome removal diverts attention from assessing and comparing the full impacts of interventions. For example, even partial reduction in tobacco use will result in a broad spectrum of outcome prevention, from heart disease to lung cancer, whereas an effective treatment for lung cancer would only reduce the burden from that disease while raising the burden from tobacco-related risks.
The preceding consideration raises another point: Because any action will have multiple consequences, a thorough analysis must consider outcomes in a *multivariate*framework that accounts for the multiple effects of actions and the competition among various outcomes. This multivariate perspective raises serious questions about the value of univariate summaries, which will be taken up after the next section.
Are socioeconomic indicators causes?
====================================
The theory outlined above is strictly a theory of effects of *actions*or *interventions*. It does not formalize all ordinary-language or intuitive uses of the words \"cause\" and \"effect\". Two naïve extreme reactions to this limitation have been common: one that denies it is meaningful to talk of causes that are not actions and so restricts \"causes\" to interventions \[[@B33]\], and one that rejects counterfactual theory outright (see the discussion of Maldonado and Greenland \[[@B23]\]). But two types of constructive reactions have also appeared. The first type generalizes the theory to encompass nonactions as causes, a prime example being the many-worlds theory \[[@B5]\]. While this generalization may best capture ordinary language, it is very controversial and not suitably operationalized for everyday use.
The second constructive response accepts the limitations of the restricted theory and instead seeks to identify potential actions within ordinary events. This approach recognizes that certain \"causes\" are best treated as intermediate outcomes; one then traces the etiology of such \"causes\" back to events with intervention potential, or else treats such \"causes\" as conditioning events and searches for actions that modify or prevent the ultimate outcomes. Earthquakes, which cause extensive destruction, provide neutral examples of unmodifiable causes. An earthquake, *y*~1~, is the outcome of a long and complex chain of events with little intervention potential under today\'s technology. Perhaps someday we will be capable of interventions that lead to dissipation of crustal stresses with less ensuing damage. But for now, mere prediction would be a major achievement and would facilitate actions to prevent damage when an earthquake occurs. An example of such an action is the enforcement of strict building codes in earthquake-prone regions.
Less neutral examples are provided by common measures of education, such as the classification \"No high-school diploma\", \"High-school diploma\", \"Some college, no degree\", \"Two-year degree\", \"Four-year degree\" and \"Graduate degree\". People believe that education leads to more income and health. But how well do differences in the observed education measure predict the effects of real interventions such as affirmative action, public-school improvements, or scholarship programs? For policy purposes, it is the implementation and evaluation of such programs that matter; the ensuing changes in education measures are only intermediates between the programs and the ultimate goals of improved social, economic and health outcomes.
The value of restricting counterfactual models to interventions is that it forces us to explain observed *associations*between risk factors and health as *outcomes*of potentially changeable events. Consider a highly charged example, \"race\", which when measured as \"white\" vs. \"black\" is strongly associated with many health events. People talk of race as a \"cause\". But to do something about racial disparities in health outcomes (which is to say, to eliminate the observed association of race and health), we must explain their origin in terms of changeable causes, such as disparities in school funding, availability of individual college funding, prevalence of racist attitudes, etc. Finding feasible interventions and estimating their costs and benefits is required to address observed disparities; asserting or denying that \"race\" is a cause does not help this endeavor.
Should different outcomes be summarized in a single number?
===========================================================
Two distinct connotations of summary measure appear extant: the first and most common presumes that the measure summarizes a single outcome variable with a single number. Classic examples include the mortality rate and the life expectancy. The second connotation, largely confined to statistics and physical sciences, allows a summary to be a vector, that is, an ordered list of numbers that summarize different dimensions of a system. An example of such a *multidimensional*or *multivariate*population summary would be the list containing life expectancy, health expectancy, health gap and the proportions of deaths due to various causes (e.g. starvation, violence, infectious disease, heart disease, stroke, cancer).
It should first be noted that all the earlier concepts and discussion apply equally to any action or outcome, whether unidimensional or multidimensional. In particular, the potential outcomes, *Y*(*x*~*j*~), may represent outcome vectors and the alternative actions, *x*~0~, *x*~1~, etc., may also be vectors; for example, *x*~0~could specify that 30%, 40% and 30% of a fixed budget be allocated to family planning, sanitation and medical supplies, respectively, and *x*~1~specifies a different allocation scheme. The chief problem in expanding to the multidimensional perspective is the limited number of dimensions that the human mind can contemplate at once. Because that limitation is a key motive for summarization, it is essential to keep track of what is lost in the dimensionality reduction that defines summarization. It also is essential to keep track of the *values*that influence (or should influence) what is kept and what is lost.
Summary measures of population health serve no good purpose when they strongly confound valuations, which vary by individual preference, culture, etc., with measures of occurrence and effect (which are presumably matters of scientific fact, albeit subject to uncertainty). For example, many individuals, in continuing to smoke, explain their behavior as stemming from a conscious preference to die sooner from cardiovascular disease or cancer than survive until mental or neurological deficit is nearly inevitable. For such individuals, measures such as healthy years of life lost due to smoking represent a conflation of someone else\'s values with the factual risks of smoking, because that summary ignores preferences among various morbidity and mortality outcomes affected by smoking. To give the individual the information necessary for personal choice, we must supply a multidimensional summary that includes lifetime risks of different diseases.
Moving to the societal level, healthy years of life lost due to smoking not only neglects the differences in resource allocation that must exist between present (actual) society and a counterfactual tobacco-free society, it also neglects differences in absolute and proportional morbidity and mortality with and without tobacco use. This neglect is addressed by measures of the economic cost of tobacco use, and by absolute and proportional morbidity and mortality comparisons. By providing all these measures, we shift to a multidimensional summary of tobacco impact.
My intention in raising these issues is not to offer a solution to a specific summarization problem. Rather, it is to remind those facing a choice among measures that candidates need not (and, for policy purposes, should not) be limited to unidimensional summaries. While our ability to think in several dimensions is limited, it can be improved with practice. That practice has proven crucial in attacking problems in physics and engineering, and there is no reason to suppose it is less important in tackling more complex social policy issues. In instances in which many different people must make informed choices based on the same scientific data, but with different values, multidimensional measures are essential if we are to provide each person and each executive body with sufficient information for rational choice.
Conflicts of interest
=====================
The author(s) declare that they have no competing interests.
Acknowledgements
================
This article is an updated version of Greenland, S. (2002), Causality theory for policy uses of epidemiologic measures, which appeared as Ch. 6.2 in Murray et al. \[[@B2]\].
| {
"pile_set_name": "PubMed Central"
} |
Background {#Sec1}
==========
Constrictive Pericarditis (CP) is a potentially curable cause of diastolic heart failure. Known and common etiologies include previous heart surgery, chest radiation, trauma and infections such as tuberculosis. Conversely, a proportion of the cases may not have well identifiable causes, thus deemed idiopathic. Early identification is cardinal, as timely pericardiectomy is associated with lower operative complications. Clinically, timely diagnosis can be challenging especially in the absence of radiological signs. With high suspicion index, careful evaluation of jugular venous pressure and its wave forms aids diagnosis. Additionally, modern echocardiography machines in experienced hands permit confirmation of clinically suspected in majority cases.
Case presentation {#Sec2}
=================
A case of C.X.Z, male aged 39-year old, and farmer by profession was wheeled into our department with severe symptoms of subacute RHF. His spouse narrated that in January,2018, he had experienced mild bilateral swelling of lower limbs (in form of stockings), and was managed on diuretics for about 7 days, after which symptoms disappeared completely. She denied him having had any cardiac surgery, chest radiation, tuberculosis or significant chest trauma. 8 months after initial symptoms, thus in October,2018, he suddenly developed chest pain, which he thought was due to long working hours in the field. On-counter remedies (pain killers) offered temporal relief. After 2-days of progressive chest pain, patient begun experiencing abdominal discomfort and observed swelling of feet after bed. On the 4th day in his illness, he developed shortness, a development that prompted him seek medical attention.
On presentation the patient through his spouse complained of breathing difficulties, abdominal fullness and swelling of lower limbs. She further narrated that, during bed time, shortness of breath worsened upon lying flat. During physical examination, patient exhibited incoherent talk, responded to various questions with same answer repeatedly. Both the neck veins (JVD\~ \> 15mmH2O) and abdomen were highly distended. Chest auscultation demonstrated a 'cardiac knock', and both S~1~ and S~2~ were muffled. Abdominal palpation revealed gross ascites. The lower extremities were cold to touch with bilateral pitting edema from knee and below. Prior and post procedure vitals are tabulated in Table.[1](#Tab1){ref-type="table"}.Table 1Pre−/post operation hemodynamic and respiratory indicesVariableRespirationPulseBlood pressureEjection fractionPre-operation32 r/m160b/m78/40 mmHg50%Post operation20r/m84b/m130/62 mmHg61%Respirations per minute,r/m.beats per minute/millimeter Mercury,mmHg.
Diagnostic Procedures {#Sec3}
---------------------
Diagnosis of localized CP was established using cogent imaging results of comprehensive transthoracic echocardiography (TTE) and computed tomography(CTA). A 4 chamber video clip (Additional file [1](#MOESM1){ref-type="media"}) of a 2D TTE examination demonstrates dyskinesia of the right ventricle(RV) due to the presence of a thickened (calcified) pericardium(cyst-like) on its anterior wall. Other visible abnormalities include: right atrium enlargement, respirophasic interventricular deviation into left ventricle and shudder, scientifically proven to be due to the differential filling rates of both ventricles in diastole. A comprehensive TTE report (not shown here) demonstrated hepatic vein distension, a patent inferior vena cava (with no respirophasic variation) and ejection fraction(EF) of 50%. The chest radiography appeared normal (Panel A). However, CTA uncovered two adjacent pads of calcification with interposed fluid: panels B and C (Fig. [1](#Fig1){ref-type="fig"})**.**Because CP is relatively common in our region,radiology personnel have gained experience in diagnosing CP with occasional application of the Swan-Ganz catheter.Fig. 1Imaging results, Panel A chest radiography, panels: B & C CTA. Panel A. Chest radiography, Panels B and C CTA with solid arrows demonstrating calcification & fluid
Additional file 12D 4CTTE video. (AVI 11827 kb)
Decortication {#Sec4}
-------------
Through standard midline sternotomy, an off-pump partial pericardiectomy was performed from cardiac surgery operation room. Using a combination of cautery, scissors and sharp hooks, we approached the pericardium from the free RV wall (junction with right atrium). After 30 to 40 min into decortication process, we judiciously punctured a plate of calcified tissue overlaying the anterior RV wall with subsequent gush of a ῾milk-like' fluid(approx.700mls). Fluid decompression led to adjustments of ventriculae pressures and cardiac out-put. The video in (Additional file [2](#MOESM2){ref-type="media"}) shows the heart beating with less obstruction. Fluid sample results and pericardium tissue pathological report are depicted in (Table [2](#Tab2){ref-type="table"}) and (Fig. [2](#Fig2){ref-type="fig"}) respectively.Table 2milk-like fluid analysis reportFig. 2Pericardium tissue pathological report
Discussion {#Sec5}
==========
Constrictive pericarditis(CP) can be viewed as a constellation of syndromes resulting from compression of the heart, etiology course and types are well discussed in other reports \[[@CR1], [@CR2]\]. However, localized CP, as a cause of acute right heart failure(RHF) is rare, and presentation with interposed fluid under-pressure is extremely odd \[[@CR3], [@CR4]\]. Our case well presented in both still images and video, demonstrate two adjacent pads responsible for pathological symptoms. When viewed from apical angle in sagittal plane, the two pads assume the shape of two jaws mouth swallow the heart, hence, the title "Heart in the 'JAWS' of a constrictor' 'We acknowledge that this particular case is worthy differentiating from pectus excavatum \[[@CR5]\].
Our case was unusual in both course and imaging presentation. Usually, CP is known to take a chronic course, and medical therapy take a palliative role in majority cases. Surprisingly enough, the right heart pump was suddenly failing in someone with a very unremarkable past medical history. The chest radiography did not show obvious radiological clue 'egg-shell' commonly observed in majority cases. At this juncture, we can only appreciate the advancement in imaging modalities that timely diagnosis was made.
Pericardiectomy(partial/complete) is associated with a myriad complications including death. In a recent European study (Busch et al.) RV dilation and EF \< 55% were cited as risk factors for(early/late) mortality after pericardiectomy for chronic CP. Despite EF of 50% the patient experienced spontaneous recovery of cardiac functions due to normal ventriculae dimensions and absence of associated lesions (tricuspid /septal). Post operation therapy was centered on pain management and liquidation of edema and ascites using low doses of Spironolactone and Furosemides. The patient was discharged home free of symptoms (NYHA-Class I) 7 days after operation.
Conclusion {#Sec6}
==========
In view of both course and presentation, when faced with acute RHF in the absence of chest radiographic signs of CP, it's suspicion of index must be heightened and advanced imaging modalities sought. Otherwise, patient could be subjected to medical therapy with catastrophic outcome. We believe this case is very unusual and could aid accurate diagnosis of similar cases.
Additional files
================
{#Sec7}
Additional file 2:Video-Immediate post fluid decompression. (FLV 1567 kb)
CP
: Constrictive pericarditis
CT
: Computed tomography angiography
EF
: Ejection fraction
RHF
: Right heart failure
TTE
: Transthoracic echocardiography
**Publisher's Note**
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
**N**one.
GJC coined the concept, drafted and edited final manuscript, ZW was main surgeon and coordinated both imaging and laboratory results, while HZ analyzed laboratory results and translation. Finally, CZ proof read and mobilized financial resources. All authors read and approved the final manuscript.
Publication fee supported by National key R&D program of China(2017YFC1308000).
Available on appropriate request from author.
Hospital Ethics committee approved publication.
Patient was informed and written consent obtained.
The authors declare that they have no competing interests.
| {
"pile_set_name": "PubMed Central"
} |
Introduction: Human Hepatocellular Carcinoma {#s1}
============================================
Human hepatocellular carcinoma (HCC) is one of the most frequent and pernicious solid tumors, ranking fifth in incidence and second in lethality worldwide ([@B1]--[@B3]). Albeit the prevalence of HCC is highest in Eastern Asia and sub-Saharan Africa, where the HBV chronic infection is endemic and the food is contaminated by the mycotoxin aflatoxin B1, its incidence is rapidly rising also in Western Europe and North America ([@B1]--[@B3]). In the latter areas, however, this escalation in HCC occurrence cannot be entirely explained by the established causal relationship linking chronic hepatitis B or C infection, or ethanol consumption, to hepatocarcinogenesis. Indeed, at least one quarter of HCC cases remains idiopathic ([@B1]--[@B3]). In the last decade, non-alcoholic fatty liver disease (NAFLD) has emerged for its potential etiopathogenetic role in liver cancer development, especially in industrialized countries. Numerous case-control studies indicate in fact that HCC patients with cryptogenic cirrhosis display clinical and demographic characteristics suggestive of NAFLD, when compared with HCC patients of viral or alcoholic etiology ([@B3]--[@B6]). In particular, it has been shown that the increased incidence of HCC in the United States over the past few decades has occurred in parallel with the epidemic of NAFLD ([@B3]--[@B6]). The latter condition is characterized by the excessive accumulation of lipids in the liver and is associated with obesity, insulin resistance, and type 2 diabetes, often evolving into HCC ([@B3]--[@B6]).
Regardless of the causative agent, most HCC patients are diagnosed with an advanced disease, precluding the employment of potentially curative therapies, including liver transplantation or partial liver resection ([@B1]--[@B3]). In addition, molecularly based treatments provided negligible benefits in terms of survival in HCC patients, with the multi-kinase inhibitors Sorafenib and Regorafenib being the only drugs able to extend the life expectancy by \~2/3 months ([@B7]--[@B9]). Consequently, new therapeutic approaches aimed at restraining the growth of advanced HCC are highly needed. For this purpose, the molecular pathogenesis of HCC should be better elucidated to identify critical targets whose inhibition might hamper liver tumor development and/or progression.
The "Lipogenic Phenotype" {#s2}
=========================
Deregulated lipid biosynthesis (commonly referred to as "*de novo* lipogenesis" or "*de novo* lipid synthesis") plays an important pathogenetic role in the development of various metabolic diseases, such as diabetes mellitus, obesity, and the metabolic syndrome. In addition, emerging evidences indicate that metabolism reprogramming, including aberrant lipogenesis, is a widespread phenomenon in most cancer types ([@B10]--[@B12]).
From the historical point of view, the scientific work of the German biochemist and Nobel Prize laureate Otto Warburg, who has been dealing with this issue for several decades since the 1920s, can be considered a pioneer work in this field ([@B13], [@B14]). The starting point was his observation that tumor cells metabolize glucose into lactate under aerobic conditions, while not using the energetically more plausible route of oxidative decarboxylation by the citric acid cycle for energy production. This observation is nowadays well-known as the "Warburg effect" or "Warburg phenomenon" ([@B13], [@B14]). One plausible explanation for this apparently paradoxical event is that glycolysis, although significantly less efficient for energy production than aerobic decarboxylation, can produce adenosine triphosphate (ATP) about 100 times faster than mitochondrial respiration would ([@B14]). Consequently, the tumor cell can provide sufficient energy for the accelerated metabolic processes along carcinogenesis. In addition, through the Warburg phenomenon, a reservoir of important metabolic intermediates available for amino acid synthesis and pentose phosphate production---indispensable prerequisites for ensuring adequate protein and DNA synthesis---is generated ([@B14]). Furthermore, elevated aerobic glycolysis results in a growth advantage for the most proliferating tumor cells within their microenvironment ([@B14]). The immediate consequence of increased glycolysis is the accumulation of the pyruvic acid (pyruvate) metabolite. While most of the pyruvate is converted into lactate and eliminated via the cell membrane, some of the pyruvate is instead converted to acetyl-CoA. In contrast to the normal cell, acetyl-coA represents the primary substrate of the *de novo* lipid synthesis in tumor cells ([@B14]).
As normal tissues can cover most of their lipid requirements via dietary lipids coming from the blood circulation, *de novo* lipogenesis does not play a significant role in the metabolism of these cell types; as a result, the expression of lipogenic enzymes is low ([@B10]--[@B12]). In striking contrast, a universal up-regulation of lipid synthesis occurs in tumor cells ([@B10]--[@B12]). Importantly, the latter phenomenon is only occasionally associated with a change in cellular morphological properties that are detectable by light microscopy (namely, lipid accumulation in tumor cells that, consequently, appear enlarged, and with a clear cytoplasm) ([@B10]--[@B12]). Most frequently, indeed, aberrant lipogenesis results in marked alterations of various molecular and metabolic processes, including intracellular signal transduction, and gene expression. At the molecular level, increased lipogenesis is primarily recognizable by the fact that numerous enzymes involved in lipid metabolism (lipogenic enzymes) display strong activity and high expression in tumor cells ([@B10]--[@B12]). In particular, this refers to the coordinated upregulation of the key enzymes involved in the conversion of glucose into fatty acids, such as ATP citrate lyase (ACLY), acetyl-CoA carboxylase (ACAC), fatty acid synthase (FASN), malate enzyme (ME), and stearoyl-CoA-desaturase 1 (SCD1). Each of these enzymes exhibits a pivotal function in the series of events leading to aberrant lipid biosynthesis. Specifically: (a) ACLY converts citrate from the citrate cycle to acetyl-CoA; (b) ACAC synthesizes malonyl-CoA starting from acetyl-CoA; (c) FASN, starting from malonyl-CoA and consuming acetyl-CoA and NADPH, synthesizes the saturated fatty acid palmitate (palmitic acid) and other saturated long-chain fatty acids; (d) ME catalyzes the production of the reducing NADPH necessary for the synthesis of long-chain fatty acids; and (e) SCD1 converts saturated fatty acids into unsaturated fatty acids, which serve as substrates for the synthesis of triglycerides, cholesterol esters, and phospholipids ([@B10]--[@B12], [@B14], [@B15]). The major steps of *de novo* lipogenesis are summarized in [Figure 1](#F1){ref-type="fig"}.
{#F1}
The requirement of lipids for proliferating tumor cells is high for several reasons. First, lipids represent the building blocks necessary for cell membrane production and, consequently, cell duplication ([@B10]--[@B12], [@B14], [@B15]). Second, the newly synthesized fatty acids are used, if needed, to provide additional energy through the β-oxidation. Third, lipids serve as anchors for selective protein transport to the membrane and as precursors for the synthesis of "lipid second messenger" molecules ([@B10]--[@B12], [@B14], [@B15]).
Based on these data, it is obviously not surprising that most epithelial tumors exhibit an increased *de novo* lipid synthesis and an associated upregulation of the lipogenic enzymes. These include carcinomas of the breast, colorectal, prostate, urinary system, ovary, upper gastrointestinal tract, lung, and oral cavity ([@B10]--[@B12]). Furthermore, it is well-established that tumor cells display increased ACLY expression and activity. Of note, suppression of ACLY by either small interfering RNA molecules (siRNAs) or the pharmacological inhibitor SB-204990 blunts the proliferation and survival of carcinoma cells both *in vitro* and *in vivo* ([@B16]). These intriguing findings are in line with the observation that ACAC, FASN, and SCD1 are up-regulated in numerous malignancies at both the transcriptional and protein level, and their inactivation by treatment with specific siRNAs or small molecular inhibitors significantly restrains tumor cell proliferation and survival ([@B10]--[@B12], [@B14], [@B15]). Altogether, these data suggest that *de novo* lipid synthesis as well as the activation of lipogenic proteins and enzymes are critical for the growth of tumor cells.
Fatty Acid Synthase In Physiology And Cancer {#s3}
============================================
As reported above, fatty acid synthase (FASN) is the critical enzyme responsible for *de novo* fatty acid synthesis ([@B10]--[@B12]). Specifically, FASN catalyzes the reaction leading to the generation of palmitate and 16-carbon long fatty acid from acetyl-CoA and malonyl-CoA ([@B10]--[@B12]). Palmitate is a 16 carbon saturated fatty acid that is a major component of cell membranes and human breast milk, and is incorporated into triglycerides for energy storage. In addition, palmitate is a substrate in the palmitoylation of membrane proteins and acts as a precursor in the synthesis of complex lipids, including cholesterol and glycerophospholipids ([@B10]--[@B12]).
FASN consists of seven functional domains: acyl carrier protein, malonyl/acetyltransferase, ketoacyl synthase, ketoacyl reductase, dehydrase, enoyl reductase, and thioesterase ([@B17], [@B18]). In humans, FASN is encoded by the *FASN* gene and composed of two identical 272 kDa multifunctional polypeptides, in which the seven domains form a single bond ([@B17]). The human *FASN* gene locus is located at chromosome 17 (17q25.3) ([@B10]).
FASN is mainly expressed in the cytosol of healthy liver, adipose, brain, cycling endometrium, and lactating mammary gland cells; in these tissues and organs, lipogenesis is a crucial physiological process ([@B10]--[@B12]).
In cancer, multiple studies have shown that FASN is strongly upregulated in tumors from breast, prostate, colorectal, bladder, ovary, and lung, especially when characterized by clinical aggressiveness, poor prognosis, and resistance to therapy. In contrast, corresponding non-tumor tissues adjacent to the tumor generally express low levels of FASN protein ([@B10]--[@B12]). However, increased FASN expression has also been detected in some benign and pre-neoplastic lesions of the prostate, breast, lung, stomach, colon, and cutaneous nevi ([@B10]--[@B12]). Furthermore, investigations conducted in breast, pancreatic, and colorectal tumors showed that cancer patients exhibit elevated levels of FASN in the serum. Once again, FASN levels in patients\' serum directly correlate with an adverse outcome ([@B10]--[@B12]).
Additional evidence linking FASN to cancer comes from experimental models. For instance, *in vitro* ectopic overexpression of *FASN* in breast cancer cells was found to promote lipogenesis along with augmented cell growth and proliferation ([@B19]). Also, transgenic overexpression of *Fasn* in mice triggered the development of prostate epithelial neoplasia, albeit it was not sufficient to induce invasive tumors *per se* ([@B20]). Further studies with immortalized prostate epithelial cells (iPrEC) suggested that, in addition to the *Fasn* expression, co-expression of androgen receptor was required for invasive adenocarcinoma development ([@B20]). Altogether, this body of evidence indicates a unique association between FASN expression and tumor development and/or progression.
Fatty Acid Synthase In Hepatocellular Carcinoma: Evidence From Human Disease And Experimental Models {#s4}
====================================================================================================
The contribution of unrestrained lipogenesis to the development of hepatocellular carcinoma (HCC) and its progression as well as the molecular mechanisms contributing to the aberrant lipid biosynthesis are starting to be understood. Despite the mounting evidence concerning the importance of aberrant lipid biosynthesis in carcinogenesis, the first studies on this phenomenon in human HCC are relative recent. In a small study ([@B21]), overexpression of the mRNA of the main lipogenic enzymes (FASN, ACAC, ACLY and SCD1) was described in HCC when compared with non-neoplastic liver counterparts. In addition, the sterol regulatory element-binding protein 1 transcription factor (SREBP1), a major inducer of lipogenesis, has been identified as a negative prognostic factor in liver cancer ([@B22]). Also, an *in vitro* study demonstrated that inhibition of FASN significantly affects the growth of human HCC cell lines in a p53-independent manner ([@B23]). Based on these intriguing observations, several studies into the pathogenetic relevance of *de novo* lipid synthesis in human HCC have been initiated, especially focusing on the molecular pathways that drive this event.
In a pioneering investigation, we analyzed the levels of the critical lipogenic proteins in a large human HCC collection ([@B24]). In particular, the HCC cohort used could be differentiated into two distinct subgroups based on patient survival after partial liver resection: a group of HCC with less aggressive biological behavior or HCCB (defined as survival longer than 3 years) and one with higher aggressive behavior or HCCP (defined as survival time shorter than 3 years) ([@B24]). Intriguingly, a simultaneous upregulation of all relevant enzymes of the lipogenic metabolism was observed in HCC when compared with non-tumorous surrounding liver tissues ([@B24]). These included the enzymes responsible for fatty acids production (FASN, ACAC, ACLY, ME, and SCD1) as well as the enzymes for cholesterol biosynthesis \[SREBP2, 3-hydroxymethylglutaryl-CoA reductase (HMGCR), mevalonate kinase (MVK), and squalene synthetase (SQS)\]. Concomitantly, their upstream inducers \[carbohydrate-responsive element-binding protein (chREBP), SREBP1, liver X receptor β (LXR-β)\] were upregulated. Of note, the highest levels of lipogenic enzymes were detected in HCC with poorer prognosis (HCCP) ([@B24]). It is noteworthy to underline that the content of the chemical end products of the respective lipid synthesis (fatty acids, triglycerides, and cholesterol) changed in an analogous manner ([@B24]). Thus, these data indicate increased lipogenesis during development and progression of HCC in humans. Subsequent investigations showed that the induction of unrestrained lipogenesis was the result of both transcriptional and post-transcriptional mechanisms ([@B24]). Specifically, in addition to the aforementioned transcription factors (chREBP, SREBP1, and LXR-β), we detected a prominent induction of the ubiquitin-specific peptidase 2a (USP2a) ([@B24]). This protein is involved in the inhibition of proteasome-induced degradation of FASN, thus inducing stabilization and increased half-life of the latter ([@B25]). Similarly, v-akt murine thymoma viral oncogene homologous (AKT) was found to inhibit the ubiquitination of SREBP1 by phosphorylation-dependent mechanisms ([@B24]). These findings indicate that presumably a complex program involving several pathways converge to increase lipid biosynthesis in human HCC.
Since it is established that the AKT/mTOR pathway is a prominent inducer of *de novo* lipogenesis in various tissues and organs ([@B26], [@B27]), our group investigated whether this also applies to human HCC. As expected, an increased induction of activated (phosphorylated) AKT, mTOR, and the mTOR effector RPS6, was detected from surrounding liver tissues to HCC, especially HCCP, when compared to normal liver ([@B24]). The importance of the AKT/mTOR signaling in lipogenesis was further substantiated in human HCC cell lines, where overexpression of myristoylated/activated AKT led to a rapid increase in cell growth and a reduction in apoptosis. This change in proliferation kinetics was paralleled by a sharp increase in lipid synthesis and up-regulation of lipogenic enzymes in AKT-overexpressing cells ([@B24]). Conversely, there was a robust inhibition of cell growth associated with a decrease in lipogenesis and a reduction in the content of lipogenic proteins when AKT was selectively suppressed in HCC cell lines ([@B24]). At the molecular level, activation of lipogenesis was dependent on an intact mTOR complex1 (mTORC1)/RPS6 signaling pathway, as the addition of the mTORC1 inhibitor rapamycin or the targeted inactivation of RPS6 by specific siRNA impaired cell growth in the same cell lines ([@B24]). The functional importance of the AKT/mTOR pathway in HCC aberrant lipogenesis and FASN induction was substantiated in a recent investigation from Zhao et al. ([@B28]). These authors confirmed the relationship between FASN and the AKT/mTOR cascade in HCC cell lines; furthermore, they identified the loss of the microRNA (miR) 1207-5p as a critical mechanism leading to unconstrained AKT/mTOR signaling pathway and FASN activity in human liver cancer ([@B28]). Alternatively, activation of the AKT/mTOR/FASN axis might be triggered by upregulation of the basigin/CD147 protooncogene, a molecular event often detected in human hepatocarcinogenesis ([@B29], [@B30]). Taken together, these data indicate that the AKT/mTOR pathway plays a leading role in the activation of lipogenesis in human HCC. The identified molecular mechanisms triggering unrestrained FASN activity and lipogenesis in HCC are summarized in [Figure 2](#F2){ref-type="fig"}.
{#F2}
In light of these important premises, we determined the requirement of FASN and *de novo* lipogenesis in hepatocarcinogenesis *in vivo*, using genetic approaches. To achieve this goal, we employed conditional FASN knockout (KO) mice ([@B31]) and various oncogene driven HCC models, such as AKT and AKT/c-Met mice. Previous data from our group showed that hydrodynamic transfection of an activated form of AKT (myristoylated/myr-AKT) triggers upregulation of FASN, aberrant *de novo* lipid synthesis, and HCC development after long latency in mice ([@B24]). To determine whether FASN expression is necessary for myr-AKT driven liver tumor development, we hydrodynamically injected myr-AKT and Cre recombinase (AKT/Cre mice) into conditional FASNfl/fl mice ([@B32]). Of note, while AKT overexpression in control mice resulted in HCC development within 22--28 weeks post-injection, none of the AKT/Cre mice exhibited pre-neoplastic and neoplastic lesions. Equivalent results were achieved following overexpression of myr-AKT in liver-specific FASN KO mice (AlbCre; FASNfl/fl mice) ([@B32]). The anti-neoplastic effect resulting from FASN ablation in AKT/Cre mice was presumably due to the downregulation of rapamycin-insensitive companion of mTOR (Rictor), the critical member of the mammalian target of rapamycin complex 2 (mTORC2) ([@B27]), which is responsible for activation of the AKT protooncogene via phosphorylation. The relevance of Rictor in this process was further demonstrated by the finding that genetic depletion of Rictor in hepatocytes prevented myr-AKT driven hepatocarcinogenesis in mice ([@B32]). The crucial role of FASN in hepatocarcinogenesis has been confirmed in a second mouse model, where myr-AKT was co-transfected with the protooncogene c-Met (AKT/c-Met mice). In this model, the co-expression of AKT and c-Met was found to dramatically accelerate HCC development in mice when compared to those transfected with AKT or c-Met alone, with all AKT/c-Met mice being required to be euthanized within 8 weeks post-injection due to high tumor burden ([@B33]). Thus, AKT, c-Met, and Cre plasmids were transfected into FASNfl/fl mice, allowing the simultaneous expression of AKT and c-Met oncogenes, while deleting FASN in the same subset of mouse hepatocytes (AKT/c-Met/Cre) ([@B33]). Once again, genetic inactivation of FASN completely blunted AKT/c-Met-driven hepatocarcinogenesis in AKT/c-Met/Cre mice, implying that although extremely aggressive, AKT/c-Met tumors fully depend on FASN activity to develop ([@B33]). Similar results were obtained more recently by Guri et al. ([@B34]). These authors generated a mouse model consisting of lack of *Tsc1* and *Pten* tumor suppressor genes, which inhibit the mTORC1 and mTORC2 pathways, specifically in the liver (termed L-dKO mouse). In these mice, liver-specific activation of the mTOR signaling cascade promoted fatty acid synthesis, liver steatosis, and HCC development. Noticeably, either treatment with the FASN inhibitor Orlistat or Fasn knockdown using adenovirus associated virus suppressed hepatocarcinogenesis in L-dKO mice ([@B34]).
Altogether, the present data indicate that FASN and related fatty acid biosynthesis play a critical pathogenetic role in hepatocarcinogenesis.
Inhibition Of Fatty Acid Synthase In Human Hepatocellular Carcinoma: Is It A Feasible Option? {#s5}
=============================================================================================
Based on the body of evidence presented before, it can be envisaged that FASN inhibition might represent a potentially effective therapeutic strategy against human HCC ([@B35]). Several FASN inhibitors have been tested against cancer in preclinical studies, including cerulenin, Orlistat, C75, Fasnall, TVB-2640, and others ([Figure 3](#F3){ref-type="fig"}). However, only TVB-2640 is currently under evaluation, alone or in combination with other medications, in clinical trials against human cancers, not comprising HCC ([Table 1](#T1){ref-type="table"}).
{#F3}
######
Current evidence on the antineoplastic properties of main FASN inhibitors in cancer.
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
**Name** **Molecular formula** **Antineoplastic activity and targeted tumor type** **Current clinical trials** **References**
--------------- ------------------------------- ------------------------------------------------------------------------------------------- ----------------------------- ------------------
Cerulenin C~12~H~17~NO~3~ Breast cancer, promyelocytic leukemia and other cells, mouse liver metastases -- ([@B35], [@B36])
C75 C~14~H~22~O~4~ Lung cancer cells, radio-sensitization in prostate cancer cells -- ([@B37]--[@B39])
Orlistat C~29~H~53~O~5~ Prostate, melanoma, breast and other cells, and xenograft tumor models -- ([@B40]--[@B45])
C93 C~13~H~15~NO~5~ Lung, ovarian and trophoblastic neoplasia cells -- ([@B46], [@B47])
Fasnall C~19~H~22~N~4~S·C~6~H~6~O~3~S Breast cancer (combination therapy) -- ([@B48])
TVB-3166 C~24~H~24~N~4~O Lung, ovarian, prostate, and pancreatic xenograft tumor models, combination with taxanes -- ([@B49], [@B50])
Compound 34 C~31~H~24~F~3~N~3~O~3~ Ovarian, prostate, prostate, lymphoma, leukemia, myeloma, lung, breast cells ([@B51])
IPI-9119 C~24~H~19~F~2~N~5~O~5~ Prostate cancer cells -- ([@B52])
GSK837149A C~23~H~22~N~8~O~5~S~2~ -- -- ([@B53])
GSK2194069 C~25~H~24~N~4~O~3~ -- -- ([@B35])
JNJ- 54302833 C~30~H~31~N~5~O~2~ -- -- ([@B35])
TVB-2640 C~27~H~29~N~5~O Numerous solid tumors, several combinations with chemotherapeutic agents under evaluation NCT03808558\ ([@B55])
NCT02223247\
NCT02980029\
NCT03179904\
NCT03032484
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
*The code of the clinical trials refers to the [ClinicalTrials.gov](https://www.ClinicalTrials.gov) repository*.
Pioneering examples of studies investigating the lipogenic dependency of cancer in recent years include those performed with small molecule FASN inhibitors like cerulenin, an antibiotic isolated from fungal extracts. Cerulenin was found to be active against numerous cancer cell lines and xenograft models; however, the highly reactive nature of the cysteine-reactive epoxide group and off-target activities hampered its clinical application ([@B36]). In particular, activation of β-oxidation and excessive energy expenditure, leading to weight-loss or anorexia, represented the major factors limiting the application of cerulenin in humans ([@B35]). Similar reasons prevented the clinical use of C-75, a synthetic inhibitor of FASN, which was also demonstrated to possess profound antineoplastic effects in experimental models, and to enhance radiation-induced apoptosis in prostate cancer cells, promoting cell cycle arrest in the G2/M phase ([@B37]--[@B39]).
Orlistat is an anti-obesity drug, which acts by blocking the absorption of free fatty acids from the gastrointestinal tract through the inhibition of pancreatic and gastric lipase that hydrolyze triglycerides ([@B40], [@B41]). Specifically, Orlistat possesses a highly reactive beta-lactone that covalently captures reactive serine residues in the FASN thioesterase domain ([@B42]). Despite its potency in restraining the growth of *in vitro* and *in vivo* cancer models ([@B43], [@B44]), the off-target activities together with the poor water solubility and gastrointestinal absorption have hindered the use of Orlistat as anti-tumor agent in patients ([@B45]).
C93 is one of the first inhibitors synthesized, which showed antineoplastic activity initially in lung cancer cell lines, and subsequently in trophoblastic neoplasias ([@B46], [@B47]), but no significant further studies were recently performed. Fasnall, a thiophenopyrimidine-based FASN inhibitor with potent and broad antitumor activity against various breast cancer models, might represent a promising alternative. Fasnall inhibits the FASN capacity to facilitate the production of phospholipids with saturated acyl chains, whereas it promotes the uptake of exogenous unsaturated fatty acids, with consequent alterations in signal transduction messages and promotion of apoptosis. Of note, Fasnall have been shown to act synergistically to prolong the survival of mouse models of breast cancer when associated with the chemotherapeutic agent carboplatin; in this study Fasnall was well tolerated, with no changes in feeding behavior or weight loss being detected in these mice, further suggesting its possible application in the clinical practice ([@B48]).
Other high potential FASN inhibitors have been recently developed. Among them, TVB-3166 is a imidazopyridine-based, orally-available, FASN inhibitor, which suppresses *de novo* palmitate synthesis *in vitro* and *in vivo*, and displays antineoplastic activity in several experimental cancer models ([@B49], [@B50]). The mechanism of action of TVB-3166 on aberrant lipogenesis resides on its property to disrupt the architecture of lipid rafts. Alterations in lipid rafts by TVB-3166 promote the mis-localization of membrane-associated oncoproteins, such as Ras, AKT, and members of the canonical Wnt/β-catenin pathway. As a consequence, TVB-3166 administration leads to the abrogation of several signaling cascades and the induction of tumor cell apoptosis ([@B49]). Lu et al. have synthesized several FASN inhibitors recently using a structure-based approach guided by X-ray crystallography approach ([@B51]). Among them, compound 34 showed a high FASN inhibitory potential and favorable pharmacological features; in addition, it strongly inhibited cell proliferation in several cancer cell lines including A2780 (ovarian), PC3M (prostate), LNCaP (prostate), OCI LY1 (lymphoma), MV4-11 (leukemia/lymphoma/myeloma), H460 (lung), A549 (lung), and MDA-MB-468 (breast), becoming an interesting candidate for future studies ([@B51]).
The synthetic drug IPI-9119, which has been recently developed, strongly inhibits FASN by promoting acylation of the catalytic serine, with high selectivity and negligible off-target activity ([@B52]). IPI-9119 was shown able to effectively block cell growth and proliferation in several cell lines, including prostatic cancer cells, reducing the proportion of S-phase cells and increased that of G0/G1 cells, and decreasing expression of cyclin A2 ([@B52]). GSK837149A was identified as a reversible low inhibitor of the FASN β-ketoacyl reductase domain, but its poor cell permeability prevented the study of its mechanism in cells ([@B53]), while other synthetic inhibitors like GSK2194069 and JNJ- 54302833 remain to be tested in pre-clinical models. In addition, several natural plant-derived polyphenols have been shown to inhibit FASN, including epigallocatechin-3-gallate (EGCG) and the flavonoids luteolin, taxifolin, kaempferol, quercetin, and apigenin ([@B54]); EGCG in a recent study reduced the growth of adenocarcinoma human lung cancer xenografts without inducing body weight loss ([@B37]). Other natural FASN inhibitors may have similar properties, and merit evaluation in future studies.
Currently, the most promising anti-FASN drug is TVB-2640, an oral, small-molecule possessing *in vitro* and *in vivo* antitumor activity associated with an acceptable non-clinical safety profile. Of note, preclinical and early efficacy data from a dose-escalation trial demonstrated a wide activity of TVB-2640 as a single agent in multiple solid tumors, including cases of stable disease ([clinicaltrials.gov/ct2/show/NCT02223247](https://www.clinicaltrials.gov/ct2/show/NCT02223247)). These encouraging results were achieved with relatively low side effects, which could be eliminated with therapy discontinuation ([@B55]). Currently four further clinical trials are testing TVB-2640 alone or in combination with other drugs in NSCLC (NCT03808558), colorectal (NCT02980029), breast cancer (NCT03179904), and astrocytomas (NCT03032484).
TVB-2640 combination treatments are based on evidences that FASN inhibitors synergize with multiple chemotherapeutic agents, such as taxanes, vinca alkaloids, 5-fluorouracil, platinum compounds, and anthracyclines. Furthermore, FASN inhibitors have been found to restore the sensitivity to chemotherapeutic drugs, including doxorubicin, and to targeted therapies, such as those including trastuzumab or lapatinib. In addition, FASN suppression might also cooperate in radio-sensitization and with antiangiogenic agents, by triggering strong tumor hypoxia because cancer cells escape antiangiogenic-driven hypoxia by upregulation of FASN-related lipogenesis ([@B38], [@B56]). These evidences strongly suggest that FASN inhibitors will play an important role in future therapeutic attempts against cancer, hopefully also against HCC.
Conclusion {#s6}
==========
HCC is a highly aggressive and frequent tumor worldwide, with its incidence rising also in low-frequency areas. Thus, these data, together with the lack of effective therapies against this tumor type, indicate that HCC represents a major health concern globally. Understanding the intricated molecular pathogenesis of this cancer entity is therefore necessary for the identification of specific targets suitable of therapeutic intervention. Recently, among the potential, novel therapeutic targets identified in HCC is FASN and the related *de novo* lipogenesis pathway. Mounting and solid evidence underscores the fact that aberrant fatty biosynthesis contributes to hepatocellular carcinogenesis in experimental models as well as in humans. Albeit several features of FASN and related lipogenesis remain to be explored, it appears clear from the data summarized in the present review article that anti-FASN-based therapies might be helpful for the treatment of HCC treatment. The use of existing drugs against FASN for the treatment of HCC (and other tumors) has been impeded by the low potency and consistent off-target effects of these molecules. However, the most recent FASN inhibitors (e.g. Fasnall, TVB-3166, and TVB-2640) seem to have overcome most of these limitations ([@B56]).
Among the critical questions that still need to be addressed for the clinical practice, is how the HCC patients can be selected for anti-FASN treatments. It is clear from HCC TCGA analysis (<https://tcga-data.nci.nih.gov/tcga/tcgaHome2.jsp>) as well as other genomic studies ([@B57]) that human HCC is a highly heterogeneous disease. Not all HCCs express FASN and its related lipogenesis genes at high levels. Consequently, some HCCs might not depend on FASN and *de novo* lipogenesis for growth. This possibility was revealed by *in vivo* mouse studies. Indeed, there was no increase in Fasn expression in mouse HCCs induced by c-Met and gain of function mutant of β-Catenin (c-Met/β-Catenin) via hydrodynamic injection. Consistently, ablation of *Fasn* did not affect HCC growth in mice ([@B35]). For this purpose, reliable biomarkers able to uncover the patients who would presumably benefit from this therapeutic strategy should be identified.
Furthermore, as *de novo* lipogenesis is an integrated part of a metabolic network, it is conceivable that disruption of fatty acid synthesis may lead to other biochemical events. These feedback biochemical and metabolic events may contribute to HCC development. For instance, in the diethylnitrosamine (DEN) induced mouse HCC model, inhibition of lipogenesis via deletion of Acac1 and Acac2 genes in the liver led to an increased HCC development ([@B58]). Mechanistically, this unexpected finding was due to the marked increased in antioxidants, including increased NADPH and reduced glutathione, which protected hepatocytes from oxidant-mediated cell death. In another example, in murine HCCs induced by overexpression of c-Met and loss of Pten (c-Met/sgPten), loss of Fasn significantly repressed HCC formation. However, over long time, HCC lesions could emerge from Fasn null genetic background. Further molecular and metabolomic analysis revealed that there was an increased cholesterol biosynthesis due to increased Srebp2 activity in the mouse liver tissues. This augmented cholesterogenesis eventually compensated for the loss of *de novo* lipogenesis, ultimately leading to HCC formation ([@B59]).
It is also important to acknowledge that two major mechanisms whereby cells acquire fatty acids required for cell growth exist: one involves FASN and its mediated *de novo* lipogenesis, while the other consists of the transport of circulating fatty acids via the "lipolytic" pathway ([@B60]). This process requires lipoprotein lipase (LPL), which releases fatty acids from lipoproteins, as well as fatty acid transporter proteins for fatty acids uptake ([@B61]). The role of exogenous fatty acids during tumor initiation and progression has been studied marginally to date. However, recent reports suggested the key role of this pathway in tumorigenesis. For instance, it was recently found that fatty acids derived from adipocytes could be transferred to melanoma cells through the fatty acid transporter protein SLC27A1. Blocking fatty acid uptake via the fatty acid transport proteins inhibitor Lipofermata significantly reduced melanoma growth and invasion ([@B62]). In HCC cells, it has been shown that LPL mediated fatty acid uptake could at least partly compensate the blockade of *de novo* lipogenesis ([@B63]). These studies indicate that presumably both *de novo* fatty acid synthesis and exogenous fatty acid uptake should be inhibited to achieve significant anti-cancer effects.
In addition, future studies are required to determine whether anti-FASN drugs can be used in combination with FDA-approved anti-HCC multi-kinase inhibitors (Sorafenib, Regorafenib, Cabozantinib), immune modulators (checkpoint inhibitors), and/or conventional chemotherapeutic drugs for the treatment of HCC. Studies to address this important point should be conducted. An alternative approach to suppress FASN in HCC (and other tumor types) could be the inhibition of FASN upstream inducers, such as USP2a and CD147. As concerns USP2a, ML364, a small molecule inhibitor of this deubiquitinase has been recently developed. Of note, ML364 administration caused cell cycle arrest in colorectal cancer and lymphoma cell lines, although the specific effect of the drug on FASN levels was not investigated ([@B64]). Preliminary results obtained in our laboratory indicate a strong growth restraint as well as downregulation of FASN in HCC cell lines treated with ML364 (Cigliano et al., unpublished observation), suggesting that inhibition of USP2a might be a promising therapy for this deadly disease. Furthermore, targeting CD147 has revealed promising results in the treatment of human HCC patients. Indeed, HCC recurrence rate was found to be significantly decreased, and the survival length of HCC patients subjected to liver transplantation prolonged, following the administration of a monoclonal antibody against CD147, in a randomized controlled trial ([@B65]).
Author Contributions {#s7}
====================
DC and XC conceived the work, designed the outline of the review, and supervised all aspects of the manuscript. All authors participated in the literature search, scrutiny, and interpretation, as well as in writing and editing all contents of the manuscript.
Conflict of Interest
--------------------
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
**Funding.** DC was supported by a grant from by the Italian Association Against Cancer (AIRC), grant number IG19175. XC was supported by a NIH grant, grant number R01CA136606.
[^1]: Edited by: Carlos Pérez-Plasencia, National Autonomous University of Mexico, Mexico
[^2]: Reviewed by: Maria-Concepcion Gutierrez-Ruiz, Universidad Autónoma Metropolitana, Mexico; Leticia Bucio Ortiz, Universidad Autónoma Metropolitana, Mexico
[^3]: This article was submitted to Cancer Metabolism, a section of the journal Frontiers in Oncology
[^4]: †These authors have contributed equally to this work
| {
"pile_set_name": "PubMed Central"
} |
INTRODUCTION
============
Streptococcus agalactiae, commonly referred as group B streptococcus (GBS), has been recognized as an emerging infection in nonpregnant adults with underlying medical conditions (mainly malignancy and diabetes mellitus).^[@R4],[@R9],[@R10],[@R17],[@R20],[@R23],[@R24],[@R32],[@R33],[@R37]^ Several reports reveal a significant increase in the disease burden in adults, a trend that might be explained by a growing number of patients with chronic medical conditions.^[@R4],[@R9],[@R10],[@R17],[@R20],[@R24],[@R32],[@R33],[@R37]^ A 2009 multistate, population-based surveillance study^[@R37]^ in the United States showed that the incidence of invasive GBS disease in nonpregnant adults doubled from 3.6 cases per 100,000 population during 1990 to 7.3 cases per 100,000 population in 2007. However, the risk is as high as 22--26 per 100,000 population among adults aged 65 years or older.^[@R24],[@R37]^ The most prevalent GBS serotypes in nonpregnant adult infections are V, Ia, II, and III; these accounted for 78.5% of invasive GBS infections in 2005--2006.^[@R37]^ GBS causes a broad spectrum of clinical manifestations; bacteremia and skin soft tissue infections are the most frequent expressions of invasive GBS disease in nonpregnant adults.^[@R35],[@R37]^ The organism can also cause respiratory, genitourinary, joint, bone, abdominal, central nervous system, and endovascular infections.
Toxic shock syndrome (TSS) is an acute illness caused mainly by exotoxin-producing strains of *Staphylococcus aureus* and *Streptococcus pyogenes*. The exotoxins typically induce pyrogenicity and enhance the lethal effects of endotoxin (gram-negative lipopolysaccharide; LPS). They can activate the immune system, bypassing the usual antigen-mediated immune response sequence, resulting in the release of large quantities of inflammatory cytokines with consequent multiorgan failure.^[@R14],[@R21],[@R22]^ Group B streptococcus toxic shock-like syndrome (TSLS) is rare, and only a few cases have been reported in the literature. We report here 2 new cases of GBS skin soft tissue infection associated with TSLS. We performed a review of the available literature and report here an overview of current knowledge about the clinical characteristics and available treatments of GBS-induced TSLS. The other objective of this study was to test the isolated strains of GBS for known streptococcal pyrogenic exotoxins (SPEs) and undescribed pyrogenic toxins.
PATIENTS AND METHODS
====================
Literature Search
-----------------
We conducted a systematic search of the literature in PubMed (National Library of Medicine, Bethesda, MD) up to January 31, 2012, for cases of GBS-induced TSLS. We searched the MEDLINE database for the following terms: "group B streptococcus," "Streptococcus agalactiae," and "toxic shock syndrome."
Case Reports
------------
### Case 1
A 61--year-old white woman with uncontrolled diabetes mellitus was admitted to Pikeville Medical Center in February 2011 with severe left leg pain 1 week after an inversion ankle sprain. The left leg was intensely tender with a 2 × 3-cm area of purplish skin discoloration. The patient had diffuse abdominal pain, nausea, several episodes of vomiting, and watery diarrhea and was initially managed with levofloxacin for possible invasive gastroenteritis. Two days later, the patient's condition deteriorated to septic shock and respiratory failure that required intubation and mechanical ventilation, intravenous fluids, vasopressors, and broad coverage antimicrobial therapy (meropenem, daptomycin, and linezolid). By that point, the patient had developed multiple areas of extensive skin erythema and hemorrhagic bullae over her left lower extremities. Laboratory values were white blood cell count (WBC) = 20,400 cells/μL (29% bands), platelets = 36,000 cells/μL, creatinine = 3.2 mg/dL, and total bilirubin levels = 3.1 mg/dL. Chest X-ray showed diffuse infiltrates, and computed tomography (CT) imaging of the lower extremities revealed soft tissue swelling, but no evidence of fluid collection or necrotizing fasciitis. Two consecutive sets of blood cultures (each set consisting of aerobic and anaerobic bottles), drawn on admission before the initiation of antibiotic therapy, grew GBS in both the aerobic and anaerobic bottles. Gram staining of tissue specimens obtained during surgical debridement demonstrated gram-positive cocci in pairs and chains. Aerobic cultures of debrided tissues showed heavy pure growth of GBS, whereas anaerobic cultures were negative. The GBS isolated strain was resistant to erythromycin but susceptible to daptomycin, levofloxacin, linezolid, penicillin, vancomycin, and clindamycin. The GBS strain was tested for macrolide-lincosamide-streptogramin B (MLS~B~) resistance; there was no blunting of the clindamycin inhibition zone proximal to the erythromycin disc suggestive of no inducible MLS~B~ resistance. Antimicrobial therapy was changed to the combination of clindamycin and ceftriaxone. Vaginal, throat, and rectal cultures were negative for GBS. The patient was extubated on day 5 and was transferred to the floor on hospital day 6. The leukocytosis, acute kidney injury, and liver dysfunction had completely resolved. Unfortunately, the patient died from acute respiratory failure secondary to mucous plug on day 8.
### Case 2
A 65-year-old white woman with uncontrolled diabetes mellitus presented to Pikeville Medical Center in March 2011 with a 1-week history of severe pain of the left gluteal area. Two days later, the patient had purulent drainage from the gluteal area along with high-grade fever and chills. The left gluteal area was erythematous and intensely tender, with a necrotic open ulcer draining purulent brownish material. The patient was profoundly hypotensive (blood pressure, 65/40 mm Hg) and required support with oxygen, intravenous fluids, vasopressors, and antimicrobial agents (piperacillin/tazobactam, vancomycin, and linezolid). Laboratory values were WBC = 35,000 cells/μL (39% bands) and creatinine = 4.6 mg/dL. CT imaging of the pelvis revealed left gluteal necrotizing fasciitis. Emergent wide debridement was performed, and the intraoperative findings were typical for necrotizing fasciitis, including loss of tissue-plane resistance and necrosis of the subcutaneous fat. Gram-staining of samples and putrid drainage obtained during the operation demonstrated gram-positive cocci in pairs. Aerobic cultures of surgical specimens showed pure heavy growth of GBS, whereas anaerobic cultures were negative. We obtained pure cultures of GBS from 2 consecutive sets of blood cultures done before commencement of antibiotic treatment (both the aerobic and anaerobic bottles). The GBS strain was resistant to erythromycin but susceptible to daptomycin, levofloxacin, linezolid, penicillin, vancomycin, and clindamycin. The isolated GBS strain was tested for MLS~B~ resistance; there was no blunting of the clindamycin inhibition zone proximal to the erythromycin disc suggestive of no inducible MLS~B~ resistance. Vaginal and rectal cultures were positive for GBS. Antimicrobial therapy was changed to the combination of clindamycin and ceftriaxone. The leukocytosis and acute kidney injury had completely resolved. After 4 weeks of antimicrobial therapy, the outcome was favorable, and the patient was discharged home.
Bacterial Materials and Methods
-------------------------------
### Pyrogenic Toxin Production and Purification
The GBS strain isolated from the patient reported in Case 1 was cultured for 18 hours stationary in 1200 mL of pyrogen-free dialyzable beef-heart medium at 37°C in the presence of 7% CO~2~^5^. Subsequently, the bacterial cells were removed by centrifugation (4000 × g; 15 min), followed by filtration (0.2 μm pore size). The resultant supernate was treated 2 hours with 4800 mL of 4°C absolute ethanol to precipitate pyrogenic toxins. The precipitate was resolubilized in pyrogen-free water at 100× concentrated relative to the original culture fluid concentration, and insoluble material was removed by centrifugation (14,000 × g; 5 min).
### Pyrogenic Toxin Assays
*Streptococcus pyogenes* bacteria (group A streptococci) are known for their production of pyrogenic toxins, notably SPEs. Pyrogenic toxins are now also recognized in group C and group G streptococci, and in many of these strains their pyrogenic toxins are related to those from group A streptococci.^[@R16]^ This raises the possibility that GBS produce known SPEs. Therefore, the 100×-concentrated GBS supernate from above was tested for SPEs A, B (cysteine protease), and C by reactivity against rabbit hyperimmune antisera raised in rabbits against individually purified SPEs.^[@R24]^ The remaining concentrated supernate from the GBS strain was diluted in PBS (0.005 M sodium phosphate, pH 7.2; 0.15 M NaCl) until 10×-concentrated and unconcentrated relative to the original culture fluid. These solutions were tested for 2 defining properties of pyrogenic toxins, pyrogenicity in rabbits that typically peaks 4 hours postintravenous inoculation and ability to amplify the lethal effects of gram-negative LPS by up to 10^6^-fold.^[@R24]^ Briefly, 3 Dutch-belted rabbits per group, either sex, were acclimated to a pyrogen test apparatus for 3 hours 1 day before use. The day of use the animals were re-acclimated to the apparatus for 1 hour with rectal thermometers inserted. The animals were injected with 1 mL/kg of the 10×-concentrated or unconcentrated GBS supernates. One group of animals was also injected with 1 μg/kg of LPS from *Salmonella typhimurium* (1/500 lethal dose 50% endpoint).^[@R23]^ Fever responses were recorded immediately before injection and hourly for a total of 4 hours. At the 4-hour time point, the rabbits treated with GBS supernates were given intravenous injections of 1 μg/kg LPS and monitored for death over a 48-hour period. The animal studies were performed in accordance with an approved University of Minnesota IACUC protocol (0908A71722); as directed in this protocol, rabbits that could not right themselves and simultaneously could not exhibit escape behavior uniformly succumb, and these animals were thus prematurely euthanized with 1 mL/kg of Beuthansia D.
Statistical Methods
-------------------
We used SAS software v. 9.1 (SAS, Inc., Cary, NC) to compare the pyrogenic characteristics of GBS supernates. Differences between fever responses in the groups of rabbits were determined using the Student t-test. P value of 0.05 was considered significant.
RESULTS
=======
Pyrogenic Toxin Assays
----------------------
The 100×-concentrated culture fluid from the GBS strain was negative when tested by antibody assay for SPEs A, B, and C. This was expected since we have not previously observed GBS to produce known group A streptococcal SPEs.^[@R15],[@R28],[@R29]^ As discussed above, 2 defining properties of pyrogenic toxins include the capacity to cause fever responses in rabbits that steadily rise and peak 4 hours postintravenous injection, and the ability to amplify the lethal effects of LPS.^[@R22]^ Fever responses in rabbits are considered significant when the average response of 3 animals exceeds 0.5°C.^[@R26]^ We therefore assayed 10×-concentrated and unconcentrated GBS culture fluid for these properties. Both 10×-concentrated and unconcentrated culture fluids were pyrogenic in rabbits, with fever peaks steadily rising for 4 hours postintravenous injection (Figure [1](#F1){ref-type="fig"}). The 10×-concentrated fluid caused an average fever response of 1.7°C at 4 hours (p \< 0.0004 by the Student t test of unpaired data when 0 and 4 h temperatures were compared). Similarly, the unconcentrated culture fluid caused a 1.0°C average fever response at 4 hours postinjection (p \< 0.0005 compared to 0 h temperatures). We cannot be certain that the GBS culture fluids would not cause fevers that rise beyond 4 hours since we were not approved by IACUC to extend measurements beyond 4 hours. However, the shape of the fever curves was typical of those expected with pyrogenic toxins. As a control for pyrogenicity, 3 rabbits also received 1 μg/kg of LPS intravenously. LPS typically causes fever responses in rabbits that peak at both 1 and 3 hours postinjection,^[@R26]^ and this was seen in the rabbits injected with LPS (see Figure [1](#F1){ref-type="fig"}). All 3 rabbits in both groups that received the GBS culture fluids succumbed within 48 hours postinjection of 1 μg/kg of LPS intravenously, consistent with the properties of pyrogenic toxin superantigens. In contrast, none of the 3 rabbits treated with 1 μg/kg of LPS intravenously alone succumbed. Collectively, these data suggest that the GBS culture fluids contained pyrogenic toxins that induced fevers and amplified the lethal effects of LPS. Although we did not assess the extent of LPS amplification, the degree of increased susceptibility seen was at least 500-fold.
{#F1}
Literature Review
-----------------
GBS-induced TSLS is rare; we identified only 11 documented cases in the literature.^[@R1],[@R2],[@R8],[@R11],[@R25],[@R30],[@R34],[@R36],[@R39],[@R40]^ In the current study, we describe 2 new cases of GBS-induced TSLS. On reviewing the previously reported cases and our 2 new cases (Table [1](#T1){ref-type="table"}), several comments can be made regarding clinical presentation and medical management. Malignancy (2/13; 15%), diabetes (3/13; 23%), and splenectomy (2/13; 15%) were the most likely underlying diseases. TSLS due to GBS initially presented as soft tissue pain in an extremity or the back in 9 patients (69%); an influenza-like syndrome characterized by fever, chills, myalgia, nausea, vomiting, and diarrhea in 8 patients (62%); and a change in the mental status in 4 patients (31%). Clinical signs of skin soft tissue infections consistent with localized swelling, erythema, and tenderness were observed in all patients, while ecchymoses, bulla formation, and sloughing of the skin were observed in 3 patients (25%). Scarlatina-like erythema was observed in 3 patients (23%). Although initial laboratory studies may reveal only mild leukocytosis, the mean percentage of bands may reach 40%-50%. Eight of the 13 patients (62%) underwent surgical debridement and/or amputation of the affected areas. Antibiotics such as clindamycin that suppress protein synthesis and therefore toxin synthesis were not given (2 patients) or were delayed (not initiated until bacterial cultures returned positive) in 7 patients (54%). Eight of the 13 patients (62%) responded to medical and surgical management, while the rest died. Our results suggest again that GBS produces 1 or more related pyrogenic toxins. Concentrated culture fluids (100×) of the GBS failed to react with antisera to the 3 best-known SPEs. Thus, it is unlikely that the GBS strain made one of the known *Streptococcus pyogenes* SPEs; instead, it apparently produced a previously undescribed pyrogenic toxin.
######
Group B Streptococcus Toxic Shock-Like Syndrome
DISCUSSION
==========
*Streptococcus agalactiae* is emerging as a cause of fulminate illness similar to *Streptococcus pyogenes*-TSLS.^[@R1],[@R2],[@R8],[@R11],[@R23],[@R28],[@R32],[@R34],[@R36],[@R37]^ Toxic shock syndrome is an acute, multisystem, toxin-mediated illness, typically resulting in shock and multiorgan failure. Bacterial pyrogenic toxins are central in the pathogenesis of TSS. These bacterial toxins trigger massive nonconventional T-cell activation, dependent only on the composition of the variable part of the β-chain of the T-cell receptor, with excessive cytokines from both T cells and antigen-presenting cells that cause tissue damage, disseminated intravascular disease, and organ dysfunction. The toxins have thus become referred to as superantigens.^[@R14],[@R21],[@R22]^ In the current study, we demonstrated that GBS produces uncharacterized pyrogenic toxin(s), explaining the ability of GBS to cause TSLS. We previously identified a novel pyrogenic toxin produced by GBS, which was isolated from a patient with TSLS, and showed that several GBS strains failed to make the known *Streptococcus pyogenes* SPEs, an observation consistent with the findings in the present study.^[@R26],[@R28]^ The GBS strain from our patient produced a pyrogenic toxin that failed to cross-react immunologically with known SPEs, likely because the protein amino acid sequences differ significantly. Future studies are needed to characterize completely this pyrogenic toxin produced by GBS strains. Whether horizontal transfer of DNA-encoding pyrogenic toxins is occurring between different GBS strains, as has been shown for SPEs transferred by bacteriophages, remains to be proven.^[@R3],[@R6]^ It is therefore plausible that such a mechanism is contributing to an increased incidence of severe invasive GBS diseases. Furthermore, GBS is capable of producing menstrual-related TSS in women with vaginal carriage of certain GBS strains. This phenomenon might be attributed to the ability of GBS pyrogenic toxins to cross vaginal mucosa.^[@R1],[@R28]^
Treatment of GBS-induced TSLS requires a multidisciplinary approach with immediate supportive measures, appropriate antimicrobial regimen, and surgical intervention. In women, vaginal examination should be performed, and any tampon or foreign body removed. As pyrogenic toxins are pivotal in TSLS, the addition of an agent with the ability to inhibit bacterial protein syntheses is important to minimize the severity of this devastating disease. Protein synthesis inhibitors such as clindamycin and linezolid enhance the phagocytosis of *Streptococcus pyogenes* and do not exhibit reduced efficacy during the stationary phase of growth; their efficacy is not diminished due to bacterial load.^[@R7],[@R12],[@R13],[@R35]^ Of additional concern is the potential for the loss of efficacy of erythromycin and clindamycin due to increasing rates of resistance expressed by GBS (32% and 15%, respectively).^[@R24]^ The MLS~B~ phenotype is mediated by the erythromycin ribosomal methylation (erm) gene and has been previously identified in certain strains of *Streptococcus pyogenes* as well as other streptococcal species.^[@R18],[@R19]^ Due to the importance of protein synthesis inhibition, it may be reasonable to consider linezolid as initial therapy until susceptibility to clindamycin is confirmed.
Conclusions
-----------
Any patient who presents with septic shock is a candidate for a prompt mucocutaneous examination and workup evaluation for TSS or TSLS. GBS produces uncharacterized pyrogenic toxins, explaining its ability to cause TSLS. For that reason, empiric antimicrobial therapy should include a protein synthesis inhibitor, preferably linezolid, while awaiting the results of susceptibility testing. Implementation of early, well-coordinated, and multidisciplinary approaches involving surgeons and infectious disease and critical care specialists is required to prevent illness progression and death due to the aggressive form of invasive GBS infection.
Abbreviations: CT = computed tomography, GBS = group B streptococcus, LPS = lipopolysaccharide, MLSB = macrolide-lincosamide-streptogramin B, SPE = streptococcal pyrogenic exotoxin, TSLS = toxicshock-like syndrome, TSS = toxic shock syndrome, WBC = white blood cell count
Financial support and conflicts of interest: Research supported by USPHS grant AI074283 from the National Institute of Allergy and Infectious Diseases, National Institutes of Health.
The authors have no conflicts of interest to disclose.
Authors had access to the study data, take responsibility for the accuracy of the analysis, and had authority over manuscript preparation and decision to submit the manuscript.
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#s1}
============
Resistance training (RT) is a safe and effective way to improve proxies of physical performance in healthy children and adolescents when appropriately prescribed and supervised.[@R1] Several meta-analyses have shown that RT has the potential to improve muscle strength and motor skills (eg, jump performance) in children and adolescents.[@R1] [@R5] However, youth athletes have different training capacities, adherence, physical demands of activities, physical conditions and injury risks compared with their non-athlete peers; so the generalisability of previous research on youth athletes is uncertain.[@R8]
To the best of our knowledge, there is only one meta-analysis available that examined the effects of RT on one specific proxy of physical performance (ie, jump performance) and in one age group (ie, youth aged 13--18 years).[@R11] It is reasonable to hypothesise that factors such as age, sex and sport may influence the effects of RT. Therefore, a systematic review with meta-analysis is needed to aggregate findings from the literature in terms of age, sex and sport-specific effects of RT on additional physical performance measures (eg, muscle strength, linear sprint performance, agility, sport-specific performance) in youth athletes.
There is also little evidence-based information available regarding how to appropriately prescribe exercise to optimise training effects and avoid overprescription or underprescription of RT in youth athletes.[@R12] The available guidelines for RT prescription are primarily based on expert opinion, and usually transfer study findings from the general population (ie, healthy untrained children and adolescents) to youth athletes. This is important because the optimal dose to elicit a desired effect is likely to be different for trained and untrained youth.[@R13]
Therefore, the objectives of this systematic literature review and meta-analysis were (1) to analyse the effectiveness of RT on proxies of physical performance in youth athletes by considering potential moderator variables, including age, sex, sport and the type of RT, and (2) to characterise dose--response relationships of RT parameters (eg, training period, training frequency) by quantitative analyses of intervention studies in youth athletes. We hypothesised that (1) RT would have a positive effect on proxies of physical performance in youth athletes, and (2) the effects would be moderated by age, sex, sport and RT type.
Methods {#s2}
=======
Our meta-analysis was conducted in accordance with the recommendations of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA).[@R14]
Literature search {#s2a}
-----------------
We performed a computerised systematic literature search in the databases PubMed and Web of Science.
The following Boolean search syntax was used: ('strength training' OR 'resistance training' OR 'weight training' OR 'power training' OR 'plyometric training' OR 'complex training' OR 'weight-bearing exercise') AND (athlete OR elite OR trained OR sport) AND (children OR adolescent OR youth OR puberty OR kids OR teens OR girls OR boys). The search was limited to: full-text availability, publication dates: 01/01/1975 to 07/31/2015, ages: 6--13; 13--18 years, and languages: English, German. The reference list of each included study and relevant review article[@R1] [@R4] [@R11] [@R15] was screened for title to identify any additional suitable studies for inclusion in our review.
Selection criteria {#s2b}
------------------
Based on the defined inclusion and exclusion criteria ([table 1](#BJSPORTS2015095497TB1){ref-type="table"}), two independent reviewers (ML and OP) screened potentially relevant articles by analysing titles, abstracts and full texts of the respective articles to elucidate their eligibility. In case ML and OP did not reach an agreement concerning inclusion of an article, UG was contacted.
######
Selection criteria
Category Inclusion criteria Exclusion criteria
-------------- ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Population Healthy young athletes (mean age of 6--18 years) Children/adolescents without an athletic background (ie, organised athletic training)
Intervention Resistance training (RT; specific conditioning method, which involves the use of a wide range of resistive loads and a variety of training types designed to enhance proxies of health, fitness and sports performance) Fewer than 6 RT sessions
Comparator Active control (ie, age-matched; conducting the same regular training as the intervention group) in order to avoid bias due to growth and maturation-related performance enhancements[@R16] Only a passive control (ie, no regular training) and/or an alternative training group as control only (eg, stable vs unstable RT)
Outcome At least one measure of muscle strength, vertical jump performance, linear sprint performance, agility and/or sport-specific performance Effects of nutritional supplements; report no means and SDs/SE for the intervention and control groups post test in the results and did not reply to our inquiries sent by email
Study design Controlled study No controlled study
Coding of studies {#s2c}
-----------------
Each study was coded for certain variables listed in [table 2](#BJSPORTS2015095497TB2){ref-type="table"}. Our analyses focused on different outcome categories. If studies reported multiple variables within one of these outcome categories, only one representative outcome variable was included in the analyses. The variable with the highest priority for each outcome is mentioned in [table 2](#BJSPORTS2015095497TB2){ref-type="table"}.
######
Study coding
----------------------------- -----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Sex Male youth athletesFemale youth athletes
Chronological age Children (boys: ≤13 years; girls: ≤11 years)Adolescence (boys:14--18 years; girls: 12--18 years)[@R18]
Biological age Prepubertal (tanner stage: I--II)Postpubertal/pubertal (tanner stage: III--V)
Sport Team sports (eg, soccer)Martial arts (eg, judo)Strength-dominated sport (eg, weight-lifting)Technical/acrobatic sports (eg, gymnastics)
Type of resistance training Machine basedFree weightsCombined machine based and free weightsFunctional trainingComplex trainingPlyometric training
Outcome categories Muscle strength (preferred one repetition maximum)Vertical jump performance (preferred countermovement jump)Linear sprint performance (preferred 20 m sprint)Agility (preferred t-agility-test)Sport-specific performance (preferred throwing, hitting and/or kicking velocities)
----------------------------- -----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
If a study solely used other tests, we included those tests in our quantitative analyses that were most similar with regard to the ones described above in terms of their temporal/ spatial structure.
Further, we coded RT according to the following training parameters: training period, training frequency, and training volume (ie, number of sets per exercise, number of repetitions per set), training intensity, temporal distribution of muscle action modes per repetition, and rest (ie, rest between sets and repetitions). Training parameters were categorised according to common classifications of RT protocols.[@R21] If a study reported exercise progression over the training period, the mean number of sets per exercise, repetitions per sets, rest between sets and training intensity were computed.
To obtain sufficient statistical power to calculate dose--response relationships, we summarised RT types as conventional RT (ie, machine based, free weights, combined machine based and free weights, functional training) and plyometric training (ie, jumping). As it is not possible to classify complex training as either conventional RT nor plyometric training,[@R22] we excluded these studies[@R23] from dose--response analyses. Our dose--response analyses were computed independent of age, sex and sport.
Assessment of risk of bias {#s2d}
--------------------------
The Physiotherapy Evidence Database (PEDro) scale was used to quantify the risk of bias in eligible studies and to provide information on the general methodological quality of studies. The PEDro scale rates internal study validity and the presence of statistical replicable information on a scale from 0 (high risk of bias) to 10 (low risk of bias) with ≥6 representing a cut-off score for studies with low risk of bias.[@R28]
Statistical analyses {#s2e}
--------------------
To determine the effectiveness of RT on proxies of physical performance and to establish dose--response relationships of RT in youth athletes, we computed between-subject standardised mean differences (SMD=(mean postvalue intervention group**−**mean postvalue control group)/pooled standard deviation). We adjusted the SMD for the respective sample size by using the term (1**−**(3/(4N-9))).[@R29] Our meta-analysis on categoric variables was computed using Review Manager V.5.3.4 (Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2008). Included studies were weighted according to the magnitude of the respective SE using a random-effects model.
At least two RT intervention groups had to be included to calculate weighted mean SMDs, hereafter refered to as SMD~wm~, for each performance category.[@R30] We used Review Manager for subgroup analyses: computing a weight for each subgroup, aggregating SMD~wm~ values of specific subgroups, comparing subgroup effect sizes with respect to differences in intervention effects across subgroups.[@R31] To improve readability, we reported positive SMDs if superiority of RT compared with active control was found. Heterogeneity was assessed using I² and χ^2^ statistics.
Owing to a low number of studies in each physical performance outcome category that completely reported information on the applied RT parameters, metaregression was precluded.[@R30] According to a scale for determining the magnitude of effect sizes in strength training research for individuals who have been consistently training for 1--5 years,[@R32] we interpreted SMD~wm~ as: trivial (\<0.35); small (0.35--0.79); moderate (0.80--1.50); large (≥1.50). The level of significance was set at p\<0.05.
Results {#s3}
=======
Study characteristics {#s3a}
---------------------
A total of 576 potentially relevant studies were identified in the electronic database search ([figure 1](#BJSPORTS2015095497F1){ref-type="fig"}). Finally, 43 studies remained for the quantitative analyses. A total of 1558 youth athletes participated, and of these, 891 received RT in 62 RT intervention groups. The sample size of the RT intervention groups ranged from 5 to 54 participants ([table 3](#BJSPORTS2015095497TB3){ref-type="table"}).
######
Included studies examining the effects of resistance training in youth athletes
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Author, year N Exp N Con Biological age Chronological age Sex Sport RT exercise TP TF TI Sets Reps Rest PEDro
--------------------------------------------------------------------------------------- ------------ ------------ ------------------ ----------------------------------------------------------------- --------- ----------------- ------------------------------------------------------------------------------------------------- ---- ---- ---- ------ ------ -------- -------
Alves 2010[@R27]\* EG I: 9\ 6 NA 17.4±0.6 M Soccer EG I (1/week): CT (eg, squats and skippings; leg extension and jumps) 6 1 85 1 6 NA 4
EG II: 8
EG II (2/week): CT (eg, squats and skippings; leg extension and jumps) 6 2 85 1 6 NA
Athanasiou 2004[@R33] 10 10 NA 13--15 M Basketball MB and FW (eg, incline press, leg extension, leg curl) 8 2 NA 3 14 NA 2
Behringer 2013[@R34] EG I: 13\ 10 (post-) pubertal EG I: 15.1±1.8; EG II: 15.5±0.9; CG: 14.6±1.8 M Tennis EG I: MB (eg, low pulley, dead lift, leg press, lateral pull down) 8 2 75 2 15 60 5
EG II: 10
EG II: PT (lower and upper body: eg, skipping, lateral barrier hop, push-ups) 8 2 NA 4 13 20
Bishop 2009[@R35] 11 11 NA EG: 13.1±1.4; CG: 12.6±1.9 ND Swimming PT (lower body: eg, hurdle jumps, DJ, jump to box) 8 2 NA 3 5 60--90 6
Brown 1986[@R36] 13 13 NA 15.0±0.7 M Basketball PT (lower body: DJ (dropping height: 45 cm) 12 3 NA 3 10 30--45 4
Cavaco 2014[@R23]\* EG I: 5\ 6 NA EG I: 13.8±0.5\ M Soccer EG I (1/ week): CT (eg, squats and linear/non-linear sprints) 6 1 85 3 6 180 5
EG II: 5 EG II: 14.2±0.5\
CG: 14.2±0.8
EG II (2/week): CT (eg, squats and linear/non-linear sprints) 6 2 85 3 6 180
Chelly 2009[@R37] 11 11 NA EG: 17±0.3; CG: 17±0.5 M Soccer FW (squats) 8 2 80 4 4 NA 4
Chelly 2014[@R38] 12 11 NA 17.4±0.5 M Handball PT (upper and lower body: eg, hurdle jumps, DJ, push-ups) 8 2 NA 4 10 NA 6
Chelly 2015[@R39] 14 13 Prepubertal 11.9±1.0 M Track and field PT (lower body: ie, hurdle jumps, DJ) 10 3 NA 5 10 NA 5
Christou 2006[@R40] 9 9 (post-) pubertal EG: 13.8±0.4; CG:13.5±0.9 M Soccer MB and FW (eg, leg press, bench press, leg extension, pec-dec) 16 2 68 3 12 150 4
DeRenne 1996[@R41] EG I: 7\ 6 NA 13.3±1.3 M Baseball EG I (1/week): MB and FW (eg, bench press, leg extension, leg curl) 12 1 88 1 10 NA 3
EG II: 8
EG II (2/week): MB and FW (eg, bench press, leg extension, leg curl) 12 2 88 1 10 NA
Escamilla 2010[@R42] 17 17 NA 12.9±1.7; CG: 12.5±1.5 M Baseball FT (upper body; elastic tubes) 4 2 NA 1 23 NA 4
Fernandez-Fernandez 2013[@R43] 15 15 NA EG: 13.2±1.6; CG: 13.2±0.5 M Tennis FT (core training; own body weight) 6 3 NA 2 17 58 5
Ferrete 2014[@R24]\* 11 13 NA EG: 9.3±0.3\ M Soccer CT (eg, squats and CMJ) 26 2 NA 3 7 NA 6
CG: 8.3±0.3
Gorostiaga 1999[@R44] 9 9 (post-) pubertal EG: 15.1±0.7; CG: 15.1±0.5 M Handball MB (eg, leg press, leg curl, bench press) 6 2 65 4 8 90 4
Gorostiaga 2004[@R45] 8 11 NA EG: 17.3±0.5; CG: 17.2±0.7 M Soccer FW (eg, squats, power clean) and PT (eg, hurdle jumps, box jumps) 11 2 NA 3 4 120 5
Granacher 2011[@R46] 14 14 (post-) pubertal EG: 16.7±0.6; CG: 16.8±0.7 M and F eg, soccer MB (eg, squats, leg press, calf raise) 8 2 35 5 10 150 6
Hetzler 1997[@R47] EG I: 10\ 10 (post-) pubertal EG I: 13.2±0.9; EG II: 13.8±0.6; CG: 13.9±1.1 M Baseball EG I (novice): MB and FW (eg, bench press, leg curl, leg press, biceps curls) 12 3 56 3 10 180 4
EG II: 10
EG II (experienced): MB and FW (eg, bench press, leg press, biceps curls) 12 3 56 3 10 180
Keiner 2014[@R48] EG I: 14\ CG I: 12\ NA EG and CG I: U17\ NA Soccer EG I: FW (eg, squats, bench press) (U17) 80 2 83 5 7 220 3
EG II: 30\ CG II: 21\ EG and CG II: U15\
EG III: 18 CG III: 17 EG and CG III: U13
EG II: FW (eg, squats, bench press) (U15) 80 2 83 5 7 220
EG III: FW (eg, squats, bench press) (U13) 80 2 83 5 7 220
Klusemann 2012[@R49] EG I: 13\ 12 NA M: 14±1; F: 15±1 M and F Basketball EG I: FT (body weight RT; supervised) 6 2 NA NA NA NA 2
EG II: 11
EG II: FT (body weight RT; video-based) 6 2 NA NA NA NA
Kotzamanidis 2005[@R50] 11 11 NA EG: 17.1±1.1; CG: 17.8±0.3 M Soccer NA (conventional RT) 13 3 87 4 NA 180 3
Martel 2005[@R51] 10 9 NA 15±1 F Volleyball PT (lower body: eg, power skips, single leg bounding; aquatic) 6 2 NA 4 NA 30 5
Matavulj 2001[@R52] EG I: 11\ 11 NA 15--16 M Basketball EG I: PT (lower body: DJ; dropping height: 50 cm) 6 3 NA 3 10 30 4
EG II: 11
EG II: PT (lower body: DJ; dropping height: 100 cm) 6 3 NA 3 10 30
Meylan 2009[@R53] 14 11 NA EG: 13.3±0.6; CG: 13.1±0.6 M Soccer PT (lower body: eg, hurdle jumps, lateral bounding, skipping) 8 2 NA 3 9 90 4
Potdevin 2011[@R54] 12 11 (post-) pubertal EG: 14.3±0.2; CG: 14.1±0.2 M and F Swimming PT (lower body: eg, DJ, hurdle jumps) 6 2 NA 3 10 NA 5
Ramirez-Campillo 2014a[@R55] EG I: 10\ 10 NA EG I: 11.6±1.4; EG II: 11.4±1.9; EG III: 11.2±2.3; CG: 11.4±2.4 M Soccer EG I: PT (lower body; vertical PT)\ 6 2 NA 3 8 60 5
EG II: 10\ EG II: PT (lower body; horizontal PT)\
EG III: 10 EG III: PT (lower body; combined vertical and horizontal PT)
6 2 NA 3 8 60
6 2 NA 2 8 60
Ramirez-Campillo 2014b[@R56] EG I: 8\ 8 NA 13.0±2.3 M Soccer EG I: PT (lower body: vertical and horizontal jumps) 6 2 NA 2 5 60 5
EG II: 8
EG II: PT (lower body: vertical and horizontal jumps; progressive PT) 6 2 NA 2 8 60
Ramirez-Campillo 2014c[@R57] 38 38 (post-) pubertal 13.2±1.8 M Soccer PT (lower body: DJ) 7 NA NA 2 10 90 5
Ramirez-Campillo 2014d[@R58] EG I: 13\ 14 Prepubertal 10.4±2.3 M Soccer EG I: PT (lower body: DJ; 30 s interest rest) 7 2 NA 2 10 30 5
EG II: 13\
EG III: 11
EGII: PT (lower body: DJ; 60 s interest rest) 7 2 NA 2 10 60
EG III: PT (lower body: DJ; 90 s interest rest) 7 2 NA 2 10 120
Ramirez-Campillo 2015a[@R59] EG I: 54\ 55 (post-) pubertal EG I: 14.2±2.2; EG II: 14.1±2.2; CG: 14.0±2.3 M Soccer EG I: PT (lower body: vertical and horizontal jumps; 24 h recovery between sessions) 6 2 NA 2 8 120 5
EG II: 48
EG II: PT (lower body: vertical and horizontal jumps; 48 h recovery between sessions) 6 2 NA 2 8 120
Ramirez-Campillo 2015b[@R60] EG I: 12\ 14 NA 11.4±2.2 M Soccer EG I: PT (lower body: bipedal jumps) 6 2 NA 6 8 NA 5
EG II: 16\
EG III: 12
EG II: PT (lower body: monopedal jumps) 6 2 NA 3 8 NA
EG III: PT (lower body: monopedal and bipedal jumps 6 2 NA 2 8 NA
Rubley 2011[@R61] 10 6 NA 13.4±0.5 F Soccer PT (lower body: eg, hurdle jumps, DJ) 14 1 NA 2 10 NA 4
Saeterbakken 2011[@R62] 14 10 NA EG: 16.6±0.3 F Handball FT (sling-training) 6 2 87 4 5 90 4
Sander 2013[@R63] EG I: 13\ CG I: 15\ NA EG and CG I: U17\ NA Soccer EG I: FW (eg, squats, bench press) (U17) 80 2 83 5 7 220 2
EG II: 30\ CG II: 25\ EG and CG II: U15\
EG III: 18 CG III: 33 EG and CG III: U13
EG II: FW (eg, squats, bench press) (U15) 80 2 83 5 7 220
EG III: FW (eg, squats, bench press) (U13) 80 2 83 5 7 220
Santos 2008[@R26] 15 10 (post-) pubertal EG: 14.7±0.5\ M Basketball CT (eg, pull over, decline press, depth jump, cone hops) 16 2 70 3 11 150 4
CG: 14.2±0.4
Santos 2011[@R64] 14 10 (post-) pubertal EG: 15.0±0.5; CG: 14.5±0.4 M Basketball PT (lower and upper body: eg, hurdle jumps, box jumps) 10 2 NA 3 10 120 5
Santos 2012[@R65] 15 10 (post-) pubertal EG: 14.5±0.6; CG: 14.2±0.4 M Basketball MB (eg, leg press, lat pull down, leg extension, pullover) 10 2 75 3 11 NA 3
Siegler 2003[@R66]† 17 17 NA 16.5±0.9; CG: 16.3±1.4 F Soccer FW (eg, squat, leg extensions, calf raises, leg curls) + PT (eg, box jumps, bouncing, skipping) 10 2 NA 3 NA NA 3
Söhnlein 2014[@R67] 12 10 NA EG: 13.0±0.9; CG: 12.3±0.8 NA Soccer PT (lower body: vertical, horizontal and lateral jumps) 16 2 NA 3 11 NA 2
Tsimachidis 2010[@R25]\* 13 13 (post-) pubertal EG: 18.0±1.2; CG:18.0±0.7 NA Basketball CT (eg, squats and sprints) 10 2 84 5 7 3 5
Weston 2015[@R68] 10 10 NA EG: 15.7±1.2; CG: 16.7±0.9 M and F Swimming FT (core training: bridge, straight-leg raise; own body weight) 12 3 NA 3 NA 60 2
Wong 2010[@R69] 28 23 NA EG: 13.5±0.7; CG: 13.2±0.6 M Soccer FW (eg, forward lunge, back half squat, biceps curls) 12 2 NA 3 9 85 2
Zribi 2014[@R70] 25 26 prepubertal EG: 12.1±0.6; CG: 12.2±0.4 M Basketball PT (lower body: DJ, hurdle jumps) 9 2 NA 8 5 180 4
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
\*Complex training study, reported training parameters referring only to strength-based exercises.
†Seperate training of free weights RT and plyometric training.
CG, control group; CMJ, counter movement jump; CT, complex training; DJ, drop jump; EG, experimental group; F, female; FT, functional training; FW, free weights; M, male; MB, machine based; N Con, number of participants in the control group; N Exp, number of participants in the experimental group; NA, not applicable; PT, plyometric training; Reps, number of repetition per set; Rest, time of rest between sets (seconds); RT, resistance training; Sets, number of sets per exercise; TF, training frequency (times per week); TI, training intensity (% of 1 repetition maximum); TP, training periods (weeks).
{#BJSPORTS2015095497F1}
There were 13 studies (21 RT intervention groups) that included children, and 29 studies (36 RT intervention groups) that included adolescents. In terms of biological maturation, only 15 studies reported Tanner stages. Three (5 RT intervention groups) of those studies examined prepubertal and 12 (15 RT intervention groups) postpubertal/pubertal youth athletes. Thirty studies (44 RT intervention groups) included boys only, whereas 4 studies (4 RT intervention groups) included girls only.
Youth athletes were recruited from team sports (soccer (20 studies; 34 RT intervention groups), basketball (9 studies; 11 RT intervention groups), baseball (3 studies; 5 RT intervention groups), handball (3 studies; 3 RT intervention groups), tennis (2 studies; 3 RT intervention groups), volleyball (1 study; 1 RT intervention group)), and strength-dominated sports (swimming (3 studies; 3 RT intervention groups), track and field (1 study, 1 RT intervention group)). No included study investigated youth athletes recruited from martial arts or technical/acrobatic sports.
Regarding the type of RT, 4 studies performed RT using machines, 4 studies using free weights, 4 studies using both machines and free weights, 5 studies performed functional RT, 5 studies performed complex training, and 19 studies applied plyometric training. Classification of studies was not always feasible due to missing information or group heterogeneity.
The RT interventions lasted between 4 and 80 weeks, with training frequencies ranging from 1 to 3 sessions per week, 1--8 sets per exercise, 4--15 repetitions per set, and 20--220 s of rest between sets. Training intensity ranged from 35% to 88% of the 1 repetition maximum (RM). Training parameters (eg, temporal distribution of muscle action modes per repetition, and rest in-between repetitions) which have gained attention in the literature[@R71] were not quantified due to insufficient data.
A median PEDro score of 4 (95% CI 4 to 5) was detected and only 4 out of 43 studies reached the predetermined cut-off value of ≥6, which can be interpreted as an overall high risk of bias of the included studies ([table 3](#BJSPORTS2015095497TB3){ref-type="table"}).
Effectiveness of RT {#s3b}
-------------------
[Table 4](#BJSPORTS2015095497TB4){ref-type="table"} shows the overall as well as age, sex, sport and training type-specific effects of RT on measures of muscle strength, vertical jump and linear sprint performance, agility and sport-specific performance.
######
Overall as well as age, sex, sport and training type-specific effects of resistance training in youth athletes
Muscle strength Vertical jump performance Linear sprint performance Agility Sport-specific performance
------------------------------------------------------ ----------------- --------------------------- --------------------------- ------------ ---------------------------- ----- ------ --------- ----- ------ --------- ----- ------ --------- -----
All 1.09 16 (23) 278 0.80 33 (47) 702 0.58 22 (34) 527 0.68 14 (25) 410 0.75 20 (27) 345
Maturity **p=NA** **p=0.60** **p=0.58** **p=0.99** **p=0.17**
Prepubertal (Tanner Stage I and II) oEG 0.91 3 (5) 76 0.65 3 (5) 76 0.58 1 (3) 37 0.27 1 (3) 37
(Post-) pubertal (tanner stage III--V) 0.61 6 (8) 90 1.15 11 (13) 261 0.51 4 (6) 169 0.57 3 (4) 149 0.72 8 (9) 135
Chronological age **p=0.43** **p=0.74** **p=0.92** **p=0.39** **p=0.05**
Children (boys ≤13 years, girls≤11 years) 1.35 3 (4) 39 0.78 10 (17) 235 0.55 9 (14) 195 0.52 6 (11) 146 0.50 6 (11) 153
Adolescence (boys 14--18 years, girls 12--18 years) 0.91 13 (17) 211 0.85 22 (28) 439 0.57 13 (18) 302 0.71 7 (12) 234 1.03 13 (15) 181
Sex **p=0.92** **p=0.54** **p=NA** **p=NA** **p=0.04**
Boys 1.21 12 (18) 220 0.85 27 (40) 615 0.63 19 (30) 474 0.74 12 (22) 374 0.72 15 (22) 288
Girls oEG 0.61 3 (3) 37 oEG -- 1.81 2 (2) 24
Sport **p=0.15** **p=0.20** **p=NA** **p=NA** **p=0.35**
Team sports 1.15 13 (20) 240 0.79 30 (44) 662 0.58 21 (33) 513 0.68 14 (25) 410 0.80 17 (24) 312
Martial arts -- -- -- -- --
Strength-dominant sports 0.58 2 (2) 24 1.22 2 (2) 26 oEG -- 0.34 3 (3) 33
Technical/acrobatic sports -- -- -- -- --
Training type **p\<0.001** **p=0.41** **p=0.12** **p=0.03** **p=0.02**
Machine based 0.36 3 (3) 36 1.45 3 (3) 38 -- -- 0.30 3 (3) 37
Free weights 2.97 2 (4) 72 0.90 3 (5) 80 0.61 3 (5) 80 1.31 1 (3) 62 --
Machine based and free weights 1.16 4 (6) 54 0.77 3 (4) 39 0.18 2 (3) 29 oEG oEG
Functional training 0.62 2 (3) 34 0.39 2 (3) 52 0.19 2 (3) 52 0.38 2 (3) 52 0.79 5 (5) 84
Complex training oEG 1.66 4 (5) 56 1.11 3 (5) 38 0.66 2 (3) 38 1.85 2 (3) 25
Plyometric training 0.39 4 (5) 56 0.81 16 (25) 406 0.64 10 (16) 300 0.62 7 (13) 249 0.74 10 (15) 190
N, total number of participants in the included experimental groups; NA, not applicable; oEG, only one experimental group; S (I), number of included studies (number of included experimental groups); SMD~wm~, weighted mean standardised mean difference; y, years.
There were moderate effects of RT on measures of muscle strength (SMD~wm~=1.09; I²=81%; χ^2^=114.24; df=22; p\<0.001; [figure 2](#BJSPORTS2015095497F2){ref-type="fig"}) and vertical jump performance (SMD~wm~=0.80; I²=67%; χ^2^=137.47; df=46; p\<0.001; [figure 3](#BJSPORTS2015095497F3){ref-type="fig"}), while there were small effects for linear sprint performance (SMD~wm~=0.58; I²=41%; χ^2^=55.74; df=33; p\<0.01; [figure 4](#BJSPORTS2015095497F4){ref-type="fig"}), agility (SMD~wm~=0.68; I²=50%; χ^2^=48.19; df=24; p\<0.01; [figure 5](#BJSPORTS2015095497F5){ref-type="fig"}) and sport-specific performance (SMD~wm~=0.75; I²=62%; χ^2^=67.81; df=26; p\<0.001; [figure 6](#BJSPORTS2015095497F6){ref-type="fig"}). By considering only the four studies with high quality (ie, low risk of bias), RT had moderate effects on measures of muscle strength (SMD=1.07; 1 study), vertical jump (SMD~wm~=0.89; 3 studies) and linear sprint performance (SMD~wm~=1.19; 2 studies); small effects on agility (SMD=0.28; 1 study); and large effects on sport-specific performance (SMD~wm~=1.73; 2 studies).
{#BJSPORTS2015095497F2}
{#BJSPORTS2015095497F3}
{#BJSPORTS2015095497F4}
{#BJSPORTS2015095497F5}
{#BJSPORTS2015095497F6}
There was no statistically significant effect of chronological and/or biological age on any proxy of physical performance. However, a tendency (p=0.05) towards larger RT effects were found for proxies of sport-specific performance in adolescents (SMD~wm~=1.03) compared with children (SMD~wm~=0.50; [table 4](#BJSPORTS2015095497TB4){ref-type="table"}). Subgroup analyses indicated that RT produced significantly larger effects (p\<0.05) on proxies of sport-specific performance in girls (SMD~wm~=1.81) compared with boys (SMD~wm~=0.72; [table 4](#BJSPORTS2015095497TB4){ref-type="table"}). Given that most included studies (n=38) examined participants competing in team sports, our subgroup analyses regarding the moderator variable 'sport' is limited and did not show any significant subgroup differences ([table 4](#BJSPORTS2015095497TB4){ref-type="table"}). Subgroup analyses demonstrated that different training types of RT produced significantly different gains in muscle strength (p\<0.001), agility (p\<0.05) and sport-specific performance (p\<0.05). Free weight RT showed the largest effects on muscle strength and agility, while for sport-specific performance, complex training produced the largest effects ([table 4](#BJSPORTS2015095497TB4){ref-type="table"}).
Dose--response relationships of RT {#s3c}
----------------------------------
### Training period {#s3c1}
There was a significant difference for the effects of conventional RT on measures of muscle strength (p\<0.001), vertical jump height (p\<0.05) and agility (p\<0.001; [figure 7](#BJSPORTS2015095497F7){ref-type="fig"}). The dose--response curves indicated that long lasting conventional RT (\>23 training weeks) resulted in more pronounced improvements in measures of muscle strength (SMD~wm~=3.40) and agility (SMD~wm~=1.31), as compared with shorter training periods (\<23 weeks). In terms of vertical jump height, a training period of 9--12 weeks appeared to be the most effective (SMD~wm~=1.20).
{#BJSPORTS2015095497F7}
### Training frequency {#s3c2}
There were no significant differences between the observed training frequencies (ie, 1, 2, 3 times per week) for RT as well as plyometric training ([figure 8](#BJSPORTS2015095497F8){ref-type="fig"}).
{#BJSPORTS2015095497F8}
### Training intensity {#s3c3}
There was a significant difference with regard to the effects of conventional RT on measures of muscle strength (p\<0.01; [figure 9](#BJSPORTS2015095497F9){ref-type="fig"}). High-intensity conventional RT (ie, 80--89% of 1 RM) resulted in more pronounced improvements in muscle strength (SMD~wm~=2.52) compared with lower training intensities (ie, 30--39%, 40--49%, 50--59%, 60--69%, 70--79% of the 1 RM).
{#BJSPORTS2015095497F9}
### Training volume (number of sets per exercise) {#s3c4}
There was a significant difference with regard to the effects of conventional RT on muscle strength (p\<0.01), and a tendency towards significance for measures of vertical jump performance (p=0.06; [figure 10](#BJSPORTS2015095497F10){ref-type="fig"}). Five sets per exercise resulted in more pronounced improvements in muscle strength (SMD~wm~=2.76) compared with fewer sets. Three sets per exercise tended to be more effective in improving vertical jump performance (SMD~wm~=1.19), as compared with four or five sets per exercise.
{#BJSPORTS2015095497F10}
For plyometric training, there was a tendency towards larger training-related effects on measures of muscle strength (p=0.09), linear sprint performance (p=0.07), as well as sport-specific performance (p=0.05) depending on the number of sets per exercise. Four sets per exercise revealed the largest effects for measures of muscle strength (SMD~wm~=0.79) and sport-specific performance (SMD~wm~=1.84), while three or four sets appear to be most effective for improving linear sprint performance (SMD~wm~=0.95).
### Training volume (number of repetitions per set) {#s3c5}
There was a significant difference in terms of the effects of conventional RT on measures of muscle strength (p\<0.05; [figure 11](#BJSPORTS2015095497F11){ref-type="fig"}). Six to eight repetitions per set produced the largest effects on muscle strength (SMD~wm~=2.42). For plyometric training, there was a tendency towards significance for proxies of sport-specific performance (p=0.05). Six to 8 repetitions per set were less effective (SMD~wm~=0.15), while 3--5 and 9--12 repetitions per set produced similar effects (SMD~wm~=0.89 and 0.93).
{#BJSPORTS2015095497F11}
### Rest between sets {#s3c6}
There was a significant difference for the effects of conventional RT on measures of muscle strength (p\<0.05; [figure 12](#BJSPORTS2015095497F12){ref-type="fig"}). Three to 4 min of rest between sets resulted in more pronounced improvements in measures of muscle strength (SMD~wm~=2.09), as compared with shorter durations of rest.
{#BJSPORTS2015095497F12}
Discussion {#s4}
==========
This systematic review with meta-analysis examined the general effects as well as the age, sex, sport and training type-specific impact of RT on proxies of physical performance in healthy young athletes. In addition, dose--response relationships of RT parameters were independently computed. The main findings were: (1) RT has moderate effects on muscle strength as well as on vertical jump performance, and small effects on linear sprint, agility and sport-specific performance in young athletes, (2) the effects of RT were moderated by the variables sex and RT type, (3) most effective conventional RT programmes to improve measures of muscle strength in healthy young athletes comprised training periods of more than 23 weeks, 5 sets per exercise, 6--8 repetition per set, a training intensity of 80--89% of the 1 RM, and 3--4 min of rest between sets.
Effects of RT on physical performance in youth athletes {#s4a}
-------------------------------------------------------
In general, RT is an effective way to improve proxies of physical performance in youth athletes, and our findings support recently published literature.[@R4] [@R17] [@R72] [@R73] We found that the main effects of RT on measures of muscle strength and vertical jump performance were moderate in magnitude, with small effects for secondary outcomes, including linear sprint performance, agility and sport-specific performance (eg, throwing velocity). The lower RT effects on secondary outcomes might be explained by the complex nature of these qualities, with various determinants contributing to the performance level. For instance, agility depends on perceptual factors and decision-making as well as on changes in direction of speed, which is again influenced by movement technique, leg muscle quality and straight sprinting speed.[@R74] Thus, muscle strength appears to be only one of several factors contributing to agility.
We recommend the incorporation of RT as an important part of youth athletes' regular training routine to enhance muscle strength and jump performance.
How age, sex, sport and training type moderate RT effects {#s4b}
---------------------------------------------------------
### Age-specific effects of RT in youth athletes {#s4b1}
Biological maturity is related to chronological age, and has a major impact on physical performance in youth athletes.[@R75] However, unlike age, growth and maturation are not linear factors.[@R76] [@R77] There is often a discrepancy between chronological age and biological maturity among youth athletes.[@R4] [@R16] [@R78]
We found no significant differences in effect sizes for any proxy of physical performance between prepubertal and postpubertal athletes. Similarly, we did not find significant differences for the effects of RT on any physical performance measure with respect to the moderator variable 'chronological age' ([table 4](#BJSPORTS2015095497TB4){ref-type="table"}). Merely, a tendency (p=0.05) towards higher sport-specific performance gains following RT in adolescents, compared with children, was identified.
Although a minimum age has been defined at which children are mentally and physically ready to comply with coaching instructions,[@R4] our subgroup analyses regarding biological and chronological age suggest that youth athletes may benefit to the same extent from RT, irrespective of age. However, it is important to note that most studies did not report the biological maturity status of the participants. Therefore, more research is needed to elucidate biological age-specific RT effects on physical performance in youth athletes and to verify our preliminary findings.
### Sex-specific effects of RT in youth athletes {#s4b2}
Previous research on the effects of RT on proxies of physical performance in youth athletes has primarily focused on boys. However, findings from male youth athletes can only partially be transferred to female youth athletes because the physiology of boys and girls (eg, hormonal status during puberty) varies. We found that male and female youth athletes show similar RT-related gains in muscle strength and vertical jump performance, but girls had significantly larger training-induced improvements in sport-specific performance (SMD~wm~=1.81) compared with boys (SMD~wm~=0.72). This suggests preliminary evidence that the RT trainability of female adolescent athletes may be at least similar or even higher compared with males. Given that girls' and boys' physiology changes differently with age and maturation,[@R76] [@R77] sex-specific effects of RT in youth athletes should be investigated with respect to biological maturity. Owing to an insufficient number of studies that examined female youth athletes and reported their biological maturity status, we were not able to include 'biological maturity' as a moderator variable in our subgroup analyses. We consider our sex-specific findings preliminary because these are based on five studies only investigating female youth athletes. More research is needed to elucidate sex-specific RT effects on physical performance in youth athletes and to verify our preliminary findings.
### Sport-specific effects of RT in youth athletes {#s4b3}
The effects of RT in elite adult athletes may be specifically moderated by the respective athlete profile of the sport performed.[@R79] [@R80] Whether this is also the case in youth athletes remains unresolved. Given that most included studies (n=38) investigated young athletes competing in team sports, our analyses with regard to the moderator variable 'sport' was limited and did not reveal any significant differences between sports disciplines ([table 4](#BJSPORTS2015095497TB4){ref-type="table"}). Therefore, further research has to be conducted to examine if youth athletes respond differently to RT programmes as per the sport practiced.
### Training type-specific effects of RT in youth athletes {#s4b4}
Various types of RT have been reported (eg, machine-based RT, free weight RT and functional RT). Each of these types has specific benefits and limitations.[@R20] [@R73] Machine-based RT may represent a safe environment for young athletes when supervision cannot be ensured, whereas supervised RT using free weights allows full range of motion that better mimics sports-specific movements.[@R20] [@R73] We found that RT programmes using free weights were most effective to enhance muscular strength and agility. In addition, complex training produced the largest effect sizes if the goal was to improve sport-specific performance. Therefore, the choice of RT types should be variable and based on the exercise goal (eg, enhancing muscle strength or sport-specific performance).
Dose--response relationships of RT in youth athletes {#s4c}
----------------------------------------------------
Planning and designing RT programmes is a complex process that requires sophisticated manipulation of different training parameters. Owing to a lack of evidence-based information on dose--response relationships following RT in youth athletes, it is quite common for established and effective RT protocols for healthy untrained children and adolescents to be transferred to youth athletes. However, this may hinder to fully recruit the adaptative potential of young athletes because the optimal dose to elicit the desired effect appears to be different in trained compared with untrained youth.[@R13] Owing to the observed limitations regarding female youth athletes and biological maturation status in the present meta-analysis, the dose--response relationships of RT in youth athletes were determined irrespective of sex and maturity.
In general, the specific configuration of RT parameters determines the underlying training stimulus and thus, the desired physiological adaptations. However, significant effects were predominantly identified for conventional RT parameters for measures of muscle strength. Therefore, it appears that gains in muscular strength may be more sensitive to the applied training parameters of the conventional RT programmes, as compared with the secondary performance outcomes (eg, linear sprint performance, agility, sport-specific performance).
### Training period {#s4c1}
The effects of short-term (\<24 weeks) RT peaked almost consistently with training periods of 9--12 weeks for both conventional RT and plyometric training. However, our subgroup analyses indicated significant differences only for conventional RT for measures of muscle strength and vertical jump performance. Nevertheless, with regard to strength gains, long-term (≥24 weeks) conventional RT was more effective in youth athletes (SMD~wm~=3.40), as compared with short-term conventional RT (SMD~wm~=0.61--1.24). Thus, it can be postulated that conventional RT programmes should be incorporated on a regular basis in long-term athlete development.[@R66] Given that continuous performance improvements are difficult to achieve particularly over long time periods, properly varying RT programmes may avert training plateaus, maximise performance gains and reduce the likelihood of overtraining.
Regular basketball practice during a detraining/reduced training period was sufficient to maintain previously achieved muscular power gains due to its predominantly power-type training drills.[@R81] Therefore, it is reasonable to hypothesise that regular training can maintain RT-based gains in muscular strength for several weeks if similar physical demands are addressed during regular training. Coaches may reduce the time spent on RT for several weeks without impairing previously achieved strength gains during competition periods when the training must emphasise motor skills and competition demands.
### Training frequency {#s4c2}
The phase of periodisation, projected exercise loads and the dose of additional physical training (ie, overall amount of physical stress) may influence training frequency.[@R21] In order to avoid overtraining and achieve maximal benefits of RT, it is important to allow the body sufficient time to recover from each RT session. However, if the rest between RT sessions is too long, adaptive processes from previous RT sessions may get lost.
Most studies performed RT two or three times per week ([figure 8](#BJSPORTS2015095497F8){ref-type="fig"}), and there was no significant difference between the observed training frequencies. To our knowledge, there is no study available that directly compared the effects of two RT sessions per week as opposed to three sessions for youth athletes. Although a reduced RT frequency of one session per week may be sufficient to maintain muscle strength gains following RT for several weeks,[@R41] [@R82] training twice per week might be preferred to achieve further gains in muscle strength in youth athletes.
### Training volume and training intensity {#s4c3}
Both volume and intensity have to be considered when prescribing RT to maximise physiological adaptations and minimise injury risk.[@R4] Different configurations of training volume and intensity result in different forms of physiological stress, which in turn induce different neural and muscular adaptations.[@R71]
Owing to the large methodological variety in dealing with training intensity during plyometric training, we were not able to consistently quantify the dose--response relationship for training intensity with regard to plyometric training.
Conventional RT programmes using average training intensities of 80--89% of the 1 RM were most beneficial in terms of improving muscle strength in youth athletes. These findings are in accordance with the position stand of the American College of Sports Medicine for strength training in adults.[@R83] The largest effect sizes for muscle strength gains in adults, trained individuals and athletes were achieved at 80--85% of the 1 RM.[@R8] [@R12] However, it should be noted that the individual percentage of 1 RM is a stress rather than a strain factor. Several studies have indicated that a given number of repetitions cannot be associated with a specific percentage rate of the 1 RM.[@R78] [@R84] Thus, to individualise RT, future studies should focus on finding a valid strain-based method to quantify RT intensity effectively.
In terms of the number of sets per conventional RT exercise, our data show similar effect size magnitudes when comparing single-set (SMD~wm~=2.41) versus multiple-set conventional RT programmes (5 sets: SMD~wm~=2.76). The primary benefit of a single-set conventional RT is time efficiency. Nevertheless, since our results for single-set conventional RT are based on two intervention groups from one study, this finding has to be interpreted with caution. Although there was no study that directly compared the effects of single-set versus multiple-set conventional RT in youth athletes, there is evidence from adult athletes that single-set conventional RT may be appropriate during the initial phase of RT,[@R85] whereas multiple-set conventional RT programmes should be used to promote further gains in muscle strength, especially in athletes.[@R86] Therefore, multiple-set conventional RT may be necessary to elicit sufficient training stimuli during long-term youth athlete development.
Regarding the applied plyometric training, 3 (for vertical jump) or 4 sets per exercise (for muscle strength, sport-specific performance) as well as 3--5 or 9--12 repetitions per set (for vertical jump, sport-specific performance) might be beneficial for youth athletes' physical performance. However, the movement quality of plyometric exercises is more important than the total session volume.[@R87] Therefore, we recommend the use of thresholds for performance variables, such as ground contact time or performance indices, to determine individualised training volume.[@R87]
### Rest between sets {#s4c4}
The duration of rest between sets and repetitions depends on parameters like training intensity and volume. The rest interval significantly affects the biochemical responses following RT.[@R71] Owing to an insufficient number of studies that reported the duration of rest between repetitions, we focused on dose--response relationships for rest between sets. Long rest periods (ie, 3--4 min of rest between sets) were most effective for improving muscle strength following conventional RT in youth athletes. This is most likely because long rest periods allow athletes to withstand higher volumes and intensities during training.
Limitations of this meta-analysis {#s4d}
---------------------------------
A major limitation is that we could not provide insights into the interactions between the reported training parameters. Our analyses are based on a variety of studies using different combinations of training parameters magnitudes (eg, training frequency, number of sets, intensity). It remains unclear if performance gains would still be maximal if, according to the present dose--response relationships, the optimum of each parameter was implemented in RT programmes.[@R81] Thus, further research is necessary to find an analytical method to provide insights into the interactions between the investigated training parameters. The modelling of training variables might help to address this limitation. Holding a set of RT variables constant while changing the effects of one specific variable could determine the unique effects of each training variable.
Further limitations of this systematic review and meta-analysis are the high risk of bias of the included studies (only 4 out of 43 studies reached a PEDro score of ≥6), the considerable heterogeneity between studies (ie, I²=41--81%), and the uneven distribution of SMDs calculated for the respective training parameters. In addition, the scale for determining the magnitude of effect sizes[@R32] is not specific for RT research in children and adolescents. Another limitation is that almost all studies failed to report RT parameters which had got recent research attention (eg, temporal distribution of muscle action modes per repetition).[@R71] Further, studies used traditional stress-based (ie, RM) instead of recent strain-based (eg, OMNI resistance exercise scale of perceived exertion[@R88]) methods to quantify RT intensity.[@R89] We were not able to aggregate the effects of moderator variables, such as sex and maturation, for the dose--response relationships due to an insufficient number of studies that specifically addressed these issues.
Summary {#s5}
=======
RT was effective for improving proxies of physical performance in youth athletes. The magnitudes of RT effects were moderate in terms of measures of muscle strength and vertical jump performance, and small with regard to measures of linear sprint, agility and sports-specific performance in youth athletes. Sex and RT type appeared to moderate these effects. However, most studies were at high risk of bias and therefore, the results should be interpreted cautiously.
A training period of more than 23 weeks, 5 sets per exercise, 6--8 repetitions per set, a training intensity of 80--89% of 1 RM, and 3--4 min rest between sets were most effective for conventional RT programmes to improve muscle strength in youth athletes. However, these evidence-based findings should be adapted individually by considering individual abilities, skills and goals. Specifically, youth coaches should not use high RT intensities before the youth athlete developed technical skills to adequately perform the RT exercises. What is already known on this topic?Resistance training is safe for children and adolescents if appropriately prescribed and supervised.Several meta-analyses have already shown that resistance training has the potential to improve muscle strength and motor skills (eg, jump performance) in healthy, untrained children and adolescents. What this study addsThis is the first systematic review and meta-analysis to examine age, sex, sport and training type-specific effects of resistance training on physical performance measures in youth athletes.The effect of resistance training was moderated by sex and resistance training type. Girls had greater training-related sport-specific performance gains compared with boys, and resistance training programmes with free weights were most effective for increasing muscle strength.Dose--response relationships for key training parameters indicate that youth coaches should aim for resistance training programmes with fewer repetitions and higher intensities to improve physical performance measures.
The authors would like to thank Dr Andrea Horn for her support during the course of the research project.
**Contributors:** ML, OP and UG performed systematic literature search and wrote the paper. ML and OP analysed the data.
**Funding:** This study is part of the research project 'Resistance Training in Youth Athletes' that was funded by the German Federal Institute of Sport Science (ZMVI1-081901 14-18).
**Competing interests:** None declared.
**Provenance and peer review:** Not commissioned; externally peer reviewed.
| {
"pile_set_name": "PubMed Central"
} |
Data may be made available upon request as we have not received approval from the Ethics Review Committee, Faculty of Medicine, University of Ruhuna, Sri Lanka, to provide the raw data (legal and ethical reasons), that is the pharmacist survey data, to journals for publication. Therefore, we are unable to release this raw data to PLOS ONE. However, the tables and results section have all the required data used to drive the conclusions. Interested researchers may make additional data access requests to the Ethics committee at: <ethics@med.ruh.ac.lk>.
Introduction {#sec005}
============
Antibiotic resistance (ABR) is growing at an alarming rate and poses a major threat to the clinical efficacy of antibiotics \[[@pone.0215484.ref001]\]. ABR increases healthcare costs, length of hospital stay, and morbidity and mortality, in both developed and developing countries \[[@pone.0215484.ref002]\].
ABR is an inescapable consequence of antibiotic use \[[@pone.0215484.ref003]\], but it increases hugely with misuse and abuse \[[@pone.0215484.ref004], [@pone.0215484.ref005]\]. Self-medication with antibiotics is one of the major factors contributing to misuse in developing countries \[[@pone.0215484.ref006], [@pone.0215484.ref007]\]. Community pharmacies are primarily responsible for self-medication with antibiotics in low and middle-income countries (LMICs) \[[@pone.0215484.ref004], [@pone.0215484.ref006], [@pone.0215484.ref008]\] where infectious diseases are a major burden \[[@pone.0215484.ref006]\]. Despite the restrictions for the use of systemic antibiotics and recommendations that such medicines be exclusively used under medical prescription, the dispensing of antibiotics without a prescription is still a common practice in this region. Over 50% of all outpatient antibiotics dispensed in most parts of the world are not prescribed by physicians \[[@pone.0215484.ref006], [@pone.0215484.ref007], [@pone.0215484.ref009]\]. In Sri Lanka, recent simulated client studies show that 61% of pharmacies dispensed antibiotics illegally when requested by product name \[[@pone.0215484.ref010]\], and 41% on symptoms-based antibiotic requests and most were supplied inappropriately for minor viral infections \[[@pone.0215484.ref011]\]. This poor dispensing practices seen in LMICs are attributed to the low level of adherence to the existing laws and regulations on good dispensing practice \[[@pone.0215484.ref010], [@pone.0215484.ref012]\], lack of knowledge and professionalism among pharmacists \[[@pone.0215484.ref012]\], demand from consumers \[[@pone.0215484.ref010], [@pone.0215484.ref012]\], profit orientation and lack of professionally qualified pharmacists \[[@pone.0215484.ref010], [@pone.0215484.ref012]\]. The lack of knowledge about antibiotics, ABR, and legal aspects of antibiotic dispensing contribute to illegal and inappropriate antibiotic dispensing practice in LMICs \[[@pone.0215484.ref012]--[@pone.0215484.ref014]\]. Therefore, community pharmacists' knowledge about antibiotics, ABR, and regulation of supply have a significant impact on their dispensing practice. The factors influencing antibiotic dispensing without a prescription may differ in various developing and developed countries, and attempts in developing interventions need to address the context and focus on the root causes specific to the part of the world.
Although knowledge and professional competency of community pharmacists in antibiotic dispensing are vital to ensure appropriate antibiotic supply to the community, limited studies (and none in Sri Lanka) have evaluated these characteristics in LMICs where antibiotic supply without a prescription is high. Therefore, this study aimed to investigate the antibiotic dispensing behaviour of Sri Lankan community pharmacies. A national survey was conducted to (i) evaluate community pharmacy staff's self-reported antibiotics knowledge and dispensing behaviour (without a prescription), and (ii) identify possible factors impacting such antibiotic dispensing behaviour.
Methods {#sec006}
=======
A cross-sectional study was conducted using a validated, self-administered and structured questionnaire among community pharmacy staffs throughout Sri Lanka between December 2016 and September 2017.
Study population and settings {#sec007}
-----------------------------
A systematic sampling of community pharmacies in Sri Lanka, comprising of all types of community pharmacies in the country, including semi-government (Rajya Osusala) and private pharmacies (single, chain and pharmacies in private hospitals) covering all nine provinces of the country was conducted. The study sample was achieved using a proportionate sampling technique in each province, as described below and elsewhere \[[@pone.0215484.ref010]\].
Sample size calculation and sampling technique {#sec008}
----------------------------------------------
The original sample size (n = 369) was estimated using standard error of proportions equation with a 95% confidence interval using margin of random error of ±5% based on the results of a previous pilot study conducted in Sri Lanka measuring knowledge about antibiotics among community pharmacists. Sample size estimation has already been reported in detail elsewhere \[[@pone.0215484.ref010]\]. The response rate in a similar survey conducted among community pharmacists in Portugal was 65% \[[@pone.0215484.ref015]\] therefore, an additional 30% was added to the original estimated sample size to address non-response rate.
Pharmacies for this survey were randomly selected from all nine provinces of the country (sampling frame). A simple random sampling technique was used in the sample selection. The proportionate sample of pharmacies for each province was derived from the list of pharmacies from the Pharmacy Directory maintained by the National Medicine Regulatory Authority (NMRA), which includes all pharmacies that had legal permission to function (as at April 2016) \[[@pone.0215484.ref016]\]. A random point in the road map in each province was taken into account to select the first pharmacy on that road as the index pharmacy. Pharmacies were approached throughout the province until the quota was filled. A similar approach was continued in all other provinces \[[@pone.0215484.ref010]\].
Data collection {#sec009}
---------------
The selected community pharmacies in each province were visited by trained research assistants who were either pharmacy undergraduates or recent graduates.
The questionnaire was distributed to either consented community pharmacist or pharmacy assistant (in the absence of a pharmacist at the time of the visit) in their preferred language (Sinhala, Tamil or English). The most senior pharmacy assistant was approached in the latter case. Community pharmacy staff either completed the questionnaire on the spot or asked the research assistant to collect it at a mutually agreed time and a day. A reminder was made over the phone a day before the agreed collection day. If the questionnaire was not ready on the agreed day, respondents were given an addressed envelope with the postal charges paid to post it to us after filling.
Ethics statement {#sec010}
----------------
This study was approved by the Ethics Review Committee, Faculty of Medicine, University of Ruhuna, Sri Lanka (Reference number 16.11.2016:3.1). Informed written consent was obtained from all participants for the self-reported survey after explaining about the study and its objectives. The consenting pharmacy staff were asked to complete the structured questionnaire described below. The questionnaire was anonymous with no personal identifiable information collected.
Survey instrument development {#sec011}
-----------------------------
The structured questionnaire ([S1 Appendix](#pone.0215484.s001){ref-type="supplementary-material"}) used in this study was developed through an extensive review of the literature \[[@pone.0215484.ref017]--[@pone.0215484.ref019]\], pertaining to knowledge, attitudes and practices of community pharmacy staff regarding antibiotics.
The face and content validity of the questionnaire was assessed using a panel of six experts from Australia and Sri Lanka, including two researchers, a pharmacologist, a clinical psychologist, a clinician and a pharmacist. The reliability, clarity, comprehension and time of completion of the modified questionnaire were assessed among ten purposely selected community pharmacists and who were excluded from the final survey. The initial questionnaire was developed in English and translated into the two primary languages spoken in Sri Lanka using forward-backwards and reconciliation process \[[@pone.0215484.ref013], [@pone.0215484.ref020]\].
Measures {#sec012}
--------
i. Socio-demographic and professional characteristics; including age, gender, geographical location, level of pharmacy education, years of working experience in the community pharmacy settings, employment type, employment status, type of pharmacy, and number of pharmacists and pharmacy assistants working at a given time in the pharmacy, were included.
ii. Knowledge was assessed using four sub-groups (concepts) including knowledge about antibiotics, ABR, use/misuse of antibiotics and legal aspects of antibiotics dispensing. Each sub-group had statements to measure the concept with "Yes", "No" and "Unsure" response options. Knowledge about antibiotics was measured using three statements. For each statement a score of one was assigned to a correct response, and these scores were summed to get an overall score (0--3). Since the index score of this variable was normally distributed, this variable was treated as a continuous variable to predict antibiotic dispensing behaviours of pharmacists \[[@pone.0215484.ref020], [@pone.0215484.ref021]\]. Knowledge about ABR (12 items), use/ misuse of antibiotics (14 items) and legal aspects of antibiotic dispensing (5 items) were similarly scored and treated as predictors for antibiotic dispensing behaviour. The distributions of the index scores of all the created score variables were normally distributed except legal aspect of antibiotic dispensing. This variable remained highly skewed even in different possible transformed status. Therefore, the six-point index score was re-coded into two levels by a median split (median = 5), poor knowledge about legal aspects of antibiotic dispensing (score ≤4), and high knowledge (score = 5), to use in subsequent analyses.
iii. Antibiotic dispensing behaviour was assessed using the following six statements: (1) I dispense antibiotics without a prescription if a patient requests it; (2) I give antibiotics without a prescription for a child with minor viral infections; (3) I give antibiotics without a prescription for an adult with minor viral infections; (4) I give antibiotics without a prescription for a child with minor bacterial infections; (5) I give antibiotics without a prescription for an adult with minor bacterial infections; (6) If I know the patient, I dispense antibiotic without a prescription on patient's request. Each of these items had five response options on a Likert scale: (1) never, (2) some time, (3) half of the time, (4) most of the time, and (5) always.
Additionally, self-reported rates of antibiotic supply without a prescription in the previous week by pharmacy staff for specific minor infections; including acute sore throat, common cold, wound infections, UTI and diarrhoea were also assessed using a Likert scale with six points (0% = 1 never, 2 = 25% of instances, 3 = 50% of instances, 4 = 75% of the instances, 5 = always and don't know).
iv. Pharmacist: Those who reported possessing a degree in pharmacy, proficiency or diploma in Pharmacy, efficiency (apprentice pharmacy training) qualification were categorised as pharmacists and the rest were considered as pharmacy assistants ([Table 1](#pone.0215484.t001){ref-type="table"}).
10.1371/journal.pone.0215484.t001
###### Socio-demographic and professional characteristics.
{#pone.0215484.t001g}
Characteristics Frequency (%)
-------------------------------------------------------------------- --------------- ------------ ----------- ---------
**Gender** .931
Male 172 (64.9) 137 (65.2) 35 (63.6)
Female 91 (34.3) 73 (34.8) 18 (32.7)
MD 2 (0.8) 0 2 (3.6)
**Age groups (Years)** .913
20--29 43 (16.2) 31 (14.8) 12 (21.8)
30--39 103 (38.9) 85 (40.5) 18 (32.7)
40--49 65 (24.5) 50 (23.8) 15 (27.3)
≥50 50 (18.9) 43 (20.5) 7 (12.7)
MD 4 (1.5) 1 (0.5) 3 (5.5)
**Geographical area** .353
Urban 189 (71.3) 147 (70.0) 42 (76.4)
Rural 76 (28.7) 63 (30.0) 13 (23.6)
**Level of Pharmacy Education** N/A
Proficiency[^a^](#t001fn002){ref-type="table-fn"} 65 (24.5) 65 (31.0) N/A
Efficiency[^b^](#t001fn003){ref-type="table-fn"} 140 (52.8) 140 (66.7) N/A
Degree[^c^](#t001fn004){ref-type="table-fn"} 5 (1.9) 5 (2.4) N/A
Pharmacy trainee[^d^](#t001fn005){ref-type="table-fn"} 11 (4.2) N/A 11 (20.0)
No pharmacy education 40 (15.0) N/A 40 (71.7)
MD 4 (1.5) 0 4 (7.3)
**Years of work experience in the community pharmacy** \< .001
≤1 21 (7.9) 8 (3.8) 13 (23.6)
2--3 27 (10.2) 25 (11.9) 2 (3.6)
4--5 34 (12.8) 30 (14.3) 4 (7.3)
\>5 178 (67.2) 144 (68.6) 34 (61.8)
MD 5 (1.9) 3 (1.4) 2 (3.6)
**Employment type** .332
Owner (pharmacist or non-pharmacists) 101 (38.1) 84 (40.0) 17 (30.9)
Employee 161 (60.8) 126 (60.0) 35 (63.6)
MD 3 (1.1) 0 3 (5.5)
**Employment status** .823
Full time 219 (82.6) 175 (83.3) 44 (80.0)
Part time 43 (16.2) 35 (16.7) 8 (14.5)
MD 3 (1.1) 0 3 (5.5)
**Type of pharmacy** .214
Rajya Osusala (Semi Government) 20 (7.5) 18 (8.6) 2 (3.6)
Private chain pharmacy 116 (43.8) 88 (41.9) 28 (50.9)
Single private pharmacy 118 (44.5) 97 (46.2) 21 (38.2)
Pharmacies in Private hospitals 11 (4.2) 7 (3.3) 4 (7.3)
**Total number of registered pharmacists working in the pharmacy** .651
None 7 (2.6) 5 (2.4) 2 (3.6)
1 179 (67.5) 143 (68.1) 36 (65.5)
≥2 72 (27.3) 59 (27.3) 13 (23.6)
MD 7 (2.6) 3 (1.4) 4 (7.3)
**Number of Pharmacist at any given time** .530
0 7 (2.6) 4 (1.9) 3 (5.5)
1 195 (73.6) 158 (75.2) 37 (67.3)
2 23 (8.7) 18 (8.6) 5 (9.1)
\>2 15 (5.8) 15 (7.3) 0
MD 25 (9.4) 15 (7.1) 10 (18.2)
MD--Missing data; N/A--Not applicable
^a^ Pharmacists with two years certificate or diploma qualification including 6 months hospital training
^b^ Pharmacists with an apprentice training program under a trained pharmacist's supervision
^c^ Pharmacists with B.Pharm or BSc pharmacy qualification
^d^ Individual registered for apprentice pharmacy program and undergoing in the training program
1. v\. Definition of antibiotics: This item was created based on responses of two statements asked from the respondents: (1) Antibiotic is an agent used to kill or inhibit the growth of microorganisms (bacteria, fungus, virus, parasites). (2) Antibiotic is an agent to kill or inhibit the growth of bacteria. Both statements had three response options: (1) Yes, (2) No, (3) Unsure. Respondents were considered having knowledge about the definition of antibiotic if they gave the correct responses to both statements.
Data analysis {#sec013}
-------------
Completed questionnaires were coded, reviewed for accuracy, entered into a database in the SPSS (Version 24.0; IBM Corporation, Somers, NY), and analysed using descriptive and inferential statistics. All categorical variables, including respondents' socio-demographic and professional characteristics and antibiotic dispensing practices, were expressed as frequencies and percentages. The knowledge scores were calculated as the mean (±SD) or frequency (%). The influence of respondents' professional and demographic factors on knowledge were tested using t-test, one-way ANOVA or chi-square test. Binary logistic regressions and multiple logistic regressions were performed to examine the association of knowledge about antibiotics, and socio-demographic and professional characteristics on reported antibiotic dispensing practice (dispensing antibiotics without a prescription in general and for specific infections).
All the predictors adjusted in the multiple regression models were entered simultaneously to predict the unique contribution of the variables to the outcomes of interest. The predictors adjusted in the models including gender (male = 1; female = 2), geographical area (urban = 1, rural = 2), age (continuous, in years), years of community pharmacy experience (continuous), employment type (owner = 1, employee = 2), employment status (fulltime = 1, part time = 2), knowledge about antibiotics (continuous, score range 0--3), knowledge about ABR (continuous, score range 0--12), knowledge about antibiotic use/ misuse (continuous, score range 0--14) and knowledge about legal aspect of antibiotic use (score 0--4 low knowledge = 0, score 5 high knowledge = 1) were entered into the models.
Odds ratios (OR), adjusted odds ratios (Adj. OR) and 95% confidence intervals (CI) were calculated where appropriate. All *p*-values presented are two tailed and level of significance was set at *p* \< .05.
Results {#sec014}
=======
In total, 267 pharmacy staff (pharmacists and pharmacy assistants) were surveyed (response rate = 72% of the original sample) and 265 participants were included in the final analyses as two questionnaires were incomplete. Of the participants included in this study, 21% (n = 55) were categorised as pharmacy assistants and 79% (n = 210) as pharmacists.
Socio-demographic and professional characteristics of the respondents {#sec015}
---------------------------------------------------------------------
[Table 1](#pone.0215484.t001){ref-type="table"} presents the difference in sample characteristics of pharmacists and pharmacy assistants. There was no significant difference observed between pharmacists and pharmacy assistants with regards to the socio-demographic and professional characteristics measured except the years of working experience. Pharmacists had worked for significantly longer period than assistants (χ^2^ (3, N = 265) = 26.85, *p* \< 0.001). About two-thirds of pharmacy staff were male (172/265, 65%), aged between 30--49 years (168/265, 63.4%) and with more than five years of community pharmacy work experience (178/265, 67.2%). More than two-thirds of community pharmacies were in urban (71.3%) areas. Only about 2% of community pharmacists had a formal university degree in pharmacy and over half of the (53%) pharmacists had an efficiency (apprentice) pharmacy training.
Self-reported knowledge about antibiotics {#sec016}
-----------------------------------------
Although all mean knowledge scores were numerically higher for the pharmacists compared to the assistants ([Table 2](#pone.0215484.t002){ref-type="table"}), only the mean knowledge score related to ABR was statistically significant t(262) = 4.98, *p* = 0.021. Furthermore, a significantly higher proportion of pharmacists were aware of the legal aspects of antibiotic dispensing compared to the assistants (χ^2^ (1, N = 265) = 8.55, *p* = 0.003) ([Table 2](#pone.0215484.t002){ref-type="table"}). However, less than a quarter of the staff (22%; 58/259) correctly defined the term "antibiotic" 24% (49/205) pharmacists and 2% (9/54) of the pharmacy assistants. The correct responses provide to the ABR related items included, knowing that "resistant bacteria can be spread in health institutions and communities" (57% of pharmacists and 56% of assistants responded correctly) and that "dispensing antibiotics shorter than normal course by a pharmacist is one of the causes of ABR" (34% of pharmacists, 29% of assistants). Similarly, items measuring misuse of antibiotics were also correctly responded by about half or less than half of the staff; these items included, "viral diseases can be treated with antibiotics" (56% of pharmacists, 53% of assistants), "Antibiotics can be used as a preventive measure to fight against future microbial attacks" (38% of pharmacists, 33% of assistants), and correct knowledge about treating acute sore throat (32% of pharmacists, 24% of assistants) and acute diarrhoea (57% of pharmacists, 55% of assistants) with antibiotics were also low.
10.1371/journal.pone.0215484.t002
###### Knowledge scores of pharmacy staff.
{#pone.0215484.t002g}
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
All the Respondents Pharmacy assistants\ Pharmacists\ t-test *p* value
Mean (±SD) Mean (±SD)
---------------------------------------------------------- --------------------------------------------------- ------------------------------------------------- --------------------------------------------------- ---------------------------------------------
Overall knowledge (n = 265)\ 26.07 (3.86) 24.96 (3.86) 26.36 (3.82) .017
(Possible score range---0--34)
Knowledge about Antibiotic (n = 263)\ 1.89 (.69) 1.87 (.64) 1.90 (.70) .801
(Possible score range---0--3)
Knowledge about ABR (n = 264)\ 9.17 (1.67) 8.64 (1.97) 9.32 (1.56) .021
(Possible score range---0--12)
Knowledge about antibiotic use/ misuse (n = 264)\ 10.42 (2.06) 10.05 (2.10) 10.51 (2.04) .144
(Possible score range---0--14)
Knowledge about legal aspect of antibiotic use (n = 265) 0.003[^b^](#t002fn002){ref-type="table-fn"}
Low knowledge 127 (49.90)[^a^](#t002fn001){ref-type="table-fn"} 36 (65.5)[^a^](#t002fn001){ref-type="table-fn"} 91 (43.3)[^a^](#t002fn001){ref-type="table-fn"}
High knowledge 138 (52.10)[^a^](#t002fn001){ref-type="table-fn"} 19 (34.5)[^a^](#t002fn001){ref-type="table-fn"} 119 (56.70)[^a^](#t002fn001){ref-type="table-fn"}
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
^a^ Proportion and percentage;
^b^Chi-square test *P* value.
Self-reported antibiotic dispensing practice {#sec017}
--------------------------------------------
As [Table 3](#pone.0215484.t003){ref-type="table"} illustrates, one in every three-pharmacy staff reported that they dispensed antibiotics without a prescription on patient request and the proportion of such dispensing practice increased to about half when the patient was known to them. Overall, 30% of the staff reported that they dispensed antibiotics without a prescription when considering all of the infections assessed, including, acute sore throat, common cold, acute diarrhoea, wound infection or uncomplicated UTI in the week prior to completing the survey. The highest reported antibiotic dispensing was for wound infections (21% of responding staff). The dispensing practice was not significantly different between pharmacists and pharmacy assistants (*p*\>0.05) in general. However, a significantly higher proportion of the pharmacy assistants reported having dispensed antibiotics without a prescription compared to the pharmacists in the past seven days for sore throat (χ^2^ (1, N = 253) = 4.20, *p* = 0.040).
10.1371/journal.pone.0215484.t003
###### Self-reported antibiotic dispensing practice.
{#pone.0215484.t003g}
---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
\ Frequency (%), N = 265
Dispensing practice items
--------------------------------------------------------------------------------------------------------------------- ------------------------ ----------- ------------ ---------------------------------------------
**Dispense antibiotics without a prescription on patient demand** 0.805
Never 179 (67.5) 36 (66.7) 143 (68.4)
Yes 84 (31.7) 18 (33.3) 66 (31.6)
MD 2 (.8)
**If I know the patient, I dispense antibiotics without a prescription on the patient's request** 0.055
Never 147 (55.5) 32 (59.3) 115 (54.8)
Yes 117 (44.2) 22 (40.7) 95 (45.2)
MD 1 (.4)
**I give antibiotics without prescription for adult patients with minor ailments caused by viral infection** 0.077
Never 233 (87.9) 51 (96.2) 182 (87.9)
Yes 27 (10.2) 2 (3.8) 25 (12.1)
MD 5 (1.9)
**Children who have viral infections, I dispense antibiotics without a prescription** 0.127[^a^](#t003fn002){ref-type="table-fn"}
Never 247 (93.2) 53 (100) 194 (94.6)
Yes 11 (4.2) 0 11 (5.4)
MD 7 (2.6)
**I give antibiotics without a prescription for adult patients with minor ailments caused by bacterial infections** 0.640
Never 216 (81.5) 43 (79.6) 173 (82.4)
Yes 48 (18.1) 11 (20.4) 37 (17.6)
MD 1 (.4)
**Children who have bacterial infections, I dispense antibiotics without a prescription** 0.555^a^
Never 245 (92.5) 49 (90.7) 196 (93.3)
Yes 19 (7.2) 5 (9.3) 14 (6.7)
MD 1 (.4)
**Proportion of dispensed antibiotics without a prescription in the last week for following minor infections**
**Acute sore throat** 0.040
Never (0%) 217 (81.9) 40 (76.9) 177 (88.1)
Yes 36 (13.6) 12 (23.1) 24 (11.9)
MD 12 (4.5)
**Common cold and cough** 0.050
Never (0%) 214 (80.8) 38 (74.5) 176 (85.9)
Yes 42 (15.8) 13 (25.5) 29 (14.1)
MD 9 (3.4)
**Wound infection** 0.940
Never (0%) 199 (75.1) 40 (78.4) 159 (77.9)
Yes 56 (21.1) 11 (21.6) 45 (22.1)
MD 10 (3.8)
**Urinary tract infections** 1.000^a^
Never (0%) 231 (87.2) 47 (92.2) 184 (90.6)
Yes 23 (8.7) 4 (7.8) 19 (9.4)
MD 11 (4.2)
**Diarrhoea** 0.190
Never (0%) 227 (85.7) 43 (84.3) 184 (90.6)
Yes 27 (10.2) 8 (15.7) 19 (9.4)
MD 11 (4.2)
**Dispensed antibiotic without a prescription in the last week for any of minor infections** 0.753
Never (0%) 179 (67.5) 35 (67.3) 144 (69.6)
Yes 80 (30.2) 17 (32.7) 63 (30.4)
MD 6 (2.3)
---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
MD--Missing data;
^a^ Fisher's exact test.
Impact of community pharmacists' knowledge, other socio-demographic and professional characteristics on antibiotic dispensing without a prescription {#sec018}
----------------------------------------------------------------------------------------------------------------------------------------------------
The pharmacists' knowledge about ABR significantly reduced the likelihood of dispensing antibiotics without a prescription for viral infections in adults (Adj. OR = .73, 95% CI: .55-.96; *p* = .027) and children (Adj. OR = .55, 95% CI: .38-.80; *p* = .002). Knowledge of the legal aspects of antibiotic dispensing also significantly decreased the likelihood of non-prescription antibiotic dispensing for unspecified patients (Adj. OR = .41, 95% CI: .21-.79; *p* = .008), the patients known to the pharmacist (Adj. OR = .29, 95% CI: .16 -.55; *p* \< .001) requesting an antibiotic and adults with bacterial infections (Adj. OR = .45, 95% CI: .20-.99; *p* = .047) ([Table 4](#pone.0215484.t004){ref-type="table"}).
Pharmacist employees were less likely to dispense antibiotics without a prescription for a direct patient request compared to owner pharmacists (Adj. OR = .43, 95% CI: .22 - .85; *p* = .015) ([Table 4](#pone.0215484.t004){ref-type="table"}). All other demographic characteristics, such as gender, geographical area, years of community pharmacy experience and employment status, once adjusted in the regression model, were not significantly associated with non-prescription antibiotic dispensing practice.
10.1371/journal.pone.0215484.t004
###### Impact of pharmacists' knowledge and socio-demographic on antibiotic dispensing practice in general.
{#pone.0215484.t004g}
Predictors Dispensing antibiotic without a prescription Adj.OR (95% CI) *p*
------------------------------------------------ ------------------------------------------------------------------ --------------------------------------------------------- -------------------------------------------------------- -------------------------------------------------------- ------------------------------------------------------- -------------------------------------------------------
Knowledge about ABR 1.12(.89, 1.40)[^NS^](#t004fn005){ref-type="table-fn"} 1.24(.99, 1.53)[^NS^](#t004fn005){ref-type="table-fn"} .73(.55, .96)[\*](#t004fn004){ref-type="table-fn"} .55(.38, .80)[\*\*](#t004fn003){ref-type="table-fn"} .92(.72, 1.18)[^NS^](#t004fn005){ref-type="table-fn"} 1.00(.66 1.50)[^NS^](#t004fn005){ref-type="table-fn"}
Knowledge about antibiotic use/ misuse 1.03(.78, 1.21)[^NS^](#t004fn005){ref-type="table-fn"} .95(.82, 1.11)[^NS^](#t004fn005){ref-type="table-fn"} 1.02(.82, 1.28)[^NS^](#t004fn005){ref-type="table-fn"} 1.02(.73, 1.41)[^NS^](#t004fn005){ref-type="table-fn"} .89(.73, 1.08)[^NS^](#t004fn005){ref-type="table-fn"} .69(.49, .96)[\*](#t004fn004){ref-type="table-fn"}
Knowledge about legal aspect of antibiotic use
Low 1 1 1 1 1 1
High .41(.21, .79)[\*\*](#t004fn003){ref-type="table-fn"} .29 (.16, .55)[\*\*\*](#t004fn002){ref-type="table-fn"} .80(.32, 2.01)[^NS^](#t004fn005){ref-type="table-fn"} 1.81(.44, 7.44)[^NS^](#t004fn005){ref-type="table-fn"} .45(.20, .99)[\*](#t004fn004){ref-type="table-fn"} .37(.09, 1.47)[^NS^](#t004fn005){ref-type="table-fn"}
Age .98(.94, 1.02)[^NS^](#t004fn005){ref-type="table-fn"} .96(.92, .99)[\*](#t004fn004){ref-type="table-fn"} .94(.88, 1.01)[^NS^](#t004fn005){ref-type="table-fn"} 1.00(.92, 1.09)[^NS^](#t004fn005){ref-type="table-fn"} .93 (.87, .98)[\*](#t004fn004){ref-type="table-fn"} 1.08(1.01, 1.15)[\*](#t004fn004){ref-type="table-fn"}
**Employment Type**
Owner 1 1 1 1 1 1
Employee .43 (.22, .85)[\*](#t004fn004){ref-type="table-fn"} 1.08(.58, 2.04) .70(.27, 1.82)[^NS^](#t004fn005){ref-type="table-fn"} .88(.21, 3.75)[^NS^](#t004fn005){ref-type="table-fn"} .69(.31, 1.57)[^NS^](#t004fn005){ref-type="table-fn"} .41(.11, 1.54)[^NS^](#t004fn005){ref-type="table-fn"}
*p* value:
\*\*\*\<0.001,
\*\*\<0.01,
\*\<0.05;
^NS^ not significant.
Other predictors adjusted in the model: Knowledge about antibiotic, gender, geographical area, years of community pharmacy experience and employment status.
As [Table 5](#pone.0215484.t005){ref-type="table"} describes, Pharmacists' knowledge about antibiotic use/ misuse significantly reduced the likelihood of dispensing antibiotics without a prescription for common cold (Adj. OR = .75, 95% CI: .60 - .94; *p* = .011) and acute diarrhoea (Adj. OR = .76, 95% CI: .58 - .99; *p* = .047). However, other knowledge items and demographic characteristics such as geographical location of the pharmacy, years of community pharmacy experience, employment status and employment type, once adjusted in the regression model were not significantly associated with antibiotic dispensing practice without a prescription for the minor infections.
10.1371/journal.pone.0215484.t005
###### Impact of pharmacists' knowledge and socio-demographic on dispensing antibiotic past week for minor infections.
{#pone.0215484.t005g}
Predictors Dispensing antibiotic without a prescription Adj. OR (95% CI) *p*
------------------------------------------------ ------------------------------------------------------------------- -------------------------------------------------------- -------------------------------------------------------- -------------------------------------------------------- -------------------------------------------------------
Knowledge about AB .81(.43, 1.54)[^NS^](#t004fn005){ref-type="table-fn"} 1.19(.65, 2.20)[^NS^](#t004fn005){ref-type="table-fn"} .73(.44, 1.21)[^NS^](#t004fn005){ref-type="table-fn"} .55(.26, 1.16)[^NS^](#t004fn005){ref-type="table-fn"} .99(.47, 2.10)[^NS^](#t004fn005){ref-type="table-fn"}
Knowledge about ABR .83(.61, 1.11)[^NS^](#t004fn005){ref-type="table-fn"} 1.01(.78, 1.43)[^NS^](#t004fn005){ref-type="table-fn"} 1.01(.79, 1.29)[^NS^](#t004fn005){ref-type="table-fn"} 1.09(.75, 1.82)[^NS^](#t004fn005){ref-type="table-fn"} .94(.66, 1.34)[^NS^](#t004fn005){ref-type="table-fn"}
Knowledge about antibiotic use/ misuse .81(.64, 1.02)[^NS^](#t004fn005){ref-type="table-fn"} .75(.60, .94)[\*](#t005fn002){ref-type="table-fn"} .89(.74, 1.07)[^NS^](#t004fn005){ref-type="table-fn"} .86(.66, 1.13)[^NS^](#t004fn005){ref-type="table-fn"} .76(.58, .99)[\*](#t005fn002){ref-type="table-fn"}
Knowledge about legal aspect of antibiotic use
Low 1 1 1 1 1
High 1.23(.46, 3.33)[^NS^](#t004fn005){ref-type="table-fn"} .94(.38, 2.29)[^NS^](#t004fn005){ref-type="table-fn"} .63(.30, 1.32)[^NS^](#t004fn005){ref-type="table-fn"} .49(.17, 1.45)[^NS^](#t004fn005){ref-type="table-fn"} .93(.31, 2.81)[^NS^](#t004fn005){ref-type="table-fn"}
Age .95(.89, 1.02)[^NS^](#t004fn005){ref-type="table-fn"} .94(.88, 1.00)[^NS^](#t004fn005){ref-type="table-fn"} .95(.90, 1.0)[^NS^](#t004fn005){ref-type="table-fn"} .97(.91, 1.31)[^NS^](#t004fn005){ref-type="table-fn"} .96(.89, 1.03)[^NS^](#t004fn005){ref-type="table-fn"}
*P* value:
\*\<0.05;
^NS^ not significant.
Other predictors adjusted in the model: gender, geographical area, years of community pharmacy experience, employment status and employment type.
Discussion {#sec019}
==========
To the best of our knowledge, this is the first nationwide study from Sri Lanka evaluating the self-reported antibiotic dispensing practice of community pharmacy staff (pharmacists and pharmacy assistants) and the impact of pharmacists' knowledge and professional and demographic characteristics on antibiotic dispensing.
This study, demonstrated that whilst pharmacists showed a greater knowledge about ABR and legal aspects of antibiotic dispensing than the pharmacy assistants, they had an overall poor knowledge of antibiotics (definition), appropriate antibiotic use and antibiotic resistance. This poor knowledge was reflected in the self-reported behaviour, with a large proportion of participants providing antibiotics to patients on request, and inappropriately in situations where patients may have had viral infections. The study findings also demonstrated that pharmacists who reported knowing about the legal requirements for supplying antibiotics with a prescription and who were knowledgeable about ABR and the use/ misuse of antibiotics, were less likely to dispense antibiotics without a prescription.
Our findings regarding poor knowledge about antibiotics are similar to findings from other studies conducted in LMICs \[[@pone.0215484.ref012]--[@pone.0215484.ref014], [@pone.0215484.ref022]\]. Apisarnthanarak *et*.*al*, in 2008, found that around 24% of Thai pharmacists had inadequate knowledge about antibiotics \[[@pone.0215484.ref014]\]. A study conducted in Indian rural pharmacies found that half of the employees could correctly define antibiotics and two-thirds of them had not heard about ABR \[[@pone.0215484.ref012]\]. The knowledge gap we found among the pharmacists can be explained by the limited professional and clinical training of pharmacists in Sri Lanka, particularly as a majority of the community pharmacists in Sri Lanka are apprentice pharmacists without a formal tertiary education or clinical training \[[@pone.0215484.ref023]\]. Thus, limited knowledge about antibiotics, their use/ misuse, as well as ABR could be a large factor influencing illegal and inappropriate supply of antibiotics without prescription to patients with minor infections presenting at community pharmacies. This educational gap presents an opportunity for strategies to be developed to address the need for further education and which can be easily addressed in the short term, both at all types of pharmacy programs and at the practising pharmacy level through the delivery of targeted continuing educational programs. Educating and upskilling pharmacists will have a long term impact on reducing the illegal and inappropriate supply of antibiotics to the public and will also ensure that pharmacists take up the role of educating not only their own staff in the pharmacy, but also the public about appropriate use of antibiotics. Overall, addressing this gap, will meet long term objectives of reduced antibiotic misuse and inappropriate use, and contribute to reduced antibiotic resistance nationally and globally \[[@pone.0215484.ref006], [@pone.0215484.ref024], [@pone.0215484.ref025]\]. Specifically, as Sri Lanka is one of the tourist destinations, millions of travellers visit Sri Lanka every year from all over the world and studies show that resistance pathogens can be transmitted globally with international travellers \[[@pone.0215484.ref025]\].
Targeted education of practising pharmacists and other staff may also specifically address other issues highlighted in this study; for example, provision of antibiotics more readily to patients known to the pharmacy staff. Familiarity with the patient appeared to be a reason why antibiotics were more likely to be provided without a prescription, with no significant difference observed between pharmacists and assistants. These findings are consistent with similar studies conducted to evaluate the inappropriate dispensing practice in LMICs \[[@pone.0215484.ref026], [@pone.0215484.ref027]\]. It is possible that pharmacy staff feel that they have a greater obligation to supply antibiotics to patients they know. It is also possible that they may feel that they are more familiar with the patients' health and medical conditions and therefore are more comfortable in providing antibiotics without a prescription. Equally, the familiarity may also allow increased opportunities for monitoring the health of the patient and evaluating the impact of the short-term supply of an antibiotic.
In addition to inadequate clinical experience and knowledge about antibiotic dispensing contributing to the illegal and inappropriate antibiotic supply reported in this study, there may be other reasons explaining the behaviour of the pharmacists and pharmacy assistants. For example, pressure from the public to supply an antibiotic, the profit motive, disregard for their legal obligations related to antibiotic supply, lax law enforcement, lack of awareness about existing antibiotic dispensing laws and fear of losing customers \[[@pone.0215484.ref013]\].
However, our present study also revealed that having pharmacists with a better knowledge of ABR and use/ misuse could reduce antibiotic supply without prescription for viral infections (appropriate practice); and better knowledge related to dispensing regulations may reduce antibiotics supply without a prescription (illegal supply) to both known and unknown patients and for bacterial infections. Recent literature from other countries has also found similar relationships \[[@pone.0215484.ref013], [@pone.0215484.ref014], [@pone.0215484.ref028]\]. Therefore, in addition to continuous education for pharmacists, revising the curricula of pharmacy programs and strict enforcement of antibiotic dispensing related regulations in the country should also be established.
However, the control of ABR cannot be the sole responsibility of health professionals and scientists. The public and other stakeholders have a major role to play as well. Public awareness should be encouraged to minimise patient demand for antibiotics and optimise appropriate use. A survey of parents of school children in Italy found that about 10% and 21% correctly knew the definition of antibiotics and ABR, respectively, and one in every third person surveyed was willing to self-medicate with an antibiotic \[[@pone.0215484.ref029]\]. Internet and social media have become popular among the public as a source of antibiotic related information \[[@pone.0215484.ref030]\]. Therefore, these sources could also be considered as one of the key tools for public education on appropriate use of antibiotics.
The government should consider ABR as a major public health issue. Policies and regulations should be put in place to enforce appropriate access, minimise the public's demand from health professionals and reduce inappropriate use of antibiotics. The media professionals should also be adequately trained on how to convey medical and scientific information in lay language to inform the populations on practices that promote the appropriate use of antibiotics. The empirical evidence may facilitate the development of educational and behavioural interventions, dispensing guides, and strengthening the policies which can be undertaken to combat against this public health issue.
Limitations {#sec020}
-----------
Like most surveys, our study also has some limitations. There is a possibility of social desirability bias on the antibiotic dispensing practice-related issues, where the respondents may give more favourable responses about their antibiotic dispensing practice \[[@pone.0215484.ref013]\]. However, to address this limitation and assess the validity of the findings, two separate pseudo-patient visits to the same pharmacies directly requesting an antibiotic product (61%) \[[@pone.0215484.ref010]\] and presenting symptoms (41%) \[[@pone.0215484.ref011]\] were conducted. The comparison of the three methods will be the focus of a future publication.
Approximately 20--30 minutes was taken to complete the questionnaire, which could have impacted the responses through survey fatigue \[[@pone.0215484.ref031]\].
Another possible limitation of this study is non response bias caused by those who did not respond. However, in our study, the response rate was very high (72%), when compared to similar research \[[@pone.0215484.ref032]\] with responses from all the different types of pharmacies in Sri Lanka. The results of this study may only be applicable in environments where the legislative framework regarding antibiotic supply and pharmacists' qualifications are comparable with Sri Lankan context.
Conclusions {#sec021}
===========
Despite Sri Lankan law prohibiting provision of antibiotics to the public without a prescription, this study revealed that pharmacy staff continue to provide antibiotics without a prescription. Pharmacists did not report an overall lower provision rate. The illegal and inappropriate antibiotics supply was associated with lower levels of pharmacists' legal and clinical knowledge about antibiotics.
There is room for improvement in community pharmacy staff's basic knowledge about antibiotics, antibiotic resistance, antibiotic use and misuse. It is important that the government strictly enforces the law regarding antibiotics supply. Furthermore, standards should be implemented and adopted on the appropriate provision and supply of antibiotics. Educational programmes and interventions should be developed and implemented to ensure that both newly qualifying pharmacists as well as currently practicing pharmacists in the community settings are well-informed about antibiotics, appropriate use and antibiotic resistance, as well as the legal requirements for supply.
Supporting information {#sec022}
======================
###### Questionnaire.
(PDF)
######
Click here for additional data file.
The authors acknowledge following people M. Bushell, Vivien Tong, Carl Schneider, Sudheera Jayasinghe, Bilesha Perera, Maneesha Prasadi, Nafas Nawfer, Mishfaq Abdullah, M. N. M. Fazlan, Shyamalie Hettigoda, M. T. M. Thanis, Nadeesha Lakmali, Fathima Zilmiya, Thivyabharathy Mohanasundaram and Mohamed Inamullah for their contribution in conducting and publishing this research. We are also thankful to the pharmacy staff who participated in the study.
[^1]: **Competing Interests:**The authors have declared that no competing interests exist.
| {
"pile_set_name": "PubMed Central"
} |
1. Introduction {#sec1}
===============
*Avena nuda* L., also named naked oats, is originated and widely separated in north and high altitude region in China. It belongs to herb of gramineous plant with annual growing. Naked oats has great value in nutrition and medicine. It contains abundant proteins with 18 kinds of amino acid, and lots of unsaturated fatty acids. Naked oats likes to grow under cool weather and has tolerance to low temperature. Naked oats is a typical temperate crop adapted to cool climates.
Low temperature is a major abiotic stress that limits the growth, productivity, and geographical distribution of agricultural crops and can lead to significant crop loss \[[@B1], [@B2]\]. To cope with low temperature, plants have evolved a variety of efficient mechanisms that allow them to adapt to the adverse conditions \[[@B3], [@B4]\]. This adaptive process involves a number of biochemical and physiological changes, including increased levels of proline, soluble sugars, and MDA, as well as enzyme activities \[[@B5]\].
Understanding the mechanisms of low temperature adaptation is crucial to the development of cold-tolerant crops. The study was designated to explore the physiological mechanism of cold tolerance in naked oats. The responses of the *Avena nuda* L. seedlings to low temperature stress were also evaluated by measuring electrolyte leakage (EL), chlorophyll content, and the concentration of MDA. We measured and compared these indices of seedlings leaves under low temperature and normal temperature. The study provided theoretical basis for cultivator and antibiotic breeding in *Avena nuda* L.
2. Materials and Methods {#sec2}
========================
2.1. Materials and Cold Treatment {#sec2.1}
---------------------------------
Naked oats cultivar Jinyan 14 (*Avena nuda* L.) was used in the experiment. Seeds were sterilized by incubation for 1 min in 75% ethanol and then washed thoroughly with sterile water. The seeds were germinated in soil in pots at 20°C under long-day conditions (16 h of cool white fluorescent light, photon flux of 70 umol m^−2^ s^−1^).
Seedlings at the four-leaf stage were subjected to cold stress. Plants were divided into three groups, one group was under normal temperature as control, the other two were subjected to low temperature processing, and each group had the stress repeated three times. Seedlings of the control group were grown at 20°C continuously. For low temperature treatments, seedlings were transferred to a temperature of 1°C and −10°C in an artificial climate box under the same light and photoperiodic conditions for 7 days.
The leaves were sampled after 0, 1, 3, 5 and, 7 d of treatment for next measurement. The leaf samples were immediately frozen in liquid nitrogen and stored at −80°C until use. Three independent biological samples for each treatment were harvested, and each replicate contained 10 plants.
2.2. Determination of Relative Electrolyte Leakage {#sec2.2}
--------------------------------------------------
For electrolyte leakage measurement, protocol was used as described \[[@B6]\]. Briefly, 100 mg leaves were placed in 25 mL distilled water, shaken on a gyratory shaker (200 rpm) at room temperature for 2 h, and the initial conductivity (*C*1) was measured with a conductivity instrument. The samples were then boiled for 10 min to induce maximum leakage. After cooling down at room temperature, electrolyte conductivity (*C*2) was measured, and the relative electrical conductivity (*C*%) was calculated based on (*C*1/*C*2) × 100. All low temperature testing experiments were repeated three times. A paired *t*-test was used to determine the difference between the cold treatment and normal condition.
2.3. Determination of Chlorophyll Content {#sec2.3}
-----------------------------------------
For estimation of total chlorophyll, protocol was followed as described \[[@B7]\]. About 100 mg of fine powder of leaf tissue was homogenized in 1 mL of 80% acetone and kept for 15 min at room temperature in dark. The crude extract was centrifuged for 20 min at 10,000 rpm at room temperature, and the resultant supernatant was used for assessing absorbance at 633 and 645 nm with a spectrophotometer. Total chlorophyll content was computed in terms of fresh weight (FW).
2.4. Determination of Proline Content {#sec2.4}
-------------------------------------
Proline concentrations in naked oats leaves were measured by the sulfosalicylic acid-acid ninhydrin method with slight modifications \[[@B8]\]. Around 100 mg of tissues were used and extracted in 5 mL of 3% sulphosalicylic acid at 95°C for 15 min. After filtration, 2 mL of supernatant was transferred to a new tube containing 2 mL of acetic acid and 2 mL of acidified ninhydrin reagent. After 30 min of incubation at 95°C, samples were kept at room temperature for 30 min and 5 mL of toluene was added to the tube with shaking at 150 rpm to extract red products. The absorbance of the toluene layer was determined at 532 nm using spectrophotometer.
2.5. Determination of Malondialdehyde (MDA) Content {#sec2.5}
---------------------------------------------------
Malondialdehyde (MDA) content in naked oats leaves was determined following the protocols as described \[[@B9]\]. Briefly, leaves were homogenized in 5 mL of 10% trichloroacetic acid containing 0.25% thiobarbituric acid. The mixture was incubated in water at 95°C for 30 min, and the reaction was stopped in an ice bath. The mixture was centrifuged at 10,000 g for 20 min, and the absorbance of the supernatant was measured at 450, 532, and 600 nm.
2.6. Determination of Peroxidase, Superoxide Dismutase, and Catalase Activity {#sec2.6}
-----------------------------------------------------------------------------
Naked oats leaves (0.5 g) were ground thoroughly with a cold mortar and pestle in 50 mmol potassium phosphate buffer (pH 7.8) containing 1% polyvinylpyrrolidone. The homogenate was centrifuged at 15,000 g for 20 min at 4°C. The supernatant was crude enzyme extraction. The activities of peroxidase (POD), superoxide dismutase (SOD), and catalase (CAT) were measured using the protocols described \[[@B10]\].
3. Results and Discussion {#sec3}
=========================
3.1. Phenotypic Changes under Cold Stress {#sec3.1}
-----------------------------------------
We investigated the phenotypic response to low temperature. During the cold treatment of 1°C, naked oats grew well as usual. Until 5 days later, the seedlings were always strong only except some leaf apexes began to get yellow. In the 7th day, most parts of seedling remained green as normal temperature as shown in [Figure 1](#fig1){ref-type="fig"}. The seedling got curl after 3-4 hours after exposure to −10°C cold stress. Some leaf began to get yellow and curled seriously in the third day, while the seedling grew slowly. Most leaves showed severe rolling and wilting in the 7th day.
No obvious differences have been found in growth and developed between normal temperature and at 1°C, indicating that the naked oats have cold tolerance at the low temperature of 1°C.
3.2. Changes of Relative Electrolyte Leakage under Cold Stress {#sec3.2}
--------------------------------------------------------------
Cold stress often causes damage to cell membranes, so, we tested the cell-membrane penetrability. Cell membrane penetrability was evaluated by the relative conductance of the cell membrane under cold stress \[[@B11], [@B12]\]. The electrolyte leakage test was performed to compare membrane integrity \[[@B13]\]. For such experiment, plants were subjected to low temperature of 1°C and −10°C. The relative electrolyte leakage of seedling leaves was increased greatly with cold treatment as shown in [Figure 2](#fig2){ref-type="fig"}. There was no significant difference under normal temperature; the average electrolyte leakage was about 9.7%. During the cold stress, the relative electrolyte leakage value gradually increased with the prolongation of low temperature stress. The electrolyte leakage of −10°C was significantly higher than that of 1°C growing plants. In the 7th day, elative electrolyte leakage increased to 66.0% under cold stress of −10°C, while 37.3% at 1°C and the normal condition still 11.0%.
When the plants were under low temperature stress, the structure of cellular membrane was damaged. The degree of cell membrane injury induced by cold stress can be reflected by intracellular electrolyte leakage rate. The relative conductance value is one of the effective indicators to indirectly evaluate plant response ability to low temperature stress \[[@B14]\]. The damage degree of cellular membrane was aggravated with the continuity of low temperature stress \[[@B12]\]. In the experiment, the electrolyte leakage in leaves of naked oats seedlings gradually increased under low temperature of 1°C stress, indicating that the damage of low temperature on cell membrane increased gradually. The electrolyte leakage increased more significantly at −10°C than 1°C, indicating that cell membrane was damaged seriously at −10°C than 1°C. At 1°C, the naked oats had some tolerance to protect the cell membrane avoiding cold damage.
3.3. Changes of Total Chlorophyll Content under Cold Stress {#sec3.3}
-----------------------------------------------------------
Chlorophyll is an extremely important and critical biomolecule in photosynthesis with function of light absorbance and light energy transformation \[[@B15]\]. Low temperature stress can influent plant photosynthesis and decrease the utilization of light \[[@B16], [@B17]\]. We examined the content of chlorophyll under low temperature of 1°C and −10°C. Compared with the control, the chlorophyll content in seedling leaves under low temperature was lower than that under room temperature. As shown in [Figure 3](#fig3){ref-type="fig"}, the change range of chlorophyll content in leaves of naked oats was 6.4--6.9 mg/g under room temperature. At 1°C, the chlorophyll content was decreased with the prolongation of low temperature stress. The chlorophyll decreased slightly from the 1st day to the 5th day, and the content was only slightly less than control, but in 7th day the chlorophyll decreased greatly to 4.6 mg/g, while at −10°C, the chlorophyll content decreased seriously. Especially in the 5th day, chlorophyll content decreased to 2.2 mg/g while 5.8 mg/g at 1°C and 6.5 mg/g at normal temperature. The results indicated that naked oats was damaged less at 1°C than at −10°C. Low temperature inhibits chlorophyll accumulations in actively growing leaves. Naked oats has some degree of cold tolerance at 1°C.
3.4. Changes of Free Proline Content under Cold Stress {#sec3.4}
------------------------------------------------------
Proline is widely distributed in plants as protection material, which is an organic osmolyte \[[@B18]\]. It plays a vital role in maintaining osmotic balance and stabilizing cellular structures in plants. Many plants accumulate free proline in response to abiotic stress of low temperature. Increased free proline content protects the plant against the stress \[[@B19]\].
The effect of cold on content of proline was investigated. There was no significant difference in proline contents without cold stress as shown in [Figure 4](#fig4){ref-type="fig"}. An increase in proline content was observed upon exposure to cold stress. Compared with the control, the free proline content in seedling leaves under low temperature was obviously higher than that under room temperature. At 1°C, the proline content increased with the prolongation of low temperature stress. At −10°C, the proline content was increased in the first five days and reached the max of 601 *μ*g/g in the 5th day, about 6 times of control. Then proline content decreased to 404 *μ*g/g in the 7th day, which was still much higher than control. At −10°C, the plant accumulated more amounts of proline than at 1°C. This indicated that the lower the temperature was, the more proline could be accumulated.
Proline plays a vital role in maintaining osmotic balance in plants. The accumulation of proline may function in preventing plants from being damaged by stress. The free proline acts as osmolytes to facilitate osmoregulation, thus protecting plants from dehydration resulting from cold stress by reducing water potential of plant cells \[[@B20]\]. In addition, proline can also function as a molecular chaperone to stabilize the structure of proteins as well as play a role in regulation of the antioxidant system \[[@B21], [@B22]\]. So increased free proline content protects the plant against the stress. The study found greater accumulation of free proline under cold stress, which may partially account for the higher tolerance of plants to cold stress. Accumulation of proline to facilitate osmo-regulation is a common adaptive mechanism for tolerance of plants to abiotic stress. The results were consistent with previous studies that proline accumulated in leaves exposed to cold, salt, and other stresses \[[@B23]\].
3.5. Changes of Malondialdehyde (MDA) Content under Cold Stress {#sec3.5}
---------------------------------------------------------------
Cold stress often causes damage to cell membranes. Malondialdehyde (MDA) is an important indicator of membrane system injuries and cellular metabolism deterioration \[[@B24]\]. So, we further measured the cell membrane penetrability. The effects of MDA contents were investigated in the seedling of naked oats. In our experiment, naked oats had a significantly higher MDA level under low temperature stress compared to the control level. As shown in [Figure 5](#fig5){ref-type="fig"}, the MDA concentrations increased with the prolongation of low temperature stress. In the seventh day, the MDA content reached to the max nearly 3 times of control at 1°C and about 4 times at −10°C. The increase in MDA content at −10°C was significantly higher than 1°C all the time.
MDA has been well recognized as a parameter reflecting damage by cold stress. Cell membrane systems are the primary sites of freezing injury in plants \[[@B25]\]. Plants subjected to low temperatures frequently suffer membrane damage, which can be evaluated by relative electrolyte leakage and MDA production. MDA is considered to be the final product of lipid peroxidation in the plant cell membrane \[[@B26]\]. MDA is also an important intermediate in ROS scavenging, and a high level of MDA is toxic to plant cells. In this experiment, in the first day the MDA only increased 0.095 *μ*mol*·*g^−1^ at 1°C, while the MDA concentrations increased to 0.122 *μ*mol*·*g^−1^ at −10°C. In the third day, MDA increased slightly to 0.115 *μ*mol*·*g^−1^ at 1°C, while MDA increased rapidly to 0.247 *μ*mol*·*g^−1^ at −10°C, which was 4 times of control. The increase of MDA at −10°C was greater than at 1°C. These results suggested that cell membrane was little damaged at 1°C at the beginning of cold stress, but was hurt seriously at −10°C. It may have contributed to the different phenotypes under cold stress. Prolonged treatment finally led to a great cell damaged and MDA accumulation. The MDA content increased slowly in first three days at 1°C, but increased rapidly to 0.20 *μ*mol*·*g^−1^ in the fifth day. It indicated that cell membranes of seedlings were injured by cold seriously in 3--5 days at 1°C.
3.6. Changes of Antioxidant Enzymes under Cold Stress {#sec3.6}
-----------------------------------------------------
Cold stress induces the accumulation of reactive oxygen species (ROS) such as superoxide, hydrogen peroxide, and hydroxyradicals \[[@B27], [@B28]\]. The elevated concentrations of ROS can damage cellular structures and macromolecules, leading to cell death \[[@B29], [@B30]\]. Plants under abiotic stress have evolved a defense system against oxidative stress by increasing the activity of ROS-scavenging enzyme. ROS can be scavenged by superoxide dismutase (SOD), peroxidase (POD), and catalase (CAT) \[[@B31]\]. SOD plays an important role in eliminating ROS induced by cold \[[@B32]\]. POD and CAT can scavenge H~2~O~2~. SOD, POD, and CAT are important enzymes protecting membrane system. Many kinds of antioxidant enzyme activities have been changed under low stress. Changes of enzyme activities are relvant to cold tolerance.
We measured the activities of these enzymes under normal and cold stress. Compared with normal, the SOD activity was higher under the cold treatment, as shown in [Figure 6(a)](#fig6){ref-type="fig"}. At 1°C, the SOD activity was increased rapidly in the first three days and remained unchanged in later days. At −10°C, the SOD activity was increased first and decreased later. In the third day, the SOD activity was increased to the max value 323 U*·*g^−1^ *·*min^−1^, which was nearly 4 times of normal temperature. Then, the SOD activity was decreased rapidly to a low value of 229 U*·*g^−1^ *·*min^−1^ in the seventh day. At this time, the seedlings were damaged seriously by low temperature. Compared with normal, the SOD activity was higher than normal under the cold treatment. Increased SOD activity contributed for cold tolerance of naked oats under cold stress. Especially in the first day, the SOD activities had increased greatly, indicated that naked oats is resisting and adapting to low temperature, and contributed to reduce the damage by cold. The SOD activity decreased later at −10°C, which may be due to the greater damage by temperature of −10°C affected the synthesis of SOD in plant.
POD activities were higher under low temperature than normal temperature as shown in [Figure 6(b)](#fig6){ref-type="fig"}. At 1°C, POD activity was slowly increased in the first three days, rapidly increased in the third to fifth days, and reached the max of 6.7 U*·*min^−1^ *·*mg^−1^ in the fifth day, more than 4 times of control. Then POD activity slightly decreased to 6.6 U*·*mg^−1^ *·*min^−1^ in the 7th day, which was almost nearly the max value. At −10°C, POD activity was rapidly increased in the first three days and reached the max of 5.2 U*·*mg^−1^ *·*min^−1^ in the third day, more than 3 times of control. Then POD activity decreased in the third to seventh days. The POD activity decreased to 4.16 U*·*mg^−1^ *·*min^−1^ in the 7th day, which was still 2.5 times more than that of control. PODs are a large family of enzymes that typically catalyze peroxide. The increased POD activities under low temperature had improved cold tolerance in some degree. POD activities decreased greatly in later days, indicating that low temperature had affected POD enzyme. It maybe due to low temperature affected RNA transcription and translation, reducing the synthesis of POD. At the same time proline content also increased under low temperature, which can degrade peroxidase. It can also decrease the POD activity.
Compared with normal, the CAT activity was higher under the cold treatment than normal temperature as shown in [Figure 6(c)](#fig6){ref-type="fig"}. At 1°C, the CAT activity was increased with the prolongation of low temperature stress. CAT activity was 0.66 U*·*g^−1^ *·*min^−1^ before of cold treatment and increased to 1.99 U*·*g^−1^ *·*min^−1^ after the seven days cold treatment. At −10°C, the CAT activity was increased rapidly in the first three days and increased slowly in later days. In the third day, the CAT activity was increased to the 1.74 U*·*g^−1^ *·*min^−1^, which was nearly 2.5 times of normal temperature. Then, the CAT activity was increased to 2.16 U*·*g^−1^ *·*min^−1^ in the seventh day. Compared with normal, the CAT activity was higher than normal under the cold treatment. Increased CAT activity contributed to cold tolerance of naked oats under cold stress.
In this experiment, the activities of these enzymes were increased in different degrees under cold stress. The results imply that higher SOD, POD, and CAT activities enhanced the capacity for scavenging ROS and contributed to enhanced tolerance of plant to cold stress. The result was similar to the SOD changes in other cold treatment plants.
4. Conclusions {#sec4}
==============
*Avena nuda* L. is an important crop and has tolerance to low temperature. In the long time of evolution, it should have some cold resistant mechanism. The cold-stress leads to complex cellular and biochemical changes such as altered membrane composition and accumulation of proline, as well as the activities of antioxidant enzymes. In the present study, we have, for the first time, investigated the physiological changes in naked oats under low temperature. As results showed, low temperature stress caused a certain degree of physiological impairment in naked oats leaves. Here, we showed that cold increased the content of proline, MDA, electrolyte leakage, SOD, CAT, and POD activities. The content of chlorophyll was decreased. The study provided theoretical basis for cultivation and antibiotic breeding in *Avena nuda* L.
This work was supported by the Shanxi Datong University Research Starting Fund for Doctor no. 2009-B-16. The seeds of naked oats were supplied by Assistant Researcher Li Gang in High Latitude Crops Institute of Shanxi Academy of Agriculture Sciences.
{#fig1}
{#fig2}
{#fig3}
{#fig4}
{#fig5}
{#fig6}
[^1]: Academic Editors: J. Huang and Z.Wang
| {
"pile_set_name": "PubMed Central"
} |
The authors confirm that, for approved reasons, some access restrictions apply to the data underlying the findings. Data contains identifying information and is available upon request from the authors.
Introduction {#s1}
============
In India, chronic conditions are a leading cause of death and disabilities and estimated to account for 67% of all the deaths in the year 2020 [@pone.0106522-Reddy1]. The national prevalence of diabetes among 20--79 years old is 8.56%. With over 65.1 million people suffering from diabetes in 2013, India has the second largest number of people living with diabetes in the world, after China. Diabetes accounted for over one million adult deaths in 2013 [@pone.0106522-InternationalDiabetes1]. Recent studies show that the major chronic conditions, including diabetes, are no longer the conditions affecting only the wealthy population but are increasingly affecting the urban poor and slum dwellers [@pone.0106522-Anand1]--[@pone.0106522-Bhojani1].
Weak health systems have been identified as major bottlenecks in effectively responding to the rising burden of chronic conditions in low- and middle-income countries (LMIC), including India [@pone.0106522-Reddy1], [@pone.0106522-Samb1]--[@pone.0106522-Dans1]. Despite recognition by global actors for the need for research on health systems [@pone.0106522-Task1], [@pone.0106522-Alliance1], little attention has been given to the role of local health systems in the delivery of care for chronic conditions. The local health system -- defined as all organizations, people and actions that primarily intend to promote, restore or maintain health at the level of cities or rural areas -- is key to health system performance. At this level, policies are adopted and implemented, responsive health services are provided and programs are applied. Recently, the integration of chronic disease prevention and management programs into district level health systems in India has been proposed [@pone.0106522-Kar2], [@pone.0106522-Directorate1].
The study, presented in this paper, is part of a larger research project to understand how local health systems can be strengthened in order to deliver better quality chronic condition care to the urban poor. A poor urban neighborhood in South India constituted the research site. The research involved the following: (i) a household survey that revealed a high prevalence of diabetes and high out-of-pocket healthcare expenditure [@pone.0106522-Bhojani1], [@pone.0106522-Bhojani2]; (ii) interviews with diabetes patients that revealed specific constraints faced in managing diabetes [@pone.0106522-Bhojani3]; and (iii) interviews with healthcare providers to better understand existing health system challenges in delivering diabetes care. This paper concerns the third aspect of this larger research framework.
Methods {#s2}
=======
We conducted a cross-sectional study with healthcare providers providing care to patients with diabetes mellitus type 2 (referred to as diabetes in the rest of the paper). We used semi-structured interviews to understand the organization of diabetes care in the local health system and the problems in diabetes management, as well as to identify feasible health service interventions from the viewpoint of the healthcare providers. The enquiry was shaped by the health system dynamics framework developed by Van Olmen et al [@pone.0106522-VanOlmen1], [@pone.0106522-VanOlmen2] to analyze (local) health systems. This analytical framework ([Figure 1](#pone-0106522-g001){ref-type="fig"}) that includes ten interactive elements establishes the building blocks of health systems as specified by the World Health Organization (WHO) [@pone.0106522-World1]. The framework also emphasizes that health systems should be geared towards outcomes and goals that are based on explicit choices of values and principles. The organization and delivery of healthcare services are considered the central processes and the immediate outputs of the health system. This framework has been helpful in analyzing local health systems in different contexts [@pone.0106522-VanOlmen2]. [File S1](#pone.0106522.s001){ref-type="supplementary-material"} provides the detailed interview guides for doctors, pharmacists, and laboratory technicians in English.
![Health systems dynamics framework.\
The health systems dynamics framework developed by Van Olmen et al [@pone.0106522-VanOlmen1] at the Department of Public Health, Institute of Tropical Medicine, Antwerp, Belgium.](pone.0106522.g001){#pone-0106522-g001}
The participants included specialist and non-specialist doctors, pharmacists and a laboratory technician working in and around Kadugondanahalli (KG Halli), a poor urban neighborhood in metropolitan city of Bangalore, the capital of the south Indian state of Karnataka. KG Halli is one of the 198 administrative units of Bangalore city with a slum area. KG Halli has a population of over 44,500 people within an area of less than a square kilometer.
KG Halli has a mixed health system with coexisting government and private healthcare sectors. Health facilities in KG Halli include two government health centers, 28 private clinics and four private hospitals. Clinics are primary care facilities managed by a single doctor who is occasionally assisted by support staff. Clinics operate on an outpatient basis. Hospitals, in addition to primary care, also provide specialist care. They provide facilities for surgery and inpatient care but greatly vary in size and services. In our study, we included all the health centers, clinics and hospitals in KG Halli that claimed to be offering care to diabetes patients and interviewed the doctors that used to treat diabetes patients at these facilities.
Our earlier study found that 85.2% of diabetes patients in KG Halli sought care from the private sector, often including health facilities located outside of the KG Halli area [@pone.0106522-Bhojani1]. We therefore decided to include health facilities that were located within a two-kilometer radius (easy-to-travel distance) from KG Halli and were used by more than 50 diabetes patients from KG Halli as per our earlier study. Additionally, there are many private laboratories and pharmacies in KG Halli. We purposely selected one of each type from the frequently used private pharmacies and private laboratories and interviewed a pharmacist and a laboratory technician. We also interviewed a pharmacist in the government health center. [Table 1](#pone-0106522-t001){ref-type="table"} provides details about the health facilities included in the study.
10.1371/journal.pone.0106522.t001
###### Profile of respondents.
{#pone-0106522-t001-1}
Respondentnumber Sex of therespondent Role of the respondentat his/her healthfacility Formal trainingof the respondent Type of therespondent's facility(based on ownership) Type of the respondent's facility(based on deliveryplatform)
------------------ ---------------------- ------------------------------------------------- ------------------------------------------------------- ------------------------------------------------------ ---------------------------------------------------------------------------------------------------
R1 Woman Non-specialist doctor Graduation in ayurveda & bridging course in allopathy Private Clinic providing primarycare on outpatient basis
R2 Man Non-specialist doctor Graduation in unani Private Clinic providing primarycare on outpatient basis
R3 Woman Non-specialist doctor Graduation in unani Private Clinic providing primarycare on outpatient basis
R4 Woman Non-specialist doctor Graduation in unani Private Clinic providing primarycare on outpatient basis
R5 Man Non-specialist doctor Graduation in ayurveda Private Clinic providing primarycare on outpatient basis
R6 Man Non-specialist doctor Graduation in ayurveda Private Clinic providing primarycare on outpatient basis
R7 Man Non-specialist doctor Graduation in allopathy Private Clinic providing primarycare on outpatient basis
R8 Man Non-specialist doctor Graduation in allopathy Private Clinic providing primarycare on outpatient basis
R9 Man Non-specialist doctor Graduation in unani Private Clinic providing primarycare on outpatient basis
R10 Man Non-specialist doctor Graduation in allopathy Government Clinic providing primarycare on outpatient basis
R11 Woman Specialist doctor Post-graduation in allopathy Government Same facility as that of R10
R12 Woman Specialist doctor Post-graduation in allopathy Private Clinic providing specialistcare on outpatient basis
R13 Man Non-specialist doctor Graduation in allopathy Private Hospital (six beds)providing primary care and limited referral care
R14 Man Non-specialist doctor Graduation in unani Private Hospital (15 beds) providingprimary care and limited referral care
R15 Man Non-specialist doctor Graduation in ayurveda Private Hospital (30 beds) providingprimary care and limited referral care
R16 Man Specialist doctor Post-graduation in allopathy Private Super-specialty hospital(600 beds) attached to amedical school providingprimary and referral care
R17 Woman Pharmacist Graduation in pharmacy Government Same facility as that of R10
R18 Woman Pharmacist Completed schooleducation till 12^th^class Private Pharmacy
R19 Woman Laboratory technician Post-graduation in laboratory technology Private Diagnostic center
After acquiring respondents' written consent, the first author, who has formal training and experience in qualitative research, conducted interviews in English or *Hindi* based on the respondents' language preference. The data privacy and anonymity of respondents was assured. Considering the limited number of diabetes care providers in KG Halli, we explained to the respondents that they may be possibly identified by local actors and about the risk associated with it as part of the consent process. For respondents' convenience, the interviews were conducted at their workplaces, usually in their consultation rooms. The interviews lasted approximately 40 to 60 minutes, occasionally interrupted by patient consultations. The interviews were tape-recorded and were transcribed verbatim by a professional transcriptionist. In addition to the interviews, the first author maintained a field diary recording the general observations made at the health facility while conducting the interviews. These observations included aspects such as physical environment of the health facility, writings and visuals displayed in the facility, and behavior of patients and staff at the facility.
We used thematic analysis. The first author coded transcripts in Nvivo software by creating respondents' profiles and using elements of the health systems dynamics framework as tree nodes. Free nodes were created to accommodate data that did not fit the tree nodes. Based on the initial coding, the research team discussed the resulting overarching themes. They discussed the relationships across and between the themes and respondents' attributes. The research team gathered expertise in relevant fields, including medicine, public health, health service research and medical anthropology. The major recurring themes were grouped into four categories representing the four out of the ten interactive elements of the health system dynamics framework (i.e., health service delivery, knowledge and information, leadership and governance, values and principles). We used these categories to present the study findings in the results section.
This study received approval from the Institutional Review Board at the Institute of Tropical Medicine, Antwerp (Belgium) and from the Technical Committee, as well as the Institutional Ethics Committee at the Institute of Public Health, Bangalore (India).
Results {#s3}
=======
In total, we conducted 19 interviews with three specialist doctors, 13 non-specialist doctors, two pharmacists and one laboratory technician. These respondents were attached to a government health center, 11 private clinics, four private hospitals, a private pharmacy and a private diagnostic center. [Table 1](#pone-0106522-t001){ref-type="table"} provides the profiles of the respondents and their respective health facilities. A non-specialist doctor working at a private clinic refused to be interviewed whereas one non-specialist doctors and one specialist doctor who had agreed to participate and worked at a private clinic and a private hospital, respectively, did not have time for interviews despite repeated attempts by the researchers. [Table 2](#pone-0106522-t002){ref-type="table"} provides summary of dominant themes defining local health system challenges in delivering quality diabetes care.
10.1371/journal.pone.0106522.t002
###### Dominant themes pertaining to local health system challenges in KG Halli.
{#pone-0106522-t002-2}
Health systemelements Dominant themes pertaining to challenges related to specific health system elements
----------------------------------- ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
**Health care delivery** **Plurality in healthcare providers and care delivery platforms:** *KG Halli, with an area* *of less than a square kilometer, had several and diverse healthcare delivery platforms* *that catered to diabetes patients, most of which belonged to the private sector. Whereas* *doctors in the government health center were formally trained in allopathy, the doctors* *in the private sector were formally trained in different systems of medicine.*
**Hospitals providing primary care:** *All the hospitals explicitly market for and provide* *basic primary care for diabetes in addition to providing the referral specialist care,* *creating a significant functional overlap with services provided by private clinics and* *the government health center.*
**Private clinics delaying referrals:** *"One thing is that no one \[doctor\] wants to leave their* *patients. If a patient goes \[referred to other facility\], he may not come back. They* *\[non-specialist doctors at clinics\] have this fear."* (R13, private hospital)
**Knowledge and** **information** **Inadequate use of the patient medical records:** *Only six of the 15 health facilities in* *this study had a system that tracked medical records of diabetes patients.*
**Periodically updating the knowledge of doctors & influence of pharmaceutical industry:** *Nearly half of all the doctors indicated that they periodically updated their clinical knowledge.* *The pharmaceutical companies had easy access to doctors for influencing their practice through* *personal periodic visits by company representatives, sponsoring of continuing medical education* *activities and provision of medical literature to doctors.*
**Lack of standard treatment protocols:** *"No, there is nothing like that \[standard treatment* *protocol\]. It depends on how we analyze it \[diabetes condition\] and accordingly treat it."*(R2, private clinic)
**Leadership and** **governance** **Poor regulation of the private sector:** *"Many doctors in this area are not qualified to* *practice \[allopathy\]. But they have been doing it. ... We have doctors who have a diploma* *in acupuncture and are practicing allopathy. Nothing is being done by the government."*(R7, private clinic)
**Poor systemic coordination:** *There was lack of coordination across different types of healthcare* *providers (government, private for-profit and not-for-profit) and across multiple health care delivery* *platforms (clinics, health centers, hospitals).*
**Widespread bribery:** *"... It \[kickbacks\] happens in 90% of cases. It's between pharmaceutical* *company and the doctor. This is rampant in this area."* (R18, private pharmacy)
**Lack of formal grievance redress platforms:** *Despite spending considerable amounts of money out of* *their pockets, the patients or community representatives had no formal functional platforms to engage* *with the formal healthcare services for expressing grievances, conveying opinions on issues or demanding accountability.*
**Values and** **principles** **Maximization of profit:** *"It \[healthcare\] has become a business nowadays."* (R6, private clinic)
**Healthcare for poor as a charity:** *"We conduct the camps to test blood sugar for free to provide some* *services for those who can't afford even sugar test."* (R16, super-specialty private hospital)
**Trust deficit among patients and providers:** *"Let the patient go to a physician. They will come back* *to you \[non-specialist doctor\] for small ailments. You should be happy because it is a circle. There* *should be no fear that if I send a patient to you, then tomorrow the patient will never come back to* *me. ... I don't think doctors have this kind of trust today"* (R1, private clinic)
1. Health Care Delivery {#s3a}
-----------------------
### Plurality in healthcare providers and care delivery platforms {#s3a1}
As enumerated in the methods section, KG Halli, with an area of less than a square kilometer, had several and diverse healthcare delivery platforms that catered to diabetes patients, most of which belonged to the private sector. All the clinics, some of the hospitals and a health center offered services for a specific duration in a day. Whereas doctors in the government health center were formally trained in allopathy, the doctors in the private sector were formally trained in different systems of medicine. In India, there are at least seven recognized systems of medicine apart from allopathy, grouped under the umbrella term AYUSH that refers to ayurveda, yoga, naturopathy, unani, siddha, homeopathy, and sowa-rigpa [@pone.0106522-Department1]. Of the 14 doctors interviewed in the private sector, four were trained in ayurveda, five in unani and six in allopathy (one non-specialist doctor had a dual training in ayurveda as well as allopathy). However, all of these doctors, irrespective of their training background, primarily practiced allopathy, which potentially compromised the competence of some of the doctors that were not formally trained in allopathy.
The doctors with training in ayurveda or unani learned to practice allopathy typically through reading literature and working early in their career in one or more of the private hospitals, where they observed practice by senior allopathic practitioners. In fact, all the allopathic hospitals but one had a duty doctor who graduated in ayurveda or unani.
> *"After my studies \[in ayurveda\], I worked in XXX \[a private allopathic hospital\] as a duty doctor. I then worked in XXX \[another private allopathic hospital\] for six months in night shifts. ... I watch the \[allopathic\] physicians and my senior doctors treat the patients. I will read the booklets. That is how I gained knowledge \[about allopathy\]"* (R15, private hospital).
When asked for their opinion, half of the doctors trained in allopathy believed that in the early and borderline cases of diabetes, ayurveda, yoga and naturopathy might play a supportive role, provided this care is done alongside provision of allopathic medicine and with strict blood glucose monitoring. Their support, at least in principle, for such 'mix' of medicines was due to the perceived harmful side effects of allopathic medication compared with the perceived safety of AYUSH practices and medications.
> *"There are some very good medications in ayurveda that can be used for diabetes treatment for long time without harm. ... However, one should control and monitor blood sugar well."* (R10, government clinic).
>
> *"For people with borderline diabetes, alternate medicines like, naturopathy, Ayurveda, yoga or homeopathy will do \[work\]. I welcome it. ... If by using it, these \[allopathic\] medications could be reduced, it is good because in allopathy, there is ill in every pill but there is no pill for every ill."* (R1, private clinic).
The doctors with training in ayurveda or unani also suggested that these systems have a very limited supportive role in early cases of diabetes and favored mixing this care with allopathy.
Despite the favorable attitude, none of the doctors exclusively trained in allopathy actually practiced mixed medicine. Two of the ayurveda and two of the unani-trained doctors occasionally, often because of patients' demands, used ayurveda or unani medications along with allopathic medications in the treatment of early diabetes. Another non-specialist doctor with dual training in ayurveda and allopathy, who treated diabetes patients using allopathic medications, occasionally referred patients to doctors practicing AYUSH systems.
### Hospitals providing primary care {#s3a2}
In the private sector, unlike the government sector, there is no policy or plan to rationalize organization of care across different levels of health services. All of the hospitals explicitly market for and provide basic primary care for diabetes in addition to providing the referral specialist care, creating a significant functional overlap with services provided by private clinics and the government health center. In fact, more than 90% of all the patients being treated at private hospitals were walk-in patients that were not referred from other health facilities. This situation also occurred at the private super-specialty hospital and led to crowding in the outpatient department of the hospital. Based on different reasoning, a specialist doctor working at this hospital justified provision of primary care; he felt that if poor patients were refused care from his hospital (that provides subsidized care) on the basis that they needed to consult primary care providers, these poor patients would altogether fall out of the healthcare net. Furthermore, this doctor suggested that his fixed service-timings and fixed salaried remuneration meant that unlike specialists at other private hospitals working on fee-for-service basis, he did not have to be selective in terms of number of patients or kind of patients he sees.
### Private clinics delaying the referrals {#s3a3}
Hospital doctors believed that non-specialist doctors at private clinics do not refer patients in time and hold onto their patients until they can't manage the patient anymore.
> *"One thing is that no one \[doctor\] wants to leave their patients. If a patient goes \[referred to other facility\], he may not come back. They \[non-specialist doctors at clinics\] have this fear."* (R13, private hospital).
Once patients were referred to hospitals, these patients were less likely to be referred back to clinics, as hospitals also provided primary care. This explains the apprehension of doctors at clinics about 'losing' patients by referring them to hospitals.
> *"No, we don't get \[patients referred back from specialists/hospitals\]. It \[referral\] is good but it depends on the specialist. ... Once they \[patients\] go there \[to specialists/hospitals\], they will call them there only."* (R3, private clinic).
2. Knowledge and Information {#s3b}
----------------------------
### Inadequate use of the patient medical records {#s3b1}
Only six of the 15 health facilities in this study had a system that tracked medical records of diabetes patients. Five of these six facilities (two private clinics, two private hospitals and a government health center) used patient-held, paper-based medical records for patients with chronic conditions such as diabetes or hypertension. This record was mainly in the form of a small booklet that could be conveniently carried by patients. In this booklet, doctors recorded information about investigations and medications during each encounter with patients. Patients were expected to bring the booklet to follow-up visits. Two of the facilities provided booklets to patients following the diagnosis, whereas patients were expected to purchase such booklets for themselves in the other three facilities. Only one of the five facilities had the corresponding facility-held, paper-based medical records for patients. Additionally, one specialist clinic used facility-held, electronic records with no corresponding patient-held records for patients with chronic conditions.
The patient-held medical records were advantageous, as they allowed for the continuity of information across health facilities/providers when patients sought care from other (than regular) facilities, including out of network or emergency facilities.
> *"If a patient is staying far from this hospital or if a patient develops acute myocardial infarction, I don't want him to waste his crucial time and come to me. He can go across to nearby health facility and show them all the treatment done till now \[through his medical records\]."* (R16, private super-specialty hospital).
The majority of the doctors, who did not use a medical record system, expressed its usefulness in improving clinical decision-making. They saw the record as useful because very few of their patients carried the loose medical prescription papers issued to them during earlier visits, making it difficult for doctors to make informed decisions. However, lack of time and the lack of human resources were reported as the common constraints for setting up and using a medical record system for patients.
> *"They \[patients\] might bring the last prescription but not all \[earlier prescriptions\] ... It \[medical records system\] will surely help but it is very difficult \[for me\] to get time to keep records."* (R8, private clinic).
All the five hospitals that were studied used facility-held, paper-based medical records for hospitalized patients. During discharge from hospitals, the patients were provided with a discharge summary. None of the facilities, including those using medical record systems, had an active follow-up or reminder system for patients. Patients were lost to follow-up. Furthermore, there was no population-level routine surveillance system for assessing prevalence of diabetes or its risk factors.
### Periodically updating the knowledge of doctors & influence of pharmaceutical industry {#s3b2}
Nearly half of all the doctors indicated that they periodically updated their clinical knowledge. The common educational tools included the continuing medical education activities (seminars, lectures) organized by professional associations and reading medical literature. Five of the doctors were members of professional associations. They considered continuing medical education as the major activity of these associations that, one or more times, included diabetes as a topic.
The pharmaceutical companies had easy access to doctors for influencing their practice through personal periodic visits by company representatives, sponsoring of continuing medical education activities and provision of medical literature to doctors. Interestingly, two of the doctors reportedly used the internet as a source of learning the latest knowledge on diabetes management.
### Lack of standard treatment protocols {#s3b3}
Despite moderate participation in continuing medical education activities, none of the doctors except one specialist were aware of any standard treatment protocol for diabetes management. The management practice for diabetes varied across the doctors, beyond the adjustments needed to accommodate for the individual needs of patients.
> *"No, there is nothing like that \[standard treatment protocol\]. It depends on how we analyze it \[diabetes condition\] and accordingly treat it."* (R2, private clinic).
A few of the doctors, especially those not trained in allopathy, used a 'trial and error' approach for deciding on the use of allopathic medications promoted by pharmaceutical companies.
> *"Once they \[pharmaceutical companies\] give \[medication\] samples, I try with patients. I will see the response, if it is good, okay, next time I will start with that. If patients don't respond to it, then I send them to other \[allopathic\] doctors."* (R4, private clinic).
Importantly, poverty in KG Halli has also shaped diabetes management practices of the doctors. Some of the doctors deviated from the knowledge-based clinical practices to adapt to the financial situation of patients even if the doctor knew the treatment would worsen the patient's health status.
> *"Most of the patients will come and ask for the \[oral\] tablets instead of insulin injections and we would give them tablets. ... Insulin is costly and they have to take all these medications. If we are not doing it, somebody else \[doctor\] will do it \[on patients' request\]."* (R14, private hospital).
3. Leadership and Governance {#s3c}
----------------------------
### Poor regulation of the private sector {#s3c1}
We used a limited interpretation of regulations by reducing them to the current toolbox of formal laws and policies. The laws formulated by governments to regulate healthcare, in the context of KG Halli, would require the following: (i) a healthcare provider to have a recognized qualification and a valid registration with the state council of her/his respective system of medicine; (ii) registration of private health facilities with the Karnataka Private Medical Establishment Act (2007) that prescribes the norms for healthcare infrastructure; (iii) a valid trade license for health facilities issued by the municipal government; and (iv) a No Objection Certificate from the Karnataka State Pollution Control Board that prescribes the norms for bio-waste management.
Of the 14 doctors working in the private sector, three knew about the regulation of bio-waste management, whereas six knew about the need for a trade license, as well as the registration of their facility, under some laws. The majority of them could not recall the name of the law or its major provisions. The doctors of only three of the private facilities (one clinic and two hospitals) were aware of all four regulations, and their facilities were in compliance with these regulations.
As mentioned earlier, the eight non-specialist private doctors, who had a degree in ayurveda or unani, primarily practiced allopathy without a degree or registration to do so. A pharmacist who ran a private pharmacy did not have the required degree. The majority of allopathy-trained doctors, who favored the mix of AYUSH with allopathy in treatment of early diabetes, were not supportive of doctors with AYUSH training that were primarily practicing allopathy.
> *"Many doctors in this area are not qualified to practice \[allopathy\]. But they have been doing it. ... We have doctors who have a diploma in acupuncture and are practicing allopathy. Nothing is being done by the government."* (R7, private clinic).
Interestingly, most of the doctors in the private sector, including those who were not complying with the prevailing regulations, found these regulations meaningful in improving healthcare services. These doctors mentioned that the poor dissemination and enforcement of these regulations by government authorities was the major reason for non-compliance with regulations by private facilities. Another concern was the delay by regulatory authorities in processing applications and granting registrations/licenses.
> *"They \[regulations\] are required and really good. But as far as the \[enforcement\] officers are concerned, I have not seen them coming down and checking it \[compliance\]. ... We applied for the registration around two years ago, still we haven't received any response."* (R2, private clinic).
### Poor systemic coordination {#s3c2}
Beyond formal regulations, coordination across different healthcare providers (government, private for-profit and not-for-profit) and across multiple healthcare delivery platforms (clinics, health centers, hospitals) is an important regulatory mechanism to steer care providers towards a coherent vision and goal in the local health system.
However, basic information, such as the number of health facilities/doctors in KG Halli and the range of services they provide, is not collected by the government or any private player. Although each private facility is expected to provide information in a prescribed format to the appropriate government health authority in their area on a monthly basis, only one of the doctors in the private sector was aware of the process and had started doing so a few months prior to this study.
Government health workers or officers had never visited most of the private health facilities in the area. Some of the private doctors were not even aware of the location of a government health facility in the area.
> *"Nobody \[from government\] comes here. Till now, in the last 20 years of my practice, I have not seen anybody from the government health service coming here."* (R8, private clinic).
Within the private sector, only a few doctors who had been practicing for many years in the area knew the other doctors in the area and had a professional interaction with them. There was no coordination between two of the government facilities, which were located close to each other in KG Halli but managed by different government authorities. All of the super-specialty government referral hospitals (outside KG Halli) were managed by the medical education department of the state government with little or no coordination with the municipal authority or the health department of the state government that are supposed to manage primary care facilities in the city. There was no coordination between the government facility in KG Halli and the private facilities for planning the organization of diabetes care.
### Widespread bribery {#s3c3}
Bribery was common at both the individual healthcare provider and organizational level. The kickbacks from the pharmaceutical companies to doctors for writing particular brands of medication were commonplace.
> *"... It \[kickbacks\] happens in 90% of cases. It's between pharmaceutical company and the doctor. This is rampant in this area."* (R18, private pharmacy).
In fact, one of the private health facilities in this study was owned by a pharmacist, who allowed three doctors to use that facility to practice without paying any rent or facility costs if these doctors directed their patients to his pharmacy housed in the same building. Interestingly, a private hospital had put a board in the patients' waiting area that stated doctors at that hospital do not insist patients buy medications from any specific pharmacy. This message was to reassure patients who knew that health facilities often associate with specific pharmacies for kickbacks. Similarly, kickbacks from private diagnostic centers to the doctors in the area were common. In fact, some of the doctors used prescription papers that had the details of a specific private pharmacy or laboratory printed at the bottom of the papers.
> *"We have to pay some 25 to 30 percent \[of cost of prescribed investigations as a kickback\] to doctors. We are giving percent to more than 20 to 25 doctors in this area."* (R19, private laboratory).
Three of the private doctors, who were approached by government regulatory authorities, reported that it was common for doctors to bribe the lower-level government officers to get the necessary license/registration for their facilities or to avoid punitive actions.
> *"The inspector had come to me. Mine is an eight feet by eight feet clinic \[smaller than the minimum space needed to run a clinic by law\]. I said, what to do sir? I am practicing here for the past 20 years. Where will I go now? Take 5000 rupees \[∼ USD 83\] sir. That drug inspector will not come for another year. Every year go on bribing them, go on practicing."* (R1, private clinic).
### Lack of formal grievance redress platforms {#s3c4}
Because the highly utilized private health sector works on a fee-for-service basis, patients are the major funders of the health system. Despite spending considerable amounts of money out of their pockets, the patients or community representatives had no functional platforms to engage with the formal healthcare services for expressing grievances, conveying opinions on issues or demanding accountability. Such engagement happened rather informally to a very limited extent as part of the doctor-patient interaction during the medical consultation. In fact, one of the doctors in the private sector felt that patients prefer to use his hospital because they could personally hold the doctors or other staff accountable because they had paid for services.
4. Values and Principles {#s3d}
------------------------
In mixed health systems, often characterized by a relative lack of stewardship, identifying common values that generally guide the health system is often difficult and not anticipated. We attempted to highlight the values and principles that were often mentioned by the healthcare providers during interviews that they believed to be important in shaping current medical practice, especially in the private sector.
### Maximization of profit {#s3d1}
Of the seven respondents who were willing to talk about the guiding factors of current medical practice in the private sector, all but one mentioned that healthcare has become a business in which the medical practice aims to maximize profits. Money, and not the patient, is at the center and guides the practice. Other respondents either refused to express their opinion or had no specific comment on this aspect. The respondents believed that in the past, doctors saw healthcare delivery more as a service to mankind that should yield a decent income for doctors. However, healthcare is increasingly becoming like any other business in which profit drives practice. This transition was seen as part of the larger societal transition in which money is becoming an important preoccupation in the lives of people in all sectors and not limited to healthcare.
> *"It \[healthcare\] has become a business nowadays."* (R6, private clinic).
An allopathic doctor referred to the very expensive medical (allopathic) education system, especially in the private sector, in which admissions to the medical schools are literally being purchased with huge sums of money, which then forces these medical graduates to 'recover' finances by charging more to the patients.
### Healthcare for poor as a charity {#s3d2}
When patients had difficulty affording healthcare, the common response from private providers was to refer them to a government facility. However, when the required treatment was available within their facilities, three of the private providers waived some of the treatment cost or gave free medications. Five of the doctors in the private sector, who expressed concern towards the poor economic conditions of patients, suggested the need to consider paying capacity of patients as a guiding factor in deciding the fees. The doctors believed in charging more money to patients who could pay and help those patients who can't afford care by charging them less. Apart from waivers in treatment cost, some doctors organized free diagnostic health camps, occasionally with limited supplies of medication, as their way 'to help' poor patients.
> *"We conduct the camps to test blood sugar for free to provide some services for those who can't afford even sugar tests."* (R16, super-specialty private hospital).
### Trust deficit among patients and providers {#s3d3}
Private doctors had strong negative opinions about government health services. Coupled with poor coordination between these sectors, this division led to low levels of trust. Implicit with the notion of referring poor patients to government hospitals and wealthier patients to private hospitals, the doctors believed strongly that government hospitals are poor facilities meant for poor patients. The overcrowding, long waiting times, scarcity of doctors, inadequate time and explanations provided to the patients, negative and even abusive attitudes of health workers, and lack of guidance to navigate chaotic set-ups in large hospitals shape the perceptions about government health facilities. Doctors in the private sector were referring poor patients to government hospitals, but they were not sure whether these patients would receive the needed treatment in the government hospitals.
> *"If the patient is very poor, we refer them to XXX \[government hospital\] or some other government hospital. Otherwise, if the patient can afford, we will send them to a private hospital. ... \\A private hospital will give good services, they will not shout at any patients."* (R14, private hospital).
However, a few autonomous super-specialty hospitals (for heart conditions, cancers) were perceived to provide similar care as the super-specialty private hospitals.
There was a lack of trust between the non-specialist and specialist doctors, which contributed to the poorly functioning referral system. As one of the non-specialist doctors working at a private clinic who struggled to make referral links work stated:
> *"Let the patient go to a physician. They will come back to you \[non-specialist doctor\] for small ailments. You should be happy because it is a circle. There should be no fear that if I send a patient to you, then tomorrow the patient will never come back to me. ... I don't think doctors have this kind of trust today."* (R1, private clinic).
The majority of doctors doubted and often blamed patients for failing to follow the prescribed treatment and lifestyle changes, and thought the patients were ignorant, unconscious, or illiterate.
> *"There are many illiterate people in this area. They are not aware of things. I also see many educated people also who don't follow (behavior change, medications). They are busy with other things and they are not conscious about it."* (R9, private clinic).
Discussion {#s4}
==========
The analytical framework {#s4a}
------------------------
Our study adds to the early experiences [@pone.0106522-Chenge1], [@pone.0106522-Pongsupap1] of applying the health system dynamics framework in LMIC. The framework was useful in shaping the research enquiry and the data analysis from a health system perspective. It helped to investigate the systemic impediments that affected the effective delivery of quality diabetes care, as well as the interconnectedness of various elements of the local health system, e.g. a specific financing strategy (fee-for-service) affecting doctors' behaviors (more patients with short consultation time, no time for record keeping) that therefore affected healthcare (less attention to prevention and patient-centeredness in care, lack of medical history affecting clinical decisions) in a context guided by changing deontological and professional values (maximization of profit from medical practice) in many of the private health facilities.
However, designing research that enables use of the full scope of this framework is difficult. The broader scope of the health system (and the framework) that involves many actors/elements and their interrelationships poses challenges in sampling the respondents and designing the tools that help capture all the relevant information. In our study, respondents were limited to diabetes care providers, who mostly owned and managed their own health facilities. This reduced somewhat the complexity of our research, but was at the same time a limitation of our study, as it provides an analysis of the local health system from the viewpoint of only one set of actors. As explained in the introduction section, our earlier work investigated the perspectives of diabetes patients [@pone.0106522-Bhojani3]. The use of indirect questioning in the interviews and recording of the observations at the health facilities helped us to better understand the issues that respondents would either hesitate to discuss or provide short answers that were difficult to be taken face value. This method particularly helped to investigate values and principles, kickbacks and patients' participation in health services.
Systemic impediments in diabetes care delivery {#s4b}
----------------------------------------------
The major gaps in organizing diabetes care identified by our study were related to the four elements of the local health system: health service delivery, information and knowledge, leadership and governance, and values and principles guiding the system.
Some of the problems identified in our study have been ailing the urban health systems in India for many years. The government of India, in one of its national five year plans developed nearly three decades ago that included a discussion of non-communicable diseases for the first time, mentioned that "the organized referral services are almost non-existent" in urban areas [@pone.0106522-Planning1]. The government health centers in urban India that were largely established to deliver family planning services in the context of the population control initiatives [@pone.0106522-TheWorld1]--[@pone.0106522-Planning6] lack provision of comprehensive primary care, including care for chronic conditions. This factor leads to (highly inefficient) overcrowding of tertiary hospitals. The multitude of government agencies providing healthcare at different levels of health services poses considerable coordination challenges. There are at least seven government agencies providing healthcare in Bangalore city, and there is very limited functional integration and no administrative integration across the agencies. The coordination between the government and the private sector is currently still a largely utopian task.
The regulation of the private health sector, which forms the dominant part of the healthcare delivery system in urban areas, is yet another challenge. Despite the acknowledgement of the need to regulate the private sector since the early independence period [@pone.0106522-Planning5], except for a few regulatory initiatives, the federal government enacted legislation to regulate the private medical establishments in 2010 [@pone.0106522-Ministryof1]. However, like our findings from the KG Halli demonstrate, the enforcement of and the compliance with the various private sector regulations remain dismal [@pone.0106522-Nandraj1]. Although the organized bodies representing the private healthcare providers have often resisted formal regulation of the private health sector [@pone.0106522-Srinivasan1], [@pone.0106522-Sheikh1], our study revealed that individual private healthcare providers (most did not comply with current regulations) perceived the existing regulations as meaningful for ensuring better healthcare services. Instead, they blamed the apathy of the regulatory agency for the poor compliance by the private sector.
However, there is more to the (local) health systems than formal regulations and health programs, many of which are formulated at higher (state or national) levels in India. As Gilson [@pone.0106522-Gilson1] states, "health systems are inherently relational and so many of the most critical challenges for health systems are relationship and behavior problems". The actors within the local health system possess discretionary powers that, through their daily practice and action, shape healthcare reforms, including leadership and healthcare delivery in local health systems. Our study revealed that the limited coordination between the government and the private healthcare sector in KG Halli happened in a context in which government reached out to private doctors (e.g., trying to convince them to adhere to the tuberculosis treatment guidelines as specified by the national tuberculosis control program or providing interested private providers with vaccines and contraceptive devices to enhance delivery of preventive and family planning services). In addition to the legislative measures to regulate the private sector, which seem to be poorly enforced and complied with, proactive interactions between government and private healthcare providers could possibly enhance coordination and rationalization of care within local health system. The presence of government-initiated disease/condition programs and insurance schemes provide 'entry points' for engagements that could, with improving relationships over time, broaden the scope beyond the programs/schemes.
The present study, as well as our earlier work [@pone.0106522-Bhojani3], revealed the poor relationships across healthcare providers, as well as between healthcare providers and patients in KG Halli, which contributed to a general lack of trust. We join Gilson [@pone.0106522-Gilson1] in arguing that the government needs to play a role beyond being the provider, funder or regulator of health services to manage the relationships and processes that influence the building of trust within local health systems. This goal could involve fostering interactions among and across government and private healthcare providers working in the area; developing a collective (health) vision for the community; and sharing of information and plans by health facilities that encourages complementary, if not joint, planning. Studies exploring the enhancement of leadership and management using similar modalities in a mixed urban local system in South Africa have shown encouraging results [@pone.0106522-Elloker1].
An important limitation of our study is that we analyzed the local health system of a relatively small poor urban neighborhood. The findings related to the health system challenges can therefore not be generalized to Bangalore city or to other areas in India. However, our findings imply the need for systems thinking in the planning of health programs for diabetes or other chronic conditions. Analytically, our study findings would help designing enquiries to understand systemic challenges in delivering care for chronic conditions in LMIC that are facing rising burden of chronic conditions and share some common challenges in their mixed health systems [@pone.0106522-Samb1], [@pone.0106522-Nishtar1]. The national program for diabetes, which is being piloted in selected districts across India, introduces some 'new' (preventive and curative) services within district health systems for diabetes patients, but there is still an overlap in care across different levels of health services [@pone.0106522-Directorate1]. The program aims to integrate diabetes care delivery into routine government health services at various levels, but it does not attempt to address many of the known systemic weaknesses in the existing government health system that are so critical to chronic condition care (e.g., poor information systems, lack of medications in government facilities or non-integration among and across government and private healthcare providers).
The government of India has recently launched the National Urban Health Mission, which is a population-based program to improve the health status of the urban population in general [@pone.0106522-Ministry1]. The programs focuses on the urban poor and disadvantaged and proposes a series of health system reforms addressing many of the gaps highlighted in our study, such as strengthening urban primary healthcare services including chronic condition care; enhancing referral links across different levels of healthcare services; and creating institutional platforms for community participation in health services.
Despite acknowledging the high utilization of large private health services by the urban poor, the mission largely focuses on the government sector. It does not include the private sector (the so-called 'elephant in the room') in discussing reforms in the government sector. The 89-page implementation framework dedicates a little over one page to mention the regulation of the health system [@pone.0106522-Ministry1]. The proposed regulations, including the development of quality standards for health services, the accreditation of health facilities and setting up of mechanisms for addressing user grievances, are all meant for the government facilities and for the limited number of private facilities that the mission might purchase. This indicates that the government does not have a strong will to regulate the private health sector. We strongly advocate that the government should take a broader and more inclusive view of the health system and augment the stewardship of the entire health system. We hope that the autonomy accorded to state and municipal governments in contextualizing the mission plan will consider the challenges as well as the potential of local health systems to enhance healthcare delivery, including diabetes care.
Conclusions {#s5}
===========
There is a lack of functional referral systems and a considerable overlap in provision of primary diabetes care across the different levels of healthcare services. Inadequate use of patient medical records and lack of standard treatment protocols affect clinical decision-making. The poor regulation of the private sector, the lack of coordination among and across the government and private healthcare providers, the widespread bribery practices and the absence of any formal grievance redress platforms, reflect weak leadership and governance. There is a huge trust deficit between patients and healthcare providers. The private sector, in which the majority of healthcare providers lack the required training, is guided by profit maximization in which healthcare for poor people is, at best, seen as charity. These systemic impediments in local health systems hinder the delivery of quality diabetes care to the urban poor. Our findings indicate the urgent need to address these systemic weaknesses in local health systems in order to integrate and improve the care for diabetes and other chronic conditions.
Supporting Information {#s6}
======================
######
**Interview guide for semi-structured interviews with healthcare providers.**
(PDF)
######
Click here for additional data file.
[^1]: **Competing Interests:**The authors have declared that no competing interests exist.
[^2]: Conceived and designed the experiments: UB ND AM SDH PK BC. Performed the experiments: UB. Analyzed the data: UB. Contributed reagents/materials/analysis tools: UB. Contributed to the writing of the manuscript: UB. Reviewed and made comments for important revisions in the draft manuscript: ND AM SDH PK BC.
| {
"pile_set_name": "PubMed Central"
} |
**ATTENTION**
The spread of Covid-19 has had a major effect on the 2020 event schedule. Many events have been postponed, rescheduled or cancelled. Readers are advised to check the websites of individual events for the latest information.
31 August--2 September 2020
Shanghai, China
**Aquatech China 2020 -- NEW DATE**
Web: [www.aquatechtrade.com/china/](http://www.aquatechtrade.com/china/){#interrefs10}
29 September--1 October 2020
Virtual
**RISE 2020 Conference -- Research, Innovation & Science for Engineered Fabrics Conference**
Web: [www.riseconf.net](http://www.riseconf.net){#interrefs20}
5--9 October 2020
Virtual
**WEFTEC 2020 -- 93rd Annual Technical Exhibition & Conference**
Web: [www.weftec.org](http://www.weftec.org){#interrefs30}
3--4 November 2020
Birmingham, UK
**Flood Expo 2020 -- NEW DATE**
Web: [www.thefloodexpo.co.uk](http://www.thefloodexpo.co.uk){#interrefs40}
30 November--2 December 2020
Kuala Lumpur, Malaysia
**ASIAWATER 2020 -- NEW DATE**
Web: [www.asiawater.org](http://www.asiawater.org){#interrefs50}
6--11 December 2020
London, UK
**ICOM 2020 -- International Congress on Membranes & Membrane Processes 2020 -- NEW DATE**
Web: [www.icom2020.co.uk](http://www.icom2020.co.uk){#interrefs60}
15--17 March 2021
Rome, Italy
**IDA 2021 International Water Reuse and Recycling Conference**
Web: <https://wrr.idadesal.org/>
20--22 April 2021
New York City, USA
**INTERPHEX 2021**
Web: [www.interphex.com](http://www.interphex.com){#interrefs80}
9--14 May 2021
Copenhagen, Denmark
**IWA World Water Congress & Exhibition -- NEW DATE**
Web: <https://worldwatercongress.org/>
13--16 June 2021
San Diego, California, USA
**AWWA ACE21 -- American Water Works Association\'s Annual Conference & Exposition**
Web: [www.awwa.org/ace/](http://www.awwa.org/ace/){#interrefs100}
14--18 June 2021
Frankfurt, Germany
**ACHEMA 2021**
Web: [www.achema.de/en/](http://www.achema.de/en/){#interrefs110}
20--24 June 2021
Singapore
**Singapore International Water Week (SIWW) -- NEW DATE**
Web: [www.siww.com.sg](http://www.siww.com.sg){#interrefs120}
| {
"pile_set_name": "PubMed Central"
} |
1. Introduction {#sec1-molecules-20-19798}
===============
The development of renewable energy technologies is pivotal for accommodating the ever increasing energy demands of the modern society. Such technologies are also important for lowering environmental pollution and greenhouse gas emissions. Towards this objective, many approaches to harvest solar energy have been investigated. The fabrication of bulk-heterojunction (BHJ) devices is one such promising strategy that has attracted considerable attention over the past two decades due to their advantages of being lightweight, low cost and their flexibility in large-area applications \[[@B1-molecules-20-19798],[@B2-molecules-20-19798],[@B3-molecules-20-19798],[@B4-molecules-20-19798],[@B5-molecules-20-19798],[@B6-molecules-20-19798],[@B7-molecules-20-19798]\]. Such devices are comprised of an interpenetrating network of organic donor and acceptor domains that is formed during their fabrication via solution processing. Conventionally, semiconducting donor polymers such as poly(3-hexylthiophene) (P3HT) and acceptors such as soluble fullerene derivatives, PC~61~BM and its C~71~ analogue (PC~71~BM), have been used to obtain a deeper understanding of device design and morphology \[[@B8-molecules-20-19798],[@B9-molecules-20-19798],[@B10-molecules-20-19798],[@B11-molecules-20-19798],[@B12-molecules-20-19798],[@B13-molecules-20-19798]\]. Apart from archetypal P3HT, conjugated polymers have also been developed and significant progress has been attained with promising BHJ architecture. Power conversion efficiency (PCE) values above 10% has been reported with such polymeric donors \[[@B14-molecules-20-19798],[@B15-molecules-20-19798],[@B16-molecules-20-19798],[@B17-molecules-20-19798]\]. In the interim, solution-processed small molecular donor-based BHJ devices have also aroused interest, mainly due to their advantages of well-defined chemical structure, convenient purification methods, such as simple column chromatography, and monodisperse molecular weight \[[@B18-molecules-20-19798],[@B19-molecules-20-19798],[@B20-molecules-20-19798],[@B21-molecules-20-19798],[@B22-molecules-20-19798],[@B23-molecules-20-19798]\]. These advantages allow and encourage researchers to exert efforts for the design and development of small molecular donors. Recently, immense efforts have been dedicated to developing small molecular-based solution-processable organic solar cells \[[@B1-molecules-20-19798],[@B21-molecules-20-19798]\] and so far, the highest PCE of 9.95% was achieved by Kan *et al.* \[[@B24-molecules-20-19798]\] which is analogous to those of the polymer-based solution-processable devices. Thus, in view of such reports and the fact that BHJ devices incorporating small molecular donors can compete with polymer-based devices, there is an overwhelming interest in developing small molecular donors.
Recent years have seen a dramatic surge not only in terms of device efficiency using small molecular donors but also in their design and efficient synthetic development. A variety of small molecule donor materials based on donor--acceptor (D--A) combinations such as D--A--D \[[@B25-molecules-20-19798],[@B26-molecules-20-19798],[@B27-molecules-20-19798]\], A--D--A \[[@B28-molecules-20-19798],[@B29-molecules-20-19798]\], D--π--A \[[@B18-molecules-20-19798],[@B19-molecules-20-19798],[@B30-molecules-20-19798]\] and star-shaped architectures \[[@B31-molecules-20-19798]\] have been reported. The D--A--D design in particular is one of the most promising and successful modules based on which various donor and acceptor units have been explored for high-performance solution-processable photovoltaic devices. A finite number of central accepting units, such as naphthalene diimide \[[@B25-molecules-20-19798]\], diketopyrrolopyrrole \[[@B27-molecules-20-19798]\], 2-pyran-4-ylidenemalononitrile \[[@B32-molecules-20-19798]\] and thiazolothiazole \[[@B33-molecules-20-19798]\] have been reported to suit the D--A--D module. Not only that the availability and selection of such accepting blocks is limited, it is furthermore imperative that the target chromophore must possess a low optical band gap, broad absorption profile, high mobility and appropriately tuned highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) energy levels using such blocks. Such requirements do possess a challenge for an organic chemist who must consider such factors while designing a new chromophore based on the D--A--D module. Therefore, it is not surprising that there exists an enormous scope for the design and development of new light-harvesting materials based on the challenging D--A--D module and is an aspiration for most of the researchers.
In our own studies of small molecule chromophores and charge transport materials based on a variety of D--A combinations, we have reported examples of successful solution-processable BHJ photovoltaic devices \[[@B18-molecules-20-19798],[@B19-molecules-20-19798],[@B25-molecules-20-19798],[@B34-molecules-20-19798],[@B35-molecules-20-19798],[@B36-molecules-20-19798],[@B37-molecules-20-19798]\]. Furthermore, we are highly interested to extend our efforts on the design and development of new chromophores that are inspired by the D--A--D module. In this study, we report the design, facile synthesis and characterization of the optoelectronic and photovoltaic properties of two small organic chromophores, **AS2** and **AS1**, (shown in [Figure 1](#molecules-20-19798-f001){ref-type="fig"}), and their direct comparison. Both materials are based on a D--A--D structural motif where a triphenyl-amine functionality has been chosen as a common donor at both ends of the central acceptor unit, 1,4-phenylenediacetonitrile (PDA), so as to get symmetrical **AS2** and **AS1**. **AS2** is a structural analogue of **AS1** where a thiophene ring has been introduced between the donor and acceptor functionalities in order to vary the optoelectronic and photovoltaic properties ([Figure 1](#molecules-20-19798-f001){ref-type="fig"}). When compared with the commonly used accepting groups, such as dicyanovinylidene, aromatizable cyanopyridone, indenedione or oxoindenemalononitrile \[[@B38-molecules-20-19798],[@B39-molecules-20-19798]\], PDA can be an acceptor of choice for extended π-conjugation over the whole molecular backbone, mainly due to its bidentate nature. It is notable to mention that the use of the PDA unit for electroluminescent conjugated polymers has been reported in the literature \[[@B40-molecules-20-19798],[@B41-molecules-20-19798]\] as well as one- and two-photon spectroscopy studies \[[@B42-molecules-20-19798],[@B43-molecules-20-19798],[@B44-molecules-20-19798],[@B45-molecules-20-19798],[@B46-molecules-20-19798]\]. However, its suitability as an acceptor in small molecular chromophores, particularly D--A--D modular, for BHJ applications is still unknown \[[@B4-molecules-20-19798],[@B21-molecules-20-19798]\]. This provides an encouragement and some strong incentive for its investigation. Owing to the exploration of PDA acceptor unit in the D--A--D module, this study continues our search for the generation of new organic chromophores for BHJ applications.
{#molecules-20-19798-f001}
2. Results and Discussion {#sec2-molecules-20-19798}
=========================
2.1. Design Strategy, Synthesis {#sec2dot1-molecules-20-19798}
-------------------------------
The materials **AS2** and **AS1** were synthesized via the Knoevenagel condensation of appropriate aldehydes with active methylene groups of the PDA acceptor unit and their chemical structures were confirmed by ^1^H- and ^13^C-NMR, mass spectrometry, and, where possible (**AS2** only), by single crystal X-ray diffraction (XRD). The synthetic methodology for synthesizing **AS1** was similar to an old literature report \[[@B42-molecules-20-19798]\], albeit dissimilar base and solvent were used to deal with a homogeneous reaction solution. Knoevenagel condensation reaction of an aldehyde with active methylene group is an efficient way of generating a double bond between a π-bridge and acceptor functionality. The use of such chemistry is a common strategy to generate organic sensitizers for dye-sensitized solar cells \[[@B47-molecules-20-19798]\]. However, the use of same strategy to develop small molecular chromophores for BHJ applications is still limited \[[@B4-molecules-20-19798],[@B21-molecules-20-19798]\]. Herein, not only we are demonstrating the Knoevenagel condensation reaction of PDA acceptor unit but also the fabrication of solution-processable BHJ devices incorporating a fullerene acceptor (PC~61~BM) and either **AS2** or **AS1** as a donor component ([Figure 1](#molecules-20-19798-f001){ref-type="fig"}). To the best of our knowledge, this is the first time PDA has been used to generate D--A--D modular small molecular chromophores for BHJ applications. Initial screening of the BHJ devices revealed that greater PCE was achieved for **AS2** (4.10% for **AS2** compared with 1.23% for **AS1**), as confirmed by the increased short-circuit current density (8.01 mA·cm^−2^ for **AS2** and 3.15 mA·cm^−2^ for **AS1**), under simulated AM 1.5 illumination (100 mW·cm^−2^).
Both materials were based on the D--A--D module and the central acceptor moiety was directly linked to the terminal donor functionalities in order to create a conjugated structure. The development of these target materials incorporates the use of two identical donor units (triphenylamine) on each side of the central core, resulting in symmetrical chromophores. Insertion of a planar, conjugated functionality, such as thiophene in **AS2**, between the donor and acceptor components of a target material can provide greater absorption over visible light spectrum when compared with otherwise structurally similar compounds \[[@B38-molecules-20-19798],[@B48-molecules-20-19798]\]. Moreover, the selection of thiophene over highly aromatic, conjugating functionalities, such as phenyl, was based on the earlier work reported by Würthner *et al.* \[[@B49-molecules-20-19798]\] and Gupta *et al.* \[[@B50-molecules-20-19798]\] where it has been demonstrated that replacement of a phenyl group with thiophene can provide significant spectral red-shifts and is advantageous for superior charge delocalization. As a result, **AS2** is deemed to exhibit a large red shift of lambda maximum when compared with the reference compound **AS1**. Both of the materials were synthesized per the reaction shown in [Scheme 1](#molecules-20-19798-f010){ref-type="scheme"} and were purified by simple column chromatography. Brick-red, needle-shaped crystals, suitable for single crystal XRD, were prepared by diffusing methanol into a dichloromethane solution of **AS2**, over approximately three days. However, no crystal growth was observed for **AS1**. Both materials were synthesized in moderate to high yields (**AS2** = 63% and **AS1** = 86%) and were highly soluble in a variety of common organic solvents, for example chlorobenzene, chloroform, and toluene. The solubility of organic p-type materials is paramount for fabricating solution-processable BHJ devices and both the materials fulfill this criterion. In fact, the solubility of **AS2** was found to be higher by 50% *w*/*v* when compared with **AS1**.
{#molecules-20-19798-f010}
The precursor aldehyde **1** of **AS2** was also crystallized by diffusion of methanol into chloroform to obtain yellow, needle-shaped crystals. Diffraction measurements were performed at 200 K on a Bruker Apex II CCD diffractometer using Mo Kα radiation(λ = 0.71073 Å) The structure were solved using dual space methods using the program SHELX-2014/7 \[[@B51-molecules-20-19798]\] using the Olex2 1.2 GUI \[[@B52-molecules-20-19798]\], with anisotropic thermal parameters for all non-hydrogen atoms. All non-hydrogen atoms were refined anisotropically by full-matric least-squares methods SHELX-2014/7. Molecular drawings were obtained using Mercury \[[@B53-molecules-20-19798]\]. The utility of **1** for dye-sensitized solar cells has been reported \[[@B54-molecules-20-19798]\], however, its use to generate small molecular chromophores for BHJ applications is seldom reported \[[@B4-molecules-20-19798],[@B21-molecules-20-19798]\]. This state of affairs encourages us and provides a strong incentive to report its synthesis and crystal growth strategy.
The compound **1** was crystallized in the monoclinic space group (P 2~1~/c) with four asymmetric units in one cell. The thiophene group and the adjacent phenyl groups are planar with pendant phenyl groups displaced around nitrogen. The packing consist of a two-fold screw axis with centers of inversion between sulfur molecules as well as a glide plane perpendicular to the thiophene plane. The packing is dominated by π-π face-to-face stacking between the thiophene and phenyl groups. The pendant phenyl groups are stabilized by π-π edge-to-face stacking with distances in the range 2.771--3.283 Å as shown in [Figure 2](#molecules-20-19798-f002){ref-type="fig"}.
{#molecules-20-19798-f002}
**AS2** co-crystallizes with chloroform in the triclinic space group (P-1) with a center of symmetry around the central phenyl group. The packing is dominated by face-to-face π-π stacking between central phenyls and thiophenes, and by edge-to-face π‑π‑stacking between pendant phenyl groups with the distance 2.772 Å as shown in [Figure 3](#molecules-20-19798-f003){ref-type="fig"}. The details of packing structure, formula, crystal size of **1** and **AS2** are described in [Table 1](#molecules-20-19798-t001){ref-type="table"}.
molecules-20-19798-t001_Table 1
######
Details of crystal data and structure refinement parameters of **1** and **AS2**.
Identification Code CCDC: 1420377 (1) CCDC: 1420378 (AS2)
--------------------------------- --------------------------------------------- -----------------------------------------------
Empirical formula C~23~H~17~NOS C~31~H~22~OS
Formula weight 355.43 442.55
Temperature/K 200(2) 200(2)
Crystal system monoclinic monoclinic
Space group P2~1~/c P2~1~
a/Å 19.916(3) 5.5812(7)
b/Å 6.6680(9) 47.120(6)
c/Å 13.4336(16) 9.2435(11)
α/° 90 90
β/° 96.613(3) 103.343(3)
γ/° 90 90
Volume/Å^3^ 1772.1(4) 2365.3(5)
Z 4 4
ρ~calc~g/cm^3^ 1.332 1.243
μ/mm^−1^ 0.194 0.158
F(000) 744.0 928.0
Crystal size/mm^3^ 0.292 × 0.076 × 0.063 0.667 × 0.137 × 0.087
Radiation MoKα (λ = 0.71073) MoKα (λ = 0.71073)
2Θ range for data collection/ 4.118 to 48.588 3.458 to 67.842
Index ranges −22 ≤ h ≤ 23, −7 ≤ k ≤ 7, −15 ≤ l ≤ 15 −8 ≤ h ≤ 8, −73 ≤ k ≤ 73, −14 ≤ l ≤ 12
Reflections collected 15255 83,300
Independent reflections 2883 \[R~int~ = 0.0707, R~sigma~ = 0.0447\] 19,180 \[R~int~ = 0.0616, R~sigma~ = 0.0596\]
Data/restraints/parameters 2883/0/235 19180/1/595
Goodness-of-fit on F^2^ 0.949 1.028
Final R indexes \[I ≥ 2σ (I)\] R~1~ = 0.0412, wR~2~ = 0.1095 R~1~ = 0.0659, wR~2~ = 0.1531
Final R indexes \[all data\] R~1~ = 0.0720, wR~2~ = 0.1274 R~1~ = 0.0972, wR~2~ = 0.1684
Largest diff. peak/hole/e Å^−3^ 0.17/−0.26 0.29/−0.39
{#molecules-20-19798-f003}
2.2. Optoelectronic Properties {#sec2dot2-molecules-20-19798}
------------------------------
The optical properties of **AS2** and **AS1** were investigated by measuring their ultraviolet--visible (UV--Vis) absorption spectra in chloroform solution and in pristine spin-cast films ([Figure 4](#molecules-20-19798-f004){ref-type="fig"}). The longest wavelength absorption maximum (λ~max~) exhibited by **AS2** in solution form was at 459 nm which was red-shifted by 23 nm when compared with the solution λ~max~ of **AS1**. Both the absorption maximum and extinction coefficient (**AS2** = 59,000 M^−1^·cm^−1^; **AS1** = 49,000 M^−1^·cm^−1^) increased with the insertion of thiophene functionality. With the insertion of thiophene ring we found enhancement to the peak molar absorptivity of \>20% of **AS2** compared with **AS1**. This enhanced profile allows a larger amount of the solar spectrum to be absorbed, thus exhibiting greater intramolecular charge transfer (ICT) transition. We observed a similar bathochromic absorption shift in the thin film spectrum of **AS2** compared with that of **AS1** ([Figure 4](#molecules-20-19798-f004){ref-type="fig"}). The strong red-shift is attributed to the extended π-conjugation within the molecular backbone of **AS2** that became possible with the insertion of thiophene functionality. This type of control over the absorption profile through the insertion of a strongly conjugated unit can help to fine tune optical energy levels, to enhance light harvesting and BHJ device performance.
Density functional theory (DFT) calculations using the Gaussian 09 suite of programs \[[@B55-molecules-20-19798]\] and the B3LYP/6-311 + G(d,p)//B3LYP/6-31G(d) level of theory indicated that the HOMO orbital densities of both **AS2** and **AS1** have a major distribution over the whole molecular backbone and the LUMO densities were delocalized through the central acceptor functionality and adjacent rings ([Figure 5](#molecules-20-19798-f005){ref-type="fig"}). Inserted thiophene rings in case of **AS2** can accommodate LUMO density with almost equal contribution as the central PDA, thus sparing the adjoining phenyl rings of the donor triphenylamine for an efficient segregation of HOMO and LUMO densities. Such separation is ideal for the ICT transition and is attributed to the presence of a strong conjugated unit, of which thiophene is an example, in the given D--A--D system. Experimental estimation of the HOMO energies was carried out using photo electron spectroscopy in air (PESA) and the LUMO energies were calculated by adding the optical band gap (film spectra) to the HOMO values (see [Figure 6](#molecules-20-19798-f006){ref-type="fig"} for energy level diagram and [Figure S1](#app1-molecules-20-19798){ref-type="app"} \[see [Supplementary Material](#app1-molecules-20-19798){ref-type="app"}\] for the PESA curve). Film spectra indicated that insertion of thiophene reduces the band gap of **AS2** by 0.24 eV when compared with **AS1**. Furthermore, the estimation of HOMO using PESA revealed that the HOMO energy level of **AS2** was raised by 0.18 eV when compared with the HOMO level of **AS1**. This is in agreement with the DFT calculations that the presence of thiophene indeed plays a crucial role for: (1) density segregation; (2) tuning the optical energy levels; and (3) theoretical and experimental band gap reduction. These measurements and calculations provide a strong rational for our design strategy that the induction of a conjugated functionality can indeed improve the optoelectronic properties of a given chromophore. The energy level diagram advised that the band gaps of these materials are all in the range required of donor materials for BHJ devices. **AS2** optical band gap is somewhat narrower in magnitude than 2.0 eV measured for the conventional P3HT. The optical and electrochemical properties of both the materials in solution and film form are summarized in [Supplementary Material Table S1](#app1-molecules-20-19798){ref-type="app"}.
{#molecules-20-19798-f004}
{#molecules-20-19798-f005}
{#molecules-20-19798-f006}
Encouraging though the optoelectronic properties are, the compounds must display thermal stability given the rigorous conditions used in device fabrication such as annealing at temperature in the excess of 100 °C. In line with this requirement and to determine the thermal stability of **AS2** and **AS1**, thermogravimetric analysis (TGA) was conducted. TGA ([Supplementary Material Figure S2](#app1-molecules-20-19798){ref-type="app"}) indicated that both **AS2** and **AS1** are thermally stable and will not degrade during the annealing of BHJ devices.
After correlating the optoelectronic properties of **AS2** and **AS1** with those of the conventional PC~61~BM acceptor (see [Figure 6](#molecules-20-19798-f006){ref-type="fig"}), we screened their potency as donor materials (p-type) in solution-processable BHJ devices under simulated sunlight and monochromatic light illumination. The blend solutions of both the materials and PC~61~BM were used to cast an active layer on top of the PEDOT:PSS surface. The BHJ device architecture used was ITO/PEDOT:PSS (38 nm)/active layer/Ca (20 nm)/Al (100 nm) where the active layer was a solution processed blend of either **AS2** or **AS1** and PC~61~BM. For **AS2**, a promising PCE of 4.10% was achieved when the film was spin-coated from a chlorobenzene solution as a 1:1 blend with PC~61~BM. By contrast, the maximum PCE obtained for a device based on **AS1** was 1.23%, when fabricated under similar conditions. The comparative current--voltage curves for the optimized blends of **AS2** and **AS1** with PC~61~BM are shown in [Figure 7](#molecules-20-19798-f007){ref-type="fig"}. High boiling solvents such as chlorobenzene are better not only from processing point of view but also for achieving smoother films without crystallization occurring on the active surfaces. Latter was particularly true as our efforts to construct BHJ devices using a low-boiling solvent, such as chloroform, afforded either uneven surfaces or minor cracks on the active surfaces.
The optimized devices based on **AS2** exhibited a decreased open circuit voltage (*V*~oc~) than the devices based on **AS1**. This in fact is consistent with the measured HOMO values where a higher HOMO for **AS2** would predict a lower *V*~oc~. On the other hand, the short circuit current density (*J*~sc~) for the devices fabricated using **AS2** was higher than the *J*~sc~ extracted from the devices based on **AS1**. This was in agreement with the observed bathochromic shift in the absorption spectrum of **AS2** compared with **AS1**. The photovoltaic cell parameters for **AS2**-based devices were 0.88 V \[open circuit voltage, *V*~oc~\], 8.01 mA·cm^−2^ \[current density, (*J*~sc~)\], 0.58 \[fill factor, (FF)\] and 4.10% \[power conversion efficiency, (PCE)\]. Initial screening of the BHJ devices based on **AS1**:PC~61~BM showed moderate device performance with a high *V*~oc~ of 0.90 V, FF of 43% *J*~sc~ of 3.15 mA/cm^2^ and an overall PCE of 1.23%. Taken as a whole, the insertion of thiophene functionality into the studied D--A--D structural motif incorporating PDA acceptor functionality resulted in significant enhancement of *J*~sc~ and PCE values by factors \>2 and \>3, respectively, thus promoting the use of a smaller conjugated functionality, such as thiophene, as an interesting structural concept for the design and development of highly efficient BHJ materials. Furthermore, it is notable to mention that **AS2** showed optimal performance with inexpensive PC~61~BM and simple device architecture, thus providing some strong incentive to apply the design concept reported in this work to the new generations of BHJ materials.
{#molecules-20-19798-f007}
The incident photon-to-current-conversion efficiency (IPCE) spectra of these BHJ devices are shown in [Figure 8](#molecules-20-19798-f008){ref-type="fig"}. The IPCE measurement of these BHJ devices was broad spectrum, typically covering most of the visible range, from 350 to 750 nm. **AS2** and **AS1** exhibited high plateaus at \~40% and \~27% for the best BHJ devices respectively. The IPCE spectrum of **AS2**, which carried a thiophene functionality, was red-shifted when compared with **AS1**, a finding that is consistent with the result of thin film absorption spectrum. The significantly higher peak IPCE of **AS2** compared to **AS1** indicated that the superior performance of **AS2** can be rationalized in terms of enhanced light-harvesting and appropriately tuned optical energy levels, thereby corroborating the design principle.
{#molecules-20-19798-f008}
To examine the physical microstructure of the blend surface, we used atomic force microscopy (AFM) in tapping mode. The actual surface morphology of the blend films of **AS2**/**AS1**:PC~61~BM (1:1 *w*/*w*) is shown in [Figure 9](#molecules-20-19798-f009){ref-type="fig"}. Physically, both the blends were found to be smooth and the root mean square roughness of 0.41 nm and 0.35 nm was observed for **AS2** and **AS1** respectively. No cracks were observed on the film surfaces when the films were spin-casted using chlorobenzene (3000 rpm). The processing of active films of BHJ devices using a high boiling solvent such as chlorobenzene is advantageous over low boiling solvent such as chloroform and is in agreement with the AFM morphology. Our attempts to fabricate BHJ devices using chloroform resulted in very poor photovoltaic performance. This was mainly due to inferior film quality. Though **AS2** exerted promising PCE in this preliminary work, ample scope still exists to explore device strategies to enhance PCE. The performance might be improved by (1) using PC~71~BM or (2) effective interlayer, such as metal oxide interlayer, which can facilitate the efficient charge extraction, and (3) devising processing methods, such as use of additives. Work towards some of such strategies is the subject of on-going work in our laboratories. The discovery of potential materials, such as **AS2**, exhibiting promising optoelectronic and photovoltaic properties opens up the way to develop D--A--D modular small organic chromophores, with the use of PDA acceptor in particular, and paves the way for such materials to be used for other organic electronic applications.
{#molecules-20-19798-f009}
3. Experimental Section {#sec3-molecules-20-19798}
=======================
3.1. Materials and Instruments {#sec3dot1-molecules-20-19798}
------------------------------
All reagents and chemicals used, unless otherwise specified, were purchased from Sigma-Aldrich (Sydney, Australia). The solvents used for reactions were obtained from Merck Speciality Chemicals (Sydney, Australia) and were used as received. Unless otherwise specified, all ^1^H- and ^13^C-NMR spectra were recorded using a Bruker AV300 spectrometer at 300 and 75 MHz or a Bruker AV400 spectrometer at 400 and 100 MHz, respectively (Bruker Corporation, Billerica, MA, USA). Chemical shifts (δ) are reported in parts per million (ppm). Thin-layer chromatography (TLC) was performed using 0.25 mm thick plates precoated with Kieselgel 60 F~254~ silica gel (Merck, Darmstadt, Germany) and visualized using UV light (254 nm and 365 nm). Melting points were measured using a MPD350 digital melting point apparatus (Gallenkamp, Sanyo, Osaka, Japan) and are uncorrected. High-resolution mass spectra experiments were carried out on a Q-Exactive FTMS (Thermo Scientific, Bremen, Germany) ionizing by atmospheric-pressure chemical ionization (APCI) from an ASAP probe. All UV-Vis absorption spectra were recorded on a Hewlett Packard (HP) 8453 diode array UV-Vis spectrophotometer (Agilent Technologies, Mulgrave Victoria, Australia). Thin films were spin-coated from chlorobenzene at a spin speed of 2000 rpm for 1 min onto cleaned glass slides. PESA measurement was recorded using a Riken Keiki AC-2 PESA spectrometer (RKI Instruments, Union City, CA, USA) with a power setting of 5 nW and a power number of 0.5. Samples for PESA were prepared on clean glass substrates. Fabrication and characterization of BHJ devices, and preparation of thin-film transistors has been reported previously \[[@B18-molecules-20-19798],[@B30-molecules-20-19798]\].
3.2. Synthesis and Characterization of Target Molecules {#sec3dot2-molecules-20-19798}
-------------------------------------------------------
### 3.2.1. Synthesis of **AS1** {#sec3dot2dot1-molecules-20-19798}
1,4-Phenylenediacetonitrile (500 mg, 3.21 mmol) was added to the mixture of 4-(diphenyl-amino)benzaldehyde (1.84 g, 6.74 mmol) in methanol (50.0 mL) at room temperature and the resulting mixture was heated at reflux overnight. The precipitated solid was collected by filtration, washed with methanol and dried under vacuum to give 1.85 g (86.3%) of **AS1** as an orange powder. m. p. 242--245 °C; HPLC (5% H~2~O/ACN): 97.6%; IR (solid film, cm^−1^) 3061, 3035 (Ar -CH str), 2250, 2206 (-CN str), 1580, 1504, 1486 (Ar C=C str), 1330, 1285, 1192, 1179; ^1^H-NMR (400 MHz, CDCl~3~): δ = 7.80--7.76 (m, 4 H), 7.68 (s, 4 H), 7.45 (s, 2 H), 7.33--7.28 (m, 8 H), 7.16--7.10 (m, 12 H), 7.05--7.02 (m, 4 H); ^13^C-NMR (400 MHz, CDCl~3~): δ = 150.2, 146.5, 141.8, 135.1, 130.8, 129.6, 126.2, 126.1, 125.8, 124.5, 120.7, 118.5, 106.6; HRMS (APCI): calculated for C~48~H~35~N~4~ \[M+H\]^+^ 667.2856; found 667.2851.
### 3.2.2. Synthesis of 5-(4-(Diphenylamino)phenyl)thiophene-2-carbaldehyde (**1**) {#sec3dot2dot2-molecules-20-19798}
A solution of 1,2-dimethoxyethane (DME, 40.0 mL) and 2M Na~2~CO~3~ (20.0 mL) is degassed with nitrogen (N~2~) for 30 min. To this solution 5-bromothiophene-2-carbaldehyde (191 mg, 1.00 mmol) and (4-(diphenylamino)phenyl)boronic acid (433 mg, 1.50 mmol) were added and the mixture was heated at 60 °C for 30 min. \[Pd(PPh~3~)~4~\] (110 mg, 0.10 mmol) was added and the resulting mixture was stirred at 90 °C overnight. The reaction mixture was extracted with diethyl ether (3 × 50 mL). The organic layers were combined, washed with brine (100 mL) and dried over anhydrous magnesium sulfate. The solvent was evaporated to afford crude yellow solid, which was crystallized from hexane and chloroform to yield 378 mg (71%) of **1** as yellow needle like crystals. m. p. 100--102 °C; IR (solid film, cm^−1^): 3313, 3023, 2794, 1956, 1887, 1720, 1659, 1584, 1527, 1485, 1465, 1324, 1261, 1226, 1177, 1153, 1075, 1054; ^1^H-NMR (300 MHz, CD~2~Cl~2~): δ = 9.87 (s, 1H), 7.75 (d, *J* = 6.8 Hz, 1H), 7.60--7.57 (m, 2H), 7.37--7.31 (m, 5H), 7.71--7.15 (m, 6H), 7.13--7.06 (m, 2H); ^13^C-NMR (300 MHz, CDCl~3~): δ = 182.9, 154.5, 149.5, 147.4, 141.8, 138.1, 129.8, 127.6, 126.5, 125.6, 124.3, 123.4, 122.6; LRMS (ESI; 2% Formic acid): *m*/*z* = 356 (M + H)^+^; HRMS (APCI): calculated for C~23~H~17~NOS \[M\]^+^ 355.1031; found 355.1026.
### 3.2.3. Synthesis of **AS2** {#sec3dot2dot3-molecules-20-19798}
A solution of sodium ehtoxide was prepared by dissolving one pellet (approx. 200 mg) of sodium hydroxide (NaOH) in ethanol (EtOH, 20 mL) with mild heating (40 °C). To this solution PDA (78 mg, 0.50 mmol) was added and allowed to dissolve. A solution of **1** (390 mg, 1.1 mmol) in EtOH (10 mL) was added dropwise and the resulting solution was refluxed for 6 h. The mixture was allowed to cool to room temperature and placed in an ice-bath for 30 min. The separated red solid was filtered off, washed with cold EtOH (50.0 mL) followed by cold hexane (50.0 mL), and dried under high vacuum at 40 °C. The solid was crystallized using CHCl~3~/hexane to afford 574 mg of **AS2** (63%) as a brick-red solid. m. p. \>240 °C (decomposed at 240 °C); IR (solid film, cm^−1^) 3674, 2988, 2208 (-CN str), 1578, 1486, 1440, 1421, 1324, 1271, 1238, 1179, 1064, 922, 829. ^1^H-NMR (300 MHz, CD~2~Cl~2~): δ = 7.77--7.76 (m, 3H), 7.65--7.63 (m, 1H), 7.60--7.59 (m, 2H), 7.36--7.31 (m, 5H) 7.18--7.15 (m, 5H), 7.12--7.08 (m, 3H); ^13^C-NMR (300 MHz, CDCl~3~) δ = 149.9, 148.6, 147.1, 136.1, 134.8, 134.7, 134.4, 129.4, 127.0, 126.7, 126.0, 124.9, 123.6, 122.8, 122.7, 118.1, 105.7; LRMS (MALDI-TOF): *m*/*z* = 830.229; HRMS (APCI): calculated for C~56~H~39~N~4~S~2~ \[M + H\]^+^ 831.2611; found 831.2602.
### 3.2.4. X-ray Crystallography {#sec3dot2dot4-molecules-20-19798}
CCDC's **1420377** for **1** and **1420378** for **AS2** contain the supplementary crystallographic data for this paper. These data can be obtained free of charge via <http://www.ccdc.cam.ac.uk/conts/retrieving.html> (or from the CCDC, 12 Union Road, Cambridge CB2 1EZ, UK; Fax: +44 1223 336033; E-mail: <deposit@ccdc.cam.ac.uk>).
4. Conclusions {#sec4-molecules-20-19798}
==============
In conclusion, we have demonstrated the first use of the PDA acceptor functionality in conjunction with a thiophene unit for the design and development of a BHJ chromophore, **AS2**, where PDA was used in the D--A--D modular arrangement. The incorporation of this strong conjugating thiophene unit helped to improve light-harvesting, photocurrent density and PCE of **AS2** when compared with an analogue, **AS1**. The incorporation of the thiophene functionality was of clear benefit in improving the BHJ performance and indicates a potential to be broadly applicable in the design and development of future high performance BHJ chromophores.
S.V.B. (RMIT) acknowledges financial support from the Australian Research Council (ARC), Australia, under a Future Fellowship Scheme (FT110100152). The CSIRO Division of Materials Science and Engineering, Clayton, Victoria is acknowledged for providing support through a visiting fellow position (A.G.). P.S. is thankful to the ARC Future Fellowship Scheme (FT130101337) at Queensland University of Technology, Brisbane, Queensland.
**Sample Availability:** Samples of the compounds **S10** and **S11** are available from the authors.
######
Click here for additional data file.
Supplementary materials can be accessed at: <http://www.mdpi.com/1420-3049/20/12/19798/s1>.
A.M.R. synthesis, characterization, properties of materials and fabrication of devices; S.L.J and C.M.P.: single crystal structure determinations; A.G. and H.P. device fabrication and characterization; A.B.: performance of DFT calculation; P.S: results and discussion part, and S.V.B.: design, supervision and analysis of data. All the authors contributed for the manuscript preparation.
The authors declare no conflict of interests.
| {
"pile_set_name": "PubMed Central"
} |
Introduction
============
Pancreatic carcinoma is one of the most frequently occurring gastrointestinal malignancies and the incidence rate has shown an upward trend worldwide ([@b1-etm-05-01-0155]). The prognosis for patients with advanced pancreatic carcinoma remains poor with a 5-year survival rate of \<5% ([@b1-etm-05-01-0155]). Among the most significant determinants of the poor prognosis associated with this malignancy are the highly aggressive loco-regional invasion and early metastasis that characterise this malignancy, such that the majority of patients present with advanced, surgically unresectable disease ([@b2-etm-05-01-0155]). Gemcitabine and erlotinib are the only agents that are approved for the treatment of pancreatic carcinoma. However, both drugs induce a poor response in patients and their use can result in patients developing multiple drug resistance ([@b3-etm-05-01-0155],[@b4-etm-05-01-0155]). Although in recent years, great progress has been observed with regard to investigations on the molecular pathogenesis of pancreatic carcinoma, the clinical treatment of pancreatic carcinoma remains a challenge. Therefore, novel therapeutic approaches to this malignancy are needed.
Gene-direct enzyme/prodrug therapy (GEPT), or suicide gene therapy, aims to improve the therapeutic efficacy of conventional cancer radio- and chemotherapy without side-effects ([@b5-etm-05-01-0155],[@b6-etm-05-01-0155]). This system has received a great deal of attention for its clinical and therapeutic potential to treat cancer. At present, a large number of enzyme/prodrug systems have been developed for GEPT, two of which are the herpes simplex virus thymidine kinase/ganciclovir (HSV-TK/GCV) and cytosine deaminase/5-fluorocytosine (CD/5-FC) ([@b7-etm-05-01-0155]). Recently, some studies ([@b8-etm-05-01-0155],[@b9-etm-05-01-0155]) have tried to enhance the therapeutic effect of suicide gene therapy by combining it with other gene therapies. However, the combination of fusion suicide gene therapy with anti-angiogenesis gene therapy for pancreatic carcinoma has yet to be reported.
Carcinoembryonic antigen-related cell adhesion molecule (CEACAM)6, also known as CD66c or NCA-90, as well as another 6 members of the CEACAM subgroup, belong to the human carcinoembryonic antigen (CEA) family ([@b10-etm-05-01-0155]). CEACAM6 overexpression was found in a wide variety of epithelial cancer types such as lung, breast, colorectal, and hepatocellular carcinomas ([@b11-etm-05-01-0155]--[@b14-etm-05-01-0155]). CEACAM6 has also become a target for pancreatic cancer therapy ([@b15-etm-05-01-0155],[@b16-etm-05-01-0155]). Overexpression of CEACAM6 was found in \>90% of invasive pancreatic adenocarcinomas ([@b16-etm-05-01-0155]). RNA interference (RNAi) is a process involving sequence-specific and post-transcriptional gene silencing. CEACAM6-specific RNAi decreases cancer cell proliferation, metastasis and angiogenesis in pancreatic cancer ([@b15-etm-05-01-0155]).
In the present study, we aimed to test the feasibility of a novel therapeutic vector system involving a combination of suicide gene therapy and antiangiogenesis gene therapy. The *in vitro* experiments on pancreatic carcinoma SW1990 cells were studied using a triple-gene vector expressing CEACAM6-shRNA and the fusion suicide gene yCDglyTK.
Materials and methods
=====================
Cell lines and cell culture
---------------------------
The SW1990 human pancreatic carcinoma cell line (CEA positive) was obtained from the Cancer Research Institution, Central South University (Hunan, China). Cells were cultured in RPMI-1640 medium (Invitrogen Inc., Carlsbad, CA, USA) with 10% fetal bovine serum at 37°C in a humidified atmosphere of 5% CO~2~ and 95% air.
Construction of the triple-gene plasmid of pcDNA3.1(-) shCEACAM6-yCDglyTK
-------------------------------------------------------------------------
pcDNA3.1(-)CV-yCDglyTK was constructed in our previous study ([@b17-etm-05-01-0155]). Expression of the fusion suicide gene yCDglyTK was regulated by CMV- enhanced CEA promoter and was expressed specifically in CEA-positive cells. Oligonucleotides encoding the corresponding small hairpin RNA (termed CEACAM6-shRNA) was generated by ligation of inserts targeting the following sequences into a hU6 promoter-contained pUC57-simple plasmid: 5′-CCG GACAGTTCCATGTATATTCAAGACGTATACATGGAAC TGTCGTTTTTT-3′ (sense: CCGGACAGTTCCATGTATA; loop: TTCAAGACG; antisense: TATACATGGAACTGT CCGG; termination code: TTTTTT). shCEACAM6 and pcDNA3.1(-)yCDglyTK were then fused by *Nhe*I and *Xba*I restriction endonuclease overnight at 37°C to form a recombinant plasmid of shCEACAM6 and the fusion suicide gene yCDglyTK .Thus, a novel triple-gene vector pcDNA3.1(*-*) shCEACAM6-yCDglyTK was developed. Double enzyme cutting and sequencing were performed to prove the accuracy of the new plasmid.
Stable transfection in vitro
----------------------------
The novel triple-gene vector, pcDNA3.1(-)shCEACAM6-yCDglyTK plasmid was mixed with Lipofectamine 2000 at the combination rate of 1/2 to 1/3. SW1990 human pancreatic carcinoma cells were seeded in 6-well plates at a density of 2×10^5^ cells per well. When the cell monolayer reached 70--80% confluence, pcDNA3.1(-) null, pcDNA3.1(-)yCDglyTK and pcDNA3.1(-)shCEACAM6-yCDglyTK were added to different 6-well plates. The following day, a 1:10 passage of the transfected SW1990 cells was performed, followed by G418 selection (400 *μ*g/ml). Approximately 3 weeks later, the resistant colonies were picked and transferred to 96-well plates. These clones were maintained in selective culture medium (with 200 *μ*g/ml of G418). Surviving colonies transfected with pcDNA3.1(-)null, pcDNA3.1(-) yCDglyTK, or pcDNA3.1(-)shCEACAM6-yCDglyTK were designated as SW/null, SW/CDTK, or SW/shCEACAM6-CDTK, respectively, and subjected to further study.
Reverse transcription polymerase chain reaction (RT-PCR) and western blot analysis
----------------------------------------------------------------------------------
Total RNA from parental SW1990 cells and three different transfected cells was extracted using TRIzol reagent (Invitrogen Inc.). The quantity and quality of RNA was assessed by absorbance at 260 and 280 nm, respectively. The RT reaction was carried out using the ReverTra Ace^®^ RT Kit (Toyobo Co., Ltd., Osaka, Japan) as per the manufacturer's instructions. Subsequently, we performed PCR on the cDNA product. For yCDglyTK, a PCR product of 707 bp was produced by the forward primer 5′-GGGAGATT AGAGGGCAAAGTGT-3′ and reverse primer 5′-ACGGCGT CGGTCACGGCATAA-3′. For CEACAM6, a PCR product of 356 bp was produced by the forward primer 5′-CGTTCAAT GTCGCAGAGGG-3′ and reverse primer 5′-CGCTGAGTA GAGTGAGGGT-3′. β-actin was used as an internal control and a PCR product of 285 bp was produced by the forward primer 5′-AGCGAGCATCCCCCAAAGTT-3′ and reverse primer 5′-GGGCACGAAGGCTCATCATT-3′.
For western blot analysis, cells were collected 72 h after transfection and lysed in loading buffer (20 mmol/l Tris-HCl, pH 7.5; 150 mmol/l NaCl, 1 mmol/l EDTA, 5 mmol/l DTT, 1% Triton X-100). The lysates were centrifuged at 12,000 × g for 15 min at 4°C. The supernatant was collected and protein concentrations were determined by the BCA protein assay. Protein (40 *μ*g) was separated via 15% SDS-PAGE and transferred to PVDF film (Amersham Pharmacia Biotech, Piscataway, NJ, USA). The films were incubated in blocking solution, consisting of 5% skimmed milk in TBS-T \[10 mM Tris-HCl (pH 8.0), 150 mM NaCl and 0.1% Tween-20\], for 1 h at room temperature, then probed with mouse anti-CEACAM6 antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA), rabbit anti-TK antibody (QED Bioscience, Inc., San Diego, CA, USA) or mouse anti-β-actin antibody (Sigma-Aldrich, St. Louis, MO, USA). This was followed by incubation with their respective peroxidase-conjugated secondary antibodies. Hybridization was visualized using the ECL chemiluminescence detection system (Kodak).
CDTK/5-FC, shCEACAM6-CDTK/5-FC system-induced cytotoxicity
----------------------------------------------------------
SW1990 cells (transfected and untransfected) were seeded in 96-well plates at a density of 8,000 cells per well and incubated at 37°C and 5% CO~2~ in humidified air for 24 h. The following day, 5-fluorocytosine (5-FC) was added into the culture medium at a final concentration of 200 *μ*g/ml. MTT assays were conducted at 24-, 48-, 72- and 96-h incubation time points to analyze cell viability. Twenty microliters of MTT solution (5 mg/ml, Sigma-Aldrich) were added to each well and incubated for 4 h. Dimethylsulfoxide (DMSO, Promega Corporation, Madison, WI, USA) was added to dissolve the blue crystal. The optical density (OD) was then determined using a multi-well plate reader (Awareness, model Stat-Fax-2100, USA) by measuring absorbance at 570 nm (OD570), with the absorbance at 690 nm as reference. The background absorbance of medium was also subtracted. Samples were assayed in triplicate, and the mean for each experiment was calculated. Cell growth curves were plotted, with culture time on the horizontal axis and OD570 on the vertical axis.
Invasion and migration assay
----------------------------
The invasion assay was performed using an 8 *μ*m pore size transwell chamber in 24-well plates (Corning Costar, Cambridge, MA, USA). SW1990 cells stably expressing pcDNA3.1(-), pcDNA3.1(-)yCDglyTK, pcDNA3.1(-) shCEACAM6-yCDglyTK or untransfected SW1900 cells in 500 *μ*l of serum-free MEM medium were loaded into the top chamber with fetal bovine serum placed in the bottom chamber as a chemoattractant. After further incubation at 37°C for 10 h, the cells on the top of the filters were removed with cotton swabs. The cells on the lower surface of the filters were fixed in 4% paraformaldehyde and stained with 0.1% crystal violet. The crystal violet was removed and the cells were washed three times with phosphate-buffered saline (PBS). The remaining crystal violet staining of the migrated cells was eluted with one wash with 33% acetic acid. The OD540 nm of the eluted crystal violet was determined as a measure of migrated cells. Each experiment was performed in triplicate. The invasion of different groups was observed under a microscope. The cell migration assay was performed in a similar mode, except that cells were seeded into the uncoated filter and incubated for 24 h. Each measurement was performed in at least three independent experiments.
Statistical analysis
--------------------
Statistical analysis was performed by one-way analysis of variance (ANOVA) test. P\<0.05 was considered to indicate a statistically significant difference. Numeric data were presented as the mean values ± standard deviation (SD).
Results
=======
Recombinant plasmid was successfully constructed
------------------------------------------------
An interfering plasmid targeting CEACAM6 was initially constructed. The CEACAM6-shRNA expression cassette was subcloned into pcDNA3.1(-)CV-yCDglyTK to construct the novel vector pcDNA3.1(*-*)shCEACAM6-yCDglyTK ([Fig. 1](#f1-etm-05-01-0155){ref-type="fig"}). In this novel triple-expressing plasmid, the CEACAM6-shRNA sequence was placed under control of the U6 promoter, while the fusion suicide gene yCDglyTK was driven by a CMV-enhanced CEA promoter. Newly constructed gene plasmid pcDNA3.1(-)shCEACAM6-yCDglyTK was identified by double-enzyme cutting and sequencing. As expected, the size of the newly constructed plasmid fragment by enzyme cutting was 7.5 kb ([Fig. 2](#f2-etm-05-01-0155){ref-type="fig"}). Sequencing results showed that the newly constructed target-combined double suicide gene plasmid was in concordance with pcDNA3.1(-) shCEACAM6-yCDglyTK (data not shown).
Recombinant plasmid was effectively delivered into SW1990 cells in vitro
------------------------------------------------------------------------
This novel vector was delivered into SW1990 cells and stably transfected cell lines were obtained by G418 selection. At the same time, SW1990 cells stably transfected with three other plasmids, pcDNA3.1(-)null, pcDNA3.1(-) yCDglyTK and pcDNA3.1(-)shCEACAM6-yCDglyTK were also established. The expression of CEACAM6 and yCDglyTK were determined via RT-PCR, western blot analysis and immunofluorescence. Compared with parent SW1990 cells and SW/null, mRNA and protein levels of CEACAM6 were significantly decreased in SW/shCEACAM6-CDTK ([Fig. 3](#f3-etm-05-01-0155){ref-type="fig"}). yCDglyTK was confirmed to be expressed in SW/CDTK and SW/shCEACAM6-CDTK cells, but not in the parent SW1990 cells and SW/null ([Fig. 4](#f4-etm-05-01-0155){ref-type="fig"}).
Recombinant plasmid pcDNA3.1(-)shCEACA M6 - yCDglyTK/5-FC system resulted in cytotoxicity in SW1990 cells
---------------------------------------------------------------------------------------------------------
After a 24-h treatment with 5-FC, the OD570 of SW/CDTK cells and SW/shCEACAM6-CDTK cells decreased significantly compared to SW1990 or SW1990/null cells (P\<0.01), as shown in the cell growth curve in [Fig. 5](#f5-etm-05-01-0155){ref-type="fig"}. As time progressed, a high proliferation rate was maintained in untransfected SW1990 and SW/null cells. After 48 h, the OD570 did not increase in SW/CDTK or SW/shCEACAM6-CDTK cells. At 72- and 96-h treatment with 5-FC, a decrease was observed for the OD570, suggesting that most cells in the process of being killed. Low shCEACAM6-CDTK cell viability was detected.
Recombinant plasmid pcDNA3.1(-)shCEACAM6-yCDglyTK inhibited SW1990 cell invasion and migration
----------------------------------------------------------------------------------------------
The motility of three different transfected cells across transwell polycarbonate membranes was evaluated. As shown in [Fig. 6A](#f6-etm-05-01-0155){ref-type="fig"}, compared with SW1990 and SW/null, the cell invasiveness and migration of SW/CDTK were attenuated to 50.41 and 80.12%, respectively (P\<0.01), while those of SW/shCEACAM6-CDTK cells were reduced more significantly to 24.61 and 45.17%, respectively (P\<0.01). By contrast, no differences were observed between SW1990 and SW/null (P\>0.05) ([Fig. 6B](#f6-etm-05-01-0155){ref-type="fig"}).
Discussion
==========
Mounting evidence suggests that combination cancer therapy has the potential to be effective in combating malignancies. Combination gene therapy has the advantages of gene therapy, elevates the therapeutic efficacy and overcomes the shortcomings of single gene therapy ([@b18-etm-05-01-0155]). Although suicide gene therapy is a potentially effective method for killing tumor cells *in vitro* and *in vivo*, the results of clinical trials indicate a need for greater efficacy ([@b19-etm-05-01-0155]). In previous studies, the co-transfer of vectors carrying different genes to enhance anti-tumor effect has been attempted ([@b20-etm-05-01-0155]). Combination of TK/CD gene therapy gene therapy with lipiodol embolism in the treatment of liver cancer may effectively inhibit cancer growth and prolong the survival time ([@b9-etm-05-01-0155]). In this study, to test the feasibility of a novel therapeutic vector system involving a combination of suicide and RNAi-based gene therapy, we initially constructed the novel vector pcDNA3.1(-)shCEACAM6-yCDglyTK, which was regulated by a U6 promoter, while the fusion suicide gene yCDglyTK was driven by a CMV-enhanced CEA promoter. Normal expression of each gene was confirmed by RT-PCR and western blot analysis. Pancreatic carcinoma cell lines stably expressing the CEACAM6 shRNA and yCDglyTK gene were then established and anti-tumor efficacy of the recombinant plasmid was evaluated *in vitro*.
Suicide genes are viral or bacterial enzymes capable of converting non-toxic prodrugs into toxic metabolites that cause tumor cell death when introduced into tumor sites. CD and HSV-TK are typical suicide genes that convert non-toxic prodrugs, such as 5-FC and GCV, into cytotoxic metabolites, such as 5-fluorouracil (5-FU) and GCV-TP, respectively ([@b20-etm-05-01-0155]--[@b23-etm-05-01-0155]). Combination of HSV-TK/GCV and CD/5-FC might have synergistic effects ([@b24-etm-05-01-0155]). Previous studies showed that yCD, a yeast-derived CD, is efficient at deaminating 5-FC to yield 5-FU, whereas bacterial CD (bCD) has a poor conversion efficiency ([@b25-etm-05-01-0155]). Moriuchi *et al*([@b26-etm-05-01-0155]) reported that TK was able to mediate the phosphorylation of 5-FU metabolites, and might reduce the cytotoxicity of CD/5-FC system. GCV had the potential to interfere with 5-FC in the yCDglyTK system, thus 5-FC was used as the only prodrug in our study. Cytotoxicity of the yCDglyTK gene in the presence of prodrugs using MTT assay was tested. The mean cell viability decreased in a time-dependent manner in yCDglyTK- and shCEACAM6-yCDglyTK-transfected SW1990 cells, although not in untransfected SW1990 cells. Our results have shown that yCDglyTK was confirmed to be expressed in SW/CDTK and SW/shCEACAM6-CDTK cells, rendering the new system efficient in delivering suicide gene into cancer cells and inducing cytotoxicity.
CEACAM6 is a single-chain GPI-anchored immunoglobulin (Ig)-like glycoprotein and is a member of the human CEA family ([@b27-etm-05-01-0155]). Jantscheff *et al*([@b28-etm-05-01-0155]) showed that CEACAM6 overexpression was associated with poor clinical outcome in colorectal cancer. CEACAM6 overexpression independently predicted poor overall survival and disease-free survival, whereas CEACAM1 or CEACAM5 was not significantly associated with these outcomes. CEACAM6 overexpression leads to morphology changes that are similar to epithelium-messenchymal-transformation ([@b29-etm-05-01-0155]), increased invasiveness ([@b29-etm-05-01-0155]), increased chemoresistance ([@b30-etm-05-01-0155]) and resistance to anoikis ([@b31-etm-05-01-0155]--[@b33-etm-05-01-0155]), whereas CEACAM6 appears to exert its pro-invasive effect in a c-Src-dependent manner, at least in part through the upregulation of MMP-9 activity ([@b34-etm-05-01-0155]). It has also been proposed that low levels of E-cadherin-mediated cell-to-cell interaction are important in tumor invasiveness and metastasis ([@b35-etm-05-01-0155],[@b36-etm-05-01-0155]). Suppressing CEACAM6 gene expression or inhibiting CEACAM6 function can reverse these effects. Inhibition of CEACAM6 function using an antibody fragment can affect cell migration, invasion and adhesion *in vitro*([@b12-etm-05-01-0155],[@b34-etm-05-01-0155]). RNAi offers a unique opportunity to silence the expression of individual genes with a high degree of specificity, allowing the roles of individual genes to be dissected ([@b37-etm-05-01-0155]). In the present study, we investigated the invasion and migration-inhibitory effects of the yCDglyTK- and shCEACAM6-yCDglyTK-transfected SW1990 cells. The invasion and migration were significantly suppressed (P\<0.01) in the two transfected groups. The inhibitory rates of shCEACAM6-yCDglyTK-transfected SW1990 cells were more prominent than those of yCDglyTK-transfected SW1990 cells. Suppression of the CEACAM6 transcripts using siRNA of CEACAM6 leading to a reduction in cancer cell invasiveness, may be associated with an increase in E-cadherin promoter activity ([@b35-etm-05-01-0155]). CEACAM6-siRNA or yCDglyTK alone has the potential to kill cancer cells and cause tumor growth delay, as well as inhibit tumor invasion and migration. However, a combination of the two genes has been shown to achieve a stronger anti-tumor effect, demonstrating a synergistic effect between CEACAM6-siRNA and yCDglyTK.
The novel recombinant plasmid was able to silence functional genome CEACAM6, inhibit tumor invasion and metastasis. However, there are potential defects with this new system. CEA protein only overexpresses in a majority of pancreatic cancer cells. In our triple-expressing vector pcDNA3.1(-) shCEACAM6-yCDglyTK, we used a CEA promoter to drive the expression of yCDglyTK, a treatment that specifically killed CEA-positive cancer cells. The novel shCEACAM6-yCDglyTK system had little effect on the CEA-negative pancreatic cancer cells. Moreover, a low level of CEACAM6 protein expression has been noted in a variety of normal human tissues, including granulocytes and epithelia from various organs ([@b38-etm-05-01-0155]) and this expression is also associated with infectious diseases ([@b39-etm-05-01-0155],[@b40-etm-05-01-0155]). The novel system may therefore not only target specific tumor tissues. Strategies aiming to improve the safety of RNAi-based gene therapy are therefore necessary.
In conclusion, the results from the present study have demonstrated that CEACAM6-targeted RNAi with suicide gene therapies had a synergistic effect. Additionally, the combination gene therapy system may be a valid and viable strategy to inhibit the proliferation, as well as attenuate the invasiveness and metastasis of pancreatic carcinoma SW1990 cells *in vitro*. The present study provides a novel gene therapy strategy that is effective, not only for pancreatic cancer, but also for other CEACAM6-expressing tumors.
This study was supported partially by the Project from Science and Technology Department, Hunan, China (no. 2010FJ4087), a grant from Science and Technology Bureau, Changsha, China (no. K1203051-31) and the Innovation Program of Central South University, China (no. YB10071).
CEACAM6
: carcinoembryonic antigen-related cell adhesion molecule
CEA
: carcinoembryonic antigen
shRNA
: short hairpin RNA
GEPT
: gene-direct enzyme/prodrug therapy
RNAi
: RNA interference
CD
: cytosine deaminase
HSV-TK
: herpes simplex virus thymidine kinase
5-FC
: 5-fluorocytosine
GCV
: ganciclovir
{#f1-etm-05-01-0155}
{#f2-etm-05-01-0155}
{#f3-etm-05-01-0155}
{#f4-etm-05-01-0155}
{#f5-etm-05-01-0155}
{#f6-etm-05-01-0155}
| {
"pile_set_name": "PubMed Central"
} |
{#sp1 .2}
{#sp2 .3}
| {
"pile_set_name": "PubMed Central"
} |
All data are contained within the paper.
Introduction {#sec005}
============
Stroke is a major cause of disability in adults \[[@pone.0149757.ref001]\]. It frequently results in hemiparesis (partial paralysis of one side of the body) which causes slow gait with kinematic anomalies \[[@pone.0149757.ref002]\],\[[@pone.0149757.ref003]\]. Methods of quantitative gait analysis are becoming increasingly used in clinical practice to aid clinical decision-making by the assessment of spatio-temporal, kinematic and kinetic parameters \[[@pone.0149757.ref004]\]. Three-dimensional analysis is the current gold standard for the biomechanical assessment of patients with abnormal gait \[[@pone.0149757.ref005]\]. This typically involves the analysis of straight-line gait, however straight-line gait does not reflect daily life situations which include curved paths, obstacle circumvention and U-turns \[[@pone.0149757.ref006]\]. Curved paths and obstacle circumvention have been studied in healthy subjects \[[@pone.0149757.ref007]\],\[[@pone.0149757.ref008]\],\[[@pone.0149757.ref009]\] and more recently in subjects with stroke \[[@pone.0149757.ref010]\],\[[@pone.0149757.ref011]\],\[[@pone.0149757.ref012]\]. The Timed Up and Go (TUG) test \[[@pone.0149757.ref013]\],\[[@pone.0149757.ref014]\] involves rising from a chair, walking 3m, turning 180°, returning, and sitting down again. It thus reflects the main aspects of gait required in daily life. It is rated according to performance time \[[@pone.0149757.ref013]\],\[[@pone.0149757.ref014]\],\[[@pone.0149757.ref015]\]. The test is useful and is quick and easy to perform, therefore it is widely used in clinical practice for the assessment of global locomotor capacity in stroke patients. However, performance time does not provide any information regarding the biomechanical behaviour of patients during the test. Moreover, several authors have recommended refining the TUG test by timing each sub-task (23), as well as carrying out a biomechanical analysis of each sub-task (24).
A recent approach to the analysis of biomechanical behavior during tasks involving curved gait is the study of trajectory. Locomotor trajectory has been evaluated in healthy subjects during imposed straight and curved walking (indicated by a line drawn on the floor) \[[@pone.0149757.ref007]\] as well as walking through doors with different spatial orientations \[[@pone.0149757.ref016]\]. The results suggest that the control of the locomotor pattern is based on the whole-body locomotor trajectory, rather than a sequence of foot pointings. To our knowledge, only one study has investigated locomotor trajectory in stroke patients \[[@pone.0149757.ref017]\]. The trajectories of patients with stroke and healthy subjects were evaluated in a virtual environment which created 5 different scenes of translational optic flow (a pattern of apparent motion of objects, surfaces, and edges in a visual scene caused by the relative motion between an observer and the scene) \[[@pone.0149757.ref017]\]. The medio-lateral and antero-posterior trajectories of the center of mass (COM) were computed while subjects were instructed to "walk straight with respect to the scene they were visualizing". Displacement of the COM was altered in the patients with motor disorders in contrast with the healthy subjects who displayed stereotypical behavior. The authors suggested that this was the result of an alteration in perception and/or a poor integration of sensorimotor information. No studies have analyzed the spontaneous trajectories of patients with stroke in a "real environment" during tasks encountered in daily life. Since many stroke patients have spatial disorders, such an analysis would be clinically relevant to guide rehabilitation, and the TUG test appears to be a pertinent test on which to base the analysis. Moreover, this test can easily be broken down into sub-tasks to analyze different locomotor task. In addition, it has been shown that perception of body verticality is altered following right hemisphere stroke \[[@pone.0149757.ref018]\], thus locomotor trajectories may differ between patients with right and left hemisphere stroke.
Several methods in the literature have been used to evaluate locomotor trajectories. The amount of deviation from either a required or an averaged trajectory appears to be particularly relevant \[[@pone.0149757.ref007]\],\[[@pone.0149757.ref016]\]. Trajectory deviation can be quantified using several parameters. The simplest is the Euclidean distance, however this method is not sufficiently accurate to compare groups with different gait velocities \[[@pone.0149757.ref019]\]. The Hausdorff Distance (HD) and Dynamic Time Warping (DTW) appear to be appropriate for the present study since these parameters can be used to compare the geometry and the spatio-temporal time series of two sequences of different lengths. HD and DTW have been used to evaluate moving objects \[[@pone.0149757.ref020]\], for handwriting recognition \[[@pone.0149757.ref021]\] and to study walking behavior \[[@pone.0149757.ref022]\],\[[@pone.0149757.ref023]\]. Since the gait of stroke patients is slower than that of healthy subjects, these parameters are pertinent \[[@pone.0149757.ref019]\],\[[@pone.0149757.ref020]\] to compare their locomotor trajectories.
The TUG test is considered to indicate a risk of falls \[[@pone.0149757.ref024]\],\[[@pone.0149757.ref025]\]. Older subjects are classified as fallers if they take 13.5sec or more to perform the test and stroke patients are considered at risk of falls if they take 15sec or more \[[@pone.0149757.ref024]\],\[[@pone.0149757.ref025]\]. However, a more recent study has suggested this test is not sufficiently accurate to discriminate fallers and non-fallers \[[@pone.0149757.ref026]\]. We thus propose to use HD and DTW to determine whether these trajectory-related parameters might permit to distinguish stroke-related fallers and non-fallers.
The aims of this study were thus: i) to analyze locomotor trajectories using HD and DTW in patients with stroke during the walking and turning sub-tasks of the TUG and to compare them with healthy subjects; ii) to determine whether trajectory parameters provide additional information to that of the conventional measure (performance time); iii) to compare the trajectory parameters of fallers and non-fallers with stroke and of patients with right and left hemisphere stroke and iv) to evaluate correlations between trajectory parameters and Berg Balance Scale scores. This study is the first to assess the locomotor trajectories of patients with stroke in real life conditions. The results should yield pertinent information for clinicians, helping to orientate rehabilitation and perhaps also to identify potential fallers. We hypothesized: 1) that the trajectories of stroke patients would deviate from those of healthy subjects, particularly during the Turn sub-task of the TUG since this task is the most challenging regarding stability, 2) that trajectory parameters would provide additional information to performance time, 3) that trajectories would differ between fallers and non-fallers and that since right hemisphere large vessel distribution stroke may alter perception of body verticality, it may also alter the locomotor trajectories and 4) that longer trajectories would be related to a poorer BBS scores since we supposed that patients with impaired balance would deviate from the optimal trajectory to ensure stability.
Methods {#sec006}
=======
Subjects {#sec007}
--------
Twenty nine patients with chronic stroke (mean age 54.2±12.2 years, 18 men), who were in- or outpatients in our department of physical medicine and rehabilitation, and twenty five healthy subjects (mean age 51.6±8.7 years, 11 men) were included. This number of subjects was sufficient to obtain a minimum statistical power of 95% with a significance level (alpha error) of 0.05, based on calculation of the effect size and statistical power using previous data published on TUG performance in stroke subjects \[[@pone.0149757.ref014]\],\[[@pone.0149757.ref027]\] \[[@pone.0149757.ref028]\]. Based on the current sample size and the results of DTW during the Turn and trajectory length, the effect sizes obtained were respectively 1.56 and 2.37 and the subsequent powers were respectively 0.99 and close to 1 which allow us to be confident in our results. Inclusion criteria were: hemiparesis following stroke, over 18 years old and able to carry out the TUG test several times consecutively without using an assistive device. Exclusion criteria were the diagnosis of other neurological or orthopedic conditions, or having undergone surgical procedures during the last 6 months. Participants' characteristics are presented in [Table 1](#pone.0149757.t001){ref-type="table"}. Patients were considered as fallers if they had fallen at least once within the last 3 months. The fallers' characteristics are presented in [Table 2](#pone.0149757.t002){ref-type="table"}. Eight patients had gait-related falls and constituted the group of fallers in this study. Six of these patients had fallen indoors (one while walking, one while walking in a narrow space, three while turning and one tripped on a rug) and 2 patients had fallen outdoors in crowded spaces. Six patients were not included in the faller group since they fell in conditions that did not involve walking (in the bathtub, on the stairs, rising from a chair, crossing an obstacle and entering a car). All patients were found to be capable of providing informed consent during the medical examination, and all gave written informed consent in accordance with the ethical codes of the World Medical Association. The study was approved by our local ethics committee (Comité de protection des personnes Ile de France XI, Ref 13005. CNIL, Ref DR-2013-283).
10.1371/journal.pone.0149757.t001
###### Subject characteristics.
{#pone.0149757.t001g}
Stroke patients (n = 29) Healthy subjects (n = 25)
------------------------------------------------------------------------ ----------------------------- ---------------------------
Age (years) 54.2±12.2 51.6±8.7
Height (m) 1.68±0.09 1.67±0.1
Weight (kg) 73.2±16.2 65.6±14.7
Gender (m/f) 18m / 11w 11m / 14w
Mean self-selected gait speeds for the walking phases of the TUG (m/s) 0.4±0.006 0.7±0.04
Time since stroke (years) 7.9±5.7 \-
Stroke etiology 19 ischemia / 10 hemorrhage \-
Hemiparetic side 12 right / 17 left \-
Falls 8 fallers related to gait \-
Modified Ashworth sum 4 \[2;7\] \-
MRC sum 23 \[19;25\] \-
Foot sensation 1 \[1;2\] \-
Toe proprioception 2 \[1;3\] \-
Barthel index 100 \[95;100\] \-
NFAC 7 \[7;7\] \-
BBS 51 \[49;52\] \-
ABC 76,3±12,9 \-
Patients with stroke had a significantly decreased gait speed compared to healthy subjects (p\<0.05)
Falls: patients were considered as fallers if they had fallen at least once within last 3 months
Spasticity: median \[interquartile range Q1;Q3\] of the sum of quadriceps, rectus femoris, hamstring and triceps surae spasticity assessed with Modified Ashworth Scale (0--4).
MRC (Medical Research Council scale): median \[interquartile range Q1;Q3\] of the sum of hip, knee and ankle flexor and extensor strength (0--5)
Foot sensation: median \[interquartile range Q1;Q3\] of the foot sensation score assessed with the Nottingham Sensory Assessment (0 = absent, 1 = impaired, 2 = normal)
Toe proprioception: median \[interquartile range Q1;Q3\] of the toe proprioception score assessed with the Nottingham Sensory Assessment (0 = absent, 1 = direction incorrect, 2 = direction ok, inaccurate position, 3 = direction ok, position accurate to 10°)
Barthel index: median \[interquartile range Q1;Q3\] Barthel score (0 to 100)
NFAC: median \[interquartile range Q1;Q3\] New Functional Ambulation Classification score (0 to 8)
BBS: median \[interquartile range Q1;Q3\] Berg Balance Scale score (0 to 56)
ABC: mean±sd Activities-specific Balance Confidence scale (0 to 100%)
10.1371/journal.pone.0149757.t002
###### Characteristics of the fallers and non-fallers.
{#pone.0149757.t002g}
Fallers (n = 8) Non-fallers (n = 21)
------------------ ------------------ ----------------------
Age (years) 59,5±11,6 52,2±12,1
Gender (m/f) 3m / 5w 15m / 6w
Hemiparetic side 2 right / 6 left 10 right / 11 left
TUG (sec) 19,7±1,8 19,1±4,9
Experimental procedure {#sec008}
----------------------
All participants performed 3 TUG tests under standardized conditions. They wore the same type of comfortable shoes \[[@pone.0149757.ref029]\], sat on a stool set to 100% of the distance from the head of the fibula to the floor \[[@pone.0149757.ref030]\] with their knees flexed to 100°, their feet placed symmetrically and their arms held out from the body \[[@pone.0149757.ref031]\],\[[@pone.0149757.ref032]\],\[[@pone.0149757.ref033]\]. Participants were instructed to rise from the stool, walk 3m, turn around a cone towards their paretic side (non-dominant side for healthy subjects), return to the stool and sit down, at their own comfortable speed. The TUG tests were recorded with a motion analysis system (Motion Analysis Corporation, Santa Rosa, CA, USA, sampling frequency 100 Hz). Thirty-four markers were fixed, by the same person, to specific bony landmarks according to the Helen Hayes marker set \[[@pone.0149757.ref034]\],\[[@pone.0149757.ref035]\],\[[@pone.0149757.ref005]\]. The marker set was used to create a 12-segment rigid-link model of the body using Dempster\'s anthropometric table which is routinely used in gait analysis \[[@pone.0149757.ref036]\],\[[@pone.0149757.ref037]\]. Markers were tracked by 8 infrared cameras and trajectories were filtered using a low-pass Butterworth filter with a cut off frequency of 6 Hz \[[@pone.0149757.ref038]\]. An open-source Biomechanical Tool Kit package for MATLAB \[[@pone.0149757.ref039]\] was used to define the phases of the gait cycle and sub-tasks of the TUG. The gait phases were defined according to Perry \[[@pone.0149757.ref003]\] and sub-tasks of the TUG were defined according to previous studies \[[@pone.0149757.ref033]\],\[[@pone.0149757.ref040]\],\[[@pone.0149757.ref041]\]. The three sub-tasks of the TUG that involve walking were analysed: the first oriented-gait sub-task (Go) which begins at toe off of the first step and ends with the first foot strike in the direction of the turn, the turning sub-task (Turn) which ends at the first foot strike lined up with the stool and the second oriented-gait sub-task (Return) ends with foot strike of the last step prior to the turn to sit \[[@pone.0149757.ref012]\].
Locomotor trajectory was evaluated by the displacement of the center of mass (COM) with the following equation 1: $$COMx = \frac{m_{1}\, x_{1} + m_{2}\, x_{2} + {\ldots.. + m}_{i}\, x_{i}}{M} = \frac{1}{M}{\sum\limits_{i = 1}^{N}{m_{i}\, x_{i}}}$$ where M = whole body mass
mi = mass of the ith segment = (whole body mass) x (mass fraction for ith segment from the anthropometrics.dat file)
xi = the x-coordinate of the center of mass for the ith segment with respect to the calibration origin
N = the number of body segments
The parameters analyzed were:
1. time to perform the Go, Turn and Return sub-tasks of the TUG, and total TUG time
2. length of the COM trajectory, HD and DTW
The trajectories of each patient and healthy subject were compared with the reference trajectory, defined as the mean of the healthy subjects' trajectories which were time-resampled \[[@pone.0149757.ref016]\].
*Trajectory length* was calculated with the following equation 2 $${Trajectory\ length} = {\sum\sqrt{(x_{i + 1}}} - x_{i})^{2} + {(y_{i + 1} - y_{i})}^{2}$$
*HD* corresponds to the geometric analysis of the trajectory. Each point of the considered subject's trajectory is assigned to the closest point of the reference trajectory and conversely, each point of the reference trajectory is assigned to the closest point of the considered subject's trajectory ([Fig 1](#pone.0149757.g001){ref-type="fig"}). HD is the greatest of all the distances from a point in one set (A) to the closest point in the other set (B). HD is thus sensitive to corner points.
{#pone.0149757.g001}
HD was calculated with the following equation 3. $$HD\ \left( A,B \right) = {\max\ \left\{ d\left( {A,B} \right),d\left( {B,A} \right) \right\}}$$ where d(A,B) and d(B,A) are the direct (minimum) Euclidean distances between two sets, A and B \[[@pone.0149757.ref023]\].
The result is in cm. The greater the distance, the higher the deviation from the reference trajectory.
*DTW* is a spatio-temporal analysis which corresponds to the path of cumulative distances that minimize the warping cost (pair of matching points) of two time series, P and Q \[[@pone.0149757.ref042]\]. The algorithm first calculates the distance between each point of the subject's trajectory and reference trajectory and then searches an optimal matching (minimal cost) between sequence points (a point of a sequence is associated with one or more points of the other sequence) ([Fig 1](#pone.0149757.g001){ref-type="fig"}). DTW correspond to the optimal path that matches the point sequences.
DTW is calculated with the following equation 4. $$DTW\ \left( Q,P \right) = \min\left\lbrack {\sum\limits_{k = 1}^{k}{d\left( {q_{ik},p_{ik}} \right)}} \right\rbrack$$ where d(*q*~*ik*~, *p*~*ik*~) is the Euclidean distance between two points in the Q and P series \[[@pone.0149757.ref043]\]. The result is in arbitrary units. Higher values indicate a larger deviation from the reference trajectory.
HD and DTW are complementary parameters since HD relates to a particular point of the trajectory (the greatest of all the distances, for the sub-task analyzed) while DTW considers the trajectory as a whole (the sum corresponding to the optimal path between the two trajectories, for the sub-task analyzed).
All parameters were calculated for the global TUG and for each sub-task using Matlab (Mathworks, Inc.).
Subjects also underwent a clinical examination as detailed in [Table 3](#pone.0149757.t003){ref-type="table"}.
10.1371/journal.pone.0149757.t003
###### Clinical examination.
{#pone.0149757.t003g}
Impairments and disabilities examined Scale
---------------------------------------------------------------------- ------------------------------------------------
Spasticity (quadriceps, rectus femoris, hamstring and triceps surae) Modified Ashworth
Strength (hip, knee and ankle flexor and extensor) Medical Research Council
Sensation and proprioception of lower limb Nottingham Sensory Assessment
Activities of daily living Barthel index
Walking independence New Functional Ambulation Classification score
Balance Berg Balance Scale
Balance confidence Activities-specific Balance Confidence Scale
Statistical analysis {#sec009}
--------------------
Performance time, DTW and HD were calculated for each sub-task of the TUG (Go, Turn and Return) as well as the total trajectory. Trajectory length was computed for the total TUG trajectory. As the parameters were not all normally distributed, medians and quartile ranges are presented and non-parametric tests were used. Mann-Whitney tests were used to compare patients and healthy subjects, fallers and non-fallers and patients with right and left hemisphere stroke. A Bonferroni correction was used (since four repeated comparisons were carried out) with an adjusted p of 0.0125. A logistic regression was performed for each sub-task of the TUG to assess the additional variance of the dependent measure (stroke/no stroke) accounted for by DTW and HD above and beyond that accounted for by TUG time and nuisance variables (sex, age, body mass index). DTW and HD were added together in the regression model. Correlations between the BBS scores and trajectory parameters were tested with Spearman's correlation for both the patients with stroke and healthy subjects, and for each sub-task (p \< 0.05 was considered as significant). All analyses were performed using Statistica (version 7.1)
Results {#sec010}
=======
Comparison of trajectory parameters between stroke patients and healthy subjects {#sec011}
--------------------------------------------------------------------------------
Results of the trajectory parameters are presented in [Table 4](#pone.0149757.t004){ref-type="table"}. [Fig 2](#pone.0149757.g002){ref-type="fig"} shows the trajectories of a patient with stroke and a healthy subject. Trajectory length, HD and DTW of the total TUG test were significantly greater in the stroke group (respectively p = 0.000001, p = 0.0001, p = 0.00004). HD and DTW were significantly greater in the stroke group during the Go (respectively p = 0.00002, p = 0.0009) and Turn (respectively p = 0.0002, p = 0.000001) sub-tasks. Both HD and DTW were greater in the patient group showing that, for a given sub-task, they deviated from the reference trajectory both at an isolated point (assessed with HD) and during the entire sub-task (assessed with DTW).
{#pone.0149757.g002}
10.1371/journal.pone.0149757.t004
###### Trajectory parameters \[medians and interquartile ranges Q1;Q3\] during the global trajectory and Go, Turn and Return sub-tasks of the TUG for both groups.
{#pone.0149757.t004g}
Stroke group Healthy group
------------------------------- ----------------------- -------------------- -------------------- -------------------- ------------------------------------------------------------ --------------------------------------------------------- --------------------------------------------------------- --------------------
HD (cm) 29.3 \[21.9;33.3\] 22.6 \[17.1;28.5\] 33.0 \[25.4;42.2\] 28.4 \[22.1;37.4\] 19.2 \[17.5;23.9\][\*](#t004fn004){ref-type="table-fn"} 15.1 \[10.5;16.7\][\*](#t004fn004){ref-type="table-fn"} 20.4 \[18.7;26.8\][\*](#t004fn004){ref-type="table-fn"} 22.8 \[20.1;27.7\]
DTW(arbitrary unit) 12983 \[10576;19958\] 4438 \[3373;6139\] 5238 \[4344;7844\] 5298 \[3561;7745\] 9023 \[7522;10969\][\*](#t004fn004){ref-type="table-fn"} 3017 \[2187;3379\][\*](#t004fn004){ref-type="table-fn"} 2252 \[1875;2638\][\*](#t004fn004){ref-type="table-fn"} 4783 \[3631;6326\]
Trajectory length (cm) 838,5 \[817.7;864.5\] \- -. \- 750.1 \[737.7;766.1\][\*](#t004fn004){ref-type="table-fn"} \- \- \-
TUG performance (time in sec) 19.4 \[15.9;21.5\] \- \- \- 9.9 \[9.5;11.5\][\*](#t004fn004){ref-type="table-fn"} \- \- \-
HD Hausdorff distance
DTW Dynamic time warping
TUG Timed Up and Go
\* significant difference between Stroke group and Healthy group for the sub-task (p\<0.05)
Additional information provided by the trajectory parameters {#sec012}
------------------------------------------------------------
The logistic regressions showed that the variance increased for the Go sub-task when the trajectory parameters were included. Indeed when all variables were included in the model the R^2^ was 0.56 and when the trajectory variables were not included (model with time and nuisance variables) the R^2^ was 0.39. The results of the predictive factors of the logistic regression for Go are presented in the appendix ([S1 Table](#pone.0149757.s001){ref-type="supplementary-material"}). For the Turn and Return sub-tasks, the trajectory parameters did not provide additional information (not selected in the multivariate model, p\<0.05). [Fig 3](#pone.0149757.g003){ref-type="fig"} presents the trajectory of two characteristic patients with similar performance times but distinct trajectories, to illustrate the additional information provided by the locomotor trajectory parameters.
{#pone.0149757.g003}
Correlation between trajectory parameters and BBS score {#sec013}
-------------------------------------------------------
There was a significant negative correlation between BBS score and trajectory length, HD and DTW during the total TUG (r between -0.53 and -0.68, p\<0.05,). BBS score was also significantly correlated with DTW during the Turn (r = -0.6, p\<0.05) but not with HD during this sub-task. No correlations were found for Go and Return.
Comparison of trajectory parameters between fallers and non-fallers {#sec014}
-------------------------------------------------------------------
DTW was significantly greater for fallers (n = 8) than non-fallers (n = 21) for the Go sub-task only (p = 0.005), no differences were found for the Turn, Return or the total TUG. There were no significant differences between fallers and non-fallers for HD and trajectory length during the total TUG or each sub-task.
Comparison of trajectory parameters between patients with right and left hemisphere stroke {#sec015}
------------------------------------------------------------------------------------------
There were no differences for the DTW and HD for the Go, the Turn, the Return, the total TUG or for the total trajectory length between patients with right (n = 17) and left (n = 12) hemisphere stroke (p\>0.05).
Discussion {#sec016}
==========
To our knowledge, this study is the first to analyze locomotor trajectories during oriented-gait involving curved paths and obstacle circumvention in stroke patients. The aims were i) to analyze the locomotor trajectories of patients with stroke during the walking and turning sub-tasks of the TUG using HD and DTW, and to compare them with healthy subjects; ii) to determine whether trajectory parameters provide additional information to the conventional measure (performance time); iii) to compare the trajectory parameters of fallers and non-fallers with stroke and of patients with right and left hemisphere stroke and iv) to evaluate correlations between trajectory parameters and BBS scores.
The results showed that, compared to healthy subjects, stroke patients had significantly longer total trajectories and larger deviations from the reference trajectory during the oriented-gait to the cone (Go) and the turning (Turn) sub-tasks. Lamontagne et al (2010) recently also found different locomotor trajectories in stroke patients compared to healthy subjects during overground walking in an environment which provided optic flow \[[@pone.0149757.ref017]\].
The results of the present study suggest that stroke patients exhibit different locomotor trajectories depending on the requirements of the sub-task. Differences in trajectory parameters between the patients with stroke and the healthy subjects during the oriented gait to the cone and the turn sub-tasks suggest that the perception of a visual target, explicitly associated with a plan to circumnavigateit, impacted the gait trajectories of the patients with stroke for reasons that remain to be determined.
Furthermore, the results of this study suggest that the analysis of locomotor trajectories is an interesting approach to the analysis of gait in patients with stroke, providing additional information to that of the conventional timed performance of specific locomotor tasks. The assessment of trajectory parameters complements timed performance, providing a more complete understanding of locomotor tasks in patients with stroke. This is supported by the results of the logistic regression analysis. Further studies are needed to determine to what extent patients with similar performance times differ in locomotor trajectory, and the factors that influence these differences.
Longer and more deviated trajectories were significantly related to poor balance during the turn sub-task. Moreover, the trajectories of the faller group were significantly more deviated than those of the non-faller group during the oriented-gait to the cone (Go). The patients' gait parameters differed significantly from those of the healthy subjects during the oriented gait to the cone and the Turn. These sub-tasks both challenge stability. In contrast, the Return appeared to be less challenging since there were no significant differences between the parameters of the patients and healthy subjects, or of the fallers and non-fallers. Thus the Go and Turn appear to be the most challenging sub-tasks of the TUG test. Hicheur et al (2007) also found that "complex" locomotor trajectories (with a large turn amplitude) induce greater deviations from the mean than "simple" trajectories (with a smaller amplitude turn) in healthy subjects \[[@pone.0149757.ref016]\]. In the present study, the HD and DTW values of the stroke group were both greater than the values of the healthy subjects during the Go and Turn sub-tasks, revealing that deviations from the trajectory occurred throughout these sub-tasks and not only at an isolated point. It is possible that these larger deviations of trajectories throughout the obstacle circumvention task and the preceding phase compensate for instability. Our results are in accordance with these obtained in other patient groups. Older adults also increase the spatial margin when walking through apertures in comparison with young subjects \[[@pone.0149757.ref044]\]. Similarly, MacLellan and Patla (2006) showed that the locomotor trajectories of healthy subjects are modified proactively and retroactively when walking on a foam mat compared to overground. They suggested that these modifications of the locomotor strategy probably minimize threats to stability \[[@pone.0149757.ref045]\]. Maintaining a consistent but minimum spatial margin between an obstacle and the self has been suggested as one of the dominant control parameters to maintain balance and avoid perturbation \[[@pone.0149757.ref046]\]. However, the hypothesis that trajectory deviations could compensate for instability cannot be affirmed by our results and further studies will be necessary to confirm or infirm this.
Finally, we expected to find differences in the trajectories of patients with right and left hemisphere stroke since right hemisphere stroke may alter the perception of body verticality \[[@pone.0149757.ref018]\]. However, our results showed that there were no differences, suggesting either that there were no significant differences between our two groups of participants in the subjective vertical (which we did not measure) or that alterations in the subjective vertical did not affect the locomotor trajectories during the TUG test in this sample of patients with moderate to good recovery. Nevertheless this assumption should be tempered since the distribution of patients with right and left hemisphere strokes was slightly asymmetrical (twelve patients with left stroke and seventeen with right stroke).
Limits and perspectives {#sec017}
-----------------------
The patients included in this study had mild impairments; therefore caution must be taken regarding generalization of the results. The lack of difference between patients with right and left hemisphere stroke should also be interpreted with caution since we did not carry out a specific assessment of subjective vertical and cognitive functions relating to spatial perception (e.g. hemi-spatial neglect). Further studies designed to assess the influence of perception on trajectory would be interesting. The analysis of the trajectories of the faller patients was not our initial objective which explains why this sub-group was small. This limits the interpretation of the data for the discrimination of fallers and non-fallers, however these preliminary results suggest that the analysis of trajectory parameters may be a relevant approach to address this issue. Further studies specifically designed to fulfil this objective are nevertheless necessary. It would also be interesting to study whether locomotor trajectories are influenced by sensory perturbations in patients with stroke. Moreover locomotor trajectory analysis could be an interesting approach to assess the impact of medical treatment (such as botulinum toxin), surgical treatment or rehabilitation on "real-life gait" instead of conventional straight-line gait analysis.
Conclusion {#sec018}
==========
This study presents an innovative approach to the quantitative analysis of locomotor trajectories in patients with stroke during oriented-gait and obstacle circumvention, based on the widely used TUG test. This approach complements timed performance since it objectively quantifies locomotor trajectory and provides additional information regarding gait alterations in the presence of an obstacle. We evaluated parameters which quantified deviation from a reference trajectory and found that the trajectory of patients with stroke was more deviated than that of healthy subjects during the turn and the phase preceding the turn. No differences were found between patients with right and left hemisphere stroke. Comparison of faller and non-faller patients also showed that trajectory parameters differed during the phase preceding the turn. These results suggest that assessing the locomotor trajectory in addition to timed performance during complex locomotor tasks such as those assessed during the TUG test (i.e preparing to circumnavigate an obstacle and turning) might be relevant in patients with stroke and might also provide a basis for estimation of fall risk.
Supporting Information {#sec019}
======================
###### Logistic regression for the Go sub-task of the TUG: predictive factors.
Caption: 1\* reference value; OR odds ratio, CI confidence interval. NS non-significant
(DOCX)
######
Click here for additional data file.
The authors wish to thank all the participants for their kind participation. We would also like to thank Johanna Robertson for her constructive criticism and for correction of the English. Many thanks to Isabelle Vaugier and Ghilas Boussaid for their help with the statistical analysis. This work was sustained by Assistance Publique---Hopitaux de Paris, Centre Innovations Clinique Garches 1429, University of Versailles Saint Quentin en Yvelines and the Garches foundation.
[^1]: **Competing Interests:**The authors have declared that no competing interests exist.
[^2]: Conceived and designed the experiments: CB NR DP. Performed the experiments: CB. Analyzed the data: CB DP. Contributed reagents/materials/analysis tools: CB DP AVH. Wrote the paper: CB NR DP AVH DB. Designed the software used in analysis: DP.
[^3]: ‡ These authors also contributed equally to this work.
| {
"pile_set_name": "PubMed Central"
} |
Introduction
============
Grain legumes and pulses in countries such as India, Thailand, Bangladesh, Sri Lanka, and Pakistan are characteristically affected by yellow mosaic disease (YMD) caused by *Mungbean yellow mosaic virus* (MYMV), *Mungbean yellow mosaic India virus* (MYMIV), *Dolichos yellow mosaic virus* and *Horsegram yellow mosaic virus*. These viruses are closely related and have distinct but overlapping host ranges. *Mungbean yellow mosaic virus* (MYMV) and *Mungbean yellow mosaic India virus* (MYMIV) occur across the Indian subcontinent affecting the majority of legume crops including blackgram (*Vigna mungo*), cowpea (*Vigna unguiculata*), dolichos (*Lablab purpureus*), horsegram (*Macrotyloma uniflorum*), lima bean (*Phaseolus lunatus*), mungbean (*Vigna radiata*), pigeon pea (*Cajanus cajan*), mothbean (*Vigna aconitifolia*), common bean (*Phaseolus vulgaris*) and soybean (*Glycine max*) ([@b17-67_16115], [@b25-67_16115]). YMD of the leguminous crops causes an estimated annual loss of US\$300 million ([@b25-67_16115]). Among legumes, soybean is an economically important crop in which YMD causes 15--75% yield loss ([@b19-67_16115]). Nucleotide sequence of the virus isolated from soybean plants in northern and central India affected by YMD showed 89% similarity with *Mungbean Yellow Mosaic India Virus* (MYMIV) and was designated as soybean isolate of MYMIV (MYMIV-\[Sb\]) by [@b23-67_16115]. This virus is transmitted by the white fly, *Bemisia tabaci* Genn. ([@b14-67_16115], [@b15-67_16115], [@b16-67_16115]) and possesses bipartite, single stranded, circular DNA genome referred as DNA A and DNA B ([@b11-67_16115]). Both the genomes encode necessary components for replication, movement and symptom development and are of 2.5--2.7 kb in size ([@b4-67_16115], [@b5-67_16115], [@b11-67_16115]). MYMIV produces typical yellow and golden mosaic patterns on the leaves of affected plants. Initially symptoms appear as small yellow specks along the veins and then spread over the leaf. In severe infections the entire leaf may become chlorotic. Since 1970s, MYMIV is posing a major threat to Indian soybean cultivation and it is reported to spread throughout India in alarming proportions ([@b26-67_16115]).
Soybean in India has become a leading oilseed crop with 41.5% and 28.6% contribution towards total oilseeds and edible oil production in the country during triennium average ending 2013--14. Beside contribution to edible oil pool, the crop is earning huge foreign exchange through export of soy meal, which has uplifted the rural economy of central India. However, the productivity of the crop which hovers around 1.2 tonne per ha in India is the major concern compared to the world average of 2.5 tonne per ha. YMD caused by MYMIV is one of the major constraints in enhancing the yield of soybean crop in India. None of the dominant varieties of central India, hub of soybean cultivation, is resistant to this virus. Therefore, it is imperative to introgress MYMIV resistance gene in elite soybean varieties with durable tolerance to MYMIV.
Various efforts have been made to understand the mechanism of natural resistance and nature of resistant gene in resistant soybean varieties. [@b29-67_16115] compared the abundance of the viral RNAs in a resistant and a susceptible variety at the early time points after agro infection. Whilst the resistant variety displayed synthesis but rapid degradation of the early viral RNAs; the degradation in the susceptible variety was delayed resulting in accumulation of those transcripts later in infection. Accumulation of the late viral transcripts and DNA replication were detectable only in the susceptible variety that indicates rapid degradation of the early viral transcripts possibly through siRNA mechanism, is one of the probable mechanisms of natural resistance against geminivirus.
There are contradictory reports on the genetic nature of *Yellow mosaic virus* resistance. It is reported to be controlled by double recessive genes in PI171443 by [@b20-67_16115] and a single dominant gene by [@b22-67_16115]. It has become imperative to find out the true nature of MYMIV resistance in soybean. Moreover, segregating material, generated for the introgression of MYMIV resistance gene in high yielding and adapted varieties, has to be screened at hot spots or under artificial conditions for development of MYMIV resistant varieties. Identification of DNA markers linked to the trait will obviate the need of screening the segregating material at hot spot or under artificial conditions. If the trait is governed by recessive gene, selection for the trait in traditional breeding methods has to be deferred because of absence of expression of the trait in heterozygous condition. Marker Assisted Selection using linked marker will help in accelerated introgression of MYMIV resistance gene in dominating but susceptible varieties of central India. Attempts have been made to identify DNA markers linked to this trait. [@b28-67_16115] did whole genome sequencing of MYMIV susceptible variety JS335 and resistant genotype UPSM534 (PI171443) to find out the genomic regions associated with resistance gene. They indicated a Single Nucleotide Polymorphism (SNP) on chr 18 (LG G) with a possible association with MYMIV resistance gene. [@b9-67_16115] reported a region on chr 17 (LG D2) in significant linkage disequilibrium with resistance gene in association mapping study. But none of the reported linkage has been validated in mapping population. Efforts were made in this study to investigate the true nature of MYMIV resistance and to map the resistance gene to find the molecular markers linked to the trait.
Materials and Methods
=====================
Development of mapping populations
----------------------------------
Two susceptible soybean genotypes viz; JS335, a dominant variety of India and NRC101, a newly developed Kunitz trypsin inhibitor free soybean genotype and two resistant genotypes viz; PI171443, donor of MYMIV resistance gene in most of the MYMIV resistant varieties released in India and SL525, a variety released for northern India, which has MYMIV resistance gene from PI171443 were used to develop mapping populations. JS335 was crossed with PI171443 to develop Recombinant Inbred Lines (RILs). F2 mapping population was developed from crossing of SL525 and NRC101. F1s & F2s of SL525 × NRC101 and advancement of RILs of JS335 × PI171443 were raised at Indore in disease free condition for development of mapping populations. RILs and F3 progeny rows were used for phenotyping. Fresh crosses were made for production of F1 seeds for phenotyping.
Phenotyping for reaction to MYMIV
---------------------------------
RILs derived from JS335 × PI171443 and F3 progeny rows of SL525 × NRC101 along with 50 F1 plants of each population were raised at Ludhiana, hot spot for YMD, along with infector rows of susceptible genotype after every two rows. A vast survey made by Indian Council of Agriculture Research-Indian Institute of Soybean Research from different locations of India has proved that MYMIV is the prevalent virus infecting soybean at Ludhiana based on PCR based assay ([@b18-67_16115]). Mapping populations were sown in June 2015. Fifty to sixty plants were raised per progeny row with 45 cm & 5 cm row to row and plant to plant distance respectively. A total of 98 F3 progeny rows, derived from the cross SL525 × NRC101 and 89 progeny rows of RILs derived from JS335 × PI171443 were tested for their reaction to MYMIV. Phenotyping was done at R5 stage when YMD was most severe. Severity of infection was scored from 0 to 9 scales. Progeny rows, which did not show any symptom in any of the plant were given a score of 0, progeny rows with any plant showing infection symptoms in 10% of the leaves were given a score of 1 and progeny rows with any of the plant showing infection symptoms in 20% of the leaves were given a score of 2 and likewise. Progeny rows with all the plants affected with a score more than 3 were classified as susceptible and those showing symptoms up to scale of 2, were classified as resistant. Progeny rows with both types of plants were classified as segregating.
DNA isolation and Molecular marker analysis
-------------------------------------------
Genomic DNA was extracted from individual F2 plants of the cross SL525 × NRC101 and bulked DNA was extracted from RILs along with their parental genotypes following cetyl trimethyl ammonium bromide procedure ([@b2-67_16115]). Parental polymorphism survey was done using 10 to 15 SSR primer pairs in such a way to get at least 6 regularly spaced polymorphic primer pair from each linkage group for each parental combination. A total of 144 regularly spaced polymorphic SSR markers from twenty linkage groups (LG) of soybean genome were used for the analysis. SSR marker sequences and their genetic position were taken from integrated linkage map published by [@b7-67_16115].
Quantified DNA was subjected to PCR amplification in 10 μl reaction mixture containing 2 μl DNA (25 ng/μl), 1 μl PCR 10× buffer, 1.1 μl MgCl~2~ (25 mM), 0.1 μl dNTPs (25 mM), 0.4 μl each forward and reverse SSR primers (30 ng/μl), 0.068 μl *Taq* DNA polymerase (3 units/μl) and 4.932 μl distilled water. DNA was denatured at 94°C for 2 min followed by 30 cycles each consisting of denaturation at 94°C for 1 min, primer annealing at 50°C for 2 min, primer elongation at 72°C for 3 min and final elongation at 72°C for 10 min in the thermocycler (MJ Research, model PTC100). Amplified products so obtained were resolved on 3% metaphore agarose gel with 50 bps ladder for allele sizing and analyzed in gel documentation system (Syngene). The SSR bands were scored manually from gel images. The SSR polymorphic between two parents were denoted S, R or H, where S band from susceptible parent only, R band from resistant parent only and H band from both the parents.
Bulked segregant analysis
-------------------------
The bulked segregant analysis (BSA) was performed following the protocol described by [@b13-67_16115]. Resistant bulk were formed by mixing equal amount of DNA from 10 progeny rows showing 0 score and susceptible bulk were formed in the similar way from 10 susceptible progeny rows showing 9 score in RILs. In F2 population derived from SL525 × NRC101, equal amount of DNA from 10 resistant homozygous plants was mixed to make resistant bulk and 10 susceptible homozygous plants was mixed to make susceptible bulk. The resistant and susceptible DNA bulks along with DNA from their parents were amplified using regularly spaced 6 to 8 polymorphic SSR primers from each linkage group to identify SSR markers linked to MYMIV resistance gene.
Data analysis and genetic mapping
---------------------------------
The SSR marker genotyping data and phenotyping data with respect to reaction to MYMIV were analyzed to construct genetic linkage map on chr 6 using Joinmap 4.0 ([@b24-67_16115]) using the Kosambi mapping function. A logarithm of odds (LOD) score of \>3 was used to identify linked loci. At each locus, segregation of allele ratio was determined by X^2^ goodness of fit to identify if the locus met the expected 1:1 or 1:2:1 ratio with a significance threshold of P = 0.05.
Results
=======
Phenotyping of the mapping populations for MYMIV reaction
---------------------------------------------------------
The parental lines involved in the crosses, their F1's, F3 progeny rows and RILs were screened for MYMIV reaction under field epiphytotic conditions with abundant white fly population and infector rows of susceptible variety. The susceptible parent genotypes viz; JS335 and NRC101 showed susceptible reaction to MYMIV with a score of \>8 while resistant parent genotypes showed resistance reaction with a MYMIV reaction score \<2. The F1 plants of both the crosses showed susceptible reaction with a score \>8 indicating the recessive nature of MYMIV resistance. A total of 98 F3 progeny rows, derived from the cross SL525 × NRC101 and 89 progeny rows of RILs derived from JS335 × PI171443 were tested for their reaction to MYMIV ([Table 1](#t1-67_16115){ref-type="table"}). The observed segregation pattern fits almost perfectly into 1 resistant: 2 segregating: 1 susceptible for F3 progeny rows, which indicates the presence of a single recessive gene in the MYMIV resistant variety SL525. Progeny rows of RILs derived from JS335 × PI171443, the donor of MYMIV resistance gene in most of the released varieties, segregated in the ratio of 1:1 again proving monogenic nature of resistance gene.
Genetic map construction
------------------------
DNA from resistant and susceptible bulks of F~2~ homozygous plants of SL525 × NRC101, and that of RILs developed from JS335 × PI171443 selected based on scoring of progeny rows along with their parents were screened with 144 regularly spaced polymorphic SSR markers from twenty linkage groups (LG) of soybean genome. The list of 144 polymorphic SSR markers, their chromosomal location, genetic positions within the chromosome and amplicon size of parental lines are presented in [Supplemental Tables 1, 2](#s1-67_16115){ref-type="supplementary-material"}. BSA identified two SSR markers, namely GMAC7L (position of 12,259,594--12,259,701 bp) and Satt322 (position of 12,336,492--12,336,709 bp), located on chr 6 (LG C2) with the genetic position of 69.68 cM and 73.18 cM respectively linked to MYMIV resistance in RILs as well as F2 mapping population. These two markers were used to genotype both the mapping populations ([Fig. 1A, 1B](#f1-67_16115){ref-type="fig"}) and all the markers within distance of 35 cM on both sides of the linked markers were also studied for parental polymorphism survey. A total of 5 SSR markers viz; Satt281, Sat_153, Satt305, Satt170 and Sat_213 other than the two linked markers identified were found to be polymorphic for RILs mapping population and 6 SSR markers viz; Satt281, Satt305, Satt170, Sat_246, Satt363 and Satt376 were found to be polymorphic for F2 mapping population. All the F2 plants and RILs were screened with these polymorphic markers. All the markers except Satt281 segregated in expected ratio of 1:1 in RILs and all markers other than Satt363 segregated in expected ratio of 1:2:1 in F~2~ mapping population. The genetic distance mapped between Sat_153 and Sat_213 in RILs is 49.7 cM as compared to genetic distance of 27.4 cM given in Soybase ([@b3-67_16115]). The genetic distance mapped between Satt281 and Satt376 is 52.2 almost same as given in Soybase (51.9 cM) ([Fig. 2](#f2-67_16115){ref-type="fig"}).
Discussion
==========
In India, breeders have developed MYMIV resistant soybean varieties through classical breeding. All the MYMIV resistant varieties till now are adapted for northern India as due to heavy infection pressure of MYMIV, nothing susceptible to the virus survives in this part of India. So, there is natural selection for MYMIV resistance while making selection in breeding populations. Lately the virus has spread to central India, which is hub of soybean cultivation. It has created havoc in the present year and is the main reason behind low productivity of soybean in the present year. Unfortunately none of the dominant variety of central India is resistant to this virus. Introgression of MYMIV resistance gene in dominant varieties is the immediate challenge. For speedy introgression of the trait, understanding the nature of genetic inheritance due to contradictory reports on its inheritance and finding molecular markers linked to the gene was the immediate challenge. First report on inheritance of the trait reported that the trait is controlled by double recessive gene ([@b20-67_16115]). Our results also prove the recessive nature of the trait, but segregation ratio obtained in our study has proven it to be monogenic. The discrepancy may result from the classification of resistant and susceptible classes. We classified the progeny rows with a MYMIV reaction score of 0 to 2 as resistant progeny rows as we observed similar score in resistant parental lines also, while the earlier group classified only those F2 plants as resistant which showed complete immune response. [@b22-67_16115] has reported dominant and monogenic nature of genetic inheritance of the trait in two resistant varieties DS9712 and DS9814. The result obtained in their study may either be due to escapes of susceptible plants due to low disease pressure or a totally different gene as ancestry of these two varieties could not be traced to known source of YMV resistance.
Efforts have been made to locate the gene on a linkage map by two groups. [@b28-67_16115] did whole genome sequencing of MYMIV susceptible variety JS335 and resistant genotype UPSM534 (PI171443) to find out the genomic regions associated with resistance gene. They indicated a SNP on chr 18 (LG G) with a possible association with MYMIV resistance gene. [@b9-67_16115] found a region on chr 17 (LG D2) in significant linkage disequilibrium with resistance gene in association mapping study. But none of the reported linkage has been validated in mapping population. We studied all the linkage maps including these two regions. We did not find any association between these genetic regions and the location of resistance gene. Association studies are known to lead to spurious associations ([@b10-67_16115]) and recent attempts to map such traits have resulted in extremely high false positive rates ([@b1-67_16115]). In our study, we have mapped MYMIV resistance gene on chr 6 (LG C2) very close to two SSR markers Satt322 and GMAC7L. The total map distance of the linkage map constructed in this study in F2 mapping population is 52.2 cM, which is very similar to map distance given in Soybase. However map distance of 49.7 cM studied in RILs mapping population is much larger as compared to 27.4 cM given in Soybase. Map positions and distances between loci on a linkage group may vary because of deletions, insertions, inversions, translocation and other chromosomal arrangement in one or both of the parents that could change distance between loci as well as their relative orders ([@b21-67_16115]).
There are many viruses that affect soybean, and resistance genes against several viruses have been extensively studied. A lot of work has been done on mapping of *soybean mosaic virus* resistance genes in soybean. Rsv1 locus was mapped on chr 13 (LG F) ([@b31-67_16115], [@b32-67_16115]), Rsv3 locus on chr 14 (LG B2) ([@b8-67_16115]) and another gene, Rsv4 locus was mapped on chr 2 (LG D1b) ([@b6-67_16115]). Six linked resistance genes, Ra, Rsc-7, Rsc8, Rsc-9, Rn1, and Rn3, were mapped on chr 7 (LG M) ([@b27-67_16115]). Rsc-11 was mapped on chr 13 (LG F) by [@b12-67_16115]. The major *Soybean dwarf virus* resistance gene *Rsdv1* was mapped on chr 5 (LG A1) ([@b30-67_16115]). The nucleotide-binding site leucien-rich repeat type resistance genes for virus have been mapped on chr 2, 8, 9 & 18 in the Phytozome database (<https://phytozome.jgi.doe.gov/pz/portal.html>). No virus resistance gene has been mapped on chr 6 (LG C2) till now.
Development of an MYMIV resistant soybean variety for central India, hub of soybean cultivation, was hampered due to difficulty in creating artificial epiphytotics for yellow mosaic disease. The only alternative available with breeder was to screen segregating populations in hotspot in main soybean crop season in the absence of reliable of molecular marker. The tightly linked marker identified in this study, will help breeder in screening MYMIV resistant soybean plants in segregating populations at their own station and selection can be carried out in off season also thus aiding in the accelerated development of MYMIV resistant cultivars in relatively shorter time span.
Supplementary Information
=========================
Authors are thankful to Director, ICAR- Indian Institute of Soybean Research, Indore for providing infrastructure to carry out the work.
{#f1-67_16115}
{#f2-67_16115}
######
χ^2^ test for segregation of MYMIV resistance gene
Type of mapping population Cross combination No. of plants MYMIV reaction[a](#tfn1-67_16115){ref-type="table-fn"} Expected ratio χ^2^ P value
---------------------------- ------------------- --------------- -------------------------------------------------------- ---------------- ------ --------- ------ ------
RIL JS335XPI171443 89 43 1 45 1:1 .101 0.75
F~2:3~ SL525XNRC101 98 23 50 25 1:2:1 .186 0.90
S denotes susceptible progeny rows, Seg denotes progenies segregating for susceptible and resistant plants and R denotes resistant progeny rows.
[^1]: Communicated by Toyoaki Anai
| {
"pile_set_name": "PubMed Central"
} |
Published: February 9, 2017
Results and Discussion {#sec1}
======================
Pharyngeal gills are a defining feature of vertebrate animals and are present as vestiges in our own embryology. However, it has been proposed that the gills of cyclostomes and gnathostomes evolved independently. This "ecto-endobranchiate hypothesis" ([Figure 1](#fig1){ref-type="fig"}) extends from the observation that the gills of cyclostomes and gnathostomes form on different regions of the branchial arches (medial versus lateral to the gill endoskeleton, respectively) and arise from distinct embryonic epithelia (endodermal versus ectodermal, respectively) \[[@bib2], [@bib3], [@bib4], [@bib5], [@bib6], [@bib7], [@bib8]\]. While the topological incongruence of gills and their skeletal support is not generally regarded as an insurmountable obstacle to homology \[[@bib7], [@bib8], [@bib14]\], the issue of distinct embryonic origins is potentially more problematic and remains unresolved.
In fishes, gills develop on pharyngeal arches, paired columns of tissue that are bound by ectodermal and endodermal epithelia and form from the walls of the embryonic foregut \[[@bib15]\]. Pharyngeal arch development begins with the iterative outpocketing of foregut endoderm in a rostral-to-caudal sequence, giving rise to a series of endodermal pouches \[[@bib16]\]. These pouches subsequently contact and fuse with overlying surface ectoderm, resulting in the perforation of gill slits and the delineation of arches ([Figure 2](#fig2){ref-type="fig"}A). The endodermal origin of gills in lamprey and hagfish is well established from classical histological studies \[[@bib12]\], but the putative ectodermal origin of gills in bony fishes is less well documented. Goette described the early development of gills in sturgeon and noted that gill filaments arise from pharyngeal ectoderm, appearing first on the outside of the pharynx, prior to perforation of the gill slits \[[@bib10]\]. Kellicott also proposed an ectodermal origin of the gills in lungfish, noting an inward migration of ectodermal epithelium following gill slit perforation but prior to gill filament differentiation \[[@bib13]\]. However, cell lineage tracing experiments have demonstrated an endodermal origin of gills in zebrafish \[[@bib17]\]. Thus, it remains unclear whether the gills of bony fishes primitively developed from endoderm or ectoderm. We therefore sought to investigate the embryonic origin of the gills in a cartilaginous fish, the little skate, *Leucoraja erinacea*---an outgroup to the bony fishes, which may permit the inference of primitive anatomical and developmental conditions in the last common ancestor of jawed vertebrates.
In the stage 22 (S22) skate embryo \[[@bib18]\], the gene encoding the developmental signaling molecule Sonic hedgehog (Shh) is expressed in the anterior endodermal domain of each developing pharyngeal pouch. Once the pouches have fused with the overlying ectoderm, this *Shh* expression domain comprises the posterior epithelium of each pharyngeal arch, where it functions to establish the anterior-posterior axis of the arch \[[@bib19]\]. We have noted that the first gill buds, which differentiate from pharyngeal arch epithelium shortly after gill slit perforation, form within this *Shh* expression domain ([Figure 2](#fig2){ref-type="fig"}B). Given the endodermal origin of this *Shh* expression domain, this suggests that these gill buds are endodermally derived ([Figure 2](#fig2){ref-type="fig"}C), though we cannot rule out that the pharyngeal *Shh* expression domain expands to include adjacent pharyngeal ectoderm upon gill slit perforation.
To directly test the embryonic origin of the gills in skate, we conducted a series of pharyngeal endodermal fate mapping experiments. We microinjected the lipophilic dye CM-DiI into the pharyngeal cavity of skate embryos at S18 (prior to the perforation of gill slits) ([Figure 3](#fig3){ref-type="fig"}A), and histological examination of labeled embryos 1 day post injection (dpi) (n = 3) revealed that this strategy allows us to focally label regions of pharyngeal endoderm without contaminating overlying ectoderm ([Figure 3](#fig3){ref-type="fig"}B). An analysis of labeled embryos at S22 (∼4--7 dpi; n = 3) allowed us to visualize the endodermal contributions to the pharyngeal arches immediately prior to ([Figure 3](#fig3){ref-type="fig"}C) and following ([Figure 3](#fig3){ref-type="fig"}D) gill slit perforation. Upon their initial delineation by adjacent gill slits, approximately 3/4 of the epithelium surrounding a pharyngeal arch derives from endoderm, with only the lateralmost 1/4 deriving from ectoderm. By S27/28, differentiated gill filaments are present on the hyoid and gill arches. Elasmobranch embryos possess both internal and transient external gill filaments, with the latter ultimately resorbing and remodeling into internal filaments \[[@bib20]\]. In labeled embryos reared to S27/28 (∼40--50 dpi; n = 7), CM-DiI-positive external gill filaments were readily apparent ([Figure 3](#fig3){ref-type="fig"}E), while histological analyses revealed CM-DiI-positive epithelium on both the external ([Figure 3](#fig3){ref-type="fig"}F) and internal gill filaments ([Figures 3](#fig3){ref-type="fig"}G and 3H). These experiments, which allow us to directly trace derivatives of pharyngeal endodermal epithelium, demonstrate the endodermal origin of gills in the little skate.
Conclusions {#sec1.1}
-----------
While the evolutionary origin of pharyngeal arches has been resolved to the deuterostome stem \[[@bib21], [@bib22], [@bib23]\], the evolutionary history of gills derived from pharyngeal arch epithelia remains contentious. Gill structures are preserved in the stem vertebrates *Myllogkunmingia* \[[@bib24]\], *Metaspriggina* \[[@bib25]\], and *Haikouichthys* \[[@bib26]\], but in the absence of consensus regarding the homology of gills within the crown group, it has remained unclear to what extent these structures may inform the nature of gills in the last common ancestor of vertebrates. Our demonstration of an endodermal origin of gills in a cartilaginous fish bolsters the homology of gills in cyclostomes and gnathostomes, and the single origin of pharyngeal gills prior to the divergence of these two ancient vertebrate lineages. Our findings, along with recent phylogenetic \[[@bib27]\] and paleontological \[[@bib25]\] advances, contribute to an emerging picture in which the crown ancestor of vertebrates was more complex---and exhibited more gnathostome-like anatomical conditions---than was previously appreciated, and are consistent with scenarios in which gills evolved along the vertebrate stem, in conjunction with a more active lifestyle \[[@bib1]\].
Experimental Procedures {#sec2}
=======================
Embryo Collection {#sec2.1}
-----------------
*L. erinacea* eggs were obtained at the Marine Biological Laboratory (Woods Hole, Massachusetts, USA) and maintained in a flow-through seawater system at ∼17°C to the desired developmental stage. Embryos were fixed in 4% paraformaldehyde in PBS overnight at 4°C, rinsed three times in PBS, dehydrated into 100% methanol, and stored at −20°C. All animal work complied with protocols approved by the Institutional Animal Care and Use Committee at the Marine Biological Laboratory.
Histology and mRNA In Situ Hybridization {#sec2.2}
----------------------------------------
For histological analysis, embryos were embedded in paraffin wax and sectioned at 7 μm as previously described \[[@bib28]\]. Sections of CM-DiI-labeled embryos were rehydrated and coverslipped with Fluoromount G containing DAPI (Southern Biotech). Sections were subsequently decoverslipped in water and stained with hematoxylin and eosin. In situ hybridization experiments for *L. erinacea Shh* (GenBank: [EF100667](ncbi-n:EF100667){#intref0010}) were performed on paraffin sections as previously described \[[@bib28]\].
Fate Mapping {#sec2.3}
------------
For CM-DiI fate mapping experiments, eggs containing embryos at S18 were windowed, and CellTracker CM-DiI (ThermoFisher Scientific; 0.5 μg/μL, prepared by diluting a 5 μg/μL ethanol stock 1:10 in 0.3M sucrose) was microinjected into the pharyngeal cavity using a pulled glass capillary needle and a Picospritzer pressure injector. Eggs were then sealed with a piece of donor eggshell and Krazy Glue and were left to develop in a flow-through seawater table for 1--50 days prior to fixation.
Author Contributions {#sec3}
====================
J.A.G. conceived the study and conducted and analyzed the cell lineage tracing experiments. J.A.G. and O.R.A.T. conducted and analyzed the mRNA in situ hybridization experiments. J.A.G. wrote the manuscript and prepared the figures with input from O.R.A.T.
This research was supported by a Royal Society University Research Fellowship (UF130182) and a grant from the University of Cambridge Isaac Newton Trust (14.23z) to J.A.G. O.R.A.T. was supported by the Wellcome Trust (PhD studentship 109147/Z/15/Z) and the Cambridge Commonwealth Trust. The authors acknowledge the support of Dr. Kate Rawlinson, Prof. Richard Behringer, Prof. Alejandro Sánchez Alvarado, Prof. Jonathan Henry, the MBL embryology community, and the staff of the MBL Marine Resources Center.
![The Ecto-endobranchiate Hypothesis\
The independent evolution of gills in cyclostomes and gnathostomes (from a gill-less common ancestor), based on their distinct embryonic origins from endoderm and ectoderm, respectively. Redrawn after Jarvik \[[@bib3]\] and Jollie \[[@bib5]\].](gr1){#fig1}
{#fig2}
{#fig3}
[^1]: Lead Contact
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#s1}
============
Adverse birth outcomes such as low birth weight, preterm birth and low Apgar scores are important predictors of infant survival and of long term health, including neurodevelopmental, emotional, and cognitive functioning and educational outcome \[[@B1]-[@B3]\].
Adverse birth outcomes have been linked to common mental disorders (CMD) in pregnant women. On average at least one in six pregnant women will experience CMD, especially depression and anxiety \[[@B4]-[@B6]\] in antenatal clinics worldwide. Maternal depression and anxiety were found to result in poor self-care and unhealthy behaviors \[[@B7]\], abnormal stress levels and hormonal alterations \[[@B8]\], and impaired immune function \[[@B9]\], all of which may directly or indirectly lead to poor birth outcomes. Important adverse birth outcomes are low birth weight (LBW), preterm birth (PTB), and low Apgar score.
Evidence linking maternal CMD symptoms to adverse birth outcomes is inconsistent and inconclusive. Studies in low-income countries, mainly from South-East Asia, found larger effects of maternal mental health on adverse birth outcomes than studies in high-income countries \[[@B10]-[@B15]\]. In high-income countries, depressed women appeared to have a higher risk of PTB than non-depressed women only if their socio-economic status (SES) was low \[[@B10]\]. In sub-Saharan Africa where stressors due to poverty and maternal ill-health prevail, CMD during pregnancy may be associated with particularly poor perinatal outcomes. Yet, the only study identified from this region did not confirm this assumption in Ethiopian women \[[@B16]\]. Overall, the comparability of results is complicated by different study designs, verification of depression and anxiety, timing of assessment and adjustment for confounders like SES.
The aim of this study was to explore the association between women's depression and anxiety symptoms during the third trimester of pregnancy and birth outcomes in a sample of African mother/child dyads from the Child Development Study (CDS) in Ghana and Côte d'Ivoire \[[@B17]\]. We recently reported a prevalence of antepartum depression of 27% and 33%, and anxiety symptoms of 11% and 17% in this sample, respectively \[[@B17]\]. Antepartum and postpartum depression was associated with the febrile illness in their offspring \[[@B18]\]. To evaluate the unique contribution of maternal depression and anxiety on birth outcomes, maternal risk factors known to be associated with impaired pregnancy course and sequelae for the neonate were eliminated by restriction to a low-obstetric risk sample \[[@B19],[@B20]\]. We hypothesized to find more unfavorable birth outcomes in children born to mothers who scored high on the general depression and anxiety symptom scales during late pregnancy.
Methods {#s2}
=======
Study design and setting {#s2.1}
------------------------
CDS is a longitudinal birth cohort of mothers and children in Ghana and Côte d'Ivoire. The primary aim of the study is to investigate the effect of maternal infectious and non-infectious disease on child development. Details of the study design have been described elsewhere \[[@B17]\]. In brief, women in their last trimester of pregnancy were consecutively recruited in two large hospitals, the Komfo Anokye Teaching Hospital in Kumasi, Ghana, and the Abobo Community Hospital in Abidjan, Côte d'Ivoire during antenatal care visits between March 2010 and December 2011. While the Komfo Anokye Teaching Hospital serves a mixed population in Ghana's second largest city, the Abobo Community Hospital provides healthcare to an underprivileged population affected by civil war. After birth, the children of included mothers were enrolled into a prospective, longitudinal birth cohort for a two-year follow-up period. Here, we report data on the association between depression and anxiety status during the third trimester of pregnancy and PTB, LBW, and low Apgar scores in newborn children to these mothers.
Ethics statement {#s2.2}
----------------
This study was conducted in accordance with the ethical principles of the Declaration of Helsinki. It was approved by the ethical committee of the Kwame Nkrumah University of Science and Technology in Kumasi, Ghana, the National Ethical Committee in Côte D'Ivoire, and the respective committee of the Chamber of Physicians in Hamburg, Germany. Generally, persons suffering from non-psychotic depression are able to understand and consent to study requirements. All participating mothers gave written informed consent.
Participants and procedures {#s2.3}
---------------------------
All women in their third trimester of pregnancy, based on gestational age assessed at the antenatal clinic, residing within a distance of ≤ 5 km around the hospitals were eligible for participation. Women under 18 years of age, with multiple pregnancies, chronic physical disease, or other risk factors like severe complications during pregnancy due to diabetes, hemorrhage, hypertension, and preeclampsia were excluded from the study (restriction to a low-obstetric risk cohort) \[[@B20]\]. The study flow is displayed in [Figure 1](#pone-0080711-g001){ref-type="fig"}. Due to political instability and resumed fighting in Côte d'Ivoire during the study period, 311 mothers did not give birth in the designated hospitals and could therefore not be followed-up. The missing mothers were, on average, more depressed, more anxious, younger, and of lower SES. 719 singletons and live births were included in the analysis. Women who scored high in the PHQ-9 and the GAD-7 were referred to local mental health professionals for further assessment and consultation. No psychotropic medications were prescribed.
{#pone-0080711-g001}
Measurements {#s2.4}
------------
### Assessment of birth outcomes {#s2.4.1}
Children's weight was measured at birth. LBW was defined as a birth weight of less than 2500 g. Gestational age was estimated according to second trimester ultrasound screening as documented in the hospital charts. PTB was defined as gestational age less than 37 0/7 weeks. The Apgar score was assessed by a trained midwife 5 minutes after birth. A low Apgar score was defined as less than 7 at 5 minutes.
### Assessment of antepartum depression and anxiety symptoms {#s2.4.2}
The explanatory variables were antepartum maternal depression and anxiety. Depression was screened for with the Patient Health Questionnaire depression module (PHQ-9) \[[@B21]\]. The PHQ-9 refers to the past two weeks and assesses the presence and severity of the nine DSM-IV depression criteria, comprising emotional, cognitive, and functional somatic symptoms \[[@B22]\]. Response options generate a continuous raw score ranging from 0 (no symptoms) to 27 (all symptoms present nearly every day); with scores 5-9 representing mild, 10--14 moderate and 15--27 moderately severe to severe depression symptoms. The PHQ-9 has been validated for use in primary care and in the general population in both high-income and low-income settings. In rural postpartum Ghanaian women, the PHQ-9 proved superior to other common depression screening measures against a semi-structured clinical interview \[[@B23]\] as reference standard. A threshold score of ≥ 10 had a sensitivity of 88% and a specificity of 88% for major depression \[[@B24]\] and was used for case classification. The reliability of the raw score was Cronbachs α = .69 for the Ghanaian women and .64 for the women from Côte d'Ivoire. Anxiety was assessed using the GAD-7, a screening questionnaire for generalized anxiety disorder (GAD) according to seven DSM-IV symptoms \[[@B25]\]. The GAD-7 has a response set similar to the PHQ-9, comprising emotional and cognitive symptoms of anxiety during the past two weeks. Item scores range from 0 to 21, with 5 - 9 representing mild, 10 - 14 moderate, and 15 - 21 severe levels of anxiety. The threshold score of ≥10 had a sensitivity of 89% and a specificity of 82% for a generalized anxiety disorder and was used for case classification. The GAD-7 has been validated in western primary care, but not in antepartum populations or in African settings, as research in antepartum anxiety is still in its early stages \[[@B26],[@B27]\]. In our study, the reliability of the raw score was Cronbachs α = .68 in both countries. Both depression assessed by the PHQ-9 and anxiety measured by the GAD-7 do not refer to a clinical diagnosis but to the result of a robust screening procedure in an epidemiological study with the above-mentioned properties.
### Assessment of socioeconomic, anthropometric, and obstetric characteristics of the women {#s2.4.3}
Other variables known to be associated with birth outcomes including SES, infant sex, previous pregnancy complications, maternal height, age, infections, hemoglobin level, malaria treatment, and caesarian section were controlled for. Maternal smoking and drinking status were not collected because both risk factors are known to be very rare among women in the two countries \[[@B28],[@B29]\]. Marital status was not included into our analyses because the study population was homogenously married. An SES score was constructed using a principal component analysis based on ownership of refrigerator, car, and bed net; toilet type, maternal education, and spouse education. Women's hemoglobin level and height were measured using standard techniques.
Statistical analysis {#s2.5}
--------------------
We first assessed correlations between birth weight, gestational age, Apgar score at 5 minutes, PHQ-9 continuous raw score, and GAD-7 continuous raw score. Children's average birth weight, gestational age and Apgar score were then compared between depressed (PHQ-9 ≥10) vs. non-depressed (PHQ-9 \<10) mothers and anxious (GAD-7 ≥10) vs. non-anxious (GAD-7 \<10) mothers using ANOVA. Univariate linear regressions were conducted to identify potential predictors associated with children's birth weight. Multivariate linear regression models were conducted to control the simultaneous confounding effects of possible predictors. Model 1 shows the role of depressive symptoms (PHQ-9 ≥10) and Model 2 shows anxiety symptoms (GAD-7 ≥10) as predictors of children's birth weight. The same analytic strategy was followed to assess predictors of PTB using logistic regression.
Results {#s3}
=======
The prevalence of depression and anxiety symptoms were 28.9% and 14.2% respectively among all women in the third trimester of pregnancy. Anxiety was more common at higher ages and in mothers from Côte d'Ivoire. The depressed mothers weighed more and had a larger BMI than the non-depressed mothers. The depressed and anxious mothers did not differ from healthy mothers in SES distribution, hemoglobin level, height, number of complications during previous pregnancies and caesarian section delivery ([Table 1](#pone-0080711-t001){ref-type="table"}).
10.1371/journal.pone.0080711.t001
###### Baseline characteristics of the study mothers by antepartum depression and anxiety status.
**Variable** **Non- depressed n = 510** **Depressed n = 207** **p** **Non-anxious n = 614** **Anxious n = 102** **p** **Total n = 717**
-------------------------------------- ---------------------------- ----------------------- --------- ------------------------- --------------------- --------- -------------------
**Socioeconomic**
Age(m± SD) 28.8 ± 5.3 29.7 ± 5.9 0.06 28.8 ± 5.4 30.6 ± 5.9 0.00 29.1 ± 5.5
Country (%) 0.22 0.04
Ghana 218 (42.8) 79 (37.8) 264 (42.9) 33 (32.0) 297 (41.3)
Côte d'Ivoire 292 (57.3) 130 (62.2) 351 (57.1) 70 (68.0) 422 (58.7)
SES (%) 0.12 0.06
Top 20% 155 (31.4) 51 (25.1) 186 (31.4) 20 (19.6) 206 (29.6)
Middle 50% 198 (40.2) 80 (39.4) 232 (39.1) 46 (45.1) 278 (39.9)
Bottom 30% 140 (28.4) 72 (35.5) 175 (29.5) 36 (35.3) 212 (30.5)
**Anthropometric and obstetric**
Hemoglobin ± SD (g/dL) 11.1 ± 1.2 10.9 ± 1.2 0.13 11.1 ± 1.1 10.8 ± 1.3 0.08 11.0 ± 1.2
Weight ± SD (kg) 70.1 ± 11.4 73.3 ± 13.9 0.00 70.9 ± 12.0 72.0 ± 13.7 0.38 71.0 ± 12.3
Height ± SD (cm) 161.0 ± 6.0 161.5 ± 6.5 0.27 161.1 ± 6.1 161.5 ± 6.2 0.52 161.1 ± 6.1
BMI ± SD 27.0 ± 4.3 28.1 ± 5.1 0.01 27.3 ± 4.5 27.5 ± 4.7 0.67 27.4 ± 4.6
Previous pregnancy complications (%) 113 (23.3) 53 (26.1) 0.43 138 (23.5) 28 (28.0) 0.33 166 (24.2)
Caesarian section(%) 94 (18.5) 37 (17.9) 0.84 111 (18.1) 20 (19.6) 0.72 131 (18.4)
PHQ-9 mean score ± SD 5.2 ± 2.5 13.4 ± 2.9 \<0.001 6.7 ± 4.1 12.4 ± 4.3 \<0.001 7.6 ± 4.6
GAD-7 mean score ± SD 4.1 ± 3.1 7.9 ± 4.3 \<0.001 4.0 ± 2.6 12.2 ± 2.4 \<0.001 5.2 ± 3.9
Notes. Depressed, PHQ-9 ≥ 10; Anxious, GAD-7 score ≥ 10; PHQ-9, Patient Health Questionnaire depression module; GAD, generalized anxiety disorder; SES, socio-economic status; BMI, body mass index
[Table 2](#pone-0080711-t002){ref-type="table"} shows the correlation coefficients between birth weight, gestational age, Apgar score at 5 minutes, PHQ-9 continuous raw score, and GAD-7 continuous raw score. The Apgar score was weakly correlated with depression and anxiety while birth weight and gestational age were not correlated with maternal mental health status. [Table 3](#pone-0080711-t003){ref-type="table"} shows the birth outcomes by maternal antepartum depression and anxiety status. The mean birth weight was 3172.1g (SD 440.6) and prevalence of LBW was 1.7%. The mean gestational age was 39.6 weeks and the proportion of PTB was 4%. We did not find any significant difference between depressed and non-depressed, and anxious and non-anxious women in terms of the average birth weight and head circumference of their children and gestational age at delivery. The average Apgar score was 8.5 in the children of anxious mothers and 8.7 in the children of non-anxious mothers (p=0.025), but the difference was not clinically meaningful.
10.1371/journal.pone.0080711.t002
###### Pearson Correlation coefficients.
Birth Weight Gestational Age Apgar score PHQ-9 score GAD-7 score
----------------- -------------------------------------------- ----------------- --------------------------------------------- -------------------------------------------- -------------
Birth weight 1
Gestational age 0.2481[\*](#nstab2.1){ref-type="table-fn"} 1
Apgar score -0.0380 -0.0414 1
PHQ-9 score 0.0445 -0.0326 -0.1059[\*](#nstab2.1){ref-type="table-fn"} 1
GAD-7 score 0.0182 -0.0309 -0.1019[\*](#nstab2.1){ref-type="table-fn"} 0.5420[\*](#nstab2.1){ref-type="table-fn"} 1
p\<0.05
10.1371/journal.pone.0080711.t003
###### Comparison of birth outcomes by maternal antepartum depression and anxiety status.
**Outcome** **Non- depressed n = 510** **Depressed n = 207** **p** **Non-anxious n = 614** **Anxious n = 102** **p** **Total n = 717**
------------------------------------------------ ---------------------------- ----------------------- ------- ------------------------- --------------------- ------- -------------------
Birth weight (g, mean ± SD) 3158.2 ± 434.8 3206.5 ± 453.9 0.18 3168.8 ± 443.5 3191.2 ± 426.5 0.64 3171.9 ± 440.9
\< 2500 g (%) 9 (1.8) 3 (1.5) 0.22 12 (2.0) 0 (0.0) 0.45 12 (1.7)
2500-2999 g (%) 163 (32.0) 52 (25.1) 186 (30.3) 29 (28.4) 215 (30.0)
3000-3499 g (%) 207 (40.6) 100 (48.3) 257 (41.9) 49 (48.0) 306 (42.7)
\>= 3500 g (%) 131 (25.7) 52 (25.1) 159 (25.9) 24 (23.5) 183 (25.6)
Gestational age at delivery (weeks, mean ± SD) 39.7 ± 1.8 39.6 ± 1.9 0.56 39.6 ± 1.9 39.5 ± 2.1 0.70 39.6 ± 1.9
\< 34 wks (%) 4 (1.1) 0 (0.0) 0.21 4 (0.9) 0 (0) 0.25 4 (0.8)
34-36 wks (%) 9 (2.5) 8 (5.4) 12 (2.8) 5 (6.5) 17 (3.4)
37-38 wks (%) 72 (20.2) 33 (22.5) 85 (20.0) 20 (26.0) 105 (20.9)
38-39 wks (%) 65 (18.3) 31 (21.1) 85 (20.0) 11 (14.3) 96 (19.1)
\> 39 wks (%) 206 (57.9) 75 (51.0) 240 (56.3) 41 (53.3) 281 (55.9)
Head circumference (cm, mean ± SD) 33.3 ± 1.8 33.3 ± 1.6 0.77 33.4 ± 1.8 33.2 ± 1.5 0.48 33.3 ± 1.7
Apgar score at 5 minutes (mean ± SD) 8.7 ± 0.6 8.6 ± 0.7 0.08 8.7 ± 0.6 8.5 ± 0.8 0.03 8.7 ± 0.6
\< 7 (%) 0 (0.0) 2 (1.0) 0.11 0 2 (2.0) 0.03 2 (0.3)
7-9 (%) 490 (96.1) 199 (96.1) 593 (96.4) 95 (94.1) 689 (96.1)
10 (%) 20 (3.9) 6 (2.9) 22 (3.6) 4 (4.0) 26 (3.6)
Notes. Depressed, PHQ-9 ≥ 10; Anxious, GAD-7 score ≥ 10
Univariate linear regression revealed no significant association between maternal depression or anxiety and children's birth weight ([Table 4](#pone-0080711-t004){ref-type="table"}). Being Ghanaian, belonging to the top 30% with respect to SES, increased maternal age, and male gender of the child were significantly associated with higher birth weight. In the multivariate analyses, neither maternal depression nor anxiety predicted children's birth weight adjusted for country, SES, maternal age and child gender. Boys were on average 147g heavier at birth than girls when country, SES, maternal age, and maternal mental disorders were held constant. One year increase in maternal age was associated with 7g increase in children's birth weight controlling for other factors. Children whose mothers belong to the top 30% SES group were 200g heavier at birth than children whose mothers belong to the lowest 30% SES group, adjusting for other factors.
10.1371/journal.pone.0080711.t004
###### Univariate and multivariate linear regression for birth weight.
Univariate Multivariate^[1](#ngtab4.1){ref-type="table-fn"}^ Multivariate^[2](#ngtab4.2){ref-type="table-fn"}^
---------------------------------- ------------ --------------------------------------------------- --------------------------------------------------- ------- -------------- --------- ------- -------------- ---------
Depression (Ref: Non-depressed) 48.4 -22.9, 119.6 0.18 52.2 -18.2, 122.6 0.15 \- \- \-
Anxiety (Ref: Non-anxious) 22.4 -70.2, 115.0 0.64 \- \- \- 17.1 -74.6, 108.7 0.72
Country (Ref: Côte d'Ivoire) 105.1 39.9, 170.2 0.00 -5.6 -89.0, 77.9 0.90 -5.2 -88.9, 78.6 0.90
SES (Ref: lowest 30%)
Middle 40% -2.2 -79.7, 75.3 0.96 15.8 -69.9, 95.6 0.70 14.1 -65.8, 94.1 0.73
Top 30% 177.8 94.7, 260.9 \<0.001 202.2 96.7, 307.7 \<0.001 199.4 93.7, 305.1 \<0.001
Child sex (Ref: Female) 144.7 80.8, 208.5 \<0.001 146.6 82.6, 210.5 \<0.001 147.6 83.5, 211.6 \<0.001
Maternal age 8.4 2.6, 14.3 0.01 7.1 1.3, 13.0 0.02 7.3 1.4, 13.2 0.02
Maternal hemoglobin -1.6 -36.4, 33.2 0.93 \- \- \- \- \- \-
Previous pregnancy complications 47.5 -29.6, 124.6 0.23 \- \- \- \- \- \-
Maternal IPTp use 14.5 -56.4, 85.4 0.69 \- \- \- \- \- \-
Maternal malaria treatment -24.0 -95.9, 47.8 0.51 \- \- \- \- \- \-
Maternal diarrhea -44.4 -122.8, 34.0 0.27 \- \- \- \- \- \-
Maternal fever 25.2 -42.7, 93.1 0.47 \- \- \- \- \- \-
Maternal urinary tract infection -23.1 -108.8, 62.5 0.60 \- \- \- \- \- \-
Maternal vaginal infection -24.7 -92.6, 43.3 0.48 \- \- \- \- \- \-
Maternal HIV infection -149.3 -369.5, 70.8 0.18 \- \- \- \- \- \-
Maternal height 5.1 -0.2, 10.4 0.06 \- \- \- \- \- \-
Delivery by caesarian section 82.9 -0.9, 166.8 0.05 \- \- \- \- \- \-
Notes: Depression, PHQ-9 score ≥ 10; anxiety, GAD-7 score ≥ 10; SES, socio-economic status; IPTp, intermittent preventive treatment in pregnancy
Multivariate linear regression using depression as predictor
Multivariate linear regression using anxiety as predictor
Univariate logistic regression revealed no significant association between maternal depression or anxiety and PTB ([Table 5](#pone-0080711-t005){ref-type="table"}). Being Ivorian, lower SES and antepartum maternal vaginal infection were associated with increased likelihood for PTB. In the multivariate analyses, neither maternal depression nor anxiety was significantly associated with children's PTB adjusted for country, SES, and maternal vaginal infection.
10.1371/journal.pone.0080711.t005
###### Univariate and multivariate logistic regression for PTB.
Univariate Multivariate^[1](#ngtab5.1){ref-type="table-fn"}^ Multivariate^[2](#ngtab5.2){ref-type="table-fn"}^
---------------------------------- ------------ --------------------------------------------------- --------------------------------------------------- ----- ---------- ------ ----- ---------- ------
Depression (Ref: Non-depressed) 1.6 0.7, 4.1 0.30 2.1 0.8, 5.6 0.15 \- \- \-
Anxiety (Ref: Non-anxious) 1.9 0.7, 5.2 0.25 \- \- \- 1.8 0.6, 5.5 0.29
Country (Ref: Côte d'Ivoire) 0.3 0.1, 0.7 0.01 0.6 0.2, 2.3 0.47 0.6 0.2, 2.5 0.52
SES (Ref: lowest 30%)
Middle 40% 0.5 0.2, 1.3 0.16 0.5 0.2, 1.6 0.25 0.5 0.2, 1.5 0.22
Top 30% 0.2 0.1, 0.8 0.03 0.3 0.1, 1.8 0.21 0.3 0.1, 1.8 0.20
Child sex (Ref: Female) 0.7 0.3, 1.8 0.46 \- \- \- \- \- \-
Maternal age 1.0 0.9, 1.1 0.67 \- \- \- \- \- \-
Maternal hemoglobin 1.1 0.6, 2.1 0.75 \- \- \- \- \- \-
Previous pregnancy complications 0.3 0.1, 1.4 0.12 \- \- \- \- \- \-
Maternal IPTp use 1.1 0.4, 3.1 0.83 \- \- \- \- \- \-
Maternal malaria treatment 1.1 0.4, 2.9 0.87 \- \- \- \- \- \-
Maternal diarrhea 0.7 0.2, 2.6 0.64 \- \- \- \- \- \-
Maternal fever 1.8 0.7, 4.5 0.21 \- \- \- \- \- \-
Maternal urinary tract infection 1.1 0.3, 3.9 0.87 \- \- \- \- \- \-
Maternal vaginal infection 3.0 1.2, 7.3 0.02 2.4 0.9, 6.7 0.10 2.3 0.8, 6.4 0.11
Maternal HIV infection 5.1 0.6, 46.1 0.15 \- \- \- \- \- \-
Maternal height 1.0 0.9, 1.1 0.72 \- \- \- \- \- \-
Delivery by caesarian section 0.6 0.2, 1.9 0.35 \- \- \- \- \- \-
Notes: Depression, PHQ-9 score ≥ 10; anxiety, GAD-7 score ≥ 10, SES, socio-economic status; IPTp, intermittent preventive treatment in pregnancy
Multivariate linear regression using depression as predictor
Multivariate linear regression using anxiety as predictor
We also assessed whether mothers with comorbidity of depression and anxiety (n = 76) differed from depressed-only (n = 132) or anxious-only (n = 27) or healthy mothers with regard to adverse birth outcomes, and did not find any significant differences.
Discussion {#s4}
==========
The results of this study suggest that, in our sample of low-obstetric risk women in sub-Saharan Africa, depression and anxiety during the 3^rd^ trimester per se are not predictive of poor birth outcomes. Although maternal depression and anxiety have been considered to substantially contribute to PTB, LBW and consecutive neonatal morbidity in low- and middle-income countries (LMICs) \[[@B10]\], we cannot confirm this association for otherwise healthy African pregnant women.
In our sample, maternal depression and anxiety symptoms were quite common while LBW and PTB were rare. Previous studies found that the prevalences of LBW and PTB were between 9 and 17% in the two countries \[[@B30],[@B31]\]. Several reasons likely account for this discrepancy: To control for confounding we restricted our sample to pregnant women without severe pregnancy complications which are known to be associated with poor birth outcomes \[[@B32]\]. The low prevalence of poor birth outcomes may suggest a strong confounding effect of pregnancy complications when examining the association between maternal mental health and birth outcomes. Also, women in Côte d'Ivoire who dropped out of the study due to political instability (n=311) scored higher on mental symptom scales and lower on SES, which may have introduced selection bias towards less severe exposure and, potentially, outcomes.
CMD during pregnancy have been evaluated in many settings, mainly focusing on depression and anxiety. Reports from high and low-income countries linking maternal CMD to perinatal outcomes showed conflicting results. A study in pregnant women from U.S. minority groups (n=712) found the adjusted odds ratio (OR) for delivering a LBW infant was 3.97 (95% confidence interval (CI) 3.80-4.15) and for PTB was 3.39 (95% CI 3.24-3.56) in adult depressed women as compared to non-depressed women \[[@B33]\]. This association was not present among adolescents. This study excluded pregnant women who had a history of non-obstetric, chronic disease (e.g. insulin-dependent diabetes mellitus or systemic lupus erythematosus), drug or alcohol abuse, or of psychiatric illness before pregnancy, but did not exclude those with pregnancy complications. Another study among mainly African-American women (n=1399) in Baltimore, Maryland, U.S., found maternal depressive symptoms independently associated with spontaneous PTB (OR 1.96; 95% CI 1.04-3.72) \[[@B34]\]. The same research group reported specific pregnancy-related anxiety to be also linked to an increased risk of spontaneous PTB in their sample, while measures of general anxiety symptoms were not applied \[[@B35]\]. In this study first and second-trimester bleeding was the only pregnancy complication adjusted for in the logistic regression model. On the other hand, studies in the U.S. \[[@B13],[@B36]\], Sweden \[[@B37]\], Norway \[[@B38],[@B39]\], and Ethiopia \[[@B40]\] did not find significant associations between maternal mental health and birth outcomes.
Data from LMICs including Pakistan \[[@B12]\], India \[[@B11]\], Brazil \[[@B41]\], and Bangladesh \[[@B15]\] found associations between depressive and/or anxiety symptoms with LBW and/or PTB, but none of these studies considered the impact of pregnancy complications, and their samples did not exclude women of high obstetric risk.
Overall, it appears that studies which controlled for pregnancy complications did not find a strong influence of maternal depression and anxiety on LBW and PTB. A systematic review on correlates of anxiety symptoms during pregnancy and the association with perinatal outcome reported that there is no evidence of a relationship between general anxiety symptoms and birth outcomes (PTB, LBW, 5 minutes Apgar score) \[[@B42]\]. Yet, a recent systematic review on the association between depression during pregnancy and PTB, LBW and intrauterine growth restriction found that antepartum depression was associated with LBW and PTB. The risk of LBW associated with antenatal depression was larger in LMICs (relative risk (RR): 2.05; 95% CI 1.43-2.93) compared to the United States (RR: 1.10; 95% CI 1.01-1.21) or Europe (RR: 1.16; 95% CI 0.92-1.47) \[[@B10]\] where effect estimates were only marginally or not significant. Moreover, the pooled effect estimates should be interpreted with caution because the differences across study populations, study designs, sample sizes, confounders, depression and/or anxiety definitions and instruments, and the timing of assessment complicated the comparison of the studies.
Considering the available data, the heterogeneity of findings is striking. We consider it likely that the literature is confounded by the detrimental effects of pregnancy complications, which have not been accounted for as distinct risk factors for PTB, LBW and other indicators of neonatal ill-health in most studies.
From an etiological perspective, pregnancy complications can evoke mood disturbances and impair maternal mental health \[[@B43]\]. On the other hand, poor maternal mental health has been found to impact on pregnancy courses through its promoting influence on risk factors for adverse birth outcomes, like unhealthy behavior and poor nutrition \[[@B7],[@B44]\], hypertension and pre-eclampsia \[[@B43]\], uterine infections \[[@B9]\], or lack of prenatal healthcare visits \[[@B45]\]. Altogether, biological pathways of how maternal antepartum depression and/or anxiety could affect birth outcomes are not well understood, and the role of pregnancy complications as confounder, mediator or effect modifier has never been extensively studied.
Almost none of the studies in LMICs tried to separate the influence of mental disorders per se from pregnancy complications, which may explain the observed larger effects in these countries. To evaluate direct effects of maternal CMD on birth outcomes, we limited the potential effect of pregnancy complications by restricting the study population to physically healthy pregnant women. We did not find significant associations between depression and anxiety during the 3rd trimester of pregnancy and poor birth outcomes.
Consistent with previous data \[[@B12],[@B15],[@B40]\], we found low maternal SES to be negatively associated with child's weight. Children whose mothers belonged to the top 30% SES group were 200g heavier at birth than children whose mothers belong to the lowest 30% SES group, adjusting for other factors. Although not necessarily associated with impaired health in a specific child, these differences in birth weight are substantial from a population perspective.
This study has limitations: First, we measured maternal depression and anxiety in the 3^rd^ trimester of pregnancy. Yet, pre-existing mental disorders that may affect fetal growth at earlier gestational stages were not captured by the nature of our study design \[[@B46]\]. Second, the anxiety measure applied is not specific for pregnancy-related anxiety but rather for general anxiety symptoms. Some researchers suggest that pregnancy- and child-directed concerns and worries could be more closely associated with perinatal outcomes than symptoms of general anxiety \[[@B47]\]. Also, the screening instruments used to assess CMD are imperfect and subject to information bias. Third, 311 mothers in Côte d'Ivoire dropping out due to political instability, may have led to selection bias. Forth, by design the results are not generalizable to women with high obstetric risk. Fifth, a lack of statistical power could explain negative results.
Despite these limitations, this study is important because it fills the data-gap on antepartum CMD and birth outcomes in resource-poor settings and in sub-Saharan Africa in particular \[[@B48]\]. To our knowledge, only one study on antepartum CMD and adverse birth outcome (LBW) in sub-Saharan Africa was conducted in Ethiopia \[[@B40]\]. Like almost all data from LMICs, this study did not control for the potential effect of pregnancy complications. Since antepartum CMD are highly prevalent in African pregnant women and pregnancy complications may have particularly severe consequences for the neonate, more studies that carefully control for potential confounders including pregnancy complications are needed to understand the effect of antepartum CMD on the health of neonates in these populations.
Conclusions {#s5}
===========
Our data suggests that maternal depression and/or anxiety in the 3^rd^ trimester of pregnancy are not independent predictors of adverse birth outcomes, such as LBW, PTB, and low Apgar scores in a low obstetric risk population with good but not unusually exceptional antenatal care. Further studies on mental disorders in pregnant women and their impact on birth outcomes are needed. Notably, the role of pregnancy complications as confounders or effect modifiers should be explored.
We thank the participating mothers, their children and the personnel at the two study sites. This work forms part of the MD, PhD, or Master theses of DB, YM, DF, JEK, EK, JB, SD, SP, LS, and LC. We thank the Johns Hopkins Libraries Open Access Promotion Fund for sponsoring the cost for this publication.
**The International CDS Study Group comprises (in addition to the authors and in alphabetical order**): Jana Baum^5^ (Hamburg, Germany), Gerd D. Burchard^5^ (Hamburg, Germany), Lisa Claussen^5^ (Hamburg, Germany), Simon Deymann^5^ (Hamburg, Germany), Heike Ewert^5^ (Hamburg, Germany), Daniel Fordjour^4^ (Kumasi, Ghana), Andreas Hahn^7^ (Hamburg, Germany), Anna Jaeger^7^ (Hamburg, Germany), Jean E. Koffi^6^ (Abidjan, Côte d'Ivoire), Esther Kra^3^ (Abidjan, Côte d'Ivoire), Jürgen May^7^ (Hamburg, Germany), Yasmin Mohammed^4^ (Kumasi, Ghana), Yaw Osei^4^ (Kumasi, Ghana), Sarah Posdzich^5^ (Hamburg, Germany), Birgit Reime^7^ (Hamburg, Germany), Lisa Schlüter^5^ (Hamburg, Germany), Egbert Tannich^7^ (Hamburg, Germany)
[^1]: **Competing Interests:**The authors have declared that no competing interests exist.
[^2]: Conceived and designed the experiments: C. Bindt SE HT EN. Performed the experiments: MK JAP MTB RH SS TF DB KAE SBN. Analyzed the data: C. Barkmann NG WL SE. Contributed reagents/materials/analysis tools: SE. Wrote the manuscript: C. Bindt NG SE.
[^3]: ¶ Information regarding the membership is available in the Acknowledgements
| {
"pile_set_name": "PubMed Central"
} |
1. Introduction {#sec1}
===============
Uterine leiomyomas are benign tumors which affect up to 80% of females in their lifetime \[[@B1]\]. Family history, nulliparity, obesity, and race have been established as risk factors for development of uterine leiomyoma \[[@B2], [@B3]\]. They are more prevalent in women of African descent and cause infertility in up to 3% of patients \[[@B4], [@B5]\]. Metastasis of uterine leiomyoma was first described by Steiner in 1939 \[[@B6]\]. Uterine leiomyomas are considered to be benign but have been reported to metastasize to various organ including the lymph nodes, bones, skeletal muscles, bone marrow, nerves, peritoneum, retroperitoneum, heart, mediastinum, and lungs \[[@B7]--[@B13]\]. Histologically, they consist of dense nodules with cells and extracellular material arranged in a trabecular pattern punctuated by foci of cystic degeneration and microcalcifications. Microscopically, the trabecular pattern consists of spindle-shaped smooth muscle cells with abundant eosinophilic cytoplasm and elongated nuclei. However, several subtypes exist: benign metastasizing leiomyoma (BML) which is most often detected in lungs as nodules and consists of densely packed smooth muscle cells; epithelioid leiomyoma, consisting of polygonal cells; cellular leiomyoma, consisting of smooth muscles and collagen; vascular leiomyoma, consisting of abundant blood vessels; leiomyoma with tubules which consists of tubular structures; lipoleiomyoma, consisting of smooth muscle cells and adipocytes; myxoid leiomyoma, consisting of smooth muscles interspersed in myxoid substance; atypical leiomyoma, consisting of atypical cells distributed in extracellular matrix \[[@B14]\].
Metastatic foci of leiomyoma have been discovered up to 24 years after hysterectomy for benign leiomyoma of the uterus and the average age of presentation is 48 years \[[@B15], [@B16]\]. BML has been postulated to be a result of hematogenous spread from a benign uterine leiomyoma. It is currently unclear if hematogenous spread is an endogenous process characteristic of the primary disease or if it is facilitated by procedures such as dilatation and curettage. BML is considered benign despite its ability to metastasize because of the lack of mitotic figures and anaplasia. Pulmonary BML is often seen in premenopausal women but is rarely seen in postmenopausal women. This case report describes the clinical presentation, pathological features, and clinical course of a patient who was diagnosed with pulmonary BML after she was found to have a hemothorax.
2. Case Summary {#sec2}
===============
The patient is a 64-year-old lady with a past medical history significant for hypertension, hyperlipidemia, hypothyroidism, and diabetes who presented with rapidly progressing dyspnea on exertion for the past two weeks. This was accompanied by a new onset progressive orthopnea and right-sided back pain of the same duration. She denied cough, hemoptysis, fever/chills, chest pain, and leg swelling. She had seen her primary care physician recently and was prescribed a trial of albuterol and was scheduled for an echocardiogram. Notably, her primary care physician documented clear lungs bilaterally, and the patient reported no improvement in her dyspnea with albuterol. She denied recent travel and had no history of a positive purified protein derivative (PPD) test. Two years ago, she was noted to have increased abdominal girth, and imaging showed uterine and ovarian masses. She subsequently underwent total abdominal hysterectomy with bilateral salpingo-oophorectomy, and pathology showed benign leiomyomas. During the same time, she was also diagnosed with multiple lung nodules and underwent a CT-guided biopsy which showed atypical cells and smooth muscle and was inconclusive. In the same year, she underwent colonoscopy which showed a benign polyp and a mammogram which showed dense irregular tissue in her right breast. Due to the discomfort experienced due to the CT-guided biopsy, she refused to have repeat evaluation of her lung nodules and was lost to follow-up. Her family history was significant for colon cancer in her mother. She worked in the office of a pediatrician and was an active smoker but denied consumption of alcohol or usage of intravenous drugs. Her medications included lisinopril, hydrochlorothiazide, levothyroxine, and an albuterol inhaler.
During her current presentation, her physical examination was remarkable for oxygen saturation of 98% on 2 L of nasal cannula, absence of jugular venous distension and any lymphadenopathy, the presence of a 2 × 3 cm irregular and mobile mass in her right breast, clear breath sounds in the left lung but audible breath sounds only in the apex of the right lung, and dullness to percussion up to two-thirds of the right lung field. The abdomen was benign, and there was no evidence of pedal edema. Her labs were significant for leukocytosis (15.7 k/mm^3^), thrombocythemia (437 k/mm^3^), and mild hyponatremia (133 mEq/L). Her chest X-ray (CXR) revealed a complete white-out of the right side when compared to her previous CXR (Figures [1(a)](#fig1){ref-type="fig"} and [1(b)](#fig1){ref-type="fig"}). A bedside ultrasound showed a complex hypoechoic space with no clear pocket for thoracentesis. A chest CT with contrast showed the entire right hemithorax filled with fluid and an irregular mass (Figures [2(b)](#fig2){ref-type="fig"} and [2(c)](#fig2){ref-type="fig"}). Placement of a pigtail catheter resulted in drainage of 1500 cc of sanguinous fluid with RBC of 1.5 million/mm^3^, glucose of 28, LDH of 966 IU/mL, and a total protein of 5.6 g/dL. The fluid cytology showed blood and inflammatory cells, but no malignant cells were detected. Flow cytometry did not reveal cells with markers characteristic of lymphoma. During the next 24 hours, the chest tube drained 2000 cc of sanguinous fluid, and her hemoglobin dropped from 12.4 to 11 g/dL. The patient reported a significant improvement in her symptoms. However, a repeat CXR showed a significant amount of fluid in the lower lung field of the right lung. A repeat chest CT was suspicious of a bleeding vessel from a necrotic mass in the right middle lobe (Figures [3(a)](#fig3){ref-type="fig"}, [3(b)](#fig3){ref-type="fig"}, and [3(c)](#fig3){ref-type="fig"}). She was taken to the OR and underwent a right-sided video-assisted thoracoscopic surgery (VATS) which was converted to right-sided open thoracotomy. During the procedure, approximately 2500 cc of sanguinous fluid was removed from the hemithorax, and the right middle lobe was removed after a large bleeding mass was noted in it. The intraoperative frozen section pathology revealed a spindle cell mass. After partial decortication, the remaining parts of the right lung was successfully reinflated. The pathological examination of the excised lung mass revealed two foci of metastasizing cellular leiomyoma with minimal cellular atypia and occasional mitoses. No necrosis or lymphovascular invasion was noted, and the bronchial and vascular margins were negative. The foci showed spindle cells and stained positive for smooth muscle actin (SMA), vimentin, desmin cytokeratin 7 (CK 7) (Figures [4(a)](#fig4){ref-type="fig"}--[4(f)](#fig4){ref-type="fig"}). The pulmonary lesions were found to be histologically similar to uterine leiomyoma and did not meet the histological criteria for sarcoma. Follow-up CXRs taken on postoperative days 1 and 16 showed gradual but complete resolution of the right-sided mass (Figures [5(a)](#fig5){ref-type="fig"} and [5(b)](#fig5){ref-type="fig"}).
3. Discussion {#sec3}
=============
BML has been observed to express estrogen and progesterone receptors and is known to regress during pregnancy, after menopause and with selective estrogen modulator therapy \[[@B17]\]. This may explain the relative paucity of medical literature on BML in postmenopausal women. Lungs are the most common site of metastasis for BML, but pulmonary BML in postmenopausal women has been described only in a handful of cases \[[@B18]--[@B22]\]. A majority of premenopausal women with pulmonary BML have multiple lung nodules, and a minority of patients have solitary lung nodules \[[@B23]\]. The monoclonal origin of BML has been established by cytogenetic studies \[[@B24]\]. It has been postulated that pulmonary BML may specifically arise from distinct subset of cells in the uterine leiomyoma with characteristic chromosomal aberrations of 19q and 22q deletions \[[@B25]\]. Most patients with pulmonary BML are asymptomatic and are diagnosed when pulmonary lesions are incidentally found on imaging. It is rarely associated with another malignancy, but association with primary lung cancer has been documented \[[@B26]\]. In the absence of tissue diagnosis, these lesions may be differentiated from primary lung cancer due to their low FDG activity on PET scans \[[@B27]\]. However, in the presence of a known tissue diagnosis of pulmonary benign leiomyoma, if some FDG avid lesions are noted amongst other non-FDG lesions on PET, leiomyosarcoma should be suspected \[[@B28], [@B29]\]. Detection of microRNA 221 (miR-221) has been recently shown to help differentiate leiomyoma from leiomyosarcoma \[[@B30]\].
Management of pulmonary BML depends on the extent of disease, presence of estrogen and progesterone receptors on the tumor, and the age of the patient. Preoperative treatment with GnRH agonists has been utilized to reduce tumor size and bleeding in women with uterine leiomyoma, but its role in treating pulmonary BML in postmenopausal women has not been clearly established. Most post-menopausal women with limited disease and with positive expression of estrogen receptors respond favorably to hormonal therapy with selective estrogen receptor modulator and aromatase inhibitors \[[@B31]\]. A combination of GnRH agonists and aromatase inhibitors has been shown to be effective in reducing tumor burden in patients with metastatic leiomyoma \[[@B32]\]. However, surgical resection may become necessary if the tumor results in complications such as the development of a hemothorax. Notably, patients with pulmonary lesions which are not resected and are not treated with hormonal therapy may develop cysts in the lung which may destroy the entire lung cavity \[[@B33]\]. Pulmonary embolism has been reported to be associated with pulmonary BML, and anticoagulation may be reasonable, especially in patients with heavy tumor burden \[[@B34]\].
Hemothorax is a rare manifestation of pulmonary BML. There has been one report of pulmonary BML noted in a nodule in the pulmonary artery which was successfully resected \[[@B35]\]. Pleural collection of blood due to pulmonary BML may also get superinfected and present as empyema \[[@B36]\]. This case illustrates the need to consider pulmonary BML as a possible etiology of hemothorax in post-menopausal women.
The study has not been presented in any form in any meeting or forum. This paper is not under consideration in any other journal. All the authors have read the paper and agreed to the content.
The Authors declare that they have no conflict of interests.
{#fig1}
{#fig2}
{#fig3}
{#fig4}
{#fig5}
[^1]: Academic Editors: K. Jamil and M. Ryberg
| {
"pile_set_name": "PubMed Central"
} |
Protocol background {#sec1}
===================
Proteins are often chemically modified by covalent conjugation to investigate the conformational dynamics, molecular interactions and enzyme kinetics in order to gain insight into their biological functions \[[@B1],[@B2]\]. Many single-molecule *in vitro* biochemical and cell-based experimental techniques primarily exploit fluorescence as a tool to investigate these biological functions. An essential requirement to these experimental approaches is to label one or more fluorophores on the desired protein with the required photophysical properties. Many labelling methods have been developed and utilized, including chemical modification of amino acid side chains through covalent or non-covalent binding, incorporation of unnatural amino acids and affinity labelling.
Fluorescein isothiocyanate (FITC) is the most widely used fluorescence labelling reagent for such experimental approaches due to its high quantum efficiency and conjugate stability \[[@B3]\]. FITC reacts with a primary amine on the protein to form a covalent amide bond. FITC-labelled protein substrates/peptides, antibodies, peptide hormones are used as specific probes in enzyme kinetics, immunocytochemistry as well as flow cytometry and identification of receptors on target cells, respectively. However, the success of fluorescence labelling is contingent upon several stringent factors including requirement of pure protein and high protein concentration. Moreover, slight protein aggregation or precipitation that is observed in majority of the cases at higher protein concentrations might lead to failure of the experiment. In addition, high background fluorescence due to the ineffective removal of unreacted FITC, prior to fluorescence studies, is also a major obstacle in this process \[[@B3]\].
Size-exclusion chromatography is commonly used for the removal of unreacted FITC molecules; however, this process is accompanied with the loss of proteins and undesired dilution of the sample, especially when a small quantity of proteins is labelled. Therefore, a specific and efficient FITC labelling technique is essential to overcome these inadequacies. Hence, we have modified the existing protocol by introducing tandem affinity purification (TAP) tags at the N- and C-termini of the target protein. The TAP method provides a high level of purification of the desired proteins without any contamination. The commercially available pMALc5E vector was modified in our lab for TAP, and it consists of two different affinity tags at the N- and C-terminus of the desired protein ([Figure 1](#F1){ref-type="fig"}A,B). Maltose-binding protein (MBP) is present at the N-terminus that allows first-step purification of the fusion protein using amylose resin. Downstream to the MBP, tobacco etch virus (TEV) protease cleavage-site (ENLYFQ/G) was introduced to facilitate the separation of MBP from the target protein. The TEV protease that is often used for easy tag removal is a highly site-specific cysteine protease found in Tobacco Etch Virus \[[@B6]\]. Due to its stringent sequence specificity and activity over a broad temperature range (4--37°C), TEV is considered a very useful reagent for cleaving fusion proteins and better than other commercially available proteases such as factor Xa, thrombin and enterokinase. Moreover, TEV is relatively easy to overproduce and purify in large quantities. The protein of interest will also have a C-terminal His~6~ tag for efficient second-step purification using immobilized metal affinity chromatography resin.
{#F1}
Having an MBP-tag has several advantages such as it enhances the expression level, improves solubility and proper folding of its fusion partner with typical yields of 10 to 30 mg fusion protein per litre culture \[[@B7]\]. Additionally, the MBP fusion system also includes ease of purification and mild elution conditions, which are highly efficient and compatible with most downstream applications. The use of the second-step purification by His~6~ at the C-terminal region removes remaining contaminants from the first purification step.
With the help of this modified method, we have efficiently labelled target protein with a significant decrease in precipitation and degradation ([Figure 2](#F2){ref-type="fig"}). Furthermore, second-step purification of labelled protein by His~6~ tag substantially removed the residual FITC and thus decreased the background fluorescence of unreacted FITC. Therefore, this optimized technique may also be used as a basis for labelling procedures with other fluorophores for high-throughput spectroscopic analysis of biological functions of macromolecules.
{#F2}
Method details {#sec2}
==============
Step 1: Cloning {#sec2-1}
---------------
For TAP tag, modify pMALc5E vector by introducing TEV recognition cleavage-site between N-terminus of MBP and MCS. Insert His~6~ tag at the C-terminal region of the vector ([Figure 1](#F1){ref-type="fig"}B).
### Materials required {#sec2-1-1}
*Escherichia coli* DH5α (New England Biolabs (NEB), Ipswich, MA, U.S.A.), pMAL c5E (NEB), Plasmid miniprep kit (Sigma chemicals, St. Louis, MO, U.S.A.), Quick-change site-directed mutagenesis kit (Stratagene, Cedar Creek, TX, U.S.A.), Ampicillin sodium salt (Sigma), Primers (Sigma), DpnI enzyme (Fermentas, Waltham, Massachusetts, United States), IPTG (Sigma), Luria-Bertani (LB) broth (Himedia, Mumbai, India).
### Method {#sec2-1-2}
Design primers ([Table 1](#T1){ref-type="table"}) for introducing TEV recognition cleavage-site using '*primer-X'* software and '*oligoanalyser'* tools.Perform insertion mutagenesis using the quick change site-directed mutagenesis kit.Incubate the amplified PCR products with DpnI enzyme at 37°C for 3 h to degrade the parental template.Note: DpnI is an endonuclease, which specifically degrades the methylated parental DNA strands.Transform the digested PCR products into *E. coli* DH5α cells. Isolate the plasmids from the transformed colonies using plasmid miniprep kit.Confirm the insertion of TEV recognition site using automated DNA sequencing. This modified vector is now represented as pMALc5E-TEV vector.Similarly, introduce His~6~ tag downstream to the MCS of pMALc5E-TEV vector using the forward and reverse primers as mentioned in [Table 1](#T1){ref-type="table"}. Confirm the insertion of His~6~ using DNA sequencing.Note: This modified pMAL vector is now represented as pMALc5E-TEV-His or TAP vector.Sub-clone the gene of interest into the MCS of TAP vector. Confirm the clone by restriction digestion and DNA sequencing.Note: In our study, we have sub-cloned phosphoprotein enriched in astrocytes (Pea15) gene. Pea15 is a 15 kDa ubiquitously expressed antiapoptotic cytosolic protein with an N-terminal death effector domain (DED) and an irregularly structured C-terminal tail. Pea15, by virtue of its DED, binds to other DED-containing proteins and inhibits Fas/TNFR1-induced extrinsic apoptosis \[[@B10]\]. This protein was labelled to be used as a substrate in an enzyme kinetics study.
###### Primer sequences for inserting TEV and His tag
Primer 5′ to 3′ sequence
----------------------- --------------------------------------------------
His~6~ forward primer CCCTGCAGGT**CACCACCACCACCA** CCACAATTAAATAA
His~6~ reverse primer TTATTTAATT**GTGGTGGTGGTGGTGGTG**ACCTGCAGGG
TEV forward primer GACAAGGTACCG**GAGAATCTTTATTTTCAGGGC**CATATG GCCG
TEV reverse primer CGGCCATATG**GCCCTGAAAATAAAGATTCTC**CGGTACCTTGTC
Nucleotide sequences of TEV cleavage-site and His tag are highlighted in bold.
Step 2: Protein expression and purification: {#sec2-2}
--------------------------------------------
In our study, all recombinant proteins were expressed in bacterial protein expression system because they are easy to culture, grow fast, homogeneous, and give high yields of recombinant protein.
Note: *E. coli* Rosetta2 DE3 host strain should be used for expression of eukaryotic proteins that contain codons rarely used in *E. coli.*If the desired protein is insoluble, co-express the gene with commercially available GroEL/GroES chaperonin plasmids or use purification from inclusion bodies \[[@B11]\].
### Materials required: {#sec2-2-1}
*E. coli* BL21(DE3) competent cells (NEB), Amylose beads (MERK, Darmstadt, Germany), Maltose (Sigma), LB broth, Lysis buffer (20 mM NaH~2~PO~4~/Na~2~HPO~4~ pH 8, 100 mM NaCl, 5 mM BME and 0.1% Triton-X 100), Buffer A (20 mM NaH~2~PO~4~/Na~2~HPO~4~ pH 8, 100 mM NaCl, and 2 mM β-mercaptoethanol or BME), Isopropyl β-D-1-thiogalactopyranoside (IPTG) (Sigma), Econo column (Bio rad, Hercules, California, U.S.A.)
### Method: {#sec2-2-2}
Transform 100 ng of pMALc5E-TEV-His construct into *E. coli* BL21(DE3) competent cells.Pick a single colony from the transformed LB agar plate and inoculate into 10 ml of LB broth containing antibiotics (100 μg/ml ampicillin) according to the host cells and plasmid construct.Incubate the culture overnight at 37°C for 16--18 h under continuous shaking at 200 rpm.After optical density (OD) reaches 0.6, inoculate 10 ml of the culture into 1 L (∼1:100 ratio of culture to media) of LB broth containing the required antibiotics and incubate at 37°C for approximately 3 h until OD reaches 0.6.Note: Time for OD~0.6~ values may vary for other culture mediums and strains.After the required incubation period, incubate the culture additionally for approximately 1 h at 18°C.Note: This step will increase the expression of recombinant protein.Induce protein expression by adding IPTG to a final concentration of 0.5 mM.Note: IPTG concentration may vary (between 0.1 and 1 mM) for other proteins.Post IPTG induction; incubate the culture at 18°C for 16 to 18 h under continuous shaking conditions.After 18 h, harvest the cultures at 10000 rpm for 40 min at 4°C. Discard the supernatant and proceed for protein purification or preserve the pellets at −80°C for future use.Note: Check protein expression before proceeding for purification.Initially, purify recombinant protein by MBP-tagged column purification method as described below.Lyse the cells by re-suspending pellet in 10 ml of lysis buffer followed by sonication and centrifugation at 10000 rpm for 20 min at 4°C. Discard the pellet and transfer supernatant (lysate) into fresh tube.Take 5 ml of amylose beads in 1 × 10 cm econo column and equilibrate the column with 5 column volumes (CV) of *Buffer A*.Note: Econo columns are high quality commercially available chromatography columns.After equilibration, add lysate and keep it for column binding at 4°C for 1 h under gentle shaking condition.After 1 h, collect the flow through and wash amylose beads with 3 CV of Buffer A.Elute bound proteins using elution buffer containing *Buffer A* + 10 mM maltose.Estimate the purity of the protein samples by loading them on 12% SDS-PAGE gel.Use \>95% pure MBP fused target protein for FITC labelling.
Step 3: Protein labelling: {#sec2-3}
--------------------------
Since the isothiocyanate group reacts with amino terminus and free primary amines in proteins, therefore, dialyse the target protein in alkaline buffer of pH 9 to generate free amino groups. This is a crucial step and should be performed prior to labelling.
Note: Do not use buffers containing free amino groups, such as tris, glycine, sodium azide etc., as they might interfere with labelling.
### Materials required: {#sec2-3-1}
FITC (Sigma), Dimethyl sulphoxide (DMSO) (Sigma), 10 kDa cutoff dialysis membrane (Thermo fisher scientific, Waltham, Massachusetts, U.S.A.), 50 mM Na~2~CO~3~/NaHCO~3~ buffer pH 9 (carbonate buffer) and 100 mM NH~4~Cl.
### Method: {#sec2-3-2}
Collect all protein elutes and pool them into a single tube and concentrate the protein to 2--4 mg/ml.Prepare 500 ml of fresh carbonate buffer (do not store for more than 3 days).Dialyse 5 ml of concentrated protein solution against 500 ml of carbonate buffer for 16 h at 4°C with slow stirring.After 16 h, collect the protein sample from the dialysis bag and transfer it to a 15 ml centrifuge tube. Centrifuge it at 4000 rpm for 10 min at 4°C to check for precipitation; later transfer the supernatant into a fresh tube.Check absorbance of the samples before and after dialysis.Note: Check for protein degradation by loading the sample on SDS-PAGE ([Figure 3](#F3){ref-type="fig"}B).Dissolve FITC in anhydrous DMSO at 1 mg/ml concentration. This should be prepared fresh for each labelling reaction.Add 5 μl of the dye to the dialysed protein (5 ml) every 30 min to get a final concentration of 100 ng FITC/1 μg protein. Perform this step by gently rocking the protein solution at 4°C for 12 h.*Precautionary measures:* FITC solution should be kept in dark or wrapped in aluminium foil. Make fresh FITC solution each time and do not re-use.Final DMSO concentration in the solution should be less than 10%.Adding FITC slowly to protein reduces precipitation of protein.All the labelling procedures should be performed under dark condition.Incubate this reaction setup in dark at 4°C for 12 h.Post 12-h incubation; centrifuge the labelled protein at 10000 rpm for 10 min at 4°C.Transfer the supernatant to a fresh tube and cover it with an aluminium foil.Perform extensive dialysis against 1 L of Buffer A for 15 h at 4°C in dark and constant gentle magnetic stirring conditions. Change the dialysis buffer every 4 h.
{#F3}
Note: This step is performed to remove unreacted FITC. Transfer the dialysed labelled protein into a fresh tube and centrifuge it at 10000 rpm for 10 min at 4°C to remove precipitate if any.
Step 4: Ni-IDA metal affinity column purification: {#sec2-4}
--------------------------------------------------
This second step of purification is performed to remove the MBP tag, residual FITC as well as to increase the purity of the labelled protein.
### Materials required: {#sec2-4-1}
TEV protease (1 mg/ml), Ni-IDA resin (MERK), Imidazole (sigma)
Note: This purification should be carried out under dark condition. To the labelled protein, add TEV protease (1 µg of TEV for 10 µg of protein) and incubate the reaction mixture at 18°C for 12 h. Check for the tag cleavage by analysing the sample on 15% SDS-PAGE.After complete cleavage, centrifuge the protein sample at 14000 rpm for 30 min at 4°C. Transfer the supernatant to a fresh falcon tube wrapped in an aluminium foil.Separate the MBP tag from the target protein by passing it through the Ni-IDA affinity resin.Take 5 ml of Ni-IDA beads in 1 × 10 cm econo column and cover it with an aluminium foil. Do not expose the resin or protein to light.Equilibrate the column in 5 CV of Buffer A.After equilibration, add labelled protein and keep it for binding at 4°C for 1 h under gentle shaking.Collect flow through and wash the column with 10 CV of Buffer A.Elute bound protein using elution buffer containing Buffer A + imidazole gradient (50--250 mM).Note: Collect the imidazole protein elutes in eppendrofs wrapped in aluminium foil. At this stage, protein elutes appear orange in colour ([Figure 3](#F3){ref-type="fig"}D) suggesting that FITC is bound to the protein.Pool pure fractions and dialyse against Buffer A in dark condition at 4°C for 12 h to remove imidazole.After 12 h, check the labelling efficiency using the following formula: $$\text{Protein concentration} = \frac{\text{A}280 - (\text{A}495 \times \text{CF})}{\varepsilon\text{ of protein}} \times \text{dilution factor}$$ Where *ε* is molar extinction coefficient, *A*~280~ is absorbance of protein at 280 nm, *A*~495~ is absorbance at 495 nm, CF is correction factor = *A*~280~/*A*~max~ $$\text{Moles of FITC per mole of protein}(\text{F}/\text{P}) = \frac{\text{Amax of labelled protein}}{\varepsilon\text{ of FITC} \times \text{Protein concetration}(\text{M})} \times \text{dilution factor}$$*ε* of FITC is 70,000 M^−1^ cm^−1^,Preserve the purified labelled proteins at −20°C for future use.Note: Purified protein can be stored for 2 to 3 months depending on the protein stability.
Step 5: Method validation {#sec2-5}
-------------------------
### Materials required: {#sec2-5-1}
Trypsin singles (Sigma), FITC labelled protein, Cytation-5 multimode plate reader (BioTek instruments, Winooski, VT, U.S.A.), Bovine serum albumin (BSA) (Sigma)
### Method: {#sec2-5-2}
The described method has been used to effectively label an MBP-tagged Pea15 for quantitative enzyme kinetics study. Using our improved procedure, we have obtained approximately 3mg/ml of ∼95% pure FITC-labelled MBP tagged Pea15 with a significant decrease in degradation, precipitation ([Figure 3](#F3){ref-type="fig"}A,B). Furthermore, MBP tag was successfully separated using the TEV protease followed by nickel affinity chromatography ([Figure 3](#F3){ref-type="fig"}C,D). This second step of purification yielded approximately 97% pure labelled Pea15. Moreover, FITC labelling efficiency was found to be \>90%. Hence, our results demonstrate that a TAP tag represents an attractive fusion construct for *in vitro* labelling of high levels of soluble recombinant proteins. Furthermore, labelled Pea15 was evaluated by monitoring the proteolytic cleavage of FITC-Pea15 using trypsin, which is a pancreatic serine protease. Trypsin specifically cleaves substrates at the carboxyl side of lysine and arginine amino acids.
#### Trypsin protease assay method: {#sec2-5-2-1}
FITC-Pea15 substrate cleavage was determined by incubating 50 nM of enzyme with 5 µM of labelled protein at 37°C for 30 min in Buffer A. Proteolytic cleavage was assessed by monitoring increase in fluorescence intensity of unquenched FITC in Cytation 5 multi-well plate reader at 485 nm excitation and 535 nm emission wavelengths \[[@B14]\]. In this protease assay, BSA was used as a negative control.
#### Result: {#sec2-5-2-2}
FITC labelled protein cleavage with a concomitant increase in fluorescence of unquenched FITC was observed in a time-dependent manner without any background FITC fluorescence ([Figure 4](#F4){ref-type="fig"}). No increase in fluorescence was observed in the negative controls. Hence, our improved protocol proved to be a substantially better method than the conventional protocol.
{#F4}
Discussion {#sec3}
==========
In the present study, we have modified the conventional FITC labelling protocol by introducing the TAP tags at the N- and C- termini of the target protein. We have also modified the commercially available pMALc5E plasmid vector to enhance protein homogeneity, concentration, and labelling efficiency. We accomplished this by introducing MBP at the N-terminus and His~6~ tag at the C-terminus regions of the target protein. In addition, we replaced the enterokinase cleavage-site (DDDDK/X) with a more specific TEV cleavage-site (ENLFYQ/G) in vector for tag separation.
Previously, our effort to label the target protein with FITC using conventional protocol resulted in excessive protein degradation, precipitation, and high residual FITC background during the dialysis and labelling processes \[[@B3]\]. This consequently led to an insufficient yield of labelled protein with lower final concentration (∼0.5 mg/ml) that limited its use for qualitative protease assays. Moreover, to remove the 45 kDa MBP tag and the residual FITC from the labelled protein, we initially attempted cleavage with TEV and subjected it to gel filtration chromatography. However, during this purification process, the overall yield of labelled protein decreased significantly and hence to circumvent these problems, we have modified the conventional labelling protocol by cloning target gene in TAP tag vector. These modifications enabled the purification of 2 mg/ml of FITC labelled protein at over 97% homogeneity from a starting culture volume of 250 ml with a significant decrease in protein precipitation and degradation.
Hence, this TAP tag fusion system offers an attractive method for purification and labelling of several difficult eukaryotic and prokaryotic proteins. To avoid interference of large MBP tag (45 kDa) on the target protein functions if any, several other small-sized tags like glutathione-S-transferase, thioredoxin (GST), streptavidin etc. can be used for TAP tag purification. Appropriate vectors coding these tags are commercially available. In addition, our improved labelling protocol can also be employed for purification and labelling of proteins from inclusion bodies even in the presence of MBP tag. These inclusion bodies can be initially purified under denaturing conditions using the His~6~ tag and subsequently refolded and labelled \[[@B15]\]. During this process, the N-terminal MBP will aid in stabilizing the target protein and eventually increase the yield of the protein. Further, refolded protein can be labelled and purified to homogeneity using the second-step purification.
In addition to FITC conjugation, this improved procedure can also be used for labelling other amine (succinimidyl esters), thiol reactive (IAEDANS, monobromobimane, maleimide etc.), and biotin conjugates for studying enzyme kinetics, protein--protein interactions, conformational changes, and dynamics. Therefore, our improved labelling procedure is well suited for high throughput spectrosopic analysis for a wide variety of studies.
The authors thank Biophysics Facility, ACTREC for providing infrastructure for the necessary experiments.
Competing Interests {#sec4}
===================
The authors declare that there are no competing interests associated with the manuscript.
Funding {#sec5}
=======
The project is funded by grant from Department of Biotechnology (DBT), Govt. of India \[grant number BT/PR10552/BRB/10/606/2008\].
Author Contribution {#sec6}
===================
K.B. and L.K.C. conceived and designed the experiments. L.K.C. and N.V. performed the experiments. L.K.C. and K.B. analysed the data. L.K.C. and K.B. wrote the paper. All of the authors read, revised and approved the manuscript submitted for publication
FITC
: fluorescein isothiocyanate
MBP
: maltose-binding protein
TAP
: tandem affinity purification
TEV
: tobacco etch virus
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#sec1}
============
Biliary atresia (BA) is the most common cause of endstage liver disease in infants with poor prognosis and high mortality.[@bib1], [@bib2] The incidence of BA in Europe and the United States is about 1:15,000--19,000 live births, whereas the Asian incidence is much higher, ranging from 1:5,000 to 1:8,000 live births in China.[@bib3], [@bib4], [@bib5], [@bib6] The etiology of BA is still unknown, but recent genetic factors have been considered to have an important role in BA.[@bib4], [@bib7] Recurrent familial cases supported the higher genetic tendency of BA.[@bib8] Researchers have identified a number of genes associated with BA, such as the migration inhibitory factor (*MIF*),[@bib9] *CD14*,[@bib10] intercellular adhesion molecule-1 (*ICAM-1*),[@bib11] *CFC1*,[@bib12] and *ITGB2* (*CD18*).[@bib13]
*CD14* is a glycosylphosphatidylinositol-anchored LPS receptor that was first reported as a differentiation marker expressed on the surface of macrophage, neutrophils, and other myeloid-lineage cells.[@bib14], [@bib15], [@bib16], [@bib17] The T/T homozygote at position 159 (rs2569190) for the *CD14* promoter polymorphism is related to the development of BA using 90 cases and 143 controls in Taiwan's population.[@bib10] Recent studies have found that hepatocytes can produce large amounts of *CD14* in acute liver injury.[@bib18] However, the role of *CD14* on the hepatocytes remains unclear in BA.
The Notch signaling pathway plays key roles in the development of the biliary system. *NOTCH2*, as a receptor on this signaling pathway, has been demonstrated that hepatocytes dedifferentiate into hepatoblasts and further differentiate into biliary epithelial cells (BECs) when the bile duct is injured, which is mainly regulated by the Notch signaling pathway.[@bib19] *NOTCH2* can keep the normal function of the intrahepatic bile duct (IHBD) in the perinatal and postnatal periods, and the low expression of this receptor leads to the abnormality of the IHBD.[@bib20] *NOTCH2* mutations cause Alagille syndrome, which involves biliary developmental defects that are characterized by neonatal jaundice, impaired differentiation of the IHBD, and chronic cholestasis.[@bib21] *CD14* and the Notch signaling receptor *NOTCH2* can co-express in hepatocytes of BA, but the interaction between these two genes and their respective impact on the bile duct differentiation process are still unknown, and the association between *CD14* and *NOTCH2* in BA has not been reported yet.
To explore the association of *CD14* (rs2569190) and *NOTCH2* (rs835576) in BA, we conducted a case-control study to verify the effects and their interaction in an independent Chinese sample.
Results {#sec2}
=======
Association of *CD14* and *NOTCH2* with BA in a Southern Chinese Population {#sec2.1}
---------------------------------------------------------------------------
*CD14* (rs2569190) and *NOTCH2* (rs835576) were selected in 506 BA cases and 1,473 controls in a Southern Chinese population. (See [Materials and Methods](#sec4){ref-type="sec"} for the detailed selection criteria.) The results showed that limited association of both SNPs was observed with BA ([Table 1](#tbl1){ref-type="table"}). The SNP rs2569190 in *CD14* showed that the frequency of the minor allele (G) was virtually the same between the control group (41.75%) and the BA patients (41.99%), giving an odds ratio (OR) of 1.01 (95% confidence interval \[CI\]: 0.87--1.17, p = 0.89). A regulatory SNP, rs835576, which was located in the 3′ UTR of *NOTCH2*, showed a slightly different frequency between the healthy controls (C: 2.4%) and the BA patients (C: 3.1%), with an OR of 1.30 (95% CI: 0.84--2.01, p = 0.23), though it failed to be statistically significant.Table 1The Association of Two SNPs Located in *CD14* and *NOTCH2* with Biliary AtresiaSNPGeneFunctionCHRBPA1/A2F_AF_UpORCI, 0.95p_adjrs835576*NOTCH2*3′ UTR1119912963C/T0.030.020.231.300.84--2.010.59rs2569190*CD14*intronic5140633331G/A0.420.420.891.010.87--1.170.57[^2]
The Genetic Epistasis between *CD14* and *NOTCH2* Was Associated with the Risk of BA Development {#sec2.2}
------------------------------------------------------------------------------------------------
The results suggested a strong effect from the epistatic interaction between the two variants in our samples (p = 8.1E−03; OR = 2.78; 95% CI: 1.32--5.88) ([Table 2](#tbl2){ref-type="table"}). Pairwise multifactor dimensionality reduction (MDR) analysis was adopted to cross-validate the epistatic interaction between these two variants. The results of cross-validation consistency (CVC) and balanced accuracy were obtained from MDR analysis of the two-locus model ([Table 3](#tbl3){ref-type="table"}). The combinations of low- and high-risk groups were classified in the model ([Figure 1](#fig1){ref-type="fig"}). The dark-shaded cells indicate the high risk of BA with genotype combination (GG:TT). In agreement with the results obtained by logistic regression from PLINK, a significant synergistical epistatic interaction (p = 0.0026) was observed, giving an OR of 1.44 (95% CI: 1.14--1.84) for this model.Table 2Logistic Regression Showing Interactive Effect of Two Regulatory SNPs Located in *CD14* and *NOTCH2*SNP 1Gene 1SNP 2Gene 2ORCI, 0.95p Valuers2569190*CD14*rs835576*NOTCH2*2.781.32--5.888.1E−03[^3]Table 3MDR Validation Showing the Interactive Effect of the Two SNPsSNP 1Gene 1SNP 2Gene 2Balanced AccuracyORCI, 0.95p Valuers2569190*CD14*rs835576*NOTCH2*0.531.441.14--1.842.6E−03[^4]Figure 1CC and GG Regulate the Lower Expression of Regulated Gene, which Is Associated with the Biliary Atresia(A and B) Fibroblast samples of minor alleles C from rs2569190 and G from rs835576 showed relatively lower expression of (A) CD14 and (B) NOTCH2, respectively. (C) CC and GG showed a genetic epistatic effect tested by MDR.
The *cis* Effect of eQTL of Two Interacting SNP Pairs with *CD14* and *NOTCH2* {#sec2.3}
------------------------------------------------------------------------------
The expression quantitative trait loci (eQTL) associations between SNPs and the potential regulated genes were closely examined. We assessed these two candidate SNPs in the Genotype-Tissue Expression (GTEx) project containing genome-wide-based, tissue-specific, regulatory eQTL variants (<http://gtexportal.org>). The genotypes of SNP rs2569190 and SNP rs835576 correlated with the expression of *CD14* in fibroblast and of *NOTCH2* in liver, respectively ([Table 4](#tbl4){ref-type="table"}). The results also further demonstrated that G, the minor allele of SNP rs2569190, is correlated with lower expression of *CD14* in fibroblast. C, the minor allele of SNP rs835576, is correlated with lower expression of *NOTCH2* in liver. In order to further confirm the expression level of these two genes in BA patients, real-time PCR results showed that the expression level of *CD14* in the BA group (n = 28) was significantly lower (0.31 ± 0.02 versus 1.00 ± 0.14, p \< 0.001) than that in the choledochal cyst (CC) group (n = 20); and the expression of *NOTCH2* in BA was also decreased (0.59 ± 0.07 versus 1.00 ± 0.19, p = 0.0262) ([Figure 2](#fig2){ref-type="fig"}). This result had been further verified by western blot, which showed that the gene-related protein expression of *CD14* was significantly decreased (209.9 ± 12.66 versus 546.3 ± 112.3, p = 0.0408) and that the expression level of *NOTCH2* also obviously declined (101.1 ± 58.55 versus 425.1 ± 72.89, p = 0.0257) in BA ([Figure 3](#fig3){ref-type="fig"}).Table 4The eQTL Effect of Two SNPs in *CD14* and *NOTCH2* with Biliary AtresiaSNPGeneRegulatory Patternp ValueTissuers2569190*CD14cis*6.40E−07transformed fibroblasts0.79liverrs835576*NOTCH2cis*4.90E−06liver[^5]Figure 2Quantitative Expression of *CD14* and *NOTCH2* in the LiverEach group of liver mRNA was detected by real-time PCR. The total expression of *CD14* and *NOTCH2* in biliary atresia (BA; n = 28) was significantly lower than that in choledochal cyst (CC; n = 20). \*p \< 0.05; \*\*p \< 0.01.Figure 3Expression of *CD14* and *NOTCH2* in Western BlotLiver tissue lysates were immunoblotted with anti-CD14 and anti-NOTCH2 antibodies. β-actin was selected as the parameter of the corresponding indicator (A). Each group included 3 samples; \*p \< 0.05, unpaired t tests (B). BA, biliary atresia; CC, choledochal cyst.
Discussion {#sec3}
==========
In the present study, we investigated the association of *CD14* (rs2569190) and *NOTCH2* (rs835576) with BA. We found that these two SNPs had a genetic interaction to increase the risk to BA, and both SNPs showed *cis*-regulatory roles in gene expression. This result suggested that *CD14* (rs2569190) could interact with *NOTCH2* (rs835576) and, together, downregulate the expression of the *NOTCH2*. The expression of *CD14* and *NOTCH2* was significantly decreased in BA patients, which could be related to the genetic susceptibility locus of these two genes.
A previous study showed that a T/T homozygote at position 159 of the *CD14* promoter polymorphism was associated with the development of BA in 90 BA patients (A allele: 61.7%, G allele: 38.3%) and 143 healthy controls (A: 52.1%, G: 47.9%), giving an OR of 1.09 (95% CI: 1.01--2.16, p = 0.043) in a southern Taiwan population.[@bib10] However, in our study, the frequency of the minor allele (G) was almost the same between 1,473 healthy controls (41.75%) and 506 BA patients (41.99%), giving an OR of 1.01 (95% CI: 0.87--1.17, p = 0.89). The allele frequency was slightly different between the two studies. Checking 1000G data from an East Asian population (EAS), the minor allele (G) frequency was found to be 42.76%, showing an allele frequency similar to that of the Southern Chinese population. The differences between the Southern Chinese population and previously reported Taiwanese individuals could be due to the sample size limit or potential subpopulation stratification. Therefore, in our subjects, we enlarged the sample sizes to gain enough convincing evidence. Another study reported that the plasma-soluble *CD14* level might serve as a biological marker and was significantly reduced in patients with the T/T and T/C genotypes when the disease progressed to liver cirrhosis, while there was no significant change in patients with the C/C genotype.[@bib10] This finding was supported by Ming-Huei Chou's research,[@bib22] which demonstrated that the hepatic *CD14* mRNA and soluble *CD14* plasma levels of patients with early-stage BA were considerably higher than those with late-stage BA. Additionally, Ahmed et al.[@bib23] also indicated that the expression of *CD14* in BA had a time-related change with overexpression in the early stage and decreased expression in the late stage. However, there may be several potential limitations in these previous studies. First, the small sample size may have reduced the statistical power of the study. Second, the individual differences in genetic susceptibility genes may lead to variable degrees of hepatic fibrosis and surgical prognosis in BA patients. Therefore, the liver tissue of BA patients should be in strict accordance with the criteria of clinical pathological staining to determine the degree and stage of hepatic fibrosis (including early and late stages), rather than simply judging by the time of Kasai surgery or liver transplantation.
In our experiments, the expression of *CD14* was significantly decreased in BA patients, similar to the findings in noted studies, but we failed to replicate the previous reported variant on *CD14* in our cohort. We suggested that the genetic epistatic effect of this previously reported variant with SNP on *NOTCH2* might boost the risk to disease by using a case-control study. The Notch signaling pathway plays a key role in the development and differentiation of hepatic stem cells into BECs in order to establish the biliary system.[@bib24], [@bib25] Mutations in Jagged1 (Jag1) and *NOTCH2* result in Alagille syndrome (AGS) in humans, which is characterized by biliary dysplasia and cholestasis,[@bib26], [@bib27] and similar findings were observed in mice,[@bib28] suggesting functional conservation. In a previous study, we also found that lower expression of *NOTCH2* could lead to malformation of the IHBD in the perinatal and postnatal periods.[@bib20] However, the specific mechanism for downregulating the *NOTCH2* expression in BA is not clear.
A regulatory SNP rs835576 located in the 3′ UTR of *NOTCH2* was previously identified to validate the susceptibility with BA. However, we found different minor alleles between healthy controls (C: 2.4%) and BA patients (C: 3.1%) with an OR of 1.30 (95% CI: 0.84--2.01, p = 0.23), which we observed similarly with the allele frequency of the SNP for Han Chinese in Beijing (HCB; C: 2.3%) reflecting the data reliability. The *NOTCH2* (rs835576) SNP showed limited evidence of associations with BA, which may be due to the relative rare minor allele frequency for the SNP and the sample size limit of the replication cohort. It is also possible that these two genes are associated with certain disease subphenotypes but not with overall susceptibility. An analysis of our patient samples did not show significant differences for this locus with any subphenotypes, although this could be due to the reduced sample sizes for each subgroup based on certain phenotypes.
Hepatocytes dedifferentiate into hepatoblasts and further differentiate into BECs when the bile duct is injured, which is mainly regulated by the Notch signaling pathway. Our experiment found that these two SNPs had a definite interaction within a haplotype to influence the risk of BA. This result suggests that *CD14* (rs2569190) may interact with *NOTCH2* (rs835576) and, together, downregulate the *NOTCH2* gene's expression. Our previous study showed that the maturation of BECs and the expression of Notch may play a role in the pathogenesis of BA, as well as the increased levels of immature BECs in patients with BA. However, newly formed immature bile ducts or ductules have no bile transport function, and bile may accumulate to form bile plugs. Together, these results suggest that *CD14* (rs2569190) may interact with *NOTCH2* (rs835576) and, together, downregulate *NOTCH2* gene's expression, which could result in immature and malfunction of BECs. The real-time PCR and western blot results showed that the expression of *CD14* and *NOTCH2* was significantly lower in BA patients compared to the control group. This further hinted that the genetic susceptibility locus of these two genes was associated with the risk of BA. Although the locations of *CD14* and *NOTCH2* in the BA liver were different, it might be related to hepatoblasts differentiated into BECs or hepatocytes in the Notch signaling pathway. These results were similar to the previous result of an epistasis test using PLINK.
However, several limits should be noted in our study. BA is a complex trait resulting from both environmental and genetic factors. Environmental factors and rare genetic variations associated with BA have not yet been identified, which limited further investigation of the gene-environment interactions. Epistasis power of the present study was examined through the Epistasis Power Calculator (<https://gump.qimr.edu.au/general/manuelF/epistasis/epipower4i.html>); based on the present sample size with the incidence rate of 1 per 10,000 infants, though the present sample size is large enough (0.98 for the case-only study, 0.97 for the case-control study), further replication in an independent cohort was still required. Replication studies from other cohorts with a larger sample size are encouraged to confirm the association. Also, the different location of *CD14* and *NOTCH2* in BA liver might be related to the Notch signal pathway and the differentiation and proliferation of hepatoblasts, but the association was not clear. Without functional experiments, it is difficult to determine whether these two SNPs are causally related to BA. Hence, cells and animal model experiments are needed to further explore the genetic function of associated interacting pairs as reported in a previous study.[@bib29], [@bib30], [@bib31]
In conclusion, the results of our study in a Chinese population verified that *CD14* (rs2569190) may interact with *NOTCH2* (rs835576) and, together, downregulate *NOTCH2* (rs835576) expression, which results in immature and malfunction of BECs in BA.
Materials and Methods {#sec4}
=====================
Study Subjects {#sec4.1}
--------------
We enrolled 506 BA cases and 1,473 controls in this hospital-based case-control study as we described previously.[@bib32] The candidate SNPs were selected upon the following criteria: SNPs rs2569190 and rs835578 were selected by the potential regulatory roles according to the GTEX portal database (<https://gtexportal.org/home/>). Furthermore, SNP rs2569190 is located on the promoter region of *CD14* and showed suggestive evidence as associated with BA.[@bib10] All subjects were recruited at the Guangzhou Women and Children's Medical Center, and the BA cases were confirmed by clinical diagnosis and pathology. 1,473 healthy children were randomly selected as the controls who had received a routine physical examination in the same hospital and matched to cases on age and gender. The exclusion criteria of the control were as follows: other types of liver diseases or autoimmune disease or children who received a liver transplantation. In addition, the liver tissues of 28 BA patients and 20 congenital CC patients (all less than 1 year old) were selected for further real-time PCR and western blot detection and comparison. This study was approved by the Institutional Review Board of the Guangzhou Women and Children's Medical Center, and the experimental process strictly abided by medical ethics. All participants have signed informed consent.
Polymorphism Analysis {#sec4.2}
---------------------
Total genomic DNA was isolated from tissue and peripheral blood samples using the TIANamp Blood DNA Kit (TianGen Biotech, Beijing, China) according to the manufacturer's protocol. SNPs (rs2569190, G/A; and rs835576, C/T) were successfully designed using the MassARRAY iPLEX Gold System (Sequenom). Moreover, 5% of samples were selected randomly for repeated assays, and the results were 100% concordant.
Real-Time qPCR {#sec4.3}
--------------
The mRNA expression levels of *CD14* and *NOTCH2* were performed by real-time PCR using the SYBR Green Supermix (Bio-Rad, 172-5124, Hercules, CA, USA). Data were collected and quantitatively analyzed on a Quant Studio 6 Flex System (Applied Biosystems, Foster City, CA, USA). The β-actin gene was used as an endogenous control to normalize for differences in total RNA in each sample. The primers were used as listed in [Table S1](#mmc1){ref-type="supplementary-material"}.
Western Blot Analysis {#sec4.4}
---------------------
Protein samples of patients' liver extracts were dissolved on 8% SDS-PAGE gels and transferred onto polyvinylidene fluoride (PVDF) membranes. The membranes were blocked with 5% BSA and incubated with CD14 (Abcam, ab106285, Cambridge, UK) and NOTCH2 (GeneTex, [GTX101593](ncbi-geo:GTX101593){#intref0025}, Irvine, CA, USA) antibodies at 4°C overnight. After three washes, the membranes were incubated with secondary antibodies at room temperature for 1 hr. After another three washes, the membranes were incubated with ECL Western Blotting Substrate for 8 min, and then the protein bands were visualized on X-ray films.
Statistical Analysis {#sec4.5}
--------------------
The allele frequency distribution included variants for the BA patients, and the controls were compared by χ^2^ test. p \< 0.05 was considered to be statistically significant. ORs and corresponding 95% CIs were calculated from the logistic regression model. Epistatic interaction between *CD14* (rs2569190) and *NOTCH2* (rs835576) was examined using PLINK[@bib33] (based on logistic regression analysis) on 506 cases and 1,473 controls. Higher order genetic interactions were detected and characterized by pairwise multifactor dimensionality reduction (MDR).
Author Contributions {#sec5}
====================
Z.L., H.L., Y.Z., and R.Z. designed and performed the study and wrote the manuscript. M.F., L.S., Y.T., X.X., H.C., H.W., and J.Z. collected the samples and information. Z.L., X.X., and Y.Z. participated in analyzing data. R.Z. and H.X. coordinated the study over the entire time. All authors reviewed the final manuscript.
Conflicts of Interest {#sec6}
=====================
The authors declare no competing financial interests.
Supplemental Information {#app2}
========================
Document S1. Table S1Document S2. Article plus Supplemental Information
This study was approved by the Institutional Review Board of the Guangzhou Women and Children's Medical Center. All the data involved in the study can be supplied upon request. This study was supported by the National Natural Science Foundation of China (grant nos. 81770510, 81771629, and 81671498) and the Science and Technology Planning Project of Guangdong Province 2017 (grant no. 2017A020214017). The authors would like to thank the Department of Pathology and Biobank, Guangzhou Women and Children's Medical Center, for providing the clinical samples.
Supplemental Information includes one table and can be found with this article online at [https://doi.org/10.1016/j.omtn.2018.10.006](10.1016/j.omtn.2018.10.006){#intref0030}.
[^1]: These authors contributed equally to the work.
[^2]: Function refers to the function role of SNP in the gene. CHR, chromosome; BP, base pair of where the SNP is located; A1/A2, minor allele/major allele; F_A, the minor allele (T) frequency in BA patients; F_U, the minor allele (T) frequency; OR, odds ratio; CI: confidence interval; p \< 0.05 was considered statistically significant; p_adj, p value adjusted by gender.
[^3]: The p value indicates the significance based on different genetic models. OR, odds ratio; CI, confidence interval; p \< 0.05 was considered statistically significant.
[^4]: The p value indicates the significance based on different genetic models. OR, odds ratio; CI, confidence interval; p \< 0.05 was considered statistically significant.
[^5]: The p value indicates the significance based on different genetic models.
| {
"pile_set_name": "PubMed Central"
} |
INTRODUCTION
============
Cervical cancer is a virus-induced disease that is caused by the integration of a human papilloma virus (HPV) DNA into the host\'s genome.[@B1]-[@B3] Infection with HPV causes disruption of the host\'s E2 gene, resulting in expression of viral oncogenes, E6 and E7. The E6 and E7 products inhibit the activities of tumor suppressors, p53 and retinoblastoma protein, respectively. This then eventually leads to the accumulation of damaged DNA and the development of cervical cancer.[@B4]-[@B5] Although cervical cancer is 1 of the causes of highest mortality in female cancer patients worldwide, it is a curable disease when diagnosed at an early stage.[@B6] However, clear prognostic factors for cervical cancer development are not yet in existence.
Mitochondria, which are cell organelles involved in the processes of cell life and death and in tumoral transformation, appear to have prominent dysfunction in cancer cells. Mitochondrial failure induces abnormal ultrastructures, deregulated metabolism, altered biochemistry and mutation of mitochondrial DNA (mtDNA) in cells.[@B7] Mitochondrial molecular chaperones play important roles in protein transport, protein complex assembly, refolding of misfolded proteins and triggering of protein degradation by proteosomes.[@B8]-[@B9] Heat shock proteins (HSPs) are molecular chaperones that are classified into families according to their molecular weight (i.e., HSP100, HSP90, HSP70, HSP60, and small HPSs). HSPs, an evolutionary conserved protein, are ubiquitous and have multiple functions in cellular homeo-stasis including gene expression regulation, DNA replication, signal transduction, differentiation, apoptosis, and cellular senescence or immortalization. They also protect cells from various stresses such as hypoxia or ischemia, as well as sudden increases in temperature.[@B10]-[@B12] In normal cells, HSP60 is mostly localized in the mitochondrial matrix and outer mitochondrial membrane, constitutively expressed under normal condition, and induced by heat shock, mitochondrial damage, and mtDNA depletion.[@B13],[@B14] Recently, other molecular roles for mammalian HSP60 have been reported. For example, human HSP60 may trigger apoptosis through caspase cascade activation by an association between HSP60/HSP10 complex and pro-caspase-3 inside the mitochondria, resulting in a subsequent release of the HSP60 into the cytoplasm.[@B15]
Overexpression of HSP60 has been reported in various tumors or cancers, such as adrenal Cushing tumors, human breast, large bowel, bronchial, exocervical, ovarian and prostate cancers.[@B16]-[@B23] Recently, an upregulation of HSP60 in preinvasive lesions of the cervix has been shown by immuhistochemistry.[@B24] Although HSP60 plays important roles in various biological events, the exact molecular roles of HSP60 are still poorly understood, and the relationship between HSP60 and invasive cervical cancer has not been reported yet. In this study, we investigated the HSP60 expression in invasive cervical cancer tissues and evaluated any prognostic significance of HSP60 in cervical cancer using proteomics, reverse transcriptase polymerase chain reaction (RT-PCR) and Western blot analyses.
MATERIALS AND METHODS
=====================
Tissue preparation
------------------
The cervical cancer patients were recruited at the outpatient clinic of the Department of obstetrics and gynecology, Gil Medical Center, Incheon, Korea. Tissue samples were collected from 20 cervical cancer patients (mean age, 47.95 ± 16.1 years) and 20 normal controls (mean age, 46.80 ± 12.9 years). Cervical biopsies were obtained from patients with a diagnosis of cervical cancer or uterine myoma as described previously.[@B25] The specimens were brought to clinicopathology department immediately after resection, and parts of the tissues were subsequently dissected, divided into several tubes, placed in liquid nitrogen and stored at -80℃ until analyzed. Cancer samples were characterized according to International Federation of Gynecology and Obstetrics (FIGO) stage as fellows: stage I, 6 case (30%); stage II, 10 case (50%); stage III, 2 case (10%); and stage IV, 2 case (10%). Histopathology of cancer samples was evaluated as fellows: squamous cell carcinoma (SCC), 16 cases (80%); adenocarcinoma, 3 cases (15%); and malignant mixed müllerian tumor (MMMT), 1 case (5%). Patients with cervical cancer underwent radiotherapy and/or chemotherapy after biopsy. Normal cervical tissues came from uterine myoma obtained from women diagnosed benign condition by Pap smear test. Our study was approved by the International Review Board on the experimental studies.
Two-dimensional gel electrophoresis
-----------------------------------
Biopsy tissues were homogenized in a R/S buffer {9 M Urea, 2% 3-\[(3-Cholamidopropyl) dimethylammonio\]-1-propanesulfonate, 50 mM dithiothreitol (DTT), 0.4% ampholyte} containing 0.5% each of protease and phosphatase inhibitor cocktail, and then ultracentrifuged at 100,000 g for 1 hour (h) at 20℃. Protein contents of the supernatants were determined using a protein assay kit (BioRad, Hercules, CA, USA), and then the samples were prepared at 1 mg/350 µL concentration with R/S buffer and stored at -80℃. One milligram (mg) of the protein sample was loaded on an immobilized pH gradient (IPG) strip (pH 4-7) in PROTEAN IEF CELL (BioRad, Hercules, CA, USA), and the strip was then covered with mineral oil and rehydrated at 50 V for 12 h. Moist wicks were placed at both ends of the strip to avoid salt contamination while the protein samples were separated by their isoelectric points (IP) at 10 KV for 15 h. The IPG strip was then soaked in equilibration buffer \[(375 mM Tris, 6 M Urea, 2% sodium dodecyl sulfate (SDS), 20% glycerol, 2.5% iodoacetamide)\] for 10 minutes (min), and then the samples which had previously been separated by IP were separated again on a 2-dimensional SDS-polyacrylamide gel (2D-PAG) \[4.0 mL of 30% acrylamide/Bis, 2.5 mL of 1.5 M Tris-HCl (pH 8.8), 0.1 mL of 10% SDS (w/v), 50 µL of 10% ammonium persulfate, 5 µL of N,N,N\'-tetramethylethylenediamine in 10 mL of 12% 2D-PAG\] by their molecular weights at 20 mA/gel for 16 h. Protein spots on 2D-gel were stained with Coomassie brilliant blue for 24-48 h,[@B26] de-stained in double distilled water (ddH~2~O), and then the stained protein spot images were analyzed using PDQuest software (BioRad, Hercules, CA, USA). The protein spots that showed an increased staining density of up to two fold compared with their controls were selected for further analysis.
Selected protein spots were individually excised into new tubes, and the Coomassie blue staining was destained with 50% acetonitrile/25 mM ammonium bicarbonate. Gel slices containing protein spots were treated with 10 mM DTT at 56℃ for 30 min, and with 55 mM iodoacetamide at room temperature for 25 min in dark. They were then digested with trypsin at 37℃ for 16 h with shaking. After pooling the supernatant into new tubes, peptide samples were dried completely using SpeedVac system for 4 h. Samples were dissolved in 50% acetonitrile/ 0.1% trifluoroacetic acid (TFA), co-crystallized by mixing with matrix (α-cyano-4-hydroxy-cinnamic acid saturated with 0.1% TFA/50% acetonitrile) and loaded on the silcon-coated 96 well microtiter sample plate (Applied Biosystems, Foster City, USA).
Peptides masses were analyzed using Voyager DE Matrix assisted laser desorption ionization-Time of flight mass spectrometry (MALDI-TOF) mass spectrometer (Applied Biosystems, Foster City, USA). Mass spectra were obtained by averaging 100 to 150 individual laser shots. Calibration of spectra was performed externally by 2 standard peptide, angiotensin 1 (m/z 1296.6853) and adrenocorticoptrophic hormone (18-39 clip) (m/z 2465. 1989). A database search for protein sequence for homology was performed using MS-Fit search algorithm (<http://prospector.ucsf.edu/prospector/4.0.8/html/msfit.htm>). Mass tolerance for the monoisotopic peptide masses were set to 50 ppm.
Proteomic analysis (i.e., two-dimensional gel electrophoresis and MALDI-TOF analysis) was carried out twice with different 2 samples in each group in order to confirm the peptide profiles analyzed.
Reverse transcription-polymerase chain reaction
-----------------------------------------------
Total RNA was extracted from cervical cancer or normal tissues with TRIzol reagent (Invitrogen, Carsbad, CA, USA) according to the manufacturer\'s instructions. The yield of total RNA was determined by measuring the absorption at 260 nm. Reverse transcription with 1 µg of total RNA was performed in a final volume of 20 µL using 200 U of Superscript II reverse transcriptase, 0.5 µg oligo dT~12-18~ as a primer, 0.5 mM dNTP Mix, 10 mM DTT, and first strand buffer (Invitrogen, Carsbad, CA, USA). The mixture of oligo dT~12-18~ primer, total RNA and diethyl pyrocarbonate (DEPC) treated ddH~2~O was heated at 70℃ for 10 min initially, and other components were then added and incubated at 42℃ for 1 h. Subsequent incubation at 70℃ for 15 min was carried to inactivate the reverse transcriptase.
To determine the relative levels of HSP60 messenger RNA (mRNA), PCR of complementory DNA (cDNA) was carried out in 20 µL mixture containing 1 µL of cDNA, 10X reaction buffer, 2.5 mM decxynucleoside triphosphate (dNTP) mix, 6 pmol each of 5\' and 3\' primer, and 1U of G-Taq DNA polymerase (Bioneer, Daejeon, Korea); Sense: 5\'-AGA TGG AGT GGC TGT GCT GA-3\', Antisense: 5\'-CAT CAT AAC CAA CTT CTG AG-3\'. The reactions were started at 94℃ for 5 min and amplified for 30 cycles of 30 sec at 94℃, 30 sec at 58℃ and 30 sec at 72℃. Final extensions were done for 7 min at 72℃ to compete polymerization. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as an internal control to confirm equal loading of the samples. The PCR products were separated on 2.5% agarose gel using NeSieve GTG agarose (FMC, Rockland, ME, USA), and analyzed by a digital image analysis system.
Western blot
------------
For Western blot analysis, cervical tissues were removed immediately, transferred to LN~2~ and stored at -80℃. The tissues were homogenized in a single detergent lysis buffer (50 mM Tris, pH 8.0; 150 mM NaCl; 1% Triton X-100; each of 0.5% protease and phosphatase inhibitor cocktail) and then centrifuged at 13,000 g for 20 min at 4℃. The supernatants transferred into new tubes were measured for protein contents using a protein assay kit (Biorad, Hercules, CA, USA), aliquoted at an 5 µg/20 µL concentration in lysis buffer, and stored at -80℃. Otherwise, they were used in the same day. The samples were mixed with loading buffer (100 mM Tris, pH 6.8; 200 mM DTT; 4% SDS; 20% glycerol; 0.2% bromophenol blue) at 1 : 1 dilution, boiled for 5 min, quickly chilled on ice, and then separated on 10% SDS-polyacrylamide Tris-glycine gels. The proteins were then transferred onto nitrocellulose membranes (Hybond-C, Amersham, Bucks, UK) and treated with 5% nonfat dry milk in 1X phosphate buffered saline-Tween (PBST) (1.46 mM NaH~2~PO~4~H~2~O; 8.05 mM Na~2~HPO~4~; 144.72 mM NaCl; 5% Tween 20) overnight at 4℃. The membranes were reacted with rabbit anti-HSP60 antibodies (Santa Cruz Biotechnology, CA, USA) at 1 : 1000 dilution for 1 h, and then reacted with HRP-conjugated goat anti-rabbit antibodies (1 : 1000 dilution; Amersham, Bucks, UK) for 1 h at room temperature. The bound antibodies were detected with chemilu-minescence according to the manufacturer\'s instructions (NEN Life Science, MA, USA), and quantified using a digital image analysis system.
Statistical analysis
--------------------
A one-way analysis of variance (ANOVA) was used to analyze the expression of HSP60 in cervical cancer and normal tissues. RT-PCR and Western blot data were subjected to post-hoc Fisher\'s protected least significant difference (PLSD) test and presented by means ± SEM. For all comparisons, the level of significance was set at *p* ≤ 0.05.
RESULTS
=======
We performed 2D proteomic analysis to evaluate the differences in protein expression between human cervical cancers and normal cervical tissues. Protein spots on the 2D gels were stained with Coomassie blue, and the staining densities of the spots were analyzed with PDQuest software. Proteomic analysis was carried out twice with different 2 samples from each group in order to confirm the peptide profiles analyzed and one of them is presented ([Figs. 1](#F1){ref-type="fig"} and [2](#F2){ref-type="fig"}). In 2D gel analysis of normal cervical tissue, 11 spots with higher staining densities then cancer tissue were selected for MALDI-TOF assay ([Fig. 1](#F1){ref-type="fig"}) and identified ([Table 1](#T1){ref-type="table"}). [Fig. 2](#F2){ref-type="fig"} reveals the 2D gel image of the cervical cancer tissue. Nineteen spots of cervical cancer tissue samples that showed up to 2 fold increased staining density compared with their normal control were selected and identified ([Table 2](#T2){ref-type="table"}). Specifically, HSP60 protein (Spot 6) which was overexpressed on cervical cancer gel displayed dramatically high 98.2% of sequence and corresponded to its intact molecular size (60 kDa) ([Table 2](#T2){ref-type="table"}).
In order to ascertain the increase of HSP60 expression in cervical cancer tissues, we investigated HSP60 mRNA expression in cervical cancer and normal cervical tissues obtained from twenty cancer patients and twenty normal controls, using semi-quantitative RT-PCR ([Fig. 3A](#F3){ref-type="fig"}). All samples were quantified and corrected for total input RNA by normalizing the expression value of GAPDH. Although HSP60 mRNA was detected in both groups, no statistically significant differences between cervical cancer and normal cervical tissues were noted ([Fig. 3B](#F3){ref-type="fig"}).
To examine the expression levels of HSP60 (60 kDa) protein, we carried out Western blot analysis in both groups ([Fig. 4A](#F4){ref-type="fig"}). Anti-actin (43 kDa) was used as an internal control and expression of HSP60 was normalized by actin protein. The results obtained from twenty cases of each group showed that the expression of HSP60 protein in cervical cancer tissues was significantly increased compared to cervical normal tissues (*p* \< 0.05) ([Fig. 4B](#F4){ref-type="fig"}).
DISCUSSION
==========
The immunohostochemical evaluation of HSP60 in precancer of the cervix has recently been reported,[@B24] however the relationship between HSP60 and invasive cervical cancer has not yet been reported. In this study, we compared expression levels of HSP60 mRNA or protein between invasive cervical cancers and cervical normal tissues.
2D gel proteomic analysis revealed that the expression of HSP60 was increased in cervical cancer tissues compared to normal tissues ([Figs. 1](#F1){ref-type="fig"} and [2](#F2){ref-type="fig"}). It is generally accepted that DNA damages which cause cancer induce cell cycle arrest, apoptosis or DNA mutation, resulting from changes in protein expression. In the present study, protein spots on the 2D gels showed different expression patterns between cervical cancer and normal cervical tissues, indicating that the protein expression of normal cervical tissues was changed by induction or inhibition of genes, and/or by degeneration or modification of proteins during carcinogenesis. HSP60 protein of cervical cancer tissues was markedly elevated, revealed by 2D gel proteomics and validated by Western blot ([Fig. 4](#F4){ref-type="fig"}). However, there was no difference in HSP60 mRNA levels between cancer and normal tissues ([Fig. 3](#F3){ref-type="fig"}). These findings suggest that serial processes after transcription become important step in cervical cancer development. Indeed, differences of HSP60 expression were detected in both 2D gels and Western blot, indicating difference in protein level, but not in mRNA level.
It has been known that HSP60 is constitutively expressed under normal conditions, and its expression is induced by stressful conditions such as heat shock, mitochondrial damage, and mtDNA depletion.[@B14] Although the expression of HSP60, which is localized in mitochondria, is directly related to mitochondrial regeneration after cell division or the increase of mitochondrial activity during normal conditions, and HSP60 is significantly less expressed in normal tissues.[@B13],[@B18],[@B27] On the other hand, the present study found that HSP60 was significantly overexpressed in cervical cancer tissues compared to normal cervical tissues. The overexpression of HSP60 in cervical cancer tissues suggests that this protein might play a different role in cervical carcinogenesis. Cappello et al.[@B22] reported that elevated expression of HSP60 may be a protective upregulation against cancer development (i.e. the blockade of apoptotic machinery that usually takes place during cancer progression). The involvement of HSP60 in the process of apoptosis and tumorigenesis is still in dispute. An antiapoptotic effect of HSP60 and down-regulation of HSP60 have been reported in cardiac myocytes and bronchial cancer, respectively.[@B28],[@B29] On the other hand, overexpression of HSP60 has been reported in breast and ovarian carcinomas and myeloid leukemia.[@B30]-[@B32] Moreover, recent studies showed upregulation of HSP60 during carcinogenesis of the large bowel and the uterine exocervix.[@B18],[@B22] Specifically, overexpression of HSP60 has been demonstrated in early cervical carcinogenesis as well as in prostate and squamous cervical cancers of both early and advanced grades.[@B18],[@B24],[@B27] These results support the possibility that the expression of HSP60 may play a role as a prognostic factor of cancer development. By employing immunohistochemistry, Castle et al.[@B24] showed up-regulation of HSPs in response to HPV infection and early cervical carcinogenesis, and discussed the role of HSP60 and HSP70 as surrogate marker for precancer development. However, the present study found the presence of HPV type 16 in 75% of our cervical cancer samples, demonstrated by PCR (data not shown) and they were shown to be invasive cancer stage I to IV, suggesting that HSP60 expression is also upregulated in invasive cervical cancer stages as well as precancer stage.
In conclusion, we compared HSP60 mRNA level and protein expression in cervical tissues diagnosed as normal and invasive cervical cancer, and confirmed upregulation of HSP60 protein in invasive cervical cancer tissues. It suggests that HSP60 is involved in cervical carcinogenesis, and can be used as a useful prognostic tool. Since the role of HSP60 in cancer progression and mechanisms involved in the regulation of HSP60 expression remain unclear, more studies are needed to elucidate the relationship between HSP60 and cancer development.
{ref-type="table"}. MALDI-TOF, matrix assisted laser desorption ionization-Time of flight mass spectrometry.](ymj-50-399-g001){#F1}
{ref-type="table"}. MALDI-TOF, matrix assisted laser desorption ionization-Time of flight mass spectrometry.](ymj-50-399-g002){#F2}
{#F3}
{#F4}
######
List of the Peptides Identified by MALDI-TOF Analysis in Human Normal Cervical Tissue
MALDI-TOF, matrix assisted laser desorption ionization-Time of flight mass spectrometry.
Eleven protein spots selected were individually analyzed. The spot number (Spot No.) corresponds to the position marked on the gel ([Fig. 1](#F1){ref-type="fig"}). Identified proteins and accession numbers were derived from the MS-Fit search algorithm.
^\*^MW (kDa) represents molecular size of protein.
^†^pI is an isoelectric point, where the net charge is zero.
^‡^D means the value obtained from the database.
^§^E indicates the value estimated from spots on the SDS gel.
^∥^Sequence coverage (%) is defined as the ratio of the portion of peptide sequence covered by the matched protein to the whole length of peptide sequence of the spot.
######
List of the Peptides Identified by MALDI-TOF Analysis in Human Cervical Cancer Tissue
MALDI-TOF, matrix assisted laser desorption ionization-Time of flight mass spectrometry.
Nineteen selected protein spots were individually analyzed. The spot number (Spot No.) corresponds to the position marked on the gel ([Fig. 2](#F2){ref-type="fig"}). Identified proteins and accession numbers were derived from the MS-Fit search algorithm.
^\*^MW (kDa) represents the molecular size of protein.
^†^pI is the isoelectric point, where the net charge is 0.
^‡^D means the value obtained from the database.
^§^E indicates the value estimated from spots on the SDS gel.
^∥^Sequence coverage (%) is defined as the ratio of the portion of peptide sequence covered by the matched protein to the whole length of peptide sequence of the spot.
| {
"pile_set_name": "PubMed Central"
} |
Plain Language Summary {#S0001}
======================
Screening for cervical cancer combined with vaccination against HPV infection, which is the main cause of cervical cancer, has the potential to eliminate cervical cancer in the future. However, only women accepting both vaccination and screening will gain full protection, while those declining both will gain little to no protection, leaving these women at higher risk of developing cervical cancer.
The purpose of this study was to examine the association between childhood HPV vaccination and cervical cancer screening, in a free-of-charge Danish setting.
Our results illustrate that women who were neither HPV vaccinated nor screened, were often non-native and had the lowest socio-economic status. Furthermore, women who were not HPV vaccinated were also more likely not to be screened; this pattern remained even when taking the influence of socio-economic status into account. However, this pattern was not similar for non-native women.
In conclusion, among native Danish women, non-attendance to HPV vaccination and non-participation in screening seem to be signs of generally poor health-preventive behavior. Where as among non-native non-western women, non-attendance in HPV vaccination and cervical cancer screening seem to be influenced by separate factors. Socio-cultural factors therefore likely influence HPV vaccination and cervical cancer screening behavior in native and non-native women in different ways.
Introduction {#S0002}
============
Organized cervical cancer screening has considerably reduced cervical cancer incidence and mortality in many western countries.[@CIT0001],[@CIT0002] Effective vaccines targeting the most oncogene human papilloma virus (HPV) types 16/18,[@CIT0003]--[@CIT0005] which cause \>70% of all cervical cancers,[@CIT0006] have been introduced in immunization programs worldwide since 2006.[@CIT0007] In Denmark, publicly funded, clinically based, three-dose HPV vaccination was introduced in the late 2008 and early 2009, and was at this time recommended to all 12--15-year-old girls born in 1993--1996.[@CIT0008]
The combination of organized cervical cancer screening and childhood HPV vaccination programs has the potential to eliminate cervical cancer. However, only women participating in both programs full protection, leaving a group of unvaccinated and unscreened women with little to no protection and thus at a higher risk of developing cervical cancer.[@CIT0009]
A large number of national and international studies have described that low socio-economic status and non-native background are associated with non-adherence to HPV vaccination[@CIT0010]--[@CIT0018] and non-participation in cervical cancer screening.[@CIT0009],[@CIT0019]--[@CIT0023] However, the relationship between socio-economic status and nativity has not been explored in the context of combined non-attendance in both HPV vaccination and cervical cancer screening.
Previous studies,[@CIT0024]--[@CIT0033] apart from one,[@CIT0034] have shown a positive association between HPV vaccination and cervical cancer screening participation. However, most of these studies have been conducted on women who were HPV vaccinated as young adults,[@CIT0029],[@CIT0031],[@CIT0034] though self-payment,[@CIT0024],[@CIT0029],[@CIT0032] or who were from highly selective[@CIT0025],[@CIT0027],[@CIT0030] or regional[@CIT0028],[@CIT0033] populations. It is thus unknown whether these previous results may be generalizable to the growing number of countries implementing national childhood HPV vaccination.
The aim of this study was to analyze if non-adherence to a childhood HPV vaccination program was associated with non-participation in cervical cancer screening program in the total population and stratified by country of origin and parental education status. Furthermore, the aim was to identify potential underlying socio-economic factors related to combined non-attendance.
Materials And Methods {#S0003}
=====================
Study Design {#S0003-S2001}
------------
A retrospective register-based nationwide-closed cohort study was conducted between 1 October 2008 and 31 December 2017.
Setting {#S0003-S2002}
-------
The study was carried out in Denmark whose population then counted approximately 5.7 million citizens.[@CIT0035] Denmark has two publicly financed national programs aiming to prevent cervical cancer: a childhood HPV vaccination program and a cervical cancer screening program.
Cervical cancer screening was introduced in Denmark in 1962 and reached national coverage in the late 1990s.[@CIT0036] In the Danish National Cervical Cancer Screening Program, women receive their first invitation when they are 23 years old and are subsequently invited every third year until the age of 49, while women aged 50--64 are invited to participate every fifth year. General practitioners (GPs) obtain a liquid-based cytology sample from the cervix during a gynecological examination. Women who do not respond to the screening invitation letter receive up to two personal reminders after 3 and 6 months, respectively. In case of non-participation, the woman receives a new invitation after 3 or 5 years, depending on her age.
HPV vaccination was implemented in the Danish childhood vaccination program in January 2009, targeting girls born in 1996. A few months earlier a start-up program targeting girls born 1993--1995 was also introduced. The HPV vaccination program is now recommended to all girls aged 12--18 years. The types of vaccine and vaccination dozes have varied over time. Parents have to consult their GP for vaccination of their daughter and if this is not done before the age of 14 years, they receive one reminder.[@CIT0008],[@CIT0037]
In the period from the time the HPV vaccine was released in 2006 until it was introduced in the childhood vaccination program almost 3 years later, all women had the opportunity to arrange HPV vaccination through self-payment (at this time approximately 500 USD).
Participants {#S0003-S2003}
------------
The study population comprises all women born in 1993 who were resident in Denmark during the entire study period. The 1993 birth cohort was chosen as it represents the ever first population offered both national childhood HPV vaccination and invited for the first cervical cancer screening through 2016--2017.
Women with a cervical cytology obtained before the age of 22.5 were excluded as these cytologies were most likely obtained due to symptoms rather than for screening purposes, potentially influencing future screening participation. We allowed a 6-month window prior to the screening invitation (age 23) for appropriate "early screening" as we assumed that these cytologies had been performed in connection with a benign gynecological examination (i.e., contraception consultation) close to the screening invitation and thus would not influence future screening participation.
Furthermore, to ensure that the exposure was restricted to free-of-charge childhood HPV vaccination, women who were HPV vaccinated outside the childhood vaccination program (self-paid vaccination) were likewise excluded as they could have a different health-promoting behavior than those vaccinated in the program.
Data Collection And Definitions {#S0003-S2004}
-------------------------------
The study population was defined in the Danish Civil Registration System,[@CIT0038] which contains information on all citizens born in or having immigrated to Denmark. In the Danish Civil Registration System, all citizens are registered with a unique ten-digit identification number, allowing direct and complete linkage to other national registries.
Data on HPV vaccination status were collected from the Danish National Health Service Register,[@CIT0039] which holds information on all tax-financed services. GPs obtain payment through a reimbursement system that includes information on the service provided and citizens receiving the service provided. Women with at least one HPV vaccination doze assigned a childhood HPV vaccination code (8328, 8329, 8330, 334, 8335, and 8336) between 1 October 2008 and the women's 22.5-year birthday were considered "vaccinated" (exposed). All other women were considered "unvaccinated" (unexposed).
Data on self-paid redeemed HPV vaccination were collected through the Danish National Prescription Registry.[@CIT0040] To identify these women, we used data on redeemed vaccine prescriptions holding the Anatomical Therapeutic Chemical Classification System codes of the two HPV vaccines available at the time (07BM01 and J07BM02). Women who were registered with at least one redeemed HPV vaccine prescription before they turned 22.5 years were defined as "self-paid vaccinated" and thus excluded from the main analyses.
Data on participation in the Danish National Cervical Cancer Screening Programme were collected from the Danish Pathology Register.[@CIT0041] The Danish Pathology Register holds individual pathology data from all pathology departments classified according to the Systematized Nomenclature of Medicine (SNOMED),[@CIT0042] which is a standardized glossary of clinical terminology used by healthcare providers for exchange of clinical health information. Cervical cytology samples were identified by the SNOMED codes T8X3\* and material type 23. Women with at least one registered cytology sample between the age of 22.5 and 24 were considered "screened." All women received a total of 18 months of follow-up. Those with no cervical cytology within this range were defined as "unscreened."
Based on the above definitions, four groups of "combined attendance" were defined by linking HPV vaccination status and screening participation: 1) vaccinated and screened (combined attenders), 2) vaccinated but not screened, 3) screened but not vaccinated, and 4) neither vaccinated nor screened (combined non-attenders). For the purpose of this study, the latter group was considered a high-risk group.
Six variables were used as proxy measures for socio-economic status (parental civil status, highest parental education, and occupation, family disposable income and area of residence, and country of origin). These data were obtained from Statistics Denmark, which offers a research service linking data from different health registers to socio-economic status data.[@CIT0043]
Socio-economic status data were obtained for the year 2009, which was the calendar year where the study population was eligible for HPV vaccination (time of exposure). Since socio-economic status data commonly used in scientific work (completed education level, occupation, income, and civil status) are not fully established at age 15,[@CIT0044] these variables were collected for the parents instead and supplemented with the participants' data on area of residence and country of origin.
Parental civil status was categorized as 1) married/cohabiting or 2) single parents. Parental educational level was defined as the highest completed education level by either parent and classified as 1) low (\<10 years), 2) middle (10--15 years), or 3) higher education (\>15 years).[@CIT0045] Family disposable household income based on the OECD-modified equivalence scale[@CIT0046] was used as an income measure. Based on tertiles, income was categorized as 1) low (lowest 33%), 2) middle (33--66%), or 3) high (highest 33%). Parental occupation was defined as the highest level of occupation by either parent and categorized as 1) working, 2) temporarily not working (including those receiving any kind of leave compensation, state education grant, or unemployment benefits), and 3) permanently not working (including early and ordinary retirement).
Area of residence was categorized according to the degree of urbanization as living in a densely, intermediate, or thinly populated area.
Native women were defined as those having Denmark as their country of origin and non-native women as those for whom Denmark was not their country of origin. Non-natives were furthermore sub-categorized according to their country of origin, into 1) western countries (EU, Andorra, Australia, Canada, Iceland, Liechtenstein, Monaco, New Zealand, Norway, San Marino, Switzerland, and the USA), and 2) non-western countries (all other countries).
Statistical Analyses {#S0003-S2005}
--------------------
Differences in sample socio-economic characteristics between the four "combined attendance groups" were tabulated and provided with 95% confidence intervals (CI) to allow comparison.
In the subsequent analyses, exposure was defined as non-adherence to HPV vaccination and outcome as non-participation in cervical cancer screening.
Logistic regression models were used to estimate the odds ratio (OR) of non-participation in cervical cancer screening with 95% CI comparing vaccinated to unvaccinated women. Unadjusted logistic regression model was performed along with a model adjusting for all six confounding socio-economic variables.
For these logistic regression models, three sensitivity analyses were performed to test the robustness of our models. Firstly, we included women screened before the age of 22.5 in the study population. Secondly, we included "self-paid" HPV-vaccinated women in the study population. Both analyses were done in order to test the hypothesis that these groups of women would have a different health-promoting behavior than the background population. Thirdly, we expanded the follow-up period by 6 months (increasing follow-up to 2 years) in order to examine whether the association found in the main analyses was present only during the relatively short follow-up period.
Finally, to determine if the association between HPV vaccination status and screening participation was modified by country of origin or parental education level, stratified and adjusted logistic regression models were used along with the Wald test for interaction.
All statistical analyses were conducted using STATA version 15 (STATA Corp., College Station, TX, USA).
Results {#S0004}
=======
A total of 31,228 women were eligible for study inclusion; a total of 6,400 (20.5%) were excluded, 1,828 (5.9%) due to self-paid HPV vaccination and 4,572 (14.6%) due to registered cervical cytology sample before the age of 22.5 years. This left 24,828 women in the main study population of whom 22,634 (91% \[95% CI: 90.8-91.5%\]) were vaccinated and 14,749 (59% \[95% CI: 58.8--60.0%\]) were screened. The mean age at HPV vaccination was 15.5 years \[95% CI: 15.4--15.5\], and the mean age at first cervical cancer screening was 23.4 years \[95% CI: 23.4--23.5\] (data not shown).
The majority of HPV-vaccinated women (13,893, 61.4% \[95% CI: 60.7--62.0%\]) participated in cervical cancer screening, whereas only 856 (39.0% \[95% CI: 36.9--41.1%\]) unvaccinated women participated in cervical cancer screening ([Figure 1](#F0001){ref-type="fig"}).
[Table 1](#T0001){ref-type="table"} shows that 55.9% \[55.3--56.6\] of the total study population was both HPV-vaccinated and screened, while 5.4% \[5.1--5.6\] was neither HPV-vaccinated nor screened. Furthermore, HPV-vaccinated and screened women (combined attenders) had more often married (58.1% \[57.3--58.8\]), working (58.1% \[57.4--58.7\]) parents with higher education (61.4% \[60.3--62.5\]) and income (65.9% \[64.7--67.1\]), and lived in thinly populated area (59.4% \[58.4--60.4\]) and had native Danish background (59.6% \[59.0--60.3\]) than the three other combined attendance groups.Table 1Socio-Economic Characteristics Of The Study Population And Their Parents, Stratified By Combined Attendance In The Two Preventive Programs Against Cervical CancerVaccinated ScreenedVaccinated Un-ScreenedUn-Vaccinated ScreenedUn-Vaccinated Un-ScreenedN ; % row (95% CI)13,893; 55.9 (55.3-56.6)8,741; 35.2 (34.6-35.8)856; 3.4 (3.2-3.6)1,338; 5.4 (5.1-5.6)**Socoeconomic Variables (n, % col)n% row95% CIn% row95% CIn% row95% CIn% row95% CIParental civil status ^a^ ** Married/cohabiting (18,762, 75.6%)10,891**58.1%(57.3-58.8)**6,479**34.5%(33.9-35.2)**563**3.0%(2.8-3.3)**829**4.4%(4.1-4.7)** Single (6,050, 24.4%)3,002**49.6%(48.4-50.9)**2,262**37.4%(36.2-38.6)**293**4.8%(4.3-5.4)**493**8.2%(7.5-8.9)** Missing (16, 0.1%)0\--0\--0\--16**100.0%(79.4-100.0)Individual area of residence ** Densely populated (5,832, 23.5%)2,914**50.0%(48.7-51.3)**2,282**39.1%(37.9-40.4)**228**3.9%(3.4-4.4)**408**7.0%(6.4-7.7)** Populated at intermediate (9,468, 38.1%)5,329**56.3%(55.3-57.3)**3,360**35.5%(34.5-36.5)**311**3.3%(2.9-3.7)**468**4.9%(4.5-5.4)** Thinly populated (9,512, 38.3%)5,650**59.4%(58.4-60.4)**3,099**32.6%(31.6-33.5)**317**3.3%(3.0-3.7)**446**4.7%(4.3-5.1)** Missing (16, 0.1%)0\--0\--0-** **16**100.0%(79.4-100.0)Individual country of origin ** Denmark (n= 22,271, 89.7%)13,282**59.6%(59.0-60.3)**7,240**32.5%(31.1-33.1)**760**3.4%(3.2-3.7)**989**4.4%(4.2-4.7)** Western country (n=225, 0.9%)95**42.2%(35.7-49.0)**84**37.3%(31.0-44.0)**20**8.9%(5.5-13.4)**26**11.6%(7.7-16.5)** Non-western country (2,316, 9.3%)516**22.3%(20.6-24.0)**1,417**61.2%(59.2-63.2)**76**3.3%(2.6-4.1)**307**13.3%(11.9-14.7)** Missing (16, 0.1%)0\--0\--0\--16**100.0%(79.4-100.0)Parents' highest education ** Higher (8,152, 32.8%)5,006**61.4%(60.3-62.5)**2,558**31.4%(30.4-32.4)**276**3.4%(3.0-3.8)**312**3.8%(3.4-4.3)** Middle (12,805, 51.6%)7,290**59.9%(56.1-57.8)**4,530**35.4%(34.5-36-2)**374**2.9%(2.6-3.2)**612**4.8%(4.4-5.2)** Low (3,644, 14.7%)1,541**42.3%(40.7-43.9)**1,549**42.5%(40.9-44.1)**195**5.4%(4.6-6.1)**359**9.9%(8.9-10.9)** Missing (226, 0.9%)56**24.8%(19.3-30.9)**104**46.1%(39.4-52.8)**11**4.9%(2.5-8.5)**55**24.3%(18.9-30.5)Family income ** High (6,102, 24.6%)4,019**65.9%(64.7-67.1)**1,734**28.4%(27.3-29.6)**191**3.1%(2.7-3.6)**158**2.6%(2.2-3.0)** Middle (9,845, 39.7%)5,895**59.9%(58.9-60.8)**3,320**33.7%(32.8-34.7)**263**2.7%(2.4-3.0)**367**3.7%(3.4-4-1)** Low (8,866, 35.7%)3,979**44.9%(43.8-45.9)**3,687**41.6%(40.6-42.6)**402**4.5%(4.1-5.0)**798**9.0%(8.4-9.6)** Missing (15, 0.1%)0\--0\--0\--15**100.0%(79.4-100.0)Parents' highest occupation ** Working (22,789, 91.8%)13,228**58.1%(57.4-58.7)**7,750**34.0%(33.4-34.6)**754**3.3%(2.1-3.5)**1,057**4.6%(4.4-4.9)** Temporarely not working (1,264, 5.1%)379**30.0%(27.5-32.6)**637**50.4%(47.6-53.2)**59**4.7%(3.6-6.0)**189**15.0%(13.0-17.0)** Permenatly not working (590, 2.4%)208**35.3%(31.4-39.3)**279**47.3%(43.2-51.4)**35**5.9%(4.2-8.2)**68**11.5%(9.1-14.4)** Missing (185, 0.8%)78**42.2%(35.0-49.6)**75**40.5%(33.4-48.0)**8**4.3%(1.9-8.0)**24**13.0%(8.5-18.7)**[^1]
On the other hand, un-vaccinated and un-screened women (combined non-attenders) had more often single (8.2% \[7.5--8.9\]), temporarily not working parents (15.0% \[13.0--17.0\]) with low education (9.9% \[8.9--10.9\]) and income (9.0% \[8.4--9.6\]), and lived in densely populated areas (7.0% \[6.4--7.7\]) and had non-native background (non-western 13.3% \[11.9--14.7\] and western 11.6% \[7.7--16.5\]).
After adjusting for socio-economic status, unvaccinated women had 2.1-fold higher odds of not participating in cervical cancer screening than vaccinated women (adjusted OR = 2.1\[95% CI: 1.9--2.3\]). The same adjusted model revealed that non-native women from non-western countries had 3.6-fold higher odds of not participating in cervical cancer screening than native women (adjusted OR=3.6 \[95% CI: 3.2--4.0\]) ([Table 2](#T0002){ref-type="table"}).Table 2Unadjusted And Adjusted Odds Ratios With 95% Confidence Intervals (CIs) For Not Participation In Cervical Cancer ScreeningOdds Ratio95% CI**Unadjusted model ** HPV vacinations status** ** VaccintedReference Un-vaccinated2.5(2.3-2.7)**Adjusted model^a^** HPV vacinations status VaccintedReference Un-vaccinated2.1(1.9-2.3) Socio-economic factors^b^ Parental civil status Married/cohabitingReference Single1.0(0.9-1.1) Individual area of residence Densley populatedReference Intermediate populated0.9(0.9-1.0) Thinly populated0.8(0.8-0.9) Individual country of origin DenmarkReference Western countries1.4(1.0-1.8) Non-western countries3.6(3.2-4.0) Parents highest education HighReference Middle1.1(1.0-1.2) Low1.3(1.2-1.5) Family income HighReference Middle1.2(1.2-1.3) Low1.5(1.4-1.6) Parents higest occupation WorkingReference Temporarely not working1.3(1.3-1.5) Permenantly not working1.2(1.0-1.4)[^2]
In the three sensitivity analyses, we found that our conclusions did not change if we included women with cervical cytology before 22.5 years, "self-paid" vaccinated women, and expanding the follow-up period.
After stratifying by country of origin, unvaccinated native women had 2.2-fold higher odds of not participating in cervical cancer screening than vaccinated native women. Western-unvaccinated women had 1.3-fold higher odds and non-western unvaccinated 1.5-fold higher odds of not participating in cervical cancer screening than vaccinated women (adjusted ORs of 2.2 \[95% CI: 2.00--2.4\], 1.3 \[95% CI: 0.6--2.8\] and 1.5 \[95% CI: 1.1--2.0\], respectively). The test for homogeneity revealed a p-value of 0.019.
After stratifying by parental education level, unvaccinated women with highly educated parents had 1.9-fold higher odds of not participating in cervical cancer screening than vaccinated, those with middle-educated parent had 2.3-fold higher odds, and those with low-educated parents had 1.8-fold higher odds (adjusted ORs of 1.9 \[95% CI: 1.6--2.3\], 2.3 \[95% CI; 2.0--2.6\], and 1.8 \[95% CI: 1.5--2.2\], respectively). The test for homogeneity, however, revealed a p-value of 0.141 ([Table 3](#T0003){ref-type="table"}).Table 3Models Testing The Interaction By Country Of Origin (Model 1) And Highest Parental Education (Model 2) On The Association Between Non-vaccination And Cervical Cancer Screening Non-participationScreening Non-ParticipationTotal nAdjusted ORs95% CIp-value^a^Interaction model 1 HPV vaccination & country of oigin Denmark22,2712.22.02.4{#ILF0001}Western countries2251.30.62.80.019Non-western countries2,3161.51.12.0Interaction model 2 HPV vaccination & parents' highest edducation High8,1521.91.62.3{#ILF0002}Middle12,8062.32.02.60.141Low3,6441.81.52.2[^3]
Discussion {#S0005}
==========
Main Findings {#S0005-S2001}
-------------
This nationwide cohort study found that the best-protected women (combined attenders) belonged to the highest socio-economic status groups and were mostly native women, while the least protected women (combined non-attenders) belonged to the lowest socio-economic status groups and were mostly non-native women from non-western countries. Furthermore, it revealed that non-adherence to childhood HPV vaccination was associated with non-participation in cervical cancer screening. Thus, even after adjusting for socio-economic factors, unvaccinated women had higher odds of not participating in cervical cancer screening than vaccinated women.
The association between non-adherence to vaccination and non-participation in screening was stronger for natives than for non-natives from both western and non-western countries, indicating that natives and non-natives may encounter different barriers for HPV vaccination and cervical cancer screening.
Strengths And Limitations {#S0005-S2002}
-------------------------
A major strength of this study is the linkage of high-quality individual data on HPV vaccination, cervical cancer screening participation, and socio-economic status. This linkage eliminates the risk of differential misclassification, e.g., caused by recall bias or social desirability bias in self-reported data. The choice of a closed cohort study design gave us complete follow-up data; however, our study accounted only for citizens living permanently in Denmark over a longer period.
Our study is the first published study to use a strong register-based design with nationwide data to link associations between HPV vaccination status in a free-of-charge childhood HPV vaccination program and later organized cervical cancer screening participation adjusted for individual and parental socio-economic status.
Our results are thus relevant to the current and future situation in Denmark as well as in other countries with organized cervical cancer screening programs implementing childhood HPV vaccination programs. However, the present cohort represents women with a high HPV vaccination adherence of 91%, which is higher than what has later been reported in Denmark for younger cohorts and what has been seen in other countries.[@CIT0047],[@CIT0048] Thus, it may be relevant to further investigate "combined attendance" in cohorts with lower HPV vaccination adherence.
Overall, when interpreting our results, it is important not to perceive the odds ratios (OR) presented in the result section as an expression of relative risk of non-participation in screening, as this interpretation would lead to an overestimation of the results. Caution is also advised when interpreting the stratified estimates among different non-native groups, as both western and non-western women were small in sample size, leading to OR estimates with broad and overlapping 95% confidence intervals. It is possible that the estimates for western and non-western women shown in [Table 3](#T0003){ref-type="table"} would have been different if the sample size had been larger.
The lack of data on indication for collecting cervical cytology was a limitation, as cytology registrations did not distinguish between cytologies taken for screening purpose and cytologies taken on medical indication. However, by excluding women with cytologies prior to screening invitation (\<22.5 yrs.), we believe that we excluded the group most likely to consist of symptomatic women from the main analysis. This minimizes the risk of information bias in our definition of outcome.
Another limitation was the lack of information on reasons for not participating in cervical cancer screening. In light of the age at first screening invitation (23 years), some women in our cohort could likely have postponed participation in screening due to pregnancy, for example, as screening during pregnancy is not recommended in Denmark. Thus, some women would falsely be defined as unscreened, when, in fact, they participated in screening at a later time according to clinical recommendations. However, relatively few women resident in Denmark are pregnant at 23 years[@CIT0049] as the mean age at first pregnancy is 29.2 years.[@CIT0050] However, non-native women from non-western countries have a slightly higher fertility (0.2 live births pr. woman higher than native women) than native women.[@CIT0043] This could mean that potentially more non-western women than native women have been misclassified as non-participants in screening in our results. Expanding the follow-up window with another 6 months did, however, not alter our main results.
Lastly, it was a limitation that our dataset had no information on non-native women, especially those from western countries, attending HPV vaccination or screening in their country of origin. Many western countries currently provide HPV vaccination and screening programs.[@CIT0051] Therefore, for western women, it is possible that the combined non-attendance is actually lower than seen in our study. In contrast, it is not likely that non-western women attend vaccination or/and screening in their home countries, and the results regarding these women would therefore be accurate despite the lack of this information.[@CIT0052]
Furthermore, in order to prevent any misclassification on HPV vaccination status that could have occurred due to errors in registration, as demonstrated in a Danish study concerning measles, mumps, and rubella vaccination coverage,[@CIT0053] we included all vaccination codes available for free HPV vaccination, and thus marked the first given vaccination code as the first given HPV vaccination.
Other Studies {#S0005-S2003}
-------------
Our main finding regarding the association between HPV vaccination and cervical cancer screening participation was in line with findings in previous studies.[@CIT0024]--[@CIT0033] Only one Australian study[@CIT0034] from 2014 found that vaccinated women had a lower screening participation than unvaccinated women. This particular study did not link HPV vaccination to screening participation at the level of the individual, which may explain the conflicting result compared to other studies. It is, however, also possible that the organization of HPV vaccination in Australia (school-based opt-out program) could weaken the association to opt-in screening. However, it seems that in this particular study, participants were above the age for participation in a school-based vaccination program and thus were vaccinated mostly through a community-based opt-in catch-up program, like in many of the other studies showing a positive association between vaccination status and screening participation.
Our study demonstrated that age at HPV vaccination did not alter subsequent screening participation patterns, as our results on childhood HPV-vaccinated women mirror results reported by previous studies conducted on varied and often older populations.[@CIT0024],[@CIT0025],[@CIT0027]--[@CIT0033] Screening behavior patterns were consistent whether the decision to be HPV-vaccinated was made by the women themselves or by their parents. It is therefore likely that a person's health-promoting behavior may be shaped during childhood and persists throughout adulthood.
In a Swedish nationwide cohort study conducted in 2018, Kreusch et al[@CIT0024] found that compared to unvaccinated women, opportunistically HPV-vaccinated women often had higher educational level, higher income, and were natives. In our study, this social inequality in health-promoting behavior was further demonstrated across both preventive programs available against cervical cancer. Thus, in our study, well-protected women (combined attenders) had the highest socio-economic status and native background, while unprotected women (combined non-attenders) had the lowest socio-economic status and non-native background.
Kreusch et al also found that educational level and native status interacted with HPV vaccination and cervical cancer screening participation.[@CIT0024] Contrary to our results, Kreusch et al demonstrated proportionally increased attendance in HPV vaccination and cervical cancer screening with increasing level of education and a stronger association between HPV vaccination and cervical cancer screening among non-native women than among native women.
In our interaction analyses, the level of education was not an effect modifier. Moreover, the association between HPV vaccination status and screening participation was stronger for native women than for non-native women. Methodological differences in the two studies (i.e., categorization of non-natives or algorithms for HPV vaccination or cervical cancer screening) could potentially explain why education and native background played a different role in the two studies. However, the most likely reason for this difference is the highly selected population in the Swedish study, in which only 13.6% were HPV vaccinated and, moreover, had paid for their vaccination themselves. It is likely that vaccinated women in this study, especially non-natives, had a particularly active health-promoting behaviour and this may explain the opposing results in the two studies.
Furthermore, in line with our interaction results regarding the influence of native background on HPV vaccination and subsequent screening, two recent Danish studies showed both lower attendance and lower effect size of offering HPV self-sampling to non-native women than to native women,[@CIT0054],[@CIT0055] supporting that health-promoting behavior may be influenced by nativity/region of origin.
Among some non-natives, factors other than socio-economic status seem to influence attendance pattern in HPV vaccination and cervical cancer screening. Besides barriers in language and lack of knowledge, lower screening participation among non-native women from the Middle-East and South Asia has been linked to socio-cultural barriers such as fatalism, fear of cancer, embarrassment, modesty, and perceiving HPV infection as related to promiscuous behavior.[@CIT0056],[@CIT0057]--[@CIT0060]
Some of these barriers are difficult to overcome, but lack of knowledge may be overcome through educational interventions and by improving women's perception on their own disease susceptibility, their perception of disease severity, and finally their perception of benefits weighed against barriers toward attending, besides improving their understanding of the healthcare system.[@CIT0061]
It therefore could be beneficial to consider adding differentiated interventions for natives and non-natives in order to equalize participation across all populations.
Conclusion {#S0006}
==========
It seems that among native Danish women, especially those with lower socio-economic status, non-adherence to HPV vaccination and non-participation in screening are both signs of a generally poor health-preventive behavior pattern. However, among non-western women, it seems that non-adherence in HPV vaccination and non-participation in cervical cancer screening are influenced by unrelated factors. It is likely that socio-cultural factors influence non-western women's HPV vaccination and cervical cancer screening behavior in different ways than native women. Therefore, a differentiated approach to native and non-native women is needed to enhance overall cervical cancer preventive behavior across different nativities. Figure 1Inclusion and exclusion flow-chart for the study population, including screening participation according to HPV vaccination status.
Data cleaning and initial analyses were performed with assistance from data manager Bo Søborg at the Department of Public Health Programmes, Randers Regional Hospital. This study was funded by the Family Hede Nielsen's Foundation and Helsefonden. The sponsors had no influence on the scientific process.
Ethics Approval And Informed Consent {#S0007}
====================================
According to Danish legislation and the Central Denmark Region Committees on Biomedical Research Ethics, the study did not require ethical approval because it was based on register data. The same institutions waived patient consent for use of register data. In accordance with Danish law and the EU's General Data Protection Regulation, the project was listed at the Central Denmark Region internal list of research projects (J. No.: 1-16-02-400-16).
Author Contributions {#S0008}
====================
All authors made substantial contributions to both design, achievement of data, analysis, and interpretation of data; took part in drafting the article or revising it critically; gave final approval of the version to be published; and agree to be accountable for all aspects of the work.
Disclosure {#S0009}
==========
Andersen B has received HPV test kits and HPV self-sampling devices from Roche and Axlab for other studies. Andersen B reports grants from Helsefonden, grants from Familien Hede Nielsens fond, during the conduct of the study; non-financial support from Roche, non-financial support from Axlab, outside the submitted work. Pedersen LK has received speaker's fee from Merck and Sanofi Pasteur in relation to lectures on HPV vaccines. The authors report no other conflicts of interest in this work.
[^1]: **Notes:** ^a^The parent(s) with whom the child lives. The purpose of the bold values is to emphasis the % values and their 95% CI.
[^2]: **Notes:** ^a^Odds ratios are adjusted for parental civil status, highest parental educational level, highest parental occupation, family disposable income, area of residence, and country of origin. ^b^Socio-economic factors used in the adjusted model, with OR for each variable's association with non-participation in cervical cancer screening.
[^3]: **Notes:** ^a^Wald test for interaction shows if the association between non-vaccination and non-screening is modified by country of origin and parental educational level.
| {
"pile_set_name": "PubMed Central"
} |
1. Introduction {#sec1}
===============
Multiple Sclerosis (MS) is a chronic autoimmune disease of the Central Nervous System (CNS) causing inflammation and neurodegeneration. The characteristic demyelination of the neurons, followed in many cases by axonal loss and gliosis, results in incapacity progression \[[@B1]\]. The incapacity status is usually assessed in MS patients with the Expanded Disability Status Scale (EDSS) \[[@B2]\]. Different types of MS have been described based on the clinical course: Relapsing-Remitting Multiple Sclerosis (RRMS), Secondary Progressive Multiple Sclerosis (SPMS), Primary Progressive Multiple Sclerosis (PPMS), Relapsing-Progressing Multiple Sclerosis (RPMS) \[[@B3]\] and Clinically Isolated Syndrome (CIS) \[[@B4]\].
There is yet no cure for MS, but several disease modifying treatments have shown to have beneficial effects on the disease progression. The approved treatments for MS in Europe include Interferon *β*-1a SC and IM, Interferon *β*-1b SC, Glatiramer Acetate SC, Natalizumab IV, oral Fingolimod, oral Dimethyl Fumarate, oral Teriflunomide, and Alemtuzumab IV.
MS has a great social and economic impact, because it affects mainly young adults \[[@B6]\], who will be more prone to unemployment \[[@B7]\] and in a big percentage will need a walking aid few years after the disease onset \[[@B8], [@B9]\]. MS is more prevalent in females \[[@B10]\] and Caucasians \[[@B11]\] and it is influenced by the geographic localization (Multiple Sclerosis International Federation).
Our aim is to contribute to the MS epidemiological knowledge in Portugal, describing the patients\' epidemiological, demographic, and clinical characteristics.
2. Methods {#sec2}
==========
The district of Braga is located in the Northwestern part of Portugal. In the district of Braga, there are two hospitals with consultation and treatment of multiple sclerosis for a resident population of 866,012 inhabitants (INE). The hospitals are Hospital de São Marcos in the city of Braga (HSM) and Hospital Senhora da Oliveira in the city of Guimarães (HSO). Braga is the youngest district of Portugal: 30.7% of the population is under 25 years, 56.4% are between 25 and 64 years, and 12.9% have more than 64 years. The population of this district is mostly lower middle class (which covers 37.1% of the population).
This is a cross-sectional study of MS patients followed in the Braga and Guimarães Hospitals. The data was collected from the clinical records, and 31/12/2009 was established as the prevalence day. At the prevalence date, only Interferon *β*-1a SC and IM, Interferon *β*-1b SC, Glatiramer Acetate SC, and Natalizumab IV were approved for MS in Europe. All the cases were assessed by a neurologist and included in the study whenever they met the Poser diagnostic criteria. Only residents of Braga district which had an office visit between the 1st of January 1997 and the 31st of December 2009 were included in this study. We requested the necessary consent to patients, Portuguese Data Privacy Committee (Comissão Nacional de Proteção de Dados), and to the Ethical Committees of both hospitals. There were no clinical records for HSM patients before 1997. In order to compare the hospitals, the year of diagnosis for HSO patients was considered to be 1997 for every patient that had records older than that.
In this study, for all MS patients, demographic characteristics (sex, age, age at diagnosis, treatment, and spatial distribution of place of residence) and clinical outcome (type of the disease, treatment, clinical severity score \[EDSS\], and year and month of diagnosis) are reported. Student\'s *t*-test and Chi-squared test were used to evaluate differences between groups. All significance tests were based on *P* \< 0.05 as the level of significance.
3. Results {#sec3}
==========
The database is composed of 345 individuals that are followed in the MS outpatient clinic of HSM and HSO. The estimated disease prevalence is 39.84 per 100,000 inhabitants (95% CI = 27.47--52.21). The female : male ratio is 1.79, with 64.06% of females MS patients and 35.94% males.
The mean age at diagnosis is approximately 35 years and varies between 13.61 and 70.55 years ([Table 1](#tab1){ref-type="table"}). At prevalence day, the mean age is 42.
HSM and HSO have a similar distribution of patients over the RRMS and SPMS forms. The number of patients with PPMS and CIS forms is too low be compared between hospitals ([Table 2](#tab2){ref-type="table"}).
The most common treatment is Interferon *β*-1b IM, with 29.86%, followed by IFN*β*-1a SC 22 *μ*g with 24.64%, Glatiramer Acetate with 17.97%, IFN*β*-1a SC 44 *μ*g with 12.17%, IFN*β*-1a IM with 7.25%, Natalizumab IV with 5.51%, and Mitoxantrone 0.29%. 2.32% of patients are without treatment ([Figure 1](#fig1){ref-type="fig"}).
Patients were divided in three groups according to their disability as quantified by the EDSS. Most patients have an EDSS of 3 or less ([Figure 2](#fig2){ref-type="fig"}).
In HSM, there were no MS patient records before 1997. HSO MS patients have records for previous years, but we decided to consider 1997 as the year of diagnosis in those cases. If we assume 866,012 as the Braga district population from 1998 to 2009, we can estimate an average annual incidence of 2.74/100,000 inhabitants in the Braga district ([Figure 3](#fig3){ref-type="fig"}).
4. Discussion {#sec4}
=============
MS patients in Portugal are almost exclusively followed in public hospitals because Disease Modifying Treatments (DMTs) are almost only offered to patients without any copayment in the National Health System (NHS) hospitals. As a result, only a very small number of patients, at national level, have access to DMTs outside the NHS hospitals. HSM and HSO are the only MS reference centers in the district of Braga, so it is expected that the vast majority of MS patients of this district be registered in one of these two hospitals. There is the possibility that some MS patients from this district are followed in other MS centers in Portugal, so the estimated prevalence may be underestimated. To more accurately determine the MS prevalence in this district, further work should be done. A cross-check between the hospitals and the primary care centers patient\'s lists would give the most accurate prevalence numbers. Additionally, MS patient\'s lists, from hospitals like Centro Hospitalar de São João, Centro Hospitalar do Porto, and Centro Hospitalar e Universitário de Coimbra should be checked to capture patients with residence in Braga district, who are followed in those hospitals.
For several years, Portugal has been considered to be a low-medium prevalence zone for MS \[[@B14]\], but recent epidemiological studies suggest that our country should instead be considered a medium prevalence zone \[[@B15], [@B16]\]. A prevalence study in the district of Santarém \[[@B16]\] determined a prevalence of 46.3 cases/100,000 in 2006 and a nationwide survey presented in 2011 by J. Pinheiro found this number to be 54/100,000 \[[@B17]\]. More recently, a prevalence study in a referral hospital and three community health clinics in a large city (Lisboa) determined a prevalence of 56.2/100,000 \[[@B18]\]. The 39.84/100,000 prevalence that we estimated in this study is probably underestimated due to the exclusion of patients who are resident in Braga district but are followed in other hospitals from cities like Oporto or Coimbra. The 2.74/100,000 annual incidence is lower than that reported for Spain and Italy \[[@B19]\], but this value is probably underestimated for the same reason and also because we considered a constant population (equal to 2009 numbers) between 1998 and 2009.
The gender ratio F : M in this study is 1.79. In both hospitals, the highest number of diagnosis was in the year 2008 (excluding 1997) with 11.88% of the patients. The MS diagnostic of HSO patients was made at a younger age. The age at diagnosis, patient distribution amongst MS forms, and disability status assessed by EDSS is also consistent with the results from previous MS studies \[[@B8], [@B19]\]. The low percentage of patients without treatment, 2.32%, is noteworthy.
5. Conclusion {#sec5}
=============
The district of Braga is a medium prevalence zone for MS with a prevalence of 39.82/100,000 inhabitants. We estimated an average annual incidence of 2.74/100,000 inhabitants between 1998 and 2009. In this study, MS patients\' demographic and clinical characteristics are consistent with what was expected in light of current scientific literature.
This research was partially supported by the Research Centre of Mathematics of the University of Minho through the FEDER Funds "Programa Operacional Fatores de Competitividade, COMPETE" and by the Portuguese Funds through FCT "Fundação para a Ciência e Tecnologia" within the Project Est-C/MAT/UI0013/2011.
MS:
: Multiple Sclerosis
CNS:
: Central Nervous System
EDSS:
: Expanded Disability Status Scale
RRMS:
: Relapsing-Remitting Multiple Sclerosis
SPMS:
: Secondary Progressive Multiple Sclerosis
PPMS:
: Primary Progressive Multiple Sclerosis
RPMS:
: Relapsing-Progressing Multiple Sclerosis
CIS:
: Clinically Isolated Syndrome
HSM:
: Hospital de São Marcos
HSO:
: Hospital Senhora da Oliveira
DMTs:
: Disease Modifying Treatments
NHS:
: National Health System.
Highlights
==========
\(i\) 345 MS patients followed in Braga district and (ii) patient\'s demographic characteristics (sex, age, age at diagnosis, treatment, and spatial distribution of place of residence) and clinical outcome (type of the disease, treatment, clinical severity score \[EDSS\], and year and month of diagnosis) were collected.
Conflict of Interests
=====================
José Figueiredo, Ângela Silva, João J. Cerqueira, Joaquim Fonseca, and Paulo A. Pereira declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interests.
{#fig1}
{#fig2}
{#fig3}
######
Age at diagnosis.
Minimum Maximum Mean Median Std. deviation 1st Q 3rd Q
------- ----------- ----------- ----------- ----------- ---------------- ----------- -----------
Total **13,61** **70,55** **35,38** **34,91** **11,58** **26,09** **42,24**
HSM 14,82 70,55 36,09 35,84 11,75 26,98 44,41
HSO 13,61 65,83 33,84 33,22 11,10 25,72 39,85
The mean age at diagnosis is 35.
######
Most of patients have the Relapsing-Remitting form of Multiple Sclerosis.
MS form HSM HSO Total
--------- --------- -------- --------- -------- --------- --------
CIS 2 0,85% 1 0,91% 3 0,87%
RRMS 190 80,85% 96 87,27% 286 82,90%
SPMS 37 15,74% 12 10,91% 49 14,20%
PPMS 6 2,55% 1 0,91% 7 2,03%
Total **235** **110** **345**
[^1]: Academic Editor: Herbert Brok
| {
"pile_set_name": "PubMed Central"
} |
Background {#Sec1}
==========
Headache disorders, including migraine, are extremely common \[[@CR1]\]. From a public-health perspective, they are also among the most disabling at population level: according to the Global Burden of Disease (GBD) study, headache disorders collectively are the third highest cause in the world of years of healthy life lost to disability (YLDs), migraine alone being sixth (third in those aged under 50 years) \[[@CR2]--[@CR5]\]. The lost-productivity and consequential financial costs are enormous \[[@CR6]\].
It might be expected that headache disorders would, everywhere, be considered important: as a personal medical problem by people directly affected by them, and as a public-health priority by health-care providers and health policy-makers. The reality is different. People with headache are under-diagnosed and undertreated not only in poorer countries with limited resources and restricted access to medical care but also, evidence indicates, in wealthy Europe and North America \[[@CR7], [@CR8]\]. We sought to verify this in Europe by analysing data from the Eurolight study.
Eurolight was an initiative supported by the European Commission Executive Agency for Health and Consumers (EHAC), and a partnership activity within the Global Campaign against Headache \[[@CR9], [@CR10]\] conducted by *Lifting The Burden* (LTB), a UK-registered non-governmental organization in official relations with the World Health Organization \[[@CR11]\]. Eurolight gathered data on headache disorders in a cross-sectional survey in 10 countries, which together represented \> 60% of the adult population (18--65 years) of the European Union (EU): Austria, France, Germany, Ireland, Italy, Lithuania, Luxembourg, Netherlands, Spain and United Kingdom (UK) \[[@CR12]\]. The survey included diagnostic enquiry regarding primary headaches and further enquiry into headache-attributed burden and utilisation of medical services and medication for headache. Here we present data on consultations and use of migraine-specific acute and preventative medications, and analyse these as indicators of adequacy of medical care for people with migraine in Europe.
Methods {#Sec2}
=======
Ethics {#Sec3}
------
The National Ethics Committee of Luxembourg gave overall approval of the protocol and provisions for data protection. Further approvals were obtained from national and/or local ethics committees wherever needed, since the methods for recruitment of participants differed between countries. In every country, prospective participants received written information explaining the project and its purpose.
Study design {#Sec4}
------------
The methods of the Eurolight project have been described in detail elsewhere \[[@CR12]\], and are summarised here.
Eurolight was a cross-sectional questionnaire-based survey of adults (18--65 years) in the EU, sampling from 10 countries. The sampling and questionnaire-distribution methods, summarised in Table [1](#Tab1){ref-type="table"}, varied between countries according to what was feasible \[[@CR12]\]: in six countries (Germany, Italy, Lithuania, Luxembourg, Netherlands, Spain), samples were population-based; in three (Austria, France, UK), general practitioners (GPs) recruited consecutive patients attending for any reason. Additional samples in Spain and Netherlands, and the only sample in Ireland, were recruited through lay organisations \[[@CR12]\].Table 1Summary of sampling and data collection methods in each country \[adapted from reference \[[@CR12]\]CountryDenominator (n)Responders (n)Responder proportion (%)Gender (% female)Target population and mode of distribution of questionnaireStudies with a general-population basisGermany300033811.357Random general-population sample from urban and rural areas, contacted by regular postItaly350050014.358Stratified general-population sample from urban and rural areas, contacted by regular postLithuania113761654.259General-population sample in and around Kaunas (urban and rural), contacted by door-to-door cold-calling and personally interviewed by trained medical studentsLuxembourg6498202331.158Stratified general-population sample contacted by regular postNetherlands-populationunknown2414incalculable50Stratified general-population sample contacted by internetSpain-workplace170099958.859Stratified sample of postal services employees, contacted by internal post by occupational health physiciansStudies conducted in general practice settingsAustriaunknown, but not \> 6000646incalculable70Consecutive patients consulting GPs or neurologists for any reason; questionnaire handed directlyFrance240087636.568Consecutive patients consulting GPs for any reason; questionnaire handed directlyUK72012817.8\*65Consecutive patients attending GPs for any reason; questionnaire handed directlyStudies among members of lay headache organizationsIrelandmembers 1500 relatives unknown195\
7313.0\
incalculable66Members of Migraine Association of Ireland and their non-biological relatives, contacted by regular postNetherlands-laymembers 500 partners unknown337\
11567.4\
incalculable57Random sample of members of Nederlandse Vereniging van Hoofdpijnpatiënten and (where existing) their non-headache-affected partners, contacted by regular postSpain-lay30027290.762Members of Asociacion Española de Pacientes con Cefalea (AEPAC) and their family; questionnaires distributed by hand via helpers of AEPAC
The survey used the same structured questionnaire in all countries \[[@CR13]\]. It had multiple parts. Demographic enquiry included age, gender, marital status and socio-economic status. Screening questions for headache (with an enquiry timeframe of the preceding year) were followed, in those responding positively, by headache-diagnostic questions based on ICHD-II \[[@CR14]\]. To avoid diagnostic confusion, participants identifying more than one headache type were asked to report only on the one that was most bothersome to them. Enquiry into health-care utilisation entailed questions on use for headache of acute and preventative medications, consultations for headache (yes or no) with nurse, GP, neurologist or headache specialist, investigations for headache (MRI, CT, X-rays of the neck, blood tests, ophthalmic examination), and admissions (number) to hospital because of headache.
Analysis and statistics {#Sec5}
-----------------------
Diagnoses were made from questionnaire responses by computerized algorithm \[[@CR15]\]. This first identified, and separated, participants reporting headache on ≥15 days/month, of whom additional questions had enquired into frequency of acute medication use. Probable medication-overuse headache (pMOH) was diagnosed when, in addition, simple analgesics were used on ≥15 days/month or medication including compound analgesics, opioids, triptans and/or ergots was taken on ≥10 days/month. A diagnosis of pMOH trumped all other diagnoses. The remainder of this group were diagnosed as "other headache on ≥15 days/month". To all others, who had headache on \< 15 days/month, the algorithm applied ICHD-II criteria for definite migraine, definite tension-type headache (TTH), probable migraine and probable TTH in that order.
Analysis focused on participants in whom migraine was diagnosed, not distinguishing between those meeting criteria for definite migraine or probable migraine \[[@CR12], [@CR15], [@CR16]\]. We identified those with migraine on \> 5 days/month as clearly eligible for preventative medication. We selected the following as indicators of adequacy of care: a) proportion receiving migraine-specific acute medications (triptans); b) proportion of those clearly eligible receiving any preventative medication; c) proportions receiving medical care through GP or specialist (neurologist or specialist in headache medicine). In our analysis, consultation with specialist(s) trumped consultation with GP(s).
Statistical analyses were performed using SPSS. We describe categorical variables in terms of frequency (n) and proportions (%), and continuous variables in terms of means ± standard deviations (SDs). We used 2 × 2 chi-squared to compare proportions receiving treatments in specialist and primary care.
Results {#Sec6}
=======
Eurolight collected 9247 correctly completed questionnaires from participants in the 10 countries (mean age 43.9 ± 13.9 years, M/F ratio 1:1.4 \[Table [1](#Tab1){ref-type="table"}\]). The principal results have been reported in detail previously \[[@CR17]\]. Here we analysed the data of 3466 people (37.6%) with migraine (definite or probable) (Table [2](#Tab2){ref-type="table"}).Table 2Sociodemographic features of participants with migraine (*N* = 3466)CountrynFemales (%)Age (years) (mean ± SD)Living with spouse or partner (%)Employed or self-employed (%)Studies with a general-population basisGermany10967.941.6 ± 11.95765Italy22170.140.2 ± 11.79471Lithuania14980.542.5 ± 12.36766Luxembourg66967.938.2 ± 11.27369Netherlands-population81560.240.5 ± 12.56869Spain-workplace40168.641.7 ± 11.17183Studies conducted in health-care settingsAustria26381.444.2 ± 13.57467France33776.043.1 ± 13.57864UK49100.0%41.7 ± 16.06463Studies among members of lay headache organizationsIreland15288.848.5 ± 13.17159Netherlands-lay19583.146.5 ± 10.97865Spain-lay10676.440.8 ± 11.07184
The demography of participants with migraine was broadly similar in all countries: they were on average approximately 40 years old, two thirds or more were female, most were married or living with a partner, and most were employed or self-employed (Table [2](#Tab2){ref-type="table"}). Across countries, one third (1175; 33.8%) reported frequent migraine (\> 5 days/month) (Table [3](#Tab3){ref-type="table"}) and therefore had clear need of preventative medication.Table 3Utilization of medical care by participants with migraine (N = 3466)CountryNUsing triptans (of all with migraine) n (%)Migraine on ≥5 days/month n (%)Using preventative medication (of those with migraine on ≥5 days/month) n (%)Consulting health professionals n (%)SpecialistGeneral practitionerNon-medicalNoneStudies with a general-population basisGermany10912 (11.0)42 (38.5)1 (2.4)7 (6.4)14 (12.8)5 (4.6)83 (76.1)Italy22114 (6.3)61 (27.6)1 (1.6)14 (6.3)21 (9.5)15 (6.8)171 (77.4)Lithuania1495 (3.4)62 (41.6)2 (3.2)16 (10.7)23 (15.4)2 (1.3)108 (72.5)Luxemburg66948 (7.2)219 (32.7)10 (4.6)39 (5.8)105 (15.7)27 (4.0)498 (74.4)Netherlands-population81575 (9.2)171 (20.8)11 (6.4)25 (3.1)106 (13.0)36 (4.4)648 (79.5)Spain-workplace40190 (22.4)153 (38.2)21 (13.7)60 (15.0)72 (18.0)28 (7.0)241 (60.1)Studies conducted in general practice settingsAustria26337 (14.1)106 (40.3)7 (6.6)46 (17.5)26 (10.0)14 (5.3)177 (67.3)France33746 (13.6)91 (27.0)4 (4.4)7 (2.1)81 (24.0)22 (6.5)227 (67.3)UK4912 (24.5)22 (44.9)2 (9.1)11 (22.4)7 (14.3)0 (0)31 (63.3)Studies among members of lay headache organizationsIreland15294 (61.8)78 (51.3)23 (29.5)34 (22.4)45 (29.6)6 (3.9)67 (44.1)Netherlands-lay195133 (68.2)120 (61.5)50 (41.7)66 (33.8)26 (13.3)21 (11.0)82 (42.1)Spain-lay10649 (46.2)50 (47.2)8 (16.0)25 (23.6)27 (25.5)10 (9.4)44 (41.5)
Utilisation of migraine-specific medications varied greatly between countries. The ranges were, for triptans, 3.4--68.2% of all participants with migraine and, for preventative medications, 1.6--41.7% of those deemed clearly eligible for them. These ranges were considerably distorted by the studies among members of lay headache organisations. In population-based samples, 3.4--22.4% of participants with migraine used triptans and 1.6--13.7% of those eligible used preventative medication. In this group, Spain-workplace was an outlier, with values in each case double those of the next highest, while Lithuania appeared to be an outlier for low use of triptans (3.4%) and Italy for low use of preventative medication (1.6% of those eligible) (Table [3](#Tab3){ref-type="table"}). Proportions were higher in GP-based samples (13.6--24.5% using triptans, 4.4--9.1% on preventative medication), and much higher among those from lay organisations (46.2--68.2% and 16.0--41.7%) (Table [3](#Tab3){ref-type="table"}).
Proportions of people receiving specific acute or preventative medications did not depend an age or socio-economic or marital status (data not shown).
These medications were not widely available without prescription; therefore, they were accessible only to a limited extent by people treating themselves or consulting a nurse rather than a physician. With regard to triptan use, not only was this highly dependent on consultation with a physician but also the proportions using triptans were greater in those seeing specialists (mean 51.3%) than in those treated by GPs (mean 35.9%; chi-squared = 62.1; *p* \< 0.001) (Fig. [1](#Fig1){ref-type="fig"}). Similarly, those clearly eligible to receive preventative medications were more likely to do so from specialists (mean 26.0%) than from GPs (mean 14.1%; chi-squared = 10.1; p \< 0.001) (Fig. [2](#Fig2){ref-type="fig"}). Participants with migraine who had consulted specialists (2.1--33.8% across all studies) were receiving the best care by these indicators; those treated by GPs (9.5--29.6%) fared less well, and the larger numbers dependent on self-medication (48.0--84.2%) appeared to be inadequately treated. Probability of consultation was itself highly dependent on the source of the sample (Table [3](#Tab3){ref-type="table"}). In the studies with a general-population basis, a minority of participants (15.8--33.0%) had done so: 3.1--15.0% had consulted a specialist (averaged across countries: 6.8%), and 9.5--18.0% had seen a GP (average: 14.4%). In studies conducted in general practice settings, 26.1--36.7% of participants had consulted, and in studies among members of lay headache organizations, 47.1--52.0% had done so, with the ranges for specialist and GP as shown in Table [3](#Tab3){ref-type="table"}. The differences reflect bias in the latter samples.Fig. 1Proportion (%) of participants with migraine using triptans, by country and by whom consulting. NL-pop: Netherlands-population; Spain-work: Spain-workplace; NL-lay: Netherlands-layFig. 2Proportion (%) of eligible participants with using migraine-preventative treatment, by country and by whom consulting. NL-pop: Netherlands-population; Spain-work: Spain-workplace; NL-lay: Netherlands-lay
Discussion {#Sec7}
==========
EU countries represent a relatively wealthy area of the world, albeit with some variation. Our study in 10 of these countries confirms that, even among these, migraine is under-treated. At best in the general-population samples (setting aside Spain-workplace), 26.1% of those with migraine in Lithuania were consulting a physician for it, triptans were in use by 11.0% in Germany, and preventative medication by 6.4%, of those who appeared to be clear candidates for it, in Netherlands. At worst, the proportions were as low as 15.8% consulting (in Italy), 3.4% using triptans (in Lithuania) and 1.6% of those eligible using preventative medication (again in Italy). The Spanish workplace-based sample fared better on all counts -- 33.0%, 20.4% and 13.7%, suggesting facilitated access to health care among this sample of employed people. Contact with GPs, and with specialists more so, was associated with higher use of these medications, which is unsurprising not only because they are in the main prescription-only but also because worse-affected people might be more likely to consult.
The samples recruited by general practitioners, and more so those from lay organisations, showed higher proportions consulting and utilising specific medications, which we attribute to the expected selection bias in these two groups. The latter did better for use of triptans (60.9% of those with migraine), but, even among these, fewer than one third (32.7%) of those clearly eligible were using preventative drugs.
While the proportions in contact with doctors and receiving migraine-specific drugs in this study were self-evidently low, at issue here is what, in a perfect world, they should be. In a resource-limited world, not everyone with migraine can receive professional health care, but neither should they: people with relatively mild and infrequent attacks who self-manage adequately with over-the-counter (OTC) remedies should not be brought into the health-care system. Published recommendations are that about 50% of people with migraine can adequately self-manage \[[@CR18]\], suggesting that the other 50% need to consult at some level. To a large extent, need for consultation is driven by the inefficacy of the OTC medications and the need for prescription drugs. Assessed by the outcome measure *sustained headache relief* (SHR), arguably an adequate if imperfect outcome \[[@CR19]\], aspirin 900--1000 mg has a predicted efficacy of 39% (headache relief at 2 h \[HR\] 52%) \[[@CR20]\] and ibuprofen 400 mg of 45% (HR 57%) \[[@CR21]\]. For paracetamol 1000 mg there are no data for SHR but HR = 56% \[[@CR22]\].
What our findings indicate is that, even in these relatively well-resourced countries, perhaps only half of those who would be expected to benefit from physician-consultation actually received it. Of those who did, proportions seen by GPs (9.5--18.0% in the population-based groups) and specialists (3.1--15.0%) were not very dissimilar. It is true that those seeing specialists might also be seeing GPs since, in such cases, we recorded only the former. But the question raised is this: is it necessary, or cost-effective, for migraine care to be thus distributed between specialists and GPs? Recommendations are that only the small minority of complicated cases should be referred from primary care to specialists \[[@CR7], [@CR18]\].
Furthermore, the quality of that care, when it was received, appeared far from perfect. If need for triptans is a reason to consult physicians, at least the majority of those doing so might be expected to receive them. While this was barely achieved (51.3%) among those seeing specialists, only 35.9% did so from GPs. Worse, while use of preventative medication by people with \> 5 migraine days/month ought by any objective standard to be close to 100%, the best we saw outside the self-selecting lay-organization members was 13.7%, and this was in an employee group with, probably, facilitated access to care.
The picture is therefore not encouraging: as reflections of reach and adequacy of headache services for headache, these findings indicate depressingly poor performance in the EU. The situation may even be worse than suggested. On the subject of bias, the Eurolight project as a whole suffered from low participation proportions -- on average, 27.5% in the non-lay samples \[[@CR17]\]. Some participation bias was likely, with the probability that, through self-selection, samples were preferentially constituted of more demanding participants, whose needs and health-care utilisation were likely to be higher than average.
Several earlier studies have found much the same. In Germany, among 7431 adults, awareness of migraine was low among those who had it, as was recognition of it by health-care providers \[[@CR23]\]. Also in Germany, in three regions of the country, a population-based study of 10,000 people found only 8% of those with migraine used triptans and only 2.3% received preventative treatment, both positively associated with socio-economic status \[[@CR24]\] and suggesting inequitable access to health care. Similar results emerged from France: among approximately 10,000 people studied, about 60% of the 1179 with migraine were unaware of the diagnosis, only 20% used triptans and only 2.3% received preventative medication \[[@CR25]\]. In Italy, a study of 2675 patients in 10 headache centres revealed that only 26.8% with migraine were previously correctly diagnosed, only 17.2% were using triptans and only 4.8% were using migraine-specific preventative medications \[[@CR26]\]. In Sweden, a recent cross-sectional study in Stockholm analysed data from a pharmaceutical registry. It found that, of patients with a diagnosis of migraine and a high attack frequency (self-reported), and utilizing triptans for acute treatment, only 4% received preventative medication \[[@CR27]\].
Some limitations of the study should be considered. First, Eurolight recruited from 10 EU countries with diverse sampling methods: some samples were population-based, others were general-practice-based and some were recruited through lay organisations \[[@CR12]\]. A degree of interest-bias was very likely \[[@CR17]\]. If this led to recruitment of more seriously affected patients, arguably our findings of under-treatment appear in even worse light. There was consistency in these findings, with more adequate medication among responders from lay groups, who were likely to be more vocal in their demands. Patients seeing specialists received better care than those treated by GPs, but important is that medical care was insufficient in *all* groups. A second limitation was that the cross-sectional Eurolight survey, while observing variations between countries, could not enquire into the reasons for under-treatment \[[@CR12]\] (which might include differences in costs, health-seeking behaviour, culture and tradition, and structure and accessibility of health-care systems).
The last 10 years have been crucial in improving knowledge of the prevalence and impact of migraine, beginning in 2007 with a systemic review of the existing literature \[[@CR28]\]. The gaps in knowledge revealed by this review have been progressively filled by population-based studies undertaken by the Global Campaign against Headache \[[@CR9], [@CR10], [@CR29]\]. The GBD studies have ranked migraine as the sixth highest cause of disability worldwide, third in both men and women aged under 50 years \[[@CR4], [@CR5]\]. The Eurolight study found that nearly one fifth of males and over a quarter of females with migraine reported the loss of \> 10% of productive days \[[@CR17]\]. It also demonstrated that the burden of migraine was not confined to attacks: there was measurable interictal burden also \[[@CR30]\]. The estimated financial costs to the EU are huge \[[@CR6]\]. Migraine remains to a very large extent to be untreated despite all this, and despite that migraine is a very treatable disorder \[[@CR7], [@CR31], [@CR32]\]. This is a failure of health care with major adverse health and economic consequences.
While some access to a wider range of drugs is achieved by consulting GPs, and rather more by seeing headache specialists, this is emphatically not a call for everyone with migraine to see specialists. Rather, it is a plea to curtail the insouciance to which migraine appears condemned \[[@CR8]\]. We identify four needs, largely to be met by education at multiple levels.
First, people with migraine should learn, through public health-education programmes, that migraine is a neurobiological disease that can often be effectively treated with correct usage of OTC drugs \[[@CR7], [@CR31], [@CR32]\]. Here, pharmacists can help, but otherwise people with migraine should consult a GP. Second, in order to relieve an otherwise insupportable load on specialists, health-care providers in general and GPs in particular need better knowledge of how to recognise, diagnose and treat migraine (along with the small number of other headache disorders that are of public-health importance) \[[@CR18], [@CR32]\]. Without this, the potential benefits of consulting GPs will be frustrated. This better knowledge will improve usage of available treatments, produce better outcomes, avoid wastage \[[@CR33]\] and, importantly, reduce overall costs \[[@CR7]\]. Third, headache services need to be structured, so that they might be delivered countrywide, efficiently and equitably to the very large number of people who stand to benefit from them \[[@CR7], [@CR18]\]. And fourth, and most important, is the urgent need for political recognition not only that the problem exists but also that it demands remedial action \[[@CR3], [@CR5], [@CR7]\].
The alternative is that large numbers of people will remain without diagnosis or best treatment, with not much hope for change, their unmitigated personal disability burdens translating into lost productivity with reduced societal output reflected in gross domestic product. The drive to produce new drugs and devices for headache \[[@CR34], [@CR35]\] offers little apparent utility if, when developed, they will not reach most patients \[[@CR36]\]. The choice between these alternatives -- invest in headache services, or do nothing -- is, surely, not a difficult decision.
Conclusion {#Sec8}
==========
In wealthy European countries, too few people with migraine consult physicians, with proportionately too many of these seeing specialists, and migraine-specific medications are used inadequately even among those who do. These findings represent yet another call for action in Europe to improve care for people with headache. Education of both health-care providers and the public should be central to this action.
Part-funding for this study was received from the European Agency for Health and Consumers of the European Commission. Financial support was also provided by *Lifting The Burden*.
We are grateful to the following who contributed to the conduct of Eurolight and/or data collection: Colette Andrée (project leader), Jose Miguel Lainez, Michel Lantéri-Minet, Daiva Rastenyte, Elena Ruiz de la Torre, Cristina Tassorelli and Lars Jacob Stovner.
ZK, CL and TJS contributed to project design and development of the methodology. ZK and CL contributed to data acquisition. Analysis and data interpretation were performed by MM, JH and ZK. The article was drafted by MM and ZK and revised for intellectual content by CL, JH and TJS. All authors reviewed and approved the final manuscript.
Competing interests {#FPar1}
===================
TJS and ZK are directors and trustees of *Lifting The Burden*, a charitable organization whose purpose, pursued in official relations with the World Health Organization, is to reduce the burden of headache worldwide. There were no conflicts of interest relating to the content of this manuscript.
Publisher's Note {#FPar2}
================
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
| {
"pile_set_name": "PubMed Central"
} |
INTRODUCTION
============
In large introductory biology classes throughout the United States, multiple-choice (MC) formats typify both formative assessments (e.g., clicker questions, concept inventories) and summative tests (e.g., midterm and final exams; see Wood \[[@B52]\] and Smith *et al.* \[[@B46]\]). While there is little doubt among educators that MC formats in general are capable of providing cost-effective, reliable, and valid inferences about student knowledge and misconceptions in many content areas, not all types of student learning outcomes may be measured using MC formats (reviewed in American Association for the Advancement of Science \[AAAS, [@B1]\] and Nehm *et al.* \[in press\]). Moreover, despite generating useful assessment information, MC tests may also produce unintended, and rarely considered, negative consequences for learners, such as the generation of false knowledge ([@B22]; [@B44]; Butler *et al.*, [@B8]; Kang *et al.*, [@B18]). Additionally, many MC tests are most conducive to detecting novice *or* expert (incorrect or correct) models of student thinking, whereas a large body of work in cognitive science indicates that many students construct mixed models of naïve and informed scientific information as they learn (e.g., Vosniadou \[2008\]; Opfer *et al*. \[[@B41]\]); right *or* wrong options---the staple of MC tests---may limit the valid measurement of student learning gains ([@B38]; [@B29]; Neumann *et al.*, [@B40]).
Collectively, these and many other limitations of MC formats should motivate biology educators to 1) develop and deploy a more diverse array of high-quality assessment methods and 2) measure a more expansive range of student knowledge, skills, and learning outcomes (e.g., AAAS \[[@B2], p. 17\]; Nehm *et al.* \[in press\]). The purpose of our study was to investigate the prospects and limitations of implementing new assessment methods in introductory biology---specifically, computerized scoring of short-answer scientific explanations. Can successful application of these innovative methods at one university be generalized, and if not, why not? What are the implications for adopting computerized-scoring programs as assessment tools in introductory biology and biology education research?
BACKGROUND
==========
Open-Response Assessments in Biology Education
----------------------------------------------
Educational researchers emphasize that assessments should be built upon and aligned with what we know about student learning and cognition (National Research Council \[NRC\], [@B26]). One major advance in our understanding of student learning is that learners do not progress directly from novice to expert levels; rather, the pathways of knowledge growth in biology (and other domains) are highly inefficient and involve integrating scientific ideas into naïve knowledge frameworks, generating heterogeneous mental models, or building coexisting models ([@B38], [@B39]; [@B49]; [@B19]; [@B13]). These so-called mixed or synthetic models may persist for long periods of cognitive development ([@B49]) and even through the years of college instruction ([@B36]).
If our biology assessments are intended to measure progress in student reasoning and build upon findings from educational research, our assessment items (whether formative or summative, MC or open response) must permit at least three general reasoning categories as options: 1) exclusively naïve answer choices; 2) assemblages of mixed or synthetic answer choices; and 3) exclusively scientific answer choices. Currently, many MC diagnostic and summative biology assessments (and concept inventories) contain option types 1 and 3 (novice and expert, respectively), despite mounting evidence in some biology domains that *most* students harbor "mixed models" of biological concepts ([@B29]). Thus, MC formats (correct *or* incorrect options) appear to be discordant with what we know about how students learn science (NRC, [@B26]; [@B38]; [@B49]). Thus, one advantage of open-response assessments is that they allow students to assemble heterogeneous knowledge elements and thereby reveal student thinking at a much more fine-grained level than "novice versus expert" assessments.
Perhaps the strongest argument for the inclusion of written assessments in introductory biology is that MC formats are not capable of measuring many desired learning outcomes for introductory biology courses (AAAS, [@B2]); a diversified assessment portfolio is needed to comprehensively capture students' learning progress (Corcoran *et al.*, [@B11]). Indeed, increasing emphasis has been directed at building assessments that mirror authentic, real-world tasks, not just those that are easily measured (NRC, [@B26]; Nehm *et al.*, in press). Many science education policy documents, for example, emphasize the importance of having students generate and evaluate scientific explanations (e.g., *Benchmarks for Science Literacy* \[AAAS, [@B1]\]; the *National Science Education Standards* \[NRC, [@B25]\]; *Taking Science to School* \[Duschl *et al*., [@B12]\]; and *Vision and Change in Undergraduate Biology Education* \[AAAS, [@B2]\]; Braaten and Windschitl \[[@B6]\]). The ability to generate scientific explanations can only be assessed using open-response formats.
One final argument for the inclusion of open-response assessments in introductory biology is that they better align with most real-world experiences than do MC formats. Increasingly, college graduates are expected to perform nonroutine tasks that cannot be automated, digitized, or outsourced (Nehm *et al.*, in press). From an educational point of view, deploying assessment tasks that model authentic problem-solving environments would help reinforce for students which types of performances are most highly valued by biology educators, and which types of evaluations they are likely to experience postgraduation (e.g., production vs. selection tasks). Overall, while MC assessments should remain in biology educators' assessment toolboxes, the many advantages of open-response formats call for their greater inclusion. Practical limitations have prevented the wider use of open-response assessments, but recent technological advances are beginning to change this situation.
Computer-Assisted Scoring Tools
-------------------------------
The increasing use of computer-assisted scoring (CAS) in many educational contexts has been motivated by the numerous constraints that characterize human scoring of constructed-response (e.g., short-answer, essay) items. Some of the most obvious limitations are the large amounts of time, money, and expertise needed to score such responses, and the consequent delayed feedback to test takers (Nehm *et al.*, in press). A more serious issue with human scoring of written responses is the persistent problem of grading subjectivity and consequent reliability threats; such problems are often introduced by the need for many different human graders to score large data sets, such as those generated in undergraduate biology courses (Nehm and Haertig, in press). Moreover, differently trained graders often disagree about the scores that should be given to a response, requiring additional training time to equalize scoring among raters. Reliable and consistent scoring of constructed-response items cannot be solved by having one human grader score all of the responses; grading fatigue and changes in scoring precision are well-known limitations of human scoring (Nehm and Haertig, in press). Thus, many long-standing problems have limited the use of open-response formats.
Fortunately, the rapid pace of developments in computer technology and text analysis software has made CAS tools more economical and accessible to educators. Consequently, many of the aforementioned limitations of human scoring have been investigated empirically using a variety of different software tools. This work has demonstrated that computer software can be "trained" to score constructed-response items as accurately and reliably as human raters ([@B42]; Yang *et al.*, [@B53]; [@B45]). Indeed, the Educational Testing Service and many other large companies now employ CAS methods in large-scale, high-stakes, standardized exams (Powers *et al.*, [@B43]). Examples of these CAS tools include C-rater ([@B47]), E-rater ([@B7]; [@B50]), and Intelligent Essay Assessor (Landauer *et al.*, [@B20]). Less work has focused on using these or similar tools for formative assessment purposes. Our research group has used two tools to perform CAS at a smaller scale, specifically grading short-answer responses within classes at individual universities: SPSS Text Analytics (SPSSTA; Urban-Lurain *et al.*, 2010; Nehm and Haertig, in press) and the Summarization Integrated Development Environment (SIDE; Nehm *et al.*, in press). SPSSTA is a commercial software package sold by a private company (IBM), whereas SIDE is a freely available software package distributed by Carnegie Mellon University ([@B23],b). Although the performance efficacy of both computer programs has been demonstrated using samples of undergraduate science students ([@B16]), the two programs differ in the ways in which they approach CAS, as well as in the methods used to perform scoring (for a review of these differences, see Haudek *et al*. \[[@B17]\] and Nehm *et al.* \[in press\]). SIDE is capable of creating scoring algorithms automatically when provided with a sufficiently large set of human-scored data for training and validation of scoring algorithms. Because we had a large sample of human-scored, short-answer responses for this study, our work on CAS was ideally suited to using SIDE.
SIDE combines a natural language processing (NLP) engine for parsing text, along with a set of machine-learning algorithms for classifying text (see Witten and Frank \[[@B51]\] for more details). Analyzing text responses using SIDE has two main steps: 1) defining the filters necessary for capturing the structure of the text and 2) specifying the summaries to be displayed and extracting the needed subsets (for details, see Mayfield and Rosé \[2010a,b\] and the supplemental material). Operationally, SIDE uses corpora of students' constructed responses that have been scored by humans to detect text patterns associated with the presence or absence of particular scientific concepts as measured by expert raters. For instance, terms such as "mutation," "genetic change," or "change in DNA" are indicative of the presence of variation, which is one of the key concepts necessary for explaining evolutionary change ([@B36]). Student responses, and the associated expert scores, are used as input to SIDE. Since we focused on five key concepts in this study, we needed to input a set of human-scoring information on whether or not the student text included a particular concept or not for each of the five key concepts of evolution that we investigated (see *Human Scoring of Explanations of Evolutionary Change*).
SIDE provides a number of interactive tools for refining and improving the accuracy of the predictions by allowing the user to examine cases where the machine-learning model misscored (either incorrectly classifying a response as containing the concept when it does not, or failing to classify a response as containing a concept when it in fact does). SIDE can save the scoring model and apply it to new text to predict human scoring ([@B23],b).
In this study, we used initial data sets, which had been scored by two biology experts, to train SIDE. We then used new, expert-scored data to validate the accuracy of the SIDE-scoring models. If this cross-validation approach is successful, this provides evidence that we can predict human scoring of new sets of student responses to the same questions with as much confidence as we would have using human raters.
Evolution: A Core Concept in Undergraduate Biology
--------------------------------------------------
Our investigations of CAS were focused on a core idea in biology: evolutionary change (AAAS, [@B2]). A large body of work, spanning more than 30 yr, has revealed a diverse array of learning difficulties with evolution in general, and natural selection in particular (reviewed in Nehm and Schonfeld \[[@B37]\] and Gregory \[[@B15]\]). Much less research has focused on psychometric issues relating to the measurement of student knowledge, although some work suggests that open-response formats and clinical interviews more validly capture student thinking about evolution than currently available MC assessments ([@B38]; Batistta *et al.*, [@B3]; Nehm *et al.*, in press). Thus, our investigation of the efficacy of CAS methods is well suited to the topic of evolution. Improvement in the measurement of students' thinking regarding evolutionary change may help to generate a deeper understanding of student learning difficulties and inform improved instructional practice.
RESEARCH QUESTIONS
==================
Our study explores three research questions: Are scoring models built using machine learning generalizable across colleges and courses (majors and nonmajors at different universities)? In other words, do scoring models built using student responses at one school function effectively at other schools?How many human-scored student responses are needed to effectively build scoring models? To what degree does sample size impact computer-scoring success?What factors limit computer-scoring efficacy, and how can these factors be mitigated to enable scoring models to be used in introductory biology courses across universities?
METHODS
=======
Sample
------
To answer our research questions, we utilized three samples of undergraduate students enrolled in biology coursework ([Table 1](#T1){ref-type="table"}): 1) nonmajors enrolled in introductory biology at Ohio State University (OSU; 264 students/1056 written explanations); 2) nonmajors enrolled in introductory biology at Michigan State University (MSU; 146 students/584 written explanations); and 3) biology majors enrolled in introductory biology at MSU (440 students/1760 written explanations). Student responses were gathered using two online survey systems (ACS and LONCAPA; for details, see [www.evolutionassessment.org](http://www.evolutionassessment.org) and [www.lon-capa.org](http://www.lon-capa.org), respectively).
######
Sample information
Ethnicity (%) Gender (%)
----- ---------- ----- --------------- ------------ ------ ---- ---- ------
OSU Nonmajor 264 79.1 14.4 6.5 42 58 20.1
MSU Nonmajor 146 66.4 13.7 19.9 40 60 19.4
MSU Major 440 79.1 11.8 9.1 42 58 19.6
^a^OSU = Ohio State University; MSU = Michigan State University.
^b^Note that *n* refers to subsampled data sets (see *Sample*).
We only included responses from individuals who completed four survey items each with responses of more than five words. The number of participants in the OSU nonmajor sample who completed all four items (see following section) was 358 (77.7% of total participants). Given the significant labor involved in scoring open-response items, we randomly selected a subset of 264 students (1056 responses) from this sample. We sampled the MSU data using the same approach The number of participants in the MSU nonmajors sample who completed all four items with more than five words was 146 (90.7% of total participants). Finally, the number of participants in the MSU biology major sample who completed all four items with more than five words was 440 (90.0% of total participants). Because of the time, money, and expertise required to score student responses, we randomly selected a subset of responses (*n* = 500) from each sample (OSU nonmajors, MSU nonmajors, and MSU majors; 1500 responses total) for the first research question. For our second research question, we scored more than twice as many responses from the OSU corpus (*n* = 1056).
Items Used to Generate Explanations of Evolutionary Change
----------------------------------------------------------
Our response corpus was composed of student explanations from four open-response items about evolutionary change. This assessment format has been employed in biology education research for more than 25 yr (e.g., Clough and Driver \[[@B10]\]) and has been shown to generate reliable and valid inferences about students' reasoning regarding evolutionary change ([@B5]; [@B38]; [@B29]). Instrument items were isomorphic ("How would biologists explain how a living X species with/without small/large Y evolved from an ancestral X species with/without large/small Y?") but differed in specific taxa and traits (i.e., X and Y).
Assessment item features, such as trait functionality and organism type, are known to influence students' reasoning regarding evolutionary change ([@B29]; Opfer *et al.*, [@B41]). Consequently, we standardized our prompts to include only animal examples and functional traits (e.g., fins, wings). Moreover, because the *familiarity* of taxa and traits is also known to influence students' reasoning regarding evolutionary change (Opfer *et al.*, [@B41]), item taxa and traits were constrained by their overall familiarity. To do so, we used the frequencies of "organism + trait" in Google rankings as a proxy for familiarity (Nehm, Beggrow, *et al.*, in press). Specifically, taxon/trait combinations included: shrew incisors, snail feet, fish fins, and fly wings.
Although students' short-answer explanations of evolutionary change varied in length, the average number of words did not differ among items (analysis of variance \[ANOVA\], *F* = 3.04, *P* \> 0.01). Specific item lengths were 1) shrew: mean = 45.5, SD = 30.3, minimum = 6, maximum = 430; 2) snail: mean = 42.9, SD = 27.9, minimum = 6, maximum = 429; 3) fish: mean = 42.5, SD = 26.3, minimum = 6, maximum = 202; 4) fly: mean = 41.6, SD = 26.5, minimum = 6, maximum = 209.
Human Scoring of Explanations of Evolutionary Change
----------------------------------------------------
Undergraduate students are known to recruit a diverse array of cognitive resources to build explanations of evolutionary change and solve evolutionary problems ([@B27]). These resources may include, for example, well-structured scientific schemas, such as natural selection; fragmented mental models built using mixtures of scientific and naïve knowledge elements; or naïve explanatory models ([@B29]). Given such diversity, it is most practical for assessment purposes to capture the existence of constituent explanatory elements in students' explanatory models (cf. Nehm and Haertig \[in press\]). For our study of automated computer scoring, two trained human raters scored all student responses for five key concepts of natural selection that were outlined by [@B36], described by [@B38], and codified in the scoring rubrics by Nehm *et al.* ([@B31]). It is important to emphasize that these concepts are central to the construct of natural selection, and necessary for explaining the operation of natural selection ([@B39]). Thus, the elements selected for scoring are not trivial or superficial aspects of reasoning regarding evolutionary change, and are associated with explanatory competence, as measured by clinical oral interviews ([@B38]). These key concepts included: 1) the presence and causes of variation (mutation, recombination, sex), 2) the heritability of variation, 3) competition, 4) limited resources, and 5) differential survival. It is important to note that scoring of short-answer explanations was dominated by the recognition of collections of key terms and short phrases, rather than elaborate grammatical expressions (see Nehm *et al.* \[[@B31]\] for details). Nevertheless, scoring was performed such that only accurate expressions counted for the "presence" of a concept; students' faulty expressions about heredity, for example, would *not* count as the presence of the key concept of heredity.
A series of studies has demonstrated that the coding rubrics used to score the short-answer explanations of evolutionary change are sufficiently clear to produce high levels of human interrater scoring agreement with moderate training ([@B36]; [@B38]; Nehm *et al.*, [@B33]; Nehm *et al.* [@B35]; [@B29]; Nehm and Haertig, in press; Nehm *et al.*, in press). In past studies, kappa agreement coefficients between human scorers with limited training ranged from 0.69 to 0.95, with an average of 0.86 (e.g., Nehm and Haertig, in press). In the present study, two biology experts (who have scored several thousand explanations of evolutionary change and have used the rubrics of Nehm *et al.* \[[@B31]\] for more than 2 yr) evaluated all student responses with very high agreement levels. The kappa reliability coefficients (*n* = 1056) for the present study were: 0.995 for variability, 1.000 for heritability, 1.000 for competition, 1.000 for limited resources, and 0.988 for differential survival. In the rare cases of disagreements between the two human raters, consensus scores were reached via deliberation. These final consensus scores were used in all subsequent analyses of human--computer scoring correspondence.
Human--Computer Correspondence Measures
---------------------------------------
We used Cohen\'s kappa to quantify the magnitude of human--computer scoring correspondence ([@B4]). Cohen\'s kappa values range from 0.0 to 1.0, and are commonly used to quantify levels of agreement among human raters, or between human and computer rating scores ([@B21]; Nehm and Haertig, in press). [@B21] introduced three general agreement levels for kappa statistics that we follow in our study: values between 0.41 and 0.60 are considered "moderate"; values between 0.61 and 0.80 are considered "substantial"; and those between 0.81 and 1.00 are considered "near perfect." We also report specific kappa values for all analyses, given the subjective nature of these categorical distinctions.
ANALYSES
========
Our first research question explores SIDE performance at detecting individual key concepts of natural selection in students' written responses. For each of our analyses, and for each key concept, two *categories* of agreement statistics are reported: 1) scoring-model *training* agreement values and 2) scoring-model *cross-validation* agreement values. The training agreement values are generated when SIDE first attempts to construct a scoring model using the corpus of human-scored student answers; that is, SIDE attempts to "learn" from the human-scoring patterns and builds a computational model that can account for these patterns. Then, SIDE examines the efficacy of this scoring model by calculating how well the model scores the *same* responses from which it learned. Kappa and percentage agreement values (which are automatically generated by the SIDE program) enable researchers to judge the strength of the machine-learning model and consider whether it is worthy of use on a new data set. In situations where the training kappa values are "substantial" (\> 0.60), the SIDE-generated scoring model is then applied to a *new* corpus of human-scored responses to determine whether the scoring model functions effectively with a new response corpus (that is, the training model is *tested*). Even if a training model performs admirably, this does not necessarily mean that it will be effective at scoring a new response corpus; both training and cross-validation performances need to be evaluated. Model cross-validation efficacy (also measured using kappa and percentage agreement statistics) must be performed manually (in our case, using SPSS, version 19.0). Cross-validation kappa values and percentage agreement values enable us to determine whether the SIDE-generated scoring models are likely to effectively score additional student responses.
In addition to exploring training and cross-validation performance of SIDE for each individual key concept, we also explored composite measures of students\' explanations of evolutionary change. Key concept score (KCS) is a composite measure of the number of scientific concepts employed in an explanatory context ([@B36]). Given that KCS has been used in prior research on learning gains ([@B36]), we examined how well SIDE performed relative to human expert scorers for KCS. Specifically, we used Pearson correlation statistics (in SPSS, version 19.0) to test for significant associations between human and computer scores of both variables (in contrast to the single-concept agreement statistics discussed above).
Our second research question examined to what degree sample size influences SIDE scoring--model performance. Specifically, we trained SIDE on two different corpora: 1) 500 responses from OSU students and 2) 1056 responses from OSU students. We then tested the two different scoring models on 1) the MSU nonmajor response corpus and 2) the MSU biology major response corpus. We calculated kappa and percentage agreement statistics to evaluate the influence of the training-sample size on SIDE scoring--model performance. All statistical tests were performed in SPSS, version 19.0.
Our third research question explored the factors that limit SIDE scoring--model performance; that is, why, in some cases, do scoring models fail to function at the desired near-perfect (kappa values \>0.80) agreement levels? Are such disagreements the products of the students (majors, nonmajors) and how they explain evolutionary change; the universities (OSU, MSU); the sample sizes; the scoring models; or combinations of these factors? This research question required examining all of the instances in which SIDE-generated scores and human-generated scores did not match, and attempting to identify the factors that contributed to score mismatches. After locating the likely source of scoring disagreements, we explored whether there were ways to mitigate these performance limitations so that future work would be more effective.
RESULTS
=======
Students' Explanations of Evolutionary Change
---------------------------------------------
To provide readers with a sense of the types of explanations of evolutionary change that undergraduate students' generate, four unedited student responses were extracted from the response corpus (see [Table 2](#T2){ref-type="table"}). As is apparent, student explanations of evolutionary change vary in length (for details, see *Items Used to Generate Explanations of Evolutionary Change*), sophistication, scientific accuracy, and scientific complexity. Adjacent to the responses in [Table 2](#T2){ref-type="table"} are two columns indicating the numbers and types of key concepts detected in each response (see scoring methods, in *Methods*, for details). Note that in the present study we investigated only the magnitudes of accurately expressed scientific concepts in student responses, not naïve ideas or misconceptions. Computer scoring of other explanatory elements is the focus of ongoing research.
######
Selected examples of students' written explanations of evolutionary change and corresponding human and computer scores
Taxon/trait/polarity Student\'s explanation of evolutionary change Human score (number of key concepts) Computer score (number of key concepts)
---------------------- ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ -------------------------------------- -----------------------------------------
Shrew incisors "Incisors may have developed on shrews due to *a genetic mutation* \[Variation\]. An offspring of a normal shrew may have had a mutated baby that had incisors, or some earlier form of incisors. The incisors would have given the new shrew an advantage *in acquiring food* \[Limited resources\] and reproducing, so it would have *a higher fitness* \[Differential survival\] leading *the incisor trait to be passed on to other generations* \[Heredity\]. As the trait will then develop with each generation due to variation involving the trait and the levels of success attached to each variant." 4 4
Snail feet "They would explain that once all the snails had small feet. Then one day there was a *mutation* \[Variation\] that produced a snail with a large foot. *The snail with a large foot was better able to produce more offspring* \[Differential survival\] in the environment *passing on his trait* \[Heredity\]." 3 3
Fish fins "There was *a random change in the DNA sequence* \[Variation\] of the fish that coded for the production of the fin. Nonrandom mating could have occurred with females selecting males with fins as partners, which disrupts HW equilibrium and leads to the evolution of the fin because *the fish with fin are better able to produce viable offspring* \[Differential survival\]." 2 2
Fly wings "The evolution of a fly species with a large wing from an ancestral fly with small wings could be through the process of natural selection or from *a random mutation* \[Variation\]." 1 1
Testing the Impact of Training Corpus on Scoring Success
--------------------------------------------------------
Our first analyses explored whether training SIDE using different human-scored corpora had an impact upon scoring-model success ([Figure 1](#F1){ref-type="fig"}). Six tests were performed ([Figure 1](#F1){ref-type="fig"}, A--F) for *each* key concept of evolution (e.g., variation, heredity, etc.). For the majority of these tests, the scoring agreements reached or exceeded near-perfect kappa values (18/30 tests) and percentage agreements above 90% (24/30 tests). Three key concepts---variation, heredity, and limited resources---were detected at near-perfect levels regardless of the training or cross-validation samples used. In contrast, competition and differential survival were very sensitive to training and cross-validation samples; in only two of the 12 tests did they reach near-perfect kappa agreement levels ([Figure 1](#F1){ref-type="fig"}, left). While raw percentage agreement values were robust for four of the five concepts (the exception being differential survival), these values do not take into account chance agreements. The dramatic difference between these two agreement statistics for competition suggests that sample size is contributing to these patterns, as we discuss in *Training Sample Sizes and Scoring Success*. Overall, the most significant factor influencing the performance of the SIDE-generated scoring models was not the training or cross-validation corpus per se, but rather, concept-specific factors.
{#F1}
Concept Frequencies in Different Samples
----------------------------------------
As shown in [Figure 2](#F2){ref-type="fig"}, human-identified frequencies of key concepts (blue bars) are in close alignment with computer-identified frequencies (red and green bars) for all samples (the different rows: OSU nonmajors, MSU majors, and MSU nonmajors). In addition, different SIDE training sets (i.e., MSU majors, MSU nonmajors) did not generate substantially different scoring frequencies of key concepts in comparison with human-generated scores (compare the different colored bars for each concept within each row). Small differences are apparent, however, among scores for differential survival in the OSU nonmajor sample (top row, right) and variation in the MSU major and nonmajor sample (middle and bottom rows, left).
{#F2}
One of the most striking patterns across all samples is that introductory biology students rarely used the concept of competition in their explanations of evolutionary change. In contrast, differential survival was used by a majority of students in all samples. Differences in the frequencies of key concept use were also apparent between samples (compare the rows in [Figure 2](#F2){ref-type="fig"}): almost twice as many responses from the MSU samples employed the concept of variation, compared with the responses in the OSU samples (bottom two rows vs. top row, left). In addition, MSU majors used the concept of limited resources more often than the other groups. Overall, [Figure 2](#F2){ref-type="fig"} demonstrates that differences in the frequencies of particular concepts vary across samples and schools, but the SIDE-generated scoring model was able to detect these differences. The extremely rare occurrence of competition ([Figure 2](#F2){ref-type="fig"}) was associated with poor model performance for this concept ([Figure 1](#F1){ref-type="fig"}).
Training Sample Sizes and Scoring Success
-----------------------------------------
Given that some key concepts were less common in the training and cross-validation data sets than other concepts (e.g., competition vs. variation), we investigated the impact of sample size on scoring-model efficacy. For *each* key concept (e.g., variation, differential survival) we performed two experiments. In the first, we trained SIDE using 500 human-scored responses from a sample of OSU nonmajors; and in the second, we trained SIDE using a sample more than twice as large: 1056 human-scored responses from OSU nonmajors. The two resulting SIDE-generated scoring models were applied separately to: 1) a corpus of MSU nonmajors' written explanations and 2) a corpus of MSU biology majors' written explanations (see *Methods*). As above, kappa agreement statistics and raw agreement percentages were calculated for all comparisons. [Figure 3](#F3){ref-type="fig"} illustrates the results of both experiments.
{#F3}
The scoring models built from the larger corpus produced higher correspondences with expert human raters in nine out of 10 tests; the exception was competition in the MSU biology major sample ([Figure 3](#F3){ref-type="fig"}). Given that the smaller training corpus (*n* = 500) produced near-perfect correspondence with human raters in most tests, doubling the training size bumped only one concept---competition in the MSU nonmajor sample---to the desired benchmark (kappa \> 0.80). Although the larger corpus did improve model performance for differential survival ([Figure 3](#F3){ref-type="fig"}), in many cases it did *not* meet our benchmark using either training corpus (*n* = 500 or 1056). In terms of raw percentage agreements, the larger corpus did not always produce improved results; in fact, the smaller corpus in many cases produced slightly higher agreement percentages. Nevertheless, in tests of model performance, 17 out of 20 comparisons of computer and human agreement reached or exceeded 90%. Additionally, using the large training data set, 9/10 analyses produced results that were detected at or above the kappa benchmark of 0.80.
Overall, in nearly all cases, doubling the training corpus improved model performance, but not substantially. The most dramatic improvement was seen in the detection of competition in the MSU nonmajor sample. Thus, the frequencies of particular concepts in the training corpus must be considered, not just overall sample size (see [Figure 2](#F2){ref-type="fig"} "Human scoring").
Explanatory Structures
----------------------
In addition to comparing individual key concept detection between human- and computer-scored explanations, it is useful to examine how students collectively assemble these concepts into explanatory structures ([Figure 4](#F4){ref-type="fig"}). One approach for representing these explanatory interrelationships is to use concept-association diagrams ([@B29]). [Figure 4](#F4){ref-type="fig"} illustrates both the frequency (the size of the circles) and co-occurrences of concepts (the thickness of the gray lines) in students' explanations of evolutionary change. For instance, approximately 20% of OSU nonmajors used both concepts of variation and differential survival in their responses (see connecting lines in [Figure 4](#F4){ref-type="fig"}). Each row in [Figure 4](#F4){ref-type="fig"} compares explanatory structures between human expert raters (left) and SIDE-scoring patterns (right) for a particular student sample (e.g., top: OSU nonmajors; bottom: MSU biology majors). As is apparent from the figure, results for students' knowledge networks are remarkably similar, regardless of whether they were scored by humans or computers.
{#F4}
Comparing the columns in [Figure 4](#F4){ref-type="fig"} also reveals that SIDE-generated scoring models can detect different explanatory structures across student samples, and these patterns closely mirror the findings reported in [Figure 2](#F2){ref-type="fig"}. The human-generated scores, for example, demonstrate that MSU majors and nonmajors used the concept of variation much more frequently than OSU nonmajors. SIDE produced the same patterns. Interestingly, human scorers also determined that MSU nonmajors used the concepts of variation and heredity more frequently than MSU majors; SIDE detected these patterns. MSU biology majors used the concept of limited resources much more frequently than MSU nonmajors, and this is also indicated in the SIDE-generated scores. While differences in the explanatory structures among majors, nonmajors, and institutions are interesting, the important point we wish to emphasize in [Figure 4](#F4){ref-type="fig"} is that SIDE-generated scores, which took minutes to generate, are in remarkable alignment with the patterns generated by humans during weeks of painstaking grading.
In addition to examining patterns of correspondence between human- and computer-generated visual representations of explanatory structure, [@B36] used KCS to quantify the number of different scientifically accurate evolutionary concepts that students use to explain evolutionary change in a prompt. [Table 3](#T3){ref-type="table"} illustrates statistically significant (*P* \< 0.001) and robust (*r* = 0.79 to 0.87) associations between human- and computer-generated KCS for all comparisons. Thus, using approaches for measuring student knowledge of evolution previously established in the literature ([@B36]; [@B29]), SIDE-generated scoring models produce patterns equivalent to those derived from human raters. Table 3.Correlation coefficients between human-scored and SIDE-scored student explanations for KCS^a^Training sampleTesting sampleHuman vs. SIDE KCD correlation (^\*\*^*P* \< 0.001)OSU nonmajor (*n* = 500)MSU nonmajor (*n* = 500)0.79^\*\*^MSU nonmajor (*n* = 500)OSU nonmajor (*n* = 500)0.80^\*\*^OSU nonmajor (*n* = 500)MSU major (*n* = 500)0.87^\*\*^MSU major (*n* = 500)OSU nonmajor (*n* = 500)0.85^\*\*^MSU major (*n* = 500)MSU nonmajor (*n* = 500)0.82^\*\*^MSU nonmajor (*n* = 500)MSU major (*n* = 500)0.82^\*\*a^In all cases, associations were strong and significant (*P* \< 0.001). KCS represents the number of different scientific concepts in a prompt. For details, see *Methods* and [@B36].
Factors Limiting Computer-Scoring Success
-----------------------------------------
Although SIDE and its machine-learning algorithms were shown to be highly effective at scoring the accuracy and complexity of undergraduates' explanations of evolutionary change, our studies revealed several limitations, which are summarized in [Table 4](#T4){ref-type="table"}. The key factors that limited the efficacy of computer scoring included: misspellings; nonadjacent key terms; very uncommon concept frequencies; and the diversity of expressions that students used to represent particular concepts. Table 4.Examples of the types of disagreements between human-scored and computer-scored explanations^a^Scoring patternCategoryExamples 1 to 5SolutionPositive computer score but negative human scoreMany key terms used, but important aspects were missing(1) "The original Shrew, who didn\'t have incisors, may have not been a *fit* species. Meaning, they may have not been *reproducing* enough to pass on their traits. Another reason would be a mutation that survived. What I mean is that a few shrews may have developed an allele mutation that resulted in the shrews developing incisors. After the mutation, it was passed on and probably *survived* because of natural *selection*, sexual *selection*, or artificial *selection* \[*more* fit, survive and reproduce *more*\]."Put a weight on core termsKey terms not adjacent, but scattered throughout the response(2) "For the fly species with wings to *survive* through natural selection it had to evolve to a species without wings. The wings of the fly were no *longer* needed so over time they grow smaller and weaker and eventually they no *longer* appeared in the offspring \[*survive longer*\]."Human augmentation of SIDE-scoring modelsNegative computer score but positive human scoreVery uncommonly used expression(3) "The fish was filling a niche in an area that required a fish with smaller fins. Generations passed and *a mutant gene* for a fish with smaller fins did well and its offspring did well through time a new species was born."Increase concept frequencies in training sample; human augmentation of SIDE-scoring modelsComplex expressions(4) "Variation of living fish species may leads \[sic\] to random mutation. It creates new sequences of DNA that will code for new or different protein. *This protein help*\[sic\] *in the creation of a new living fly species with wings in this situation. This species then reproduce* \[sic\] *and evolve through some time*. It may help to explain how a new living fly species with wings evolve \[sic\] even though it is originally from an ancestral fly species that lacked wings."Human augmentation of SIDE-scoring modelsSpelling errors and spacing errors(5) preditor \[predator\], servive \[survive\], springoffs \[offspring\], foodso \[food so\]Incorporate a spell-check program during data collection^a^Categories: types of scoring problems; examples: specific student responses; solutions: approaches used to correct the computer--human disagreement.
Spelling and spacing errors produce human--computer score disagreements. While our human raters easily understood what students were attempting to explain when they wrote "preditor" \[predator\], the computer was not able to do so. Misspelled words such as "servive" \[survive\], "springoffs" \[offspring\], and "foodso" \[food so\] also produced misclassifications in our study (see [Table 4](#T4){ref-type="table"}).
We also found that when student responses included key terms suggestive of a concept, but the words constructing the concept were scattered throughout the written response (conveying a very different meaning), the SIDE-scoring models often mistakenly linked these words together and considered the concept to be present. For example, in [Table 4](#T4){ref-type="table"}, example 2, the scoring model identified text elements characteristic of the concept of differential survival ("survive longer") to be present, even though the response included the words "survive" and "longer" in separate sentences.
Complex expressions, or very long sentences, also posed problems for the software. For example, one student (correctly) explained differential survival in the following way: "the fish with large fins is not suitable for the living environment in that specific area, so the fish with smaller fins survive and reproduce." Because expressions like this were rare in the student response corpora, and did not contain explicit language about differential survival, the scoring models failed to detect them. Fortunately, expressions like this were uncommon in the samples.
Low concept frequencies also prevented the machine-learning algorithm from building successful scoring models; without enough positive cases to analyze, the computer failed to annotate new cases appropriately. Competition serves as an example of a concept that was very rarely used by students to explain evolutionary change; only about 10 instances of competition were found among 500 written explanations. In addition to low concept frequencies, unusual expressions also lead to misclassifications. The term "mutant gene," while clearly identifiable as a "cause of variation" by a biologist, was too rare to be incorporated into the scoring models ([Table 4](#T4){ref-type="table"}).
We found that the concept of differential survival was influenced by the frequencies with which particular samples used particular terms. While the kappa values between scoring models built using the OSU nonmajor and the MSU major sample were nearly 0.80 (near-perfect), the kappa values between scoring models built by the OSU nonmajor and MSU nonmajor samples and between the MSU major and MSU nonmajor samples did not meet this benchmark (0.60). The cause of this pattern appears to be that the language patterns that MSU nonmajors used were somewhat different from the language patterns that the OSU nonmajor and MSU major samples used to describe differential survival. For example, the term "differential" (such as "differential reproduction rate" or "differential survival success") was observed 12 times among 500 responses in MSU nonmajor sample, whereas the term "differential" was observed only three times among 500 responses in the OSU nonmajor sample and only once among 500 responses in the MSU major sample. Consequently, the program incorporated "differential" as a diagnostic term for the MSU nonmajor scoring model, but not for the other samples.
DISCUSSION
==========
While CAS is becoming increasingly common throughout the educational hierarchy (Nehm *et al.*, in press), biologists have been slow to make use of this technological innovation. Two recent studies by Nehm and Haertig (in press) and Nehm *et al.* (in press) tested the efficacy, respectively, of SPSSTA, version 3.0 ([@B14]) and SIDE ([@B23],[@B24]). Using large samples of undergraduate biology students in single classes at one university, they demonstrated that both of these analytical tools are capable of generating assessment scores equal in precision to those by trained, expert raters (biologists with PhDs). Overall, Nehm *et al.* (in press) suggested that when clear scoring rubrics have been developed, and student ideas on a particular topic are well established, SIDE is much more powerful and cost effective than SPSSTA (Haudek *et al.*, [@B17]; Nehm *et al.*, in press). Since both of these factors apply to our present study, we chose SIDE as our CAS tool. For biologists who have not developed robust grading rubrics, or who have not comprehensively investigated student thinking about a topic, SPSSTA will be a more appropriate starting point (Haudek *et al.*, [@B17]).
Prior studies of SIDE did not investigate several questions that arise when biologists apply scoring models beyond a single instructor, course, or college. First, are scoring models generalizable across colleges and courses (e.g., major vs. nonmajor)? That is, will successful scoring models built at one university work at another? Second, how much human scoring is needed to build a robust scoring model, and can human scoring of additional student responses compensate for scoring-model limitations across courses and colleges? Finally, what factors might account for scoring models that function effectively in a class at one university but fail at a similar class in another? Can these failures be fixed?
It is important to emphasize that CAS tools---including machine learning---are not capable of comprehending the meanings of students' lexical responses. Programs such as SIDE simply note the presence or absence of particular words (or word pairs) in response corpora, build large matrices of word combinations, and apply sophisticated machine-learning algorithms to predict human-scoring patterns ([@B23],[@B24]). Consequently, machine-learning tools are very sensitive to language, but not its meaning(s). Expert human raters, in contrast, can effortlessly comprehend diverse linguistic expressions and understand their equivalence (e.g., "some live and some die" is equivalent to "differential survival"); in contrast, computers view different text as indicative of different information. For this reason, mundane text differences, such as spelling (color vs. colour; fecundity vs. fedunctity \[sic\]) impact scoring-model success.
Depending upon the scientific key concept for which a scoring model is built (e.g., variation, differential survival, etc.), different lexical expressions are used in different frequencies. Indeed, different populations of students---such as biology majors and nonmajors---may use characteristically different linguistic expressions to represent biological concepts. Some word combinations in some samples will be more diagnostic and predictive of key concepts than in others. Because of these concept-specific and sample-specific issues, we discuss our specific results relating to sample source (university; majors vs. nonmajors) and sample size (500 responses vs. 1056 responses) separately for each concept for which a scoring model was developed: variation, heredity, competition, and differential survival.
For key concept 1, variation, we found that SIDE scoring--model success was not sensitive to sample source ([Figure 1](#F1){ref-type="fig"}). That is, regardless of which response corpus was used to train SIDE (i.e., OSU vs. MSU students; majors vs. nonmajors), the scoring models generated excellent agreement with trained expert raters and near-perfect kappa values (\> 0.80). However, we did find that scoring models for variation were somewhat sensitive to sample size (that is, whether 500 or 1056 responses were used to build the scoring models). In comparison with the key concept of heredity, for example, in which a doubling of the training-sample size had almost no impact upon kappa values (adding 0.04 to 0.05), a doubling of the sample size for variation produced meaningful increases in kappa values (adding 0.14 to 0.80).
The explanation for the increase in kappa values with increasing training sample size for variation (but not heredity) appears to be related to the diversity and frequency of linguistic expressions that students used to represent these biological concepts. Although the most common term used by students to represent variation was "mutation," various other terms were also used, such as "different alleles," "genetic change," or "error in DNA." If only a few students used particular written expressions when linguistically representing the concept of variation (such as "genetic makeup"), then such expressions would be unlikely to be included in the machine-learning model, and downstream disagreements between human and computer scores would result. The frequencies of particular expressions, and their associations with other terms, influence scoring-model success. Indeed, we found that doubling the training sample for variation increased the frequencies of particular terms to a threshold at which they were included in the scoring models, producing improved kappa agreement statistics. For example, the matrix included 268 words for the *n* = 500 sample, while the matrix included 386 words for the *n* = 1056 sample. Matrix size is associated with differences in scoring-model performances.
For the concept of heredity, computer-scoring success was very stable and very successful regardless of sample size or source ([Figures 1](#F1){ref-type="fig"} and [3](#F3){ref-type="fig"}). Biology majors and nonmajors from different colleges and classes appear to use a consistent and detectable array of linguistic expressions to represent heredity concepts (e.g., [Table 1](#T1){ref-type="table"}).
The third concept we investigated was competition. Unlike the previous results, computer-scoring success for competition was sensitive to both sample source and sample size ([Figures 1](#F1){ref-type="fig"} and [3](#F3){ref-type="fig"}). Given our findings for variation and heredity, this result is surprising; the Nehm *et al.* ([@B31]) scoring rubric indicates that a very small set of terms is typically used to detect competition (e.g., compete, competition, competes). When we examine the frequency of students who used this concept, we find that only 1--2% of students used competition in their explanations of evolutionary change. Indeed, only 10 to 20 responses (out of 1000) included linguistic expressions relating to competition. Statistically, the probability that the algorithm will include such rare occurrences is low. Two solutions may be used to tackle the problem of rare responses: first, to amass a larger corpus of responses; or second, to use a special function in SIDE that allows users to augment the model and weight particular terms (see Nehm *et al.* \[in press\] for details). It is difficult for SIDE to build scoring models for extremely rare concepts.
The next concept we studied was limited resources. We found that the scoring models for this concept were stable in relation to both sample source and sample size ([Figure 3](#F3){ref-type="fig"}). Kappa values were near-perfect (\> 0.80) for the small data sets (*n* = 500) across samples, although there were some minor deviations. Overall, regardless of course or college, it appears that students commonly use consistent language patterns to represent this evolutionary concept, and scoring models for this concept work very well.
The final concept that we studied was differential survival. Similar to our findings for competition, differential survival was sensitive to both sample source and sample size. The comparatively weak performance of the differential survival scoring models was not a result of low response frequencies (as we observed with competition); large percentages of students utilized this idea in their explanations of evolutionary change (e.g., 60.3%; [Figure 2](#F2){ref-type="fig"}). Scoring problems in this case were a product of students' highly variable language use. This is in line with the scoring rubrics of Nehm *et al.* ([@B31]), which were built using different student samples and also note the diverse expressions with which students represent this evolutionary concept (e.g., "increase their survival," "survived better," "the species dies while others survive"). Because we also found that SIDE-scoring models were sensitive to sample source, different linguistic expressions may have been related to instructor discourse patterns. If, for example, students are imitating instructors' language (cf. Nehm *et al.* \[[@B35]\]), and different instructors use different phrases to represent biological ideas, then the sample source will impact scoring-model efficacy (as we found). Although the scoring model built using the largest sample (*n* = 1056) demonstrated relatively good kappa values (e.g., 0.69, 0.89; see [Figure 3](#F3){ref-type="fig"}), the highly variable ways of communicating the concept of differential survival appears to have limited scoring-model performance.
Generalizing Our Findings to Other Samples and Populations
----------------------------------------------------------
Very few studies in biology education have examined the similarities and differences between different student populations' short-answer explanations of biological phenomena, including evolutionary change. In two studies of primarily underrepresented biology students (many of whom were English-language learners) from a minority-serving institution in the eastern United States, [@B36] and [@B38] used short-answer, constructed-response assessments similar to those in the present study to reveal students' thinking patterns regarding evolutionary concepts. [@B38] reported that their findings were generally similar to those of primarily white student populations documented in the literature. Our current findings---from primarily white, midwestern undergraduates in large, public, research universities---are also very similar to those documented in these prior studies (see [Table 1](#T1){ref-type="table"}). This suggests that undergraduate biology students, regardless of racial and ethnic background, may utilize a large but relatively constrained set of concepts when conceptualizing evolutionary change. Nevertheless, such conjecture should be tested using diverse student samples from different geographic regions of the country. Until such work is completed, we cannot with confidence argue that machine-learning tools will be effective for assessing *all* introductory biology students.
Implications for Introductory Biology Faculty
---------------------------------------------
Our study has produced robust, automated, and generalizable scoring models capable of detecting most (but not all) of the core evolutionary concepts emphasized in standards documents, curricula, and textbooks (Nehm *et al.*, [@B34]). Biology educators can make use of our work by downloading the free software package SIDE (see [@B23],[@B24]) and incorporating our scoring models (freely available from the senior authors) to evaluate their students' written explanations of evolutionary change. Using a PC computer with an i7 processor, scoring 1000 written responses takes seconds to a few minutes (depending on the concept) and produces high levels of accuracy that are comparable with consensus scores generated by two trained biologists (see [Figure 4](#F4){ref-type="fig"}).
In addition to a user\'s manual ([@B23]), and details on the workings of SIDE ([@B24]), learning how to use SIDE is illustrated in a series of video tutorials (freely available at <http://evolutionassessment.org>). Given that this emerging form of assessment research is new, it is important to emphasize that the software is not packaged in a user-friendly format, and like other technological tools (e.g., clickers, new operating systems, new software), effort is required to learn to use it.
Our research to date has only validated a small set of biological concepts, and introductory biology instructors are likely to want to know how their students interpret a broader array of concepts in evolution (and other content areas). We are continuing to build scoring models for other concepts, such as naïve ideas or misconceptions of evolution (Ha and Nehm, unpublished results). We speculate that improved technology and advanced research on machine-learning assessment will enable more and more concepts to be detected in students' written responses.
National partnerships among introductory biology educators could make future work on machine learning more efficient and cost effective. Indeed, all biology educators, regardless of whether they view automated scoring as beneficial or not, could help move the field forward by collecting large corpora of students' written responses to different prompts across subject areas (genetics, matter and energy transformation, cell biology; Haudek *et al*., [@B17]); this would help those researchers interested in using and refining machine-learning methods. Additionally, faculty from minority-serving institutions, or those teaching large English language--learning populations, are needed to expand our knowledge base on how scientific language is used to communicate core concepts in biology.
Perhaps the most significant implication of our work for introductory biology educators is that evaluating students' written work, especially in large classes, is not impossible. This is significant from an assessment standpoint, as we contend that the process of asking students to communicate their understanding of scientific phenomena is a worthwhile activity, regardless of whether automated methods will be employed to assess these responses (e.g., Chi *et al.* \[1994\]). When analyzing students' written responses, we have been surprised by students' limited capacity to communicate and explain core scientific concepts (such as evolution)---particularly those students who perform admirably on MC assessments (cf. Nehm and Schonfeld \[2008\]). Without providing students practice and feedback in communicating their scientific understanding, we cannot expect this situation to improve.
Future work is needed to expand our concept of what constitutes a sound explanation of evolutionary change. Quantifying students' use of necessary and sufficient scientific elements (key concepts) as a benchmark for competency, as we have done, captures only one facet of short-answer scientific explanations (cf. Braaten and Windschitl \[[@B6]\]). Logic, persuasion, and argumentation skills are also important dimensions of scientific explanation, but they were not investigated in our study. Expanding our assessment framework will likely stimulate discussions about what facets of scientific explanation are most important for fostering scientific literacy.
Implications for Biology Education Researchers
----------------------------------------------
Research in the use of machine learning (and text analysis in general) in biology education is only beginning (Haudek *et al.*, [@B17]; Nehm and Haertig, in press; Nehm *et al.*, in press); much remains to be learned. A community of practice on text analysis in STEM education has recently been established (see Haudek *et al.* \[[@B17]\] and <http://aacr.crcstl.msu.edu>), providing a forum for researchers interested in learning more about these innovative assessment methods. Our current study has uncovered several findings likely to be of interest to researchers motivated to pursue this line of work.
First, even though we collected large response corpora, some concepts were nevertheless quite rare, limiting model performance. A large sample (*n* = 500) does not guarantee high concept frequency. In many instances, we were surprised by which concepts were used by students (and which were not). Second, we documented several factors that caused problems for machine-learning methods (e.g., misspellings; linguistic diversity) that nevertheless can be addressed by using a spell-checker during data gathering and weighting text expressions prior to analysis. Third, the diversity of linguistic expressions associated with concepts was highly variable (and generally unpredictable a priori), impacting scoring success. Some concepts were easily detected by the software, whereas others were not. Overall, the process of building automated scoring models is effortful and requires clear scoring rubrics and thousands of carefully evaluated responses.
Supplementary Material
======================
###### Supplemental Material
We thank Kristen Smock and Judith Ridgway, OSU, and Donna Koslowsky and Tammy Long, MSU, for assistance with data gathering; Luanna Prevost for comments on the manuscript; and the Automated Analysis of Constructed Response (AACR) group for discussions. R.H.N. thanks Elijah Mayfield and Caroline Rosé and the faculty of the Carnegie Mellon Pittsburgh Science of Learning Center summer school for help with machine-learning methods. We also thank two reviewers for helpful suggestions for improving the manuscript. We thank the National Science Foundation (NSF; REESE grant 090999 to principal investigator \[PI\] R.H.N.) for funding M.H. and collaborative NSF CCLI 1022653 to PIs Jenny Knight, R.H.N., and M.U.-L. for support of the AACR group. Any opinions, findings, and conclusions or recommendations expressed in this publication are those of the authors and do not necessarily reflect the view of the NSF. The research was conducted under OSU IRB Protocol \# 2008B0080 (R.H.N., PI).
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#Sec1}
============
Maize (*Zea mays*), also known as corn, is an important cereal crop grown under varied climatic conditions. China is the second-largest producer of maize after the United States. Maize can be processed into human food, animal feed, and industrial products. It is a high-production crop of national importance. A soil environment conducive to growth is required for the optimum growth and yield of maize. Tillage methods in cropping systems have been a part of most agricultural systems throughout history^[@CR1]^. The tillage method has a significant effect on the yield components of maize^[@CR2]^. Tillage changes soil physical properties, such as its water-holding capacity, pore size distribution, bulk density, and aggregation. Tillage allows more organic matter to be degraded by microorganisms, while no-tillage systems promote the establishment and strengthening of macroaggregates^[@CR3]--[@CR5]^. An appropriate tillage system is necessary to provide an appropriate environment for seed germination, weed control, regular moisture availability and the reduction of surface runoff through increased infiltration^[@CR6]--[@CR8]^. Adel El Titi^[@CR1]^ explained the importance of understanding the structures and functions of soil ecosystems under different tillage practices as a crucial requirement for applying farming concepts. However, intensive maize production has recently resulted in some adverse effects to the soil, such as nitrate leaching^[@CR2]^.
Phosphorus is an essential element for maize growth. The two vital processes in the transfiguration and translocation of phosphorus elements in the soil are geochemical and biological^[@CR9]^. Soil phosphorus (P) exists dynamically as dissolvable, labile, and non-labile P, and the chemical equilibrium between labile and non-labile P is weaker than the balance between dissolvable and labile P^[@CR10]^. Phosphorus fertilizer use has played an essential role in increasing crop productivity^[@CR11]^. Phosphorus, mainly organic phosphorus, is only available for plant uptake after hydrolysis by the enzyme phosphatase^[@CR12]^. Adequate phosphorus fertilization can affect maize plant height, leaf area index, kernel number, leaf photosynthesis, and plant growth^[@CR13],[@CR14]^. However, the excessive phosphate fertilizer used in agricultural production becomes a source of surface water pollution. Therefore, to maintain sustainable agricultural development and protect the environment, it is vital to establish proper tillage and fertilization methods in order to reduce fertilizer loss in agricultural production, the availability of phosphorus is dependent on the production system applied^[@CR15]^. Tillage methods can change the phosphorus retention parameters in the near-surface zone^[@CR15]^. The spatial distribution of phosphorus can also affect its availability, especially in rhizosphere and non-rhizosphere soil. Rhizosphere soil refers to the narrow soil, which is directly influenced by root secretions and their affiliated soil microorganisms^[@CR16]^. The zone of soil around the roots where the soil properties, soil microorganisms, and plant roots interact is abundant in organic compounts^[@CR17]--[@CR20]^. Dynamic changes in plant nutrients, biology, and soil chemistry take place in the rhizosphere. The enzyme activity in rhizosphere soils is generally higher than that in non-rhizosphere soil^[@CR21]^. Various techniques for studying the chemical changes in rhizosphere soil have been established for annual crops, grasses, and legumes^[@CR22]^. Guo^[@CR23]^ demonstrated that the continuous cropping of maize and soybean in different tropical soils treated with a high dose of P fertilizer greatly reduced the labile and moderately labile inorganic (Pi) fractions. P in the soil has unique properties, such as low solubility and high fixation by soil particles. Therefore, the available P for crops is controlled by two factors: (1) the bioavailability and acquisition of P based on rhizosphere processes and (2) the limited availability and acquisition of P in terms of plant root architecture as well as mycorrhizal association^[@CR24]^. To estimate the various P fractions and P changes in soil, different fractionation methods with multiple chemical sequential extractions can be used^[@CR25],[@CR26]^. (Safari Sinegani)^[@CR27]^ reported that plants significantly decreased the levels of all inorganic P fractions in the rhizosphere soil compared to those in non-rhizosphere soil. The reduction was not equal for each fraction, and the percentage of apatite-P increased in the rhizosphere soil. (Yong-Fu)^[@CR28]^ also noted a significant decrease in P fractions, pH and phosphatase activities in the rhizosphere. Soluble P in solution in the rhizosphere should be exchanged 20 to 50 times per day by P delivery from bulk soil to the rhizosphere to meet crop requirements^[@CR29]^. Numerous studies have focused on the changes in the phosphorus fractions associated with soil chemical properties under different types of fertilizers^[@CR30]^.
Previously, many studies have been conducted on how different tillage systems can be used as tools for increasing the crop yields. There is little concentration about which tillage method is suitable for maintaining a supply of available P for maize, thereby decreasing the fertilizer application and improving the environment. We hypothesized that tillage systems have an effect on phosphorus and its forms in the rhizosphere and non-rhizosphere soil. Therefore, the objective of this study was to assess the effects of different tillage systems on the distribution of phosphorus and its different forms in the rhizosphere and non-rhizosphere soil under maize (*Zea mays L*.) in northeastern China.
Results {#Sec2}
=======
The available P content in the maize rhizosphere and non-rhizosphere soil under different tillage systems {#Sec3}
---------------------------------------------------------------------------------------------------------
The available soil P was evaluated under the different tillage methods in the rhizosphere and non-rhizosphere soil. The available P significantly decreased with the age of the maize., which indicated that much more nutrients were taken by plant roots with maize growing. Then, the available P continually declined with crop growth. The rhizosphere region showed a lower content of available P than the non-rhizosphere region (Fig. [1](#Fig1){ref-type="fig"}). Compared to the rhizosphere soil, the non-rhizosphere soil had 132.9%, 82.5%, 259.8%, and 148.4% more available P in the CR, PR, CN and PN treatments, respectively, at the maturity stage. The CN treatment had the lowest amount of available P and the PR treatment had the highest amount of available P throughout the growth stage, which indicated that soil disturbance was beneficial to the increase of available phosphorus content. (Fig. [1](#Fig1){ref-type="fig"}).Figure 1Comparison of available P in the maize rhizosphere and non-rhizosphere under different tillage methods. The bars represent the standard error of the three replicates. The letters above the columns represent significant differences among the four tillage methods.
The total content of P in the maize rhizosphere and non-rhizosphere soil under different tillage systems {#Sec4}
--------------------------------------------------------------------------------------------------------
The total P concentration in the soil decreased with the maize growing up in both soils (Fig. [2](#Fig2){ref-type="fig"}). At the early stage of plant growth, the level of total P in both soils was high. Afterward, it decreased with time until harvesting. In comparison to the rhizosphere soil, the non-rhizosphere soil had 62%, 71.4%, 80.9% and 78.3% more total P in the CR, PR, CN and PN treatments, respectively. The CN treatment had the lowest amount of total P throughout the growth stages (Fig. [2](#Fig2){ref-type="fig"}).Figure 2Comparison of total P in the maize rhizosphere and nonrhizosphere soil under different tillage methods. The bars represent the standard error of the three replicates. The letters above the columns represent significant differences among the four tillage methods.
Alkaline phosphatase content in the maize rhizosphere and non-rhizosphere soil under different tillage systems {#Sec5}
--------------------------------------------------------------------------------------------------------------
There were significant differences in the alkaline phosphatase level between the maize growth stages. At the seedling stage, all treatments contained a higher level of alkaline phosphatase. This level decreased with the age of the maize until the maturity period, which marked the lowest amount of alkaline phosphatase in both soils studied. Compared to the rhizosphere soil, the non-rhizosphere soil had 106.4%, 57.8%, 71.8% and 54% more alkaline phosphatase in the CR, PR, CN and PN treatments, respectively. The CN treatment showed the lowest amount of alkaline phosphatase over the whole growing cycle. All these results showed that there was an increase in chemical phosphatase activity in both regions of the studied soil, and the CN treatment showed lower alkaline phosphatase levels compared to the other treatments.
The distribution of P fractions in the maize rhizosphere and non-rhizosphere soil under different tillage systems {#Sec6}
-----------------------------------------------------------------------------------------------------------------
Figure [4(a--e)](#Fig4){ref-type="fig"} shows the different levels of P fractions in the maize rhizosphere and non-rhizosphere soil at the seedling and maturity stages of maize growth. The amounts of Ca~2~-P (Fig. [4a](#Fig4){ref-type="fig"}), Ca~8~-P (Fig. [4b](#Fig4){ref-type="fig"}), Fe-P (Fig. [4d](#Fig4){ref-type="fig"}) and O-P (Fig. [4e](#Fig4){ref-type="fig"}) were high at the seedling stage and decreased at the maturity stage. There were significant differences in the amounts of Ca~2~-P, Ca~8~-P, Fe-P, and O-P under the different tillage methods at different growth stages. The seedling stage had the maximum amounts of Ca~2~-P, Ca~8~-P, Fe-P and O-P fractions compared with those amounts in the other treatments (Fig. [4a](#Fig4){ref-type="fig"}). The rhizosphere soil displayed lower amounts of Ca~2~-P, Ca~8~-P, Fe-P and O-P than the non-rhizosphere soil in the two soil regions studied (Fig. [4b](#Fig4){ref-type="fig"}).Figure 3Comparison of alkaline phosphatase levels in the maize rhizosphere and nonrhizosphere soil under different tillage methods. The bars represent the standard error of the three replicates. The letters above the columns represent significant differences among the four tillage methods.Figure 4(**a--e**) The concentrations of P fractions in the maize rhizosphere and non-rhizosphere soil at the seedling and maturity stages of maize growth.
The Pearson correlation coefficient (Table [1](#Tab1){ref-type="table"}) showed that the available P in the soil was significantly positively correlated with Ca~2~-P(r = 0.871), Ca~8~-P (r = 0.910), Al-P(r = 0.696), Fe-P(r = 0.759), and occluded-P(r = 0.844). There was no significant relationship between the total P and the P fractions found. The total P measured in the soil had no relationship with the evaluated P fractions.Table 1The correlation coefficients (r) of the relationships between selected Pfractions and total phosphorus (mg kg^−1^).Total PAvailable PCa~2~-PCa~8~-PAl-PFe-PO-PTotal P1.000Available P0.1791.000Ca~2~-P−0.1030.871\*\*1.000Ca~8~-P0.0420.910\*\*0.927\*\*1.000Al-P0.5260.696\*\*0.549\*0.672\*\*1.000Fe-P0.0020.759\*\*0.945\*\*0.875\*\*0.593\*1.000O-P−0.0480.844\*\*0.976\*\*0.877\*\*0.4980.940\*\*1.000\* and \*\* indicate significance at p \< 0.05 and p \< 0.01, respectively.
Discussion {#Sec7}
==========
Dynamics of soil available P and total P during maize growth {#Sec8}
------------------------------------------------------------
The rhizosphere soil is dominated by microbes and enzymes^[@CR31]^. The rhizosphere is modified by plant roots, and has more microbial communities than bulk soil or non-rhizosphere soil^[@CR32]^. The application of any compound used as fertilizer to a plant of interest, such as maize, can influence the microbial activity and community structure in the rhizosphere soil^[@CR16]^. There were significant differences in the amounts of available and total P during the growth stages of maize (Figs. [1](#Fig1){ref-type="fig"} and [2](#Fig2){ref-type="fig"}). For the different treatments, it appeared that the contents of total and available P were higher during the seedling stage in both the rhizosphere and the non-rhizosphere soil. Both P types continued to decrease with plant growth until the harvesting period. This decrease in nutrients was probably due to uptake by the growing crop.The nutrient concentration in non-rhizosphere soil was significantly higher than that in rhizosphere soil during the full summer maize growing season^[@CR33]^, which was consistent with our results. However, it was similar to our findings that the addition of P and K fertilizers significantly increased the levels of microelements (P, K, Ca, and Mg) in rhizosphere soil. Additionally, the fertilizers transformed the pH and electrical conductivity of the rhizosphere soil compared to those in the control^[@CR34]^. Rhizosphere and non-rhizosphere soil are essential soil regions that connect the soil environment to plant root systems. A lack of soil disturbance during soil preparation can influence the availability of phosphorus nutrients for adsorption by maize. The application of phosphorus significantly altered the amount of available P in non-rhizosphere soil under all maize tillage methods applied in this research.
The available P and total P content in the maize rhizosphere and non-rhizosphere soils under different tillage systems {#Sec9}
----------------------------------------------------------------------------------------------------------------------
Phosphorus (P) is an essential mineral nutrient for plant growth and development, and it is affected by different systems of physical soil manipulation^[@CR35],[@CR36]^. The mechanical management of soil during tillage may increase the contact chances between soil solution- or fertilizer-derived P and soil particles. It exposes soil particles, enabling the establishment of fixed, insoluble P compounds^[@CR37]^. Tillage methods may influence the availability and distribution of plant nutrients (N, P, K)^[@CR38]--[@CR41]^. No-tillage systems affect some chemical parameters associated with soil acidity that may influence P accessibility, plant growth, and yield^[@CR42]^. In our investigation, rhizosphere soil had lower available and total P than non-rhizosphere soil.The low amount of P in the rhizosphere was due to the rapid uptake of P by plant roots, which was facilitated by the low solubility and mobility of P in the soil^[@CR24]^. The continuous no-tillage treatment (CN) had the lowest concentration of phosphorus in the maize rhizosphere soil, while plowing-rotary tillage (PR) had the highest level of phosphorus in both soil regions. This indicated that less soil disturbance promoted the availability and uptake of phosphorus by the plants; the no-till method maintains a favorable soil environment, and therefore, phosphate nutrients are readily available for uptake by the roots (Figs. [1](#Fig1){ref-type="fig"} and [2](#Fig2){ref-type="fig"}). The continuous no-tillage practice significantly improved the soil's ability to hold phosphates in the non-rhizosphere soil and thus resulted in an increasing amount of soil P under the CN treatment compared to that under the rotary tillage systems. Similar results were also reported by Zhang^[@CR43]^, who suggested that short-term no-tillage improved P availability in surface soils at 0--20 cm. The addition of phosphorus through manure application also increased levels of different phosphorus forms and the phosphorus saturation of the near-surface soil zone in a no-tillage system^[@CR44]^. These findings indicate that the no-tillage method in spring maize has significant effects on phosphorus availability and improve our understanding of P management practices.
Alkaline phosphatase activity in the soil under different tillage systems {#Sec10}
-------------------------------------------------------------------------
Alkaline phosphatase is secreted by bacteria, fungi, and earthworms, and it works catalytically at a high pH of 7.00. The abundance of different mineralized phosphate groups can be predicted using alkaline phosphatase. Alkaline phosphatase speeds up or slows down the cleavage of ester-phosphate bonds, making P available in soils^[@CR45],[@CR46]^. Acid and alkaline phosphatases in the rhizosphere and bulk soils of legumes have various functions, including Nfixation in beans^[@CR47]^. In the rhizosphere and non-rhizosphere soils, alkaline phosphatase was more active in the continuous no-tillage treatment than in the other three treatments. Continuous plowing-rotary tillage was observed to have the lowest alkaline phosphatase activity. The tillage method had a significant effect on the alkaline phosphatase balance in the soil (Fig. [3](#Fig3){ref-type="fig"}). The level of alkaline phosphatase decreased with the age of the maize, and the lowest amount of alkaline phosphatase in both soils was observed at the maturity stage. In comparison to the non-rhizosphere soil, the rhizosphere soil showed slightly lower alkaline phosphatase concentrations. The CN treatment showed the lowest alkaline phosphatase level of all treatments over the whole growing cycle. Compared with the CN treatment, there was an increase in alkaline phosphatase activity in the other three treatments. This suggested that no-tillage may influence the availability of water and improve the amount and variety of organisms in the soil. Therefore, no-tillage increased the activity of alkaline phosphatase in the soil. It was also found that alkaline phosphatase was more active at the tasseling stage than at the other maize growth stages evaluated in all the treatments (Fig. [3](#Fig3){ref-type="fig"}). Alkaline phosphatase works in the presence of phosphate nutrients. During the reproductive growth stage, it is necessary to increase the availability of phosphate in the soil. NTSM (no-tillage with corn straw return) and NTG (no-tillage with grass) increased P content and phosphatase enzyme activity and provided a basis for using this method to improve P availability and decrease the application of fertilizer to soils^[@CR36]^. A similar observation also indicated that alkaline phosphatase activity in rhizosphere soil was significantly higher than that in non-rhizosphere bulk soil^[@CR47]^. The use of phosphate fertilizer may increase alkaline phosphatase activity in the rhizosphere soil. Overall, the continuous no-tillage method facilitated more alkaline phosphatase activities in the maize rhizosphere soil. The soil under plowing with rotary tillage had the lowest soil enzyme components compared to that under the other tillage methods; therefore, the CN treatment could be used as a strategy for soil health and productivity, resulting in a sustainable agricultural system.
The partitioning of P fractions in the maize rhizosphere and non-rhizosphere under different tillage systems {#Sec11}
------------------------------------------------------------------------------------------------------------
The Hedley sequential-phosphorus (P) fractionation method has been used worldwide to investigate the effects of land-use and management systems on soil P. In natural environments, vegetation varieties, composition, and percent of vegetation cover significantly affect all P fractions. Most P fractions increase with the level of phosphorus applied^[@CR25]^. There were significant correlations between total P and Ca~2~-P, Ca~8~-P, and Al-P, and the relative abundances of P forms were in the order of Ca~10~-P \> Ca~8~-P \> Al-P \> Fe-P \> Ca~2~-P \> Occl-P^[@CR48]^.Compared to those under conventional tillage, the amounts of organic matter and phosphorus in the top few centimeters under no-tillage were higher^[@CR49]^. The evaluation of the relationship between P availability indicators and inorganic P fractions showed the abundances of the different P forms, which were in the order of Ca~2~-P \< Fe-P \< Al-P \< Occluded-P \< Ca~8~-P \< Ca~10~-P. Total P was positively correlated with Olsen P and exchangeable P^[@CR50]^. In both soils, the amounts of the different P fractions were high at the seedling stage and then decreased gradually until maturity. P fractions are associated with phosphate nutrients, which are required by the plant during its whole life cycle; this is why the P fractions were abundant at the seedling stage (Fig. [4a--e](#Fig4){ref-type="fig"}). The levels of the P forms in the maize rhizosphere and non-rhizosphere soil followed this arrangement: CN \< CR \< PN \< PR treatments. However, the non-rhizosphere soil had higher levels of the different P fractions than the rhizosphere soil. The no-tillage method caused less soil disturbance, resulting in a healthy soil environment with good soil aeration and soil moisture levels. However, our analysis showed that available P was positively correlated with the Ca~2~-P, Fe-P and Al-P, Ca~8~-P, and occluded-P fractions. In contrast, the total P had no relationship to the P fractions. In view of these results, the P fractions in the maize rhizosphere and bulky soil were enhanced by phosphorus addition. This suggested that the studied phosphorus forms could influence the uptake of phosphorus by spring maize. On the other hand, the non-rhizosphere soil had higher levels of the different P fractions than the rhizosphere soil (Fig. [4a--e](#Fig4){ref-type="fig"}). The amounts of available P were significantly correlated with the concentrations of Ca~2~-P(r = 0.871), Ca~8~-P (r = 0.910), Al-P(r = 0.696), Fe-P(r = 0.759), and occluded-P(r = 0.844) (Table [2](#Tab2){ref-type="table"}). Available phosphorus, soluble phosphorus, DCP-P, occluded-P, Al-P, and Fe-P in non-rhizosphere soil were lower than in rhizosphere soil^[@CR27]^. For this experiment, a significant difference among the treatments was found only for occluded P. This was in agreement with several studies demonstrating that no-tillage systems were able to maintain higher levels of available P for maize and reduced phosphorus loss^[@CR15],[@CR51]^. Generally, P fractions were dominant in the non-rhizosphere soil under a continuous no-tillage approach.Table 2Basic physical and chemical properties of the tested soil.CharacteristicValueOrganic matter16.97 g kg^−1^pH7.10Available P34.64 mg kg^−1^Total P462.08 mg kg^−1^Soil bulk density1.61 g cm^−3^Soil textureBlack Sandy LoamSoil typeBlack Chernozem
Materials and Methods {#Sec12}
=====================
Experimental design {#Sec13}
-------------------
The experiment was conducted in 2018 in the experimental field of Qingshan Town, Taobei District in Baicheng City, Jilin Province, China (45°41′N latitude, 122°55′E longitude), at an altitude of approximately 200 m. The land use in the plot was a continuous maize monoculture system. This region has a temperate continental monsoon climate with a mean yearly temperature of 5.2 °C and 399.8 mm of average rainfall per annum, and most of the precipitation occurs between April and August. The maize variety used in this region was Xiangyu 998, which is a primary native medium-maize variety. The type of soil was a sandy, loamy chernozem (classified according to the Canadian soil classification system). The chemical properties of the 0--20 cm soil layer are provided in Table [2](#Tab2){ref-type="table"}.
The experiment was conducted over three years (2016--2018) of continuous rotary tillage (CR), continuous no-tillage (CN), plowing-rotary tillage (PR) and plowing-no tillage (PN). There were three replicates per treatment in 12 subplots of 500 m^2^. The crop residues from the previous-year harvest were left in the field as straw return. During soil preparation in the spring, the specifics of the treatments were as follows:(CR) rotary tillage every year, the soil tilling depth was approximately 10--12 cm, no-tillage seeder sowing and fertilizer; (CN)seeder sowing and fertilizer at approximately 20 cm soil depth, direct use of a no-tillage seeder for sowing and fertilization without other treatments;(PR) plowing at a depth of approximately 20 cm in the first year and the same treatment as CR in the last two years; and(PN) plowing and tillage at approximately 20 cm depth in the first year and the same treatment as CN in the second year.
The maize was planted in May at a density of 65000 plants ha^−1^ and was harvested in late October 2018. In each treatment, the combined basal fertilizer applied was (N:P~2~O~5~:K~2~O = 26%:11%:11%) with 800 kg ha^−1^ applied during sowing by a planter machine. Other field management practices were carried out conventionally.
Soil sampling {#Sec14}
-------------
Soil sample collection was conducted in May 6, June 24, August 8, October 2, 2018 at four maize growing stages, namely, the seedling stage, elongation stage, tasseling stage, and maturity stage. The samples were divided into rhizosphere soil and non-rhizosphere soil. The soil used for determining the P form was collected at the seedling and maturity stages. Three whole plant roots, including apical and older roots, were dug out from each prominent subplot (at the seedling stage, approximately eight plants were involved). The soil was removed carefully and systematically in a soil area of 28 cm (14 cm on each side of the plant base in the interrows direction) × 35 cm (10 cm in the narrow interrows and 25 cm in the wide interrows) and a depth of 40 cm. Following the careful removal of the unsecured soil from the roots (collected as non-rhizosphere soil), the remaining firmly held earth was shaken gently over a clean paper sheet. After carefully hand-picking out the visible thin roots (except for root hairs), this soil was collected as rhizosphere soil. The collected soil samples were ground down into fine particles and sieved through a 3 mm sieve.
Soil determination {#Sec15}
------------------
The soil samples were air-dried, and the P concentration was determined on neutralized extracts using the molybdate-colorimetric method of^[@CR52]^ at 88 nm. Several consecutive P fractionation procedures have been used to determine the forms of P and distributions of P forms in the soil^[@CR53]^. Alkaline phosphatase and a series of inorganic forms of soil P (Al-P, O-P, Fe-P,Ca~2~-P, and Ca~8~-P) were determined sequentially according to the method by Jiang and Gu (1989)^[@CR54]^ as follows:(i).Soil (1 gram)Shake for 1 hour in 50 mL 0.25 mol/L NaHCO~3~ solution (pH = 7.50)Centrifuge and remove supernatant for P determination of Ca~2~-P(ii).Residual from step (i)Wash twice with 95% alcohol, shake with 50 mL 0.5 mol/L NH~4~Ac solution (pH = 4.20, let sit for 4 hours, shake again for 1 hourCentrifuge and remove supernatant for P determination of Ca~8~-P(iii).Residual from step (ii)Wash twice with saturated NaCl, shake with 50 mL 0.5NH~4~F solution for 1 hourCentrifuge and remove supernatant for P determination of Al-P(iv).Residual from step (iii)Wash twice with saturated NaCl, shake with 50 mL 0.1 mol/L NaOH-0.1 mol/L Na~2~CO~3~ solution (pH = 8.20) for 2 hours, let sit for 16 hours, shake again for 2 hoursCentrifuge and remove supernatant for P determination of Fe-P(v).Wash twice with saturated NaCl, shake with 40 mL 0.3 mol\\L sodium citrate solution plus 1gNa~2~S~2~O~6~ heated at 80 °C for 15 min
Centrifuge and remove supernatant for P determination of Occl-P
Statistical analysis {#Sec16}
--------------------
The data were analyzed, and the differences were compared using SPSS Statistics 23.0 (SPSS, Inc., Chicago, IL, USA). The means were compared by Duncan's test at the 0.05 significance level. Figures were created in Origin Pro 8.0. The means and standard errors from the statistical analysis were brought into Origin Pro 8.0, and the figures were created using the column tool.
Conclusion {#Sec17}
==========
As the growth period continued, the available P, total P, and P fractions in the soil gradually decreased. The absorption of plant nutrients from rhizosphere soil was high under the continuous no-tillage method. The use of continuous no-tillage methods can effectively increase the content of available nutrients in the soil. The establishment of no-tillage methods can not only increase the amount of available P in the soil but can also effectively maintain a continuous soil P supply throughout the whole maize growth period. The continuous no-tillage treatment promoted the absorption of P nutrients by maize, and the non-rhizosphere soil had sufficient P to exchange with the rhizosphere region for plant uptake.
**Publisher's note** Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
The authors are grateful to the National Key R&D Program of China (2018YFD0300201) and the National Key R&D Program of China (2016YFD0300807) for providing financial support to conduct our research.
X.Z. designed the research, X.T. and S.J. conducted the field and laboratory experiments, X.T., S.J. and M.F. analyzed the data, X.T. and F.M. wrote the paper, and L.Z., X.G., H.W. and B.S. reviewed the paper. All authors discussed the results and approved the manuscript.
The authors declare no competing interests.
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#s1}
============
MRI is considered the gold standard imaging modality for the central nervous system. Although MR images provide high resolution and good sensitivity standard MR sequences still suffer of low specificity especially in the case of brain tumors during or after treatment. Tumor progression, inflammation and infection can induce similar changes in enhancement which are impossible to differentiate. Treatment related changes, including radiotherapy-induced changes, add another level of complexity to the interpretation of the brain images which remains challenging notwithstanding the increased utilization of advanced imaging methodologies. Previous studies suggest that nearly half of glioblastoma multiforme (GBM) patients with radiological deterioration after standard chemoradiation do not suffer from tumor recurrence but from pseudoprogression induced by the treatment [@pone.0052008-Fink1]--[@pone.0052008-Brandsma2]. Treatment decision, as whether to operate on a patient with radiographic deterioration, continue chemoradiation or change to another non-surgical treatment is a daily struggle involving interdisciplinary teams of neurosurgeons, neuro-oncologists and neuro-radiologists which are often unable to reach unanimous interpretation of the patient's status. Therefore, reliable distinction between these conditions is essential for appropriate patient management and for appropriate patient selection for clinical trials [@pone.0052008-Young1]--[@pone.0052008-VandenBent1].
It has been previously suggested [@pone.0052008-Forsyth1]--[@pone.0052008-Kim1] that the histological tumor fraction (i.e., tumor burden) comprises a subcomponent of the total enhancement seen on contrast-enhanced MRI and represents a potentially useful predictor of survival in patients with recurrent brain tumors. In fact, studies suggest that histological quantification of tumor burden provides more meaningful prognostic information than simply reporting the presence of tumor [@pone.0052008-Hu1]--[@pone.0052008-Tihan1].
Brain metastases are the most common intracranial tumor in adults, occurring in approximately 10% to 30% of adult cancer patients. It is believed that the annual incidence is rising (due to better treatment of systemic disease and improved imaging modalities). The prognosis of patients diagnosed with brain metastases is generally poor.
10.1371/journal.pone.0052008.t001
###### Biopsied samples and map characteristics.
{#pone-0052008-t001-1}
Sample \# Patient \# delayed enhancement population Histological description Tumor type
----------- ------------------ ------------------------------------------------------------------------------------------------------ ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- ------------------------------------------------------
1 1 Mixed regions of red, blue and green populations One cellular region consisted of small cells with no mitoses. Proliferation was seen in 5% of the cells by Ki67 staining, implying active tumor. Other regions showed post radiation changes. One region was of brain parenchyma with no obvious abnormalities GBM post chemoradiation
2 1 Cortical region of blue population and deeper white matter region of red population Subcortical infiltrating zone of active tumor with rare mitosis and a deeper, white matter region of post radiation changes. Ki67 staining of the active tumor zone showed proliferation in 3--5% of the cells GBM post chemoradiation
3 1 Region of blue population Active tumor consisting of a hypercellular area of small cells. Ki67 staining showed proliferation in 10--12% of the cells GBM post chemoradiation
4 1 Mixed regions of blue and red populations Regions of active tumor consisting of a hypercellular area of small cells and regions of post radiation changes. Ki67 staining in the active tumor region showed proliferation in 10--12% of the cells GBM post chemoradiation
1 4 Region of red population Radiation necrosis GBM post chemoradiation
2 4 Cortical region of blue population and white matter region of red population Cortical region shows active tumor accumulating focally beneath the meninges. Focal proliferation of blood vassals and palisading necrosis are identified as well. Most of the deeper white matter region show radiation necrosis GBM post chemoradiation
1 11 Region of blue population Highly cellular tumor with small regions of tumor necrosis with and without pseudo palisading regions of proliferating blood cells Secondary GBM post chemoradiation
2 11 Region of blue population Highly cellular tumor Secondary GBM post chemoradiation
3 11 Region of blue population Highly cellular tumor with small regions of tumor necrosis with and without pseudo palisading regions of proliferating blood cells Secondary GBM post chemoradiation
4 11 Region of blue population Highly cellular tumor with large proliferating vessels and small regions of tumor necrosis with palisading regions of proliferating blood cells Secondary GBM post chemoradiation
1 13 Region of blue population Tumor consisting of atypical cells, mitoses and proliferating vessels Newly diagnosed GBM
1 14 Region of blue population Tumor consisting of atypical cells and numerous mitoses GBM post chemoradiation
2 14 Border between a red regionand a blue region A region of necrosis with scanty nuclear dust and a region of tumor with pleomorphism and small regions of palisading necrosis GBM post chemoradiation
1 16 Border between a red regionand a blue region Necrotic region including necrotic blood vessels and nuclear dust and a cellular tumor region with atypical cells, mitoses and vascular proliferation Newly diagnosed GBM
2 16 Region of blue population Tumor region with small foci of palisading necrosis Newly diagnosed GBM
1 17 Region of blue population Highly cellular tumor with small necrotic foci Newly diagnosed GBM
1 19 Region of blue population Regions of high cellularity and of low cellularigy typical of oligodendroglioma tumors Newly diagnosed analplastic oligodendroglioma
1 20 Region of red population on theborder of a blue population Mostly necrotic tissue including necrotic blood vessels. Small foci of tumor are present Newly diagnosed analplastic oligodendroglioma
2 20 Region of blue population on theborder of a red population Most of the tissue is tumor. One small area of necrosis at the periphery of the section Newly diagnosed analplastic oligodendroglioma
3 20 Region of blue population on theborder of surrounding brain Mostly tumor tissue bordered by brain tissue infiltrated by tumor Newly diagnosed analplastic oligodendroglioma
1 21Metastasis \#1 Sample taken from a blue region bordered by a green region on one side and a red region on the other A sample showing a region of morphological active tumor bordered by normal cerebellum tissue on one side and necrotic tissue on the other Sample taken from NSCLC cerebellar brain metastasis
2 21Metastasis \#1 Sample taken from a red region bordered by a blue region A sample showing a large necrotic region borders by tumor Sample taken from NSCLC cerebellar brain metastasis
1 27 Mixed area of blue and red regions A mixture of active tumor regions and necrotic regions Sample taken from NSCLC cortical brain metastasis
2 27 Blue region with small red foci on the border of surrounding brain Several small samples of active tumor, tumor necrosis and edematous brain Sample taken from NSCLC cortical brain metastasis
1 29 Red region on the border of surrounding brain Radiation induced gliotic brain tissue Sample taken from breast cerebellar brain metastasis
2 29 Mixed blue and red region on the border of surrounding brain cerebral tissue and mixed regions of tumor and radiation necrosis Sample taken from breast cerebral brain metastasis
3 29 Mixed blue, green and red regions tumor, cerebral tissue and radiation necrosis with ecstatic blood vessels Sample taken from breast cerebral brain metastasis
4 29 A red region bordered by surrounding brain on one side and a blue region on the other Gliotic brain and radiation necrosis with a small tumor mass Sample taken from breast cerebral brain metastasis
1 30 Red region surrounded by a blue rim small regions of tumor (∼30% of the sample) within larger region of necorosis (∼70% of the sample) Sample taken from breast cortical brain metastasis
2 30 Blue region on the border of surrounding brain small tumoral region on the border of normal cortex Sample taken from breast cortical brain metastasis
3 30 Blue region on the border of surrounding brain Highly cellular tumor adjacent to normal cortex Sample taken from breast cortical brain metastasis
4 30 Red region bordered by small blue region on one side and surrounding brain on the other mostly necrosis with small foci of tumor and adjacent normal brain Sample taken from breast cortical brain metastasis
List of 32 biopsied samples, delayed enhancement subtraction map characteristics and histological evaluation of 9 patients with primary brain tumors and 4 patients with brain metastases.
Stereotactic radiosurgery (SRS) is a radiotherapy technique which permits the delivery of a single large dose of radiation to the tumor while minimizing irradiation of adjacent normal tissue. It is applied to treat both benign and malignant tumors as well as for vascular lesions and functional disorders. Among the reported complications of SRS is radiation-induced necrosis which, similarly to pseudoprogression, can be difficult to differentiate both clinically and radiologically from recurrent tumor at the treatment site. The incidence of radiation induced necrosis may vary between 5% to 11% according to the volume of the treated lesion and the applied dose [@pone.0052008-Shaw1].
The difficulty in interpretation of these post treatment images prevents appropriate treatment in nearly half of treated GBM patients and 5--11% of treated brain metastases patients [@pone.0052008-Fink1]--[@pone.0052008-Brandsma2], [@pone.0052008-Shaw1]. The ability to clearly differentiate tumor from non-tumoral tissues is crucial for appropriate patient management in GBM and brain metastases as well as in other brain tumors undergoing surgery, chemotherapy and/or radiation therapy.
Conventional MRI is currently unable to provide reliable distinction between tumor progression and treatment effects such as pseudoprogression and radiation necrosis [@pone.0052008-Young1]--[@pone.0052008-Rane1]. MRS can distinguish residual or recurrent tumors from pure treatment-related necrosis, but not from mixed necrosis and tumor tissue [@pone.0052008-Rock1]. Diffusion weighted MRI (DWMRI) and diffusion-tensor MRI has also been assessed for differentiating tumor/necrosis after RT [@pone.0052008-Chan1]--[@pone.0052008-Xu1], however, the specificity of DWMRI is less than MRS. It has been suggested that combining DWMRI with MRS may improve the differentiation [@pone.0052008-Ricci1]. FDG-PET has been shown to be useful in differentiating necrosis from recurrence, but the reported sensitivity and specificity of FDG PET in the brain are low [@pone.0052008-Tsuyuguchi1]. There is limited, but increasing evidence that PET with amino acid tracers may contribute to the differentiation between treatment-related necrosis from tumor recurrence [@pone.0052008-Nelson1]. Whether these techniques will also allow a reliable distinction between pseudoprogression and real progression is yet to be determined.
{#pone-0052008-g001}
{#pone-0052008-g002}
{#pone-0052008-g003}
A number of fast acquisition MRI techniques have been applied to study microvasculature parameters in this context. The two most commonly used methods are dynamic contrast-enhanced MRI (DCE MRI) and dynamic susceptibility-weighted contrast MRI (DSC MRI) [@pone.0052008-Lacerda1]--[@pone.0052008-Barboriak1]. DCE MRI measures the changes in T1 relaxation associated with disrupted BBB following contrast administration using parameters such as fractional blood volume (fBV) and permeability (Kps or Ktrans). DSC MRI uses echo planar sequences with a rapid bolus of gadolinium-based contrast agents to assess changes in T2\* within the vasculature and interstitial space. Typical calculated parameters are the relative peak height (rPH), relative cerebral blood volume (rCBV) and the percentage recovery (%REC) or recirculation factor (RF).
{#pone-0052008-g004}
10.1371/journal.pone.0052008.t002
###### Non-biopsied tumors histology and map characteristics.
{#pone-0052008-t002-2}
Patient \# delayed enhancement population Histological description[\*](#nt103){ref-type="table-fn"} Tumor type
------------------- ---------------------------------------------------------------------------------------- -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- ----------------------------------------------------------------------------------------
3 Overall enhancing lesion consists of71% blue population and22% red population Cellular tumor with many mitoses and regions of "geographic necrosis". In some regions of the tumor it is also possible to see proliterative blood vesselsSome regions surrounding the tumor (not in all slices and not all around the tumor) depict brain tissue infiltrated by a small number of tumor cells. The main findings in these regions around the tumor are abnormal proliferation of blood vessels and many histiocytes GBM post chemoradiation
21 Metastasis \#2 lesion consisting of blue (58%)regions and red (31%) regions Metastatic carcinoma showing squamoid features with extensive areas of tumor necrosis NSCLC cortical brain metastasis
22 lesion consisting of blue (56%)regions and red (35%) regions Several samples showing regions of tumor and regions of radiation necrosis. Significant cauterize artifacts are noticed as well Adenoid Cystic Carcinoma Cortical brain metastasis
23 Metastasis \#1 Metastasis consisting of a centralred (42%) region surrounded by athick blue rim (47%) Central slice shows a large necrotic region surrounded by significant regions of morphologically active tumor NSCLC mediall brain metastasis Pathology report addressing metastases resected unblock
23 Metastasis \#2 lesion consisting of blue (56%)regions and red (33%) regions Morphologically active tumor is present in the histological samples NSCLC cortical brain metastasis
24 lesion consisting of a large blue (56%)region surrounded by a red rim (34%) Active tumor was found Melanoma midline brain metastasis
25 lesion consisting of a blue central region(61%) surrounded by a red (33%) region Active tumor was found Breast cortical brain metastasis
26 lesion consisting of thin blue rim(40%) within a larger red (51%) mass Large mass of radiation necrosis. Small fociof active tumor were found after ki67 staining Breast peri-ventricular brain metastasis
28 lesion is mostly red (53%) with smallelongated blue regions (27%) Mostly necrotic samples with small foci of active tumor NSCLC medial brain metastasis
List of tumors with no stereotactic biopsies, delayed enhancement subtraction map characteristics and histological evaluation of 1 patient with primary brain tumor and 7 patients with brain metastases.
Pathology report is based on all samples obtain from the neurosurgeons unrelated to the pre-surgical maps.
Parametric maps that are derived from DCE and DSC data have been proposed as noninvasive methods for assessing response to therapy. Treatment induced changes typically shows decreased rCBV, whereas recurrence shows high rCBV [@pone.0052008-Aronen1]--[@pone.0052008-Kim2]. Still, most of these studies show some degree of overlap between the two disease entities. DSC was recently applied by Gahramanov et al [@pone.0052008-Gahramanov1] which demonstrated the feasibility for differentiating pseudoprogression from real tumor progression using ferumoxytol. Narang et al [@pone.0052008-Narang1] applied DCE to a cohort of 29 patients with gliomas and brain metastasis suspected for treatment-induced necrosis or recurrent/progressive tumor and demonstrated the feasibility of predicting real progression. Barajas et al [@pone.0052008-Barajas1] applied DSC MRI for differentiating tumor progression from radiation necrosis in GBM patients undergoing external beam radiation therapy. Their analysis showed that rPH and rCBV were significantly higher in patients with recurrent GBM than in patients with radiation necrosis while the %REC values were significantly lower.
Using fast acquisition techniques has the disadvantages of low spatial resolution and high sensitivity to susceptibility artifacts. Due to these limitations, significant results are mostly obtained by averaging over the entire enhancing lesion while important information regarding the location and shape of small active tumor regions may be limited. This effect stands out especially in the case of GBM, where psuedoprogression hardly ever describes a status with no residual tumor (unlike the case of brain metastases and radiation necrosis where complete tumor resolution is more likely to occur), as we know that the vast majority of all GBM patients, including those experiencing pseudoprogression, eventually recur.
More encouraging results were obtained using delayed T1-weighted MRI (T1-MRI) permeability methods, which image beyond the first pass circulation of contrast, sometimes as long as 10--15 min. Hazle et al were able to reliably distinguish between recurrence, radiation necrosis, and a combination of both factors [@pone.0052008-Hazle1]. They found that radiation necrosis and tumor enhance at different rates, enabling significant differentiation between recurrent tumor, radiation necrosis and mixed radiation necrosis and tumor (p\<0.001). Diehn et al [@pone.0052008-Diehn1] showed that using intra-tumoral and peri-tumoral MRI information it was possible to predict activation of hypoxia and proliferation gene-expression programs, respectively. Furthermore, the intratumoral distribution of gene-expression patterns was found to predict patient outcome.
{#pone-0052008-g005}
{#pone-0052008-g006}
In the current study we apply a novel methodology based on delayed contrast extravasation MRI for calculating delayed enhancement subtraction maps, depicting unique temporal enhancement characteristics with high resolution and high sensitivity to subtle BBB disruption [@pone.0052008-Israeli1]. In order to confirm the application of our maps for differentiating tumor tissues from non tumoral tissues we compared pre-surgical maps of patients with primary or metastatic brain tumors with histological assessment of resected tissue samples.
Materials and Methods {#s2}
=====================
Patients and Treatment {#s2a}
----------------------
The study was conducted after approval of the local ethics committee at Sheba Medical Center. Written informed consent was obtained from all patients.
Ten patients with primary brain tumors and 10 patients with brain metastases were recruited prior to surgery and scanned by conventional and delayed contrast extravasation MRI. In all 20 patients pre-surgical delayed enhancement subtraction maps were compared with histological findings. In addition, the application of our maps for prediction of progression was studied in a small cohort of 13 newly diagnosed GBM patients undergoing standard chemoradiation and followed up to 19.7 months post therapy.
Primary Brain Tumors Group {#s2b}
--------------------------
The primary brain tumor patients consisted of 8 patients with histologically confirmed glioblastoma (World Health Organization \[WHO\] grade IV astrocytoma) of which 3 were newly diagnosed patients, 4 progressed following chemoradiation (60 Gy in 30 daily fractions, 5 days a week, with concomitant daily temozolomide of 75 mg/m^2^ for 42 days. Chemoradiation was followed by adjuvant temozolomide of 150--200 mg/m^2^ daily for 5 days every 28 days.) and one patient with secondary GBM was recruited 2 years after resection of an Anaplastic Astrocytoma (WHO grade III tumor) and chemoradiation. Two patients with newly diagnosed Anaplastic Oligodendrioma (WHO grade III tumor) were recruited as well.
{#pone-0052008-g007}
{#pone-0052008-g008}
The mean age of this group of patients was 58.6±3.6 with a range of 37--80. Nine of the 10 patients were men.
Brain Metastases Group {#s2c}
----------------------
The brain metastases patients consisted of 4 patients with breast cancer metastases, 4 with non small cell lung cancer (NSCLC) metastases, 1 with malignant melanoma metastases and 1 patient with adenoid cystic carcinoma metastases. All patients were treated with a single dose of 18--20 Gy (to the 80% isodose line) LINAC based SRS 13.1±2.9 months prior to resection.
The mean age of this group of patients was 50.2±4.0 with a range of 30--67. Four of the 10 patients were men.
10.1371/journal.pone.0052008.t003
###### Newly diagnosed GBM cohort.
{#pone-0052008-t003-3}
Patient \# 1^ST^ Surgery 2^nd^ Surgery Samples comparedwith histology Time to progression \[months\] Treatment at progression
------------ --------------- --------------- -------------------------------- --------------------------------- --------------------------
1 GTR STR 4 7.5 Surgery+Bevacizumab
2 GTR -- -- Not reached (19.7) --
3 GTR GTR 8 6.5 surgery
4 GTR GTR 2 6.2 Surgery+Bevacizumab
5 STR -- -- 3.6 Bevacizumab
6 GTR GTR -- 11.6 Surgery+ Bevacizumab
7 GTR -- Not reached) 12.5 --
8 STB -- Died from unrelated disease (3) --
9 STB -- Not reached (5.2) --
10 GTR -- 5.0 Bevacizumab
12 STB -- 4.1 Bevacizumab
15 GTR -- 3.6 Bevacizumab
18 GTR -- Not reached (5.2) --
List of newly diagnosed GBM patients with post chemoradiation treatment follow-up.
Column \#2: Patients were diagnosed with GBM prior to initiation of chemoradiation by histological analysis of either gross tumor resection (GTR), sub-total resection (STR) or stereotactic biopsy samples (STB).
Column \#4: Stereotactic samples were taken from locations determined using the late enhancement subtraction maps calculated from the pre-surgical MRIs.
Column \#5: In cases where progression was not reached, the duration of follow-up is listed in parenthesis.
{#pone-0052008-g009}
Newly Diagnosed GBM Group {#s2d}
-------------------------
In an attempt to explore additional possible applications of our calculated maps, 13 newly diagnosed GBM patients were scanned by conventional and delayed extravasation MRI 3--4 weeks after chemoradiation and every 2 months thereafter. Patients with clinical deterioration underwent additional MRI exams at the discretion of the physician. These patients were followed for 2--19.7 months. Eight of the later have progressed while 5 remained progression free.
The mean age of this group of patients was 54.3±4.1 with a range of 28--72. Eight of the 13 patients were men.
Inclusion criteria for all patients included WHO performance status of 2 or less and adequate hematologic, renal, and hepatic function. Exclusion criteria included contraindications to MRI and stable condition to undergo an MRI exam.
MRI Data Acquisition {#s2e}
--------------------
MR images were acquired using a clinical General Electric 3.0 T MRI machine (GE Medical Systems, Waukesha, WI, USA) with the HD12 operating system, gradients intensity of up to 4.3 Gauss/cm and the standard GE phased array head-coil. MR sequences included T2\* perfusion-weighted MRI (PWI), Fast spin-echo T2-weighted MRI, T2 FLAIR and echo-planar diffusion-weighted MRI (DWMRI). High resolution spin-echo T1-weighted MR images (T1-MRIs) were acquired before and at 3 time points after contrast injection: 2.6±0.1 min (immediately after the PWI sequence), 15.4±0.4 min and 75.3±0.7 on average. These times are referred to as the 2 min, 15 min and 75 min time points throughout the text. In order to minimize the burden on the patients, they were scanned up to 30 min after contrast injection, and were then asked to return for a short scan 75 min after contrast injection. T1-MRI was acquired with TE/TR = 22/240 ms, field of view 26×19.5 cm, 5/0.5 mm slice thickness and 512×512 pixels. A standard single dose (0.2 ml/Kg) of Gd-DOTA (Dotarem, 0.5 mmol/mL, Guerbet, 95943 Roissy CdG Cedex, France) was injected intravenously using an automatic injection system 6 seconds after starting the PWI sequence.
MRI Data Analysis {#s2f}
-----------------
All image analysis was performed using MatLab (version R2006b, The MathWorks, Inc. Natick, MA, US).
The overall goal of the analysis was to obtain delayed subtraction maps, where the T1-MRIs of the 1^st^ series post contrast were subtracted from the T1-MRIs of later series. These maps depict spatial distribution of contrast accumulation/clearance in the tissue, blood vessels and CSF. For example, in case of normal blood vessels, due to clearance of contrast agent from the blood system, there is no increase in contrast accumulation in the late scans; therefore, the subtraction maps show negative values (blue in the maps). The signal decay of the blood vessels is faster than that of the tissue (where the signal is averaged over the tissue and microvasculature), therefore, blood vessels have lower values than tissue. In case of contrast accumulation, i.e. regions where contrast clearance is slower than contrast accumulation, the maps show positive values (red in the maps).
In order to increase the sensitivity to small changes it was essential to perform image pre-processing consisting of corrections for intensity variations and whole body image registration.
### Correcting for intensity variations {#s2f1}
Signal intensity homogeneity throughout the image and between slices depends on various parameters including: strength and homogeneity of the static magnetic field, oscillating excitation field, gradients, sensitivity of the receiving coil and various parameters of the sampled tissue. An intensity correction was performed on each image separately by calculating an intensity variation map consisting of the large scale intensity variations. The later map was then subtracted from the original image.
### Rigid body and elastic/local registration {#s2f2}
Rigid body registration was performed using least squares approach and 6 parameter (rigid body) spatial transformation with the SPM5 (Statistical parametric mapping) MatLab routine (an academic software kit by "Wellcome Trust Centre for Neuroimaging"). Since head movements affected the magnetic fields thus inducing distortions in the MRIs, it was necessary to add local/elastic registration. The registration was performed by dividing each slice to a grid of 20×20 mm volumes. Each square volume was allowed to move freely in x-y-z till the sum of the absolute values of the intensity difference between the 2 time points reached a minimum. The resulting three 3D translation matrices were smoothed using circular smearing and interpolated to obtain translation values per pixel. These high resolution matrices were then applied to register T1-MRIs of the second time point to the location of the first time point.
### Subtraction maps {#s2f3}
Following the pre-processing, subtraction maps were calculated by simply subtracting the processed images of the series acquired 2 min after the contrast injection from a series acquired later on.
### Enhancing lesion volume {#s2f4}
The enhancing portion of the lesions, depicted on conventional contrast-enhanced MRI, was calculated from the spin-echo contrast-enhanced T1-MRIs acquired 2 min after contrast injection. Regions of interest (ROIs) were defined over the entire enhancing region in each slice. A threshold, determined from intensity distribution histograms of the tumor and surrounding regions, was applied to the ROIs to include only enhancing portions of the tumor. The number of pixels in the enhancing portions of the ROIs were counted and multiplied by the volume of a single pixel. The resulting volume was referred to as the state of the art parameter for assessment of tumor volume [@pone.0052008-Wen1].
Histology {#s2g}
---------
A total of 32 stereotactic tissue samples and 8 resected lesions acquired from 10 patients with primary brain tumors and 10 patients with brain metastases were histologically examined and compared with our delayed enhancement subtraction maps. Stereotactic locations for tissue samples were determined prior to surgery using the calculated subtraction maps which were co-registered to the conventional T1-MRIs, for 9 patients with primary brain tumors (20 tissue samples) and 4 patients with metastatic tumors (12 tissue samples) ([Table 1](#pone-0052008-t001){ref-type="table"}). Existence/absence of morphologically active tumor in the histological analysis of 8 metastatic tumors acquired from 7 patients was compared with the pre-surgical maps as well ([Table 2](#pone-0052008-t002){ref-type="table"}). Eight additional samples were obtained from one GBM patient, chosen by the neurosurgeon during surgery as representative samples (patient \#3, [Table 2](#pone-0052008-t002){ref-type="table"}).
All samples were marked by the neurosurgeon during resection and then fixed and stained by H&E according to the routine hospital procedure.
Histological interpretation was performed by the hospital neuro-pathologist. Morphologically active tumor was defined as demonstrating one or more of the following characteristics: hyper cellularity, small cells, mitosis, high Ki67, pseudo-palisading necrosis and vascular proliferation.
Non-tumoral abnormal tissue was defined as demonstrating one or more of the following characteristics: radiation changes including large, widely spaced atypical astrocytes, blood vessels hyalinization, fibrinoid material in vessels, proliferating small vessels and non palisading tumor necrosis.
Progression (GBM Patients Only) {#s2h}
-------------------------------
Disease progression was diagnosed using standard MRI sequences. Each case was presented to the hospital tumor board, consisting of a neuro-oncologist, a neurosurgeon and a neuro-radiologist. The physicians were blinded to the delayed enhancement subtraction maps so that progression was diagnosed according to the RANO (revised Mcdonald's) criteria. When disease progression was determined, the patient's treatment was changed (surgery or Bevacizumab). The time from the first MRI follow-up to disease progression is listed for all patients in [Table 3](#pone-0052008-t003){ref-type="table"}. In those patients with no disease progression, the time of the last follow-up was listed.
Time to Progression (TTP) (GBM Patients Only) {#s2i}
---------------------------------------------
TTP was defined as the time from the end of chemoradiation till progression was determined.
Tumor Growth Rate (GBM Patients Only) {#s2j}
-------------------------------------
The feasibility of applying the delayed enhancement subtraction maps for prediction of TTP was demonstrated by studying the correlation between initial tumor growth rate and TTP. Initial tumor growth was calculated as (V−Vo)/Vo, where V was the tumor volume at the second follow-up and Vo was the tumor volume calculated from the first follow-up MRI (3 weeks after the end of treatment). Initial tumor growth rate was calculated by dividing the tumor growth by the time that passed between the first and second MRI follow-ups.
Statistical Methods {#s2k}
-------------------
Results for averaging over a group of values are presented as average±standard error. Correlations were assessed by performing t tests and calculating two-tailed p-values, unless otherwise stated. Prediction of TTP was assessed using the logrank test [@pone.0052008-Peto1]. In view of the small number of patients involved, the p-values were calculated based on the permutation distribution [@pone.0052008-Moses1]. In this method one calculates the logrank statistic for each possible allocation of the patients to the two predictor groups "high" and "low" preserving the number of patients in each group, and then calculates the p-value as the proportion of allocations that yield a logrank statistic equal to or higher than the one observed with the study data.
Results {#s3}
=======
Delayed Enhancement Subtraction Maps {#s3a}
------------------------------------
Enhancement subtraction maps were calculated using the data acquired 15 (15 min maps) and 75 min (75 min maps) after contrast injection. Two primary enhancement patterns were found in our maps: One characterized by slower contrast clearance than contrast accumulation at the delayed time point relative to the 2 min time point (positive signal, colored red in our maps) and the other by faster clearance than accumulation (negative signal, colored blue in our maps). Normal brain regions, due to the intensity variation correction, had an average value of zero (green). Examples of 15 and 75 min maps are shown in [Figure 1](#pone-0052008-g001){ref-type="fig"}. It can be seen that the fast (blue) region in the 75 min map is depicted more clearly and over larger regions than in the 15 min map. Examples of the signals intensities of the fast and slow regions as a function of time after contrast injection are shown as well, demonstrating the different rates of contrast accumulation and clearance of these regions.
Histology, Tumor Versus Non Tumoral Tissues {#s3b}
-------------------------------------------
Thirty two stereotactic samples acquired from 9 patients with primary brain tumors and 4 patients with brain metastases undergoing surgery were compared with our pre-surgery calculated maps. The samples were taken from fast (blue) regions according to the maps, slow (red) regions, and regions consisting of mixed fast and slow components. Histological evaluation confirmed for all samples the discrimination between fast regions, determined to consist of morphologically active tumor, and slow regions, consisting of non-tumoral tissues. Regions consisting of both fast and slow components in the maps consisted of tumor and non-tumoral tissues in the histological samples ([Figures 2](#pone-0052008-g002){ref-type="fig"}, [3](#pone-0052008-g003){ref-type="fig"}, [4](#pone-0052008-g004){ref-type="fig"}, [Table 1](#pone-0052008-t001){ref-type="table"}).
Patient \#3 ([Table 2](#pone-0052008-t002){ref-type="table"}) with a newly diagnosed GBM underwent a second resection 6 months after chemoradiation due to clinical deterioration. The patient died 10 days post-surgery. The maps calculated from his last MRI scan, showed that the fast component reached 71±3% of the enhancing lesion volume. Histological analysis was performed for 8 samples taken from 2 main regions of the lesion. In both regions the tumor load was estimated by the neuro-pathologist to cover ∼70% of the examined samples, in agreement with our calculated maps.
Eight additional metastases acquired from 7 patients, showed a fast component in our maps and were confirmed by histology to consist of morphologically active tumor ([Figures 3](#pone-0052008-g003){ref-type="fig"}, [4](#pone-0052008-g004){ref-type="fig"}, [Table 2](#pone-0052008-t002){ref-type="table"}).
Histology, Blood Vessels {#s3c}
------------------------
In an attempt to find a morphological explanation for the difference between the vessels function of fast and slow regions in our maps, we examined the morphological appearance of blood vessels in these regions. Typical fast population vessel morphology consisted of proliferating endothelial cells, dilated lumen, peri vascular fibrosis and glomeruloid vessels. The outline of the vessels lumens in these regions seemed to be undamaged. Vessels in the slow regions, on the other hand, presented different stages of vessel necrosis with significantly damaged lumens. In most vessels a silhouette of the vessel wall could still be recognized, but only rarely residual endothelial cells could still be detected. In some cases scattered blood cells were seen in various distances from the necrotic vessel. Examples are shown in [Figures 3](#pone-0052008-g003){ref-type="fig"}, [4](#pone-0052008-g004){ref-type="fig"}, [5](#pone-0052008-g005){ref-type="fig"}.
Significance of Long Delays {#s3d}
---------------------------
The volumes and intensities of the fast population, calculated from the 75 min maps, were found to be significantly different then those calculated from the 15 min maps (data calculated from 30 MRI exams of the 30 recruited patients): r = 0.91, p\<0.0001 and r = 0.79, p\<0.0001, respectively (Wilcoxon matched-pairs signed-ranks test). The average ratio between the volumes of the fast population calculated from the 75 min maps and the volumes calculated from the15 min maps was 2.0±0.3 and the average ratio between the intensities of the fast population at the two time points was 1.8±0.1, suggesting increased sensitivity to tumor tissues at the longer delays. There was no significant difference in this increased sensitivity between the primary brain tumors and the brain metastases groups.
Correlation with Conventional MRI {#s3e}
---------------------------------
The volume of the fast population was found to correlate significantly with the enhancing lesion volume (representing the conventional tumor volume): r = 0.94, p\<0.0001 (based on 30 acquired MRI exams). According to our maps, in this cohort of patients 48.4±1.9% (on average) of the enhancing lesion on conventional MRI **did not** represent morphologically active tumor ([Figure 6](#pone-0052008-g006){ref-type="fig"}). For example, according to our maps only 55%, 38%, 39%, 56% and 83% of the enhancing lesions presented in [Figures 1](#pone-0052008-g001){ref-type="fig"}, [2](#pone-0052008-g002){ref-type="fig"}, [6](#pone-0052008-g006){ref-type="fig"}, [7A and 7D](#pone-0052008-g007){ref-type="fig"} respectively, consisted of the fast population. There was no significant difference in the average percentage of the fast population between the primary brain tumors and the brain metastases groups.
Significant correlation was also found between fast population volume×intensity and rCBV calculated from PWI (r = 0.69, p\<0.005), suggesting that rCBV may be a dominant characteristic of the fast population. Still, the relatively low r value implies that there may be other contributions to the fast component. Examples of agreement and disagreement between the vessel function maps and rCBV are shown in [Figure 7](#pone-0052008-g007){ref-type="fig"}.
Correlation with TTP (GBM Patients) -- Preliminary Results {#s3f}
----------------------------------------------------------
In an attempt to explore the feasibility of applying our maps for predicting time to progression 13 newly diagnosed patients were followed by conventional and delayed-contrast extravasation MRI. The follow-up periods of each patient are listed in [Table 3](#pone-0052008-t003){ref-type="table"}. Four parameters were studied as possible predictors for progression: Initial fast volume, initial enhancing volume, initial fast growth rate and initial enhancing growth rate. Kaplan-Meier curves of TTP in patients above and below the median of each predictor are shown in [Figure 8](#pone-0052008-g008){ref-type="fig"}. The p values were calculated with the log rank test using the permutation distribution. It can be seen that the initial fast growth rate provides a near-significant difference between the two groups of patients, in contrast to the other predictors, suggesting that this predictor may be a candidate for prediction of TTP. A larger study consisting of more patients and longer follow-ups is required in order to assess the application of our maps for this purpose.
Examples of Progression and Pseudoprogression in GBM Patients Post Chemoradiation {#s3g}
---------------------------------------------------------------------------------
Within the cohort of GBM patients with follow-up, for 7 of the 8 progressing patients, progression was determined in the first MRI follow-up in which significant increase in the fast component volume was noticed. For one patient progression was determined 5 weeks after the increase in the fast component. While an increase in the fast component preceded progression in all patients, significant increase of the enhancing lesion volume was not necessarily followed by progression. An example of a patient who experienced significant increase in the enhancing lesion volume 4.2 months post treatment but remained progression free for an additional 7.4 months is shown in [Figure 9A, B, C](#pone-0052008-g009){ref-type="fig"}. An example of a patient who experienced significant increase in the fast volume 6.5 months post treatment (with no significant increase in the enhancing lesion volume) is shown in [Figure 9D, E, F](#pone-0052008-g009){ref-type="fig"}. This patient was determined to progress 6.5 months post treatment and underwent surgery.
Discussion {#s4}
==========
GBM is the most common and most aggressive type of primary brain tumor in humans. The current standard of care for newly diagnosed GBM is surgical resection (when possible) followed by radiotherapy with concomitant and adjuvant temozolomide chemotherapy. The rate of early treatment induced radiological changes which mimic tumor progression -- pseudoprogression - increased significantly since the addition of chemotherapy to the treatment regimen. Due to the increasing occurrence of brain metastases and the expending use of radiosurgery to treat them, the rate of treatment-induced radiation necrosis is rising as well. Conventional MR imaging is currently unable to provide reliable distinction between tumor recurrence and treatment effects. Clinically, this question has important consequences; therefore, a reliable distinction between these conditions is crucial.
A significant advantage of using the delayed enhancement and clearance rates instead of the commonly studied early rates (DCE and DSC) is the ability to apply sequences with lower temporal resolution, such as high resolution spin-echo T1-MRI. These sequences nearly completely avoid susceptibility artifacts while providing high signal-to-noise ratios, high resolution and high sensitivity to contrast variations. By measuring the clearance of the contrast agent from the tissue at these long delays it is reasonable to assume that the sensitivity to physiological parameters is increased, providing additional information unattainable when using short acquisition times. This assumption is supported by the two fold increase in sensitivity to tumoral (fast) tissue obtained by increasing the delay from 15 to 75 minutes. In this context it is important to note that the patients are not held in the scanner for these long times. The patients are scanned for 30 min post contrast injection and then asked to return for an additional short scan of the 75 min point.
Recent PWI studies suggest that high rCBV values are associated with tumor recurrence and low values with treatment effects. Still, up to date, there are no published studies supporting consensuses regarding which PWI-based analytic method best estimates histological tumor fraction as a predictor of progression or overall survival in GBM patients post treatment [@pone.0052008-Hu1]. Our preliminary results show significant correlation of the fast component with high rCBV values, suggesting that rCBV may be a dominant characteristic of this population. Still, in some patients we observed low rCBV values in fast regions of the subtraction maps. This may be explained by distortion of the PW images in this population of post-surgery patients due to close vicinity of the tumor to surgical screws. Distortion may also be induced by hemorrhages which are frequent in these tumor types. In some cases the contradiction between our maps, depicting a blue component representing morphologically active tumor (confirmed histologically), and low rCBV values may be explained by the low resolution of the PW images impeding the sensitivity to small tumoral regions in contrast to the high resolution/sensitivity of our maps. On the other hand, these contradicting cases may imply existence of other contributions to the fast component, such as increased vessel permeability. A study designed to mathematically model the signal time curves for determining the dominant factors contributing to the different enhancement populations is ongoing.
The most pronounced effect depicted in the delayed enhancement subtraction maps is the clear differentiation between fast and slow populations, where the terms fast and slow refer to the clearance rate of the contrast agent between the early time point (2 min) and the delayed time point (75 min). The common feature of vessels morphology in the fast regions was found to be the undamaged vessels lumens, while vessels in the slow regions presented significantly damaged lumens. Therefore, one explanation for the difference between the 2 populations may be that vessels in the fast regions provide efficient contrast clearance from the tissue, while the damaged lumens in the slow regions are unable to clear the accumulating contrast efficiently, resulting in contrast accumulation.
The association between the fast/slow components of the maps and tumor/non-tumoral tissues as determined by histological analysis is currently based on 32 biopsy samples obtained from 9 patients with primary brain tumors and 4 patients with brain metastases. Additional confirmation was obtained from 1 GBM patient and 8 metastases resected from 7 patients. These data suggest that our subtraction maps, calculated from delayed contrast extravasation MRI, enable clear differentiation between tumor and non-tumoral tissues in various types of brain tumors. Additional data is needed to establish the reliability of this conclusion. This study is ongoing.
The manner of which our maps may be applied for differentiating progression from pseudoprogression in GBM patients post chemoradiation was demonstrated in [Figure 9](#pone-0052008-g009){ref-type="fig"}. In our maps, progression is reflected by a significant increase in the fast component of the enhancing lesion while pseudoprogression is reflected by an increase in the slow component with no significant increase in the fast component. Therefore, using our maps in routine MRI follow-up may aid the physician in determining progression versus pseudoprogression in patients presenting an increase in the enhancing lesion on T1-MRI. An increase in the fast component volume, suggesting progression, implies that a change in the current therapy should be employed. No significant increase in the fast volume component, suggesting pseudoprogression, implies that the patient is responding to the current therapy and thus continuation is preferred, if possible.
The maps may be applied in a similar manner to patients following SRS. Patients with growing volumes of the slow component would be recommended for follow-up, if possible, while patients with growing volumes of the fast component would be recommended for treatment such as surgery or repeated SRS.
Due to improved treatment protocols and extended survival early and late radiation-induced neuro-toxicity of patients with brain tumors undergoing radiation-based therapies has become a major concern and efforts to minimize unnecessary exposure of surrounding normal brain without compromising treatment efficacy are of increasing interest [@pone.0052008-Rane2]--[@pone.0052008-Abe1]. The ability of our maps to depict morphologically active tumor regions with high resolution may thus be applied for optimizing radiation treatment planning by localizing the treatment to the fast/blue regions in our maps. This methodology may be similarly applied in the post surgery scenario to enable differentiation between post surgical changes and tumor remnants thus allowing depicting and treating residual tumor post surgery.
High resolution depiction of tumoral tissues may also be beneficial in the planning of surgical resections, especially in the case of close proximity to functionally eloquent brain regions. In these cases determination of the exact extent of the tumor might be crucial for the decision whether microsurgical tumor removal might be warranted. Our maps may also be applied for guiding stereotactic biopsies for molecular classification of the tumor. In this case our maps may aid in preventing the acquisition of biopsies with a high amount of necrosis or out of the infiltration zone with 'contamination' of the specimen by normal brain tissue, both potentially leading to false-negative results [@pone.0052008-Tonn1].
Novel approaches for local drug delivery including injections, infusions, trans-nasal delivery, convection enhanced delivery, local BBB disruption and various types of polymeric implants may also benefit from the application of our maps for planning and monitoring the treatments.
Increased tumor vascularity has been shown to correlate with both shortened survival and higher grade of malignancy in gliomas. Consequently, anti-angiogenic agents such as Bevacizumab, a monoclonal antibody targeting vascular endothelial growth factor, are now commonly employed to treat progressive malignant gliomas. The wide-spread use of these agents has added a layer of complexity to the evaluation and characterization of malignant gliomas as these agents have been shown to rapidly and markedly decrease contrast enhancement on contrast-enhanced T1-MRI [@pone.0052008-Russell1]--[@pone.0052008-Vredenburgh1]. We have recently initiated a study in which our subtraction maps are applied to recurrent GBM patients treated by Bevacizumab in an attempt to provide improved depiction of the tumor thus enabling better patient follow-up and understanding of Bevacizumab mechanism of action in GBM patients.
As up to 45% of suspected low-grade gliomas turn out to be high grade gliomas, histological diagnosis is mandatory before any therapeutic decision [@pone.0052008-Kunz1]. As even these partially anaplastic lesions may not display contrast enhancement on conventional contrast-enhanced T1-MRI, additional imaging information is warranted. We hope that the high sensitivity of our maps may provide additional valuable information for delineation of the tumor borders as well as for targeting stereotactic biopsies.
Additional studies are required in order to assess the reliability of these applications.
In summary, the difficulty in interpretation of post treatment images prevents appropriate treatment in nearly half of treated GBM patients and a significant number of patients with brain metastases. The presented results demonstrate the feasibility for a straightforward application of our high resolution delayed enhancement subtraction maps in the daily clinical scenario. The ability to clearly differentiate tumor from non-tumoral tissues may provide the physician with a clear understanding of the patient current situation thus enabling improved patient management.
The authors thank our patients and their families for participating in this important study. We thank Prof Zvi Ram for many productive discussions and Sharona Salomon and Shoshana Farfuri for their invaluable help in managing the regulatory aspects of the study.
[^1]: **Competing Interests:**Leor Zach, David Guez, David Last, Dianne Daniels and Yael Mardor are authors on a pending patent titled VESSEL FUNCTION MAPS (Oct 2011).
[^2]: Conceived and designed the experiments: LZ YG YM. Performed the experiments: LZ DG DL DD YG ON CH DN RS ZRC YM AT. Analyzed the data: LZ DG DL DD YG ON DN YM. Wrote the paper: LZ DG DL DD YM.
| {
"pile_set_name": "PubMed Central"
} |
1. Introduction {#sec1}
===============
We showed that tracheal intubation with atracurium significantly decreased vocal cord injuries compared with tracheal intubation without muscle relaxants (8% versus 42%) \[[@B1]\]. Tracheal intubation with atracurium at maximum neuromuscular block, however, did not decrease vocal cord injuries compared with tracheal intubation two minutes after injection of atracurium; the overall incidence was 27%, that is, higher than described in the literature (up to 12%) \[[@B2]\]. Maybe vocal cord injuries did not only occur during tracheal intubation but also during surgery and during removal of the tracheal tube.
Volatile anesthetics increase neuromuscular block of muscle relaxants; anesthesia induction with desflurane increased neuromuscular block compared with a total intravenous anesthesia \[[@B3]\]. Thus, sevoflurane as part of the anesthesia would increase neuromuscular block; moreover, sevoflurane would lengthen neuromuscular block; vocal cords, therefore, would be longer relaxed. We speculated that sevoflurane would cause less vocal cord injuries than propofol during surgery and after removal of the tracheal tube.
After surgery, we assessed hoarseness, sore throat, and vocal cord injuries---by stroboscopy---up to 72 hours. We expected that the patients receiving sevoflurane would have had a lower incidence and severity of hoarseness and vocal cord injuries than the patients receiving an intravenous anesthesia with propofol.
2. Methods {#sec2}
==========
The study was performed at the University Hospital of Rostock, Germany, between August 2010 and October 2011. Ethical approval for this study (registration number: A 2010 29) was provided by the Institutional Review Committee (Ethik-kommission der Universitat Rostock, Rostock, Germany) on 10 May 2010. The study was registrated at ClinicalTrialsGov under number NCT01616966.
After obtaining written informed consent, we studied 65 adult patients, aged 18--80 yr, ASA I--III, undergoing orotracheal intubation for surgery of the ear. All patients were examined by stroboscopy one day before surgery and were excluded from the study when preexisting pathologies of the vocal cords were found. Exclusion criteria were obesity (defined as body mass index (BMI) \> 40 kg/m²), an allergy against the study drugs, and patients with a known or suspected difficult airway (Mallampati score 3 or 4 and a mouth opening \< 3.5 cm).
2.1. Randomization and Monitoring {#sec2.1}
---------------------------------
Patients were randomized---by a study nurse---in two groups, according to the randomization list, which was prepared via a computer-generated randomization program, as follows \[[@B4]\]: SEVO group, receiving sevoflurane during anesthesia, or TIVA group, receiving an intravenous anesthesia with propofol. We used neuromuscular monitoring to have comparable muscular block at time of intubation in both groups; neuromuscular monitoring was performed with the TOF Watch SX device (Organon Teknika, Eppelheim, Germany). We used neuromonitoring to have comparable depth of anesthesia between the study groups; neuromonitoring was performed with the BIS Vista brain monitoring system (Aspect Medical Systems, Norwood, MA, USA).
The acceleromyographic measurements were done according to the guidelines of Good clinical research practice in pharmacodynamic studies of neuromuscular blocking agents from 2007 \[[@B5]\]. We used a calibrated TOF watch SX. After recalibration of the acceleromyography, rocuronium 0.45 mg/kg was injected over 5 s. Time to a train-of-four (TOF) ratio of 1.0 (defined as the time from start of injection of rocuronium until the TOF ratio reached 1.0) was measured.
The bispectral index (BIS) monitoring was applied according to the manufacturer\'s guidelines; BIS Quatro sensor electrodes were placed on the patient\'s forehead. The BIS values were noted continuously every minute during surgery. The target BIS was between 40 and 50. In the TIVA group, propofol infusion was increased from 4.0 mg/kg/h by 1.0 mg/kg/h than the BIS value was above 50; propofol infusion was decreased by 1.0 mg/kg/h than the BIS value was under 40 (minimum dosage was 4.0 mg/kg/h). In the SEVO group, sevoflurane was increased by 0.1 vol.% than the BIS value was above 50; sevoflurane was decreased by 0.1 vol.% than the BIS value was under 40 (minimum concentration was 1.0 vol.%).
2.2. Induction and Maintenance of Anesthesia {#sec2.2}
--------------------------------------------
One hour before the beginning of surgery, the patients received midazolam 7.5 mg orally. Induction was standardized for both groups. Remifentanil 0.4 *μ*g/kg/min was applied continuously for 2 min; afterwards, propofol 2.0 mg/kg was administered. After calibration of the TOF Watch SX, rocuronium 0.45 mg/kg was given. If the neuromuscular block was incomplete, rocuronium 0.15 mg/kg was added. Tracheal intubation was performed, when maximum block was achieved. All tracheal intubations were performed by the same anesthesiologist to control interindividual differences.
Maintenance of anesthesia was standardized; TIVA group: propofol 4.0 mg/kg/h and remifentanil 0.25--0.35 *μ*g/kg/min; SEVO group: sevoflurane 1.0 vol.% in a 50% oxygen-air mixture and remifentanil 0.25--0.35 *μ*g/kg/min. Fifteen minutes before the expected end of surgery, all patients received piritramide 0.05--0.1 mg/kg i.v. The patient\'s tracheas were extubated, when the TOF ratio was 1.0 and patients opened their eyes or began to cough; afterwards, the patients were moved to the postanesthesia care unit (PACU).
2.3. Assessment of Hoarseness, Sore Throat, and Vocal Cord Injuries {#sec2.3}
-------------------------------------------------------------------
Hoarseness was defined as an acoustic quality that was different than the previous voice quality of the patient \[[@B6]\]. Sore throat was defined as continuous throat pain \[[@B7]\]. In the PACU and 24 h, 48 h, and 72 h after surgery, an investigator blinded to the group assignment of the patients assessed the incidence and severity of hoarseness and sore throat (see Appendix). If hoarseness or sore throat persisted over 72 h, a daily follow-up examination was performed until complete restitution. Vocal cord injuries were assessed by laryngostroboscopy by an ear-nose-throat physician who was unaware of the patient\'s group assignment (see Appendix); all examinations were performed by the same ear-nose-throat physician.
Compared to indirect laryngoscopy, laryngostroboscopy can detect functional disorders of the vocal cords, such as dysfunction of the mucosal wave. All patients were examined by stroboscopy 24 h after surgery. When hoarseness was lasting longer than 48 h, a second stroboscopy was performed 72 h after surgery.
2.4. Assessment of Intubating Conditions, Intubating Variables, and Extubating Conditions {#sec2.4}
-----------------------------------------------------------------------------------------
The intubating conditions were assessed according to the consensus conference on Good Clinical Research Practice (GCRP) in Pharmacodynamic Studies of Neuromuscular Blocking Agents \[[@B5]\]. In addition, the following variables were assessed: Cormack\'s grades, time for intubation, the number of intubation attempts (see Appendix), and the extubating conditions---excellent = no coughing, good = slight coughing, and poor = sustained coughing during removing of the tracheal tube. The following factors were standardized: tube size (men: ID = 8.0 mm; women: ID = 7.0 mm), type of tube, and cuff inflation with air (cuff pressure was ≤25 mmHg) \[[@B1], [@B8]\].
2.5. Statistical Analysis {#sec2.5}
-------------------------
Statistical analysis was performed using the SigmaStat for Windows Version 3.5, (Systat Sotware Inc., San Jose, California, USA). Demographic data were analyzed using Mann-Whitney *U*-test or *t*-test; results are presented as mean (SD) or as median (range). Comparisons between groups were performed using the *χ*² test, Fisher\'s exact test, or Kruskal-Wallis\' ANOVA test. Results were considered statistically significant, when *P* \< 0.05.
The sample size calculation was based on the study by Maruyama et al. \[[@B9]\]. The required number of patients for our study groups was calculated on the assumption that 55% of the patients suffered from hoarseness after a total intravenous anesthesia \[[@B9]\] and 16% after an anesthesia with desflurane \[[@B1]\]. For an 80% power and an *α* = 0.05, 52 patients (26 patients in each group) were needed. To compensate for possible drop-outs, we enrolled 59 patients (10% more than needed).
3. Results {#sec3}
==========
Six patients (10%) had preexisting pathologies at the vocal folds; consequently, these patients were excluded from the study. 59 patients were randomized for the study, 30 patients in the TIVA group, and 29 patients in the SEVO group. One patient---from the TIVA group---was excluded because of a Cormack grade 3 ([Figure 1](#fig1){ref-type="fig"}). There were no significant differences in the patient\'s characteristics ([Table 1](#tab1){ref-type="table"}). Dosages of propofol, remifentanil, and rocuronium and end-tidal concentrations of sevoflurane are shown in [Table 2](#tab2){ref-type="table"}.
3.1. Intubating Conditions, Intubating Variables, and Extubating Conditions {#sec3.1}
---------------------------------------------------------------------------
The patient\'s tracheas were extubated at a TOF ratio of 1.0. Patients in the SEVO group had a significantly longer neuromuscular block compared with the TIVA group ([Table 3](#tab3){ref-type="table"}; [Figure 2](#fig2){ref-type="fig"}). Tracheal intubation was successful in all patients of both groups. Intubating variables and intubating conditions were comparable (not significant). Overall 39 patients (67%) had coughing during removal of tracheal tubes; 17 (59%) patients in the SEVO group versus 22 (76%) patients in the TIVA group; *P* = 0.26 ([Table 3](#tab3){ref-type="table"}).
3.2. Hoarseness, Sore Throat and Vocal Cord Injuries {#sec3.2}
----------------------------------------------------
The overall incidence (TIVA and SEVO groups together) of hoarseness was 10% (6 patients); the overall incidence of sore throat was 15% (9 patients). The incidence was comparable between groups ([Table 4](#tab4){ref-type="table"}); the severity was comparable, too (data not shown). The overall incidence of vocal cord injuries after surgery was 24% (14 patients). There was no significant difference between the TIVA and SEVO groups: 5 versus 9 patients (*P* = 0.36). No patient suffered from hoarseness or had vocal cord injuries longer than 3 days. Vocal cord injuries were edema in 7 patients and erythema in 7 patients ([Figure 3](#fig3){ref-type="fig"}); there were no hematoma and no granuloma. The majority of vocal cord injuries were unilateral (left: 3; right: 6 patients); 5 patients had bilateral vocal cord injuries (1 patient; SEVO group versus 4; TIVA group; *P* = 0.35). A dysfunction of the mucosal wave was found in 20 patients: 8 patients in the SEVO group and 12 patients in the TIVA group.
Postoperative nausea and vomiting (PONV) were not observed in the PACU. The BIS values are shown in [Figure 4](#fig4){ref-type="fig"}.
4. Discussion {#sec4}
=============
We showed that both anesthesia techniques---anesthesia with sevoflurane and intravenous anesthesia with propofol---may be used. The rate of vocal cord injuries was comparable between the two anesthetics. The risk for postoperative residual curarization, however, was higher in the sevoflurane group.
The incidence of hoarseness and sore throat was 10% in the TIVA group; this is lower than reported by Maruyama et al. \[[@B9]\]. He found hoarseness in 55% and sore throat in 50% of patients after a total intravenous anesthesia. The reason is, in our opinion, that the general anesthesia was maintained with propofol, fentanyl, and ketamine in his study. In our study, general anesthesia was maintained with propofol and remifentanil---a potent ultra short-acting synthetic opioid drug---which was applied continuously. Therefore, the depth of anesthesia, measured by BIS monitoring, was stable during surgery (between 29 and 42). BIS values between 45 and 60 are recommended for a balanced anesthesia \[[@B10]\]. An anesthesia with fentanyl as a bolus technique provides unstable consciousness states during surgery compared with remifentanil.
The trauma causing laryngeal injury can occur on several occasions \[[@B11]\]: during tracheal intubation, during surgery, and during tracheal extubation. We standardized the anesthesia induction and tracheal intubation to control the risk factors for laryngeal injury.
During surgery, the head was moved only slightly by the surgeon. The incidence of vocal cord injuries was 24% in the present study. Laryngeal injuries following tracheal intubation vary between 4% and 12% \[[@B12]--[@B14]\]; without muscle relaxant it was 42% \[[@B1]\]; with double-lumen tube it was 44% \[[@B15]\]; with intravenous anesthesia it was 27% \[[@B2]\]. We found only minor injuries such as erythema and edema at the vocal folds. The majority of the vocal cord injuries were unilateral with 65%, indicating a slight injury during tracheal intubation or resulting from the movement of the head during surgery of the ear. Bilateral vocal cord injuries---as a typical sign for an injury during removal of the tracheal tube---were observed only in 35%. It is difficult to determine, when the vocal cord injuries occurred. We performed tracheal intubation, when maximum neuromuscular block was achieved; therefore, we hope that the baseline injuries---caused by tracheal intubation---are comparable in both groups. Bilateral injuries occur when the vocal cords beat against the tracheal tube during surgery or during removal of the tracheal tube. The incidence of coughing at tracheal extubation was comparable between groups (59% SEVO group versus 76% TIVA group; *P* = 0.26); the incidence of bilateral injuries was comparable, too (1 patient SEVO group versus 4 TIVA group; *P* = 0.35).
Inhalational agents may increase the neuromuscular block by a central mechanism; this may explain why the time of muscle relaxation in the SEVO group was increased compared with the TIVA group. Time to a TOF ratio of 1.0 was significantly longer in the SEVO group; doses of rocuronium were comparable between groups. Five patients in the TIVA group needed neostigmine to achieve a TOF ratio of 1.0; eleven patients in the SEVO group, however, needed neostigmine (*P* = 0.14); this is not significantly related, but this is clinically relevant. The risk for residual curarization is increased with sevoflurane. Residual curarization in the postanesthesia care unit is associated with an increased incidence of critical respiratory events, such as hypoxemia \[[@B16]\]. In our study, all patients were monitored with a calibrated TOF Watch SX; therefore, patient\'s tracheas were extubated, when the TOF ratio had reached 1.0.
We demonstrated that sevoflurane did not decrease laryngeal morbidity; the risk for residual curarization, however, was higher compared with total intravenous anesthesia. Neuromuscular monitoring is---especially under sevoflurane anesthesia---necessary to detect residual neuromuscular blockade.
Assessment of Hoarseness, Sore Throat, and Vocal Cord Injuries {#secA}
==============================================================
\(A\) *Hoarseness* \[[@B1]\]. Do you have any hoarseness at all since your operation? If the answer was no, hoarseness was graded 0 = none; if the answer was yes, hoarseness was recorded as follows: 1 = noticed by patient, 2 = obvious to observer, and 3 = aphonia.
\(B\) *Sore Throat* \[[@B7], [@B17]\]. Do you have any sore throat? If the answer was no, sore throat was graded 0 = no sore throat; if the answer was yes, sore throat was recorded as follows: 1 = mild (pain with deglutition), 2 = moderate (pain present constantly and increasing with deglutition), and 3 = severe (pain interfering with eating and requiring analgesic medication).
\(C\) *Vocal Cord Injuries* \[[@B1], [@B2], [@B11]\]. Location: unilateral (left or right vocal cord) or bilateral (both vocal cords).
Type of injury: erythema = redness of the mucosa with surrounding inflammatory swelling, edema = swollen mucosa at the vocal folds, hematoma = bleeding into vocal cord, granuloma = granulation tissue remains as chronic, localized, one rounded tissue, and dysfunction of the mucosal wave = abnormal movement of the vocal cords.
\(D\) *Intubating Variables*. These variables are as follows.Time for intubation = time in seconds from the initial inserting of the laryngoscope into the patient\'s mouth until blocking of the cuff.Number of intubation attempts.
{#fig1}
{#fig2}
{#fig3}
{#fig4}
######
Demographic data, duration of surgery, and anesthesia.
SEVO group (*n* = 29) TIVA group (*n* = 29) *P* value
------------------------------ ----------------------- ----------------------- -----------
Age (yr) 47 (17) 49 (15) 0.76
Weight (kg) 75.0 (14.2) 81.2 (16.8) 0.13
Height (cm) 172.2 (8.9) 173.5 (9.8) 0.61
Body mass index (kg/m²) 25.2 (4.1) 26.7 (3.1) 0.13
Gender ratio (female/male) 15/14 11/18 0.43
Smoking 10 (34%) 12 (41%) 0.78
Reflux 3 (10%) 4 (14%) 1.00
Duration of surgery (min) 75 (37) 73 (33) 0.87
Duration of anesthesia (min) 102 (36) 94 (34) 0.33
Values are mean (SD) or numbers (%). SEVO group: anesthesia with sevoflurane and remifentanil. TIVA group: anesthesia with propofol and remifentanil.
######
Doses of rocuronium, propofol, remifentanil, and end-tidal concentrations of sevoflurane and administration of neostigmine.
SEVO group (*n* = 29) TIVA group (*n* = 29) *P* value
------------------------------------------- ----------------------- ----------------------- -----------
Propofol for induction of anesthesia (mg) 150 (100--280) 170 (90--260) 0.37
Remifentanil (*µ*g/kg/min)\* 0.256 (0.015) 0.259 (0.026) 0.98
Propofol (mg/kg/h)\* --- 4.9 (0.8)
Sevoflurane (vol.%)^†^ 1.3 (0.3) ---
Rocuronium (mg) 30 (25--55) 40 (25--70) 0.17
Administration of neostigmine (*n*) 11 5 0.14
Values are mean (SD), median (range), or numbers. SEVO group: anesthesia with sevoflurane and remifentanil. TIVA group: anesthesia with propofol and remifentanil. \*Mean dosage during maintenance of anesthesia. ^†^Mean end-tidal concentration of sevoflurane during anesthesia.
######
Neuromuscular measurements, intubating variables, and coughing during removal of the tracheal tube.
---------------------------------------------------------------------------
SEVO group\ TIVA group\ *P* value
(*n* = 29) (*n* = 29)
---------------------------------- -------------- ------------- -----------
Neuromuscular measurements
Time till TOF ratio = 1.0 (min) 71 (38--148) 52 (21--74) \<0.001
Time without relaxation (min) 14 (2--151) 34 (0--118) 0.03
Intubating variables
Cormack grades 1/2 17/12 19/10 0.78
Time for intubation (s) 16 (9--170) 16 (9--26) 0.49
Attempts (*n*) 1/2/3 25/3/1 29/0/0 0.12
Extubating conditions
Coughing 17 (59%) 22 (76%) 0.26
---------------------------------------------------------------------------
Values are median (range) or numbers (%). Time till TOF ratio = 1.0 (min) = time from start of injection of rocuronium until the TOF ratio reached 1.0. Time without relaxation = time from TOF ratio of 1.0 till tracheal extubation. SEVO group = anesthesia with sevoflurane and remifentanil. TIVA group = anesthesia with propofol and remifentanil.
######
Incidence of hoarseness, sore throat, and vocal cord injuries.
Hoarseness Sore throat Vocal cord injuries
------------ ------------ ------------- --------------------- --------- --------- ------ --------- --------- ------
PACU 2 2 1.00 2 1 1.00 --- --- ---
At 24 h 1 1 1.00 5 2 0.42 9 5 0.36
At 48 h 0 0 --- 2 1 1.00 --- --- ---
At 72 h 0 0 --- 1 1 1.00 0 0 1.00
\>72 h 0 0 --- 0 1 1.00 --- --- ---
Patients\* 3 (10%) 3 (10%) 1.00 6 (21%) 3 (10%) 0.47 9 (31%) 5 (17%) 0.36
Values are shown as numbers of patients (%).\*Patients = number of patients with hoarseness, sore throat, or vocal cord injuries (without dysfunction of mucosal wave). PACU: postanesthesia care unit. SEVO group: anesthesia with sevoflurane and remifentanil. TIVA group: anesthesia with propofol and remifentanil.
[^1]: Academic Editor: Kouichiro Minami
| {
"pile_set_name": "PubMed Central"
} |
1. Introduction {#sec1}
===============
NR1I2 (nuclear receptor subfamily 1 group I member 2) was discovered in 1998 and named as PXR (pregnane X receptor) or PAR (the receptor activated by pregnane) based on its activation by endogenous pregnanes 21-carbon steroids. Besides, human PXR is known as SXR (steroid and xenobiotic receptor)[@bib1], [@bib2]. PXR is enriched in small intestine, duodenum, liver, rectum, colon and gallbladder, while its expression in other organs/tissues is either low or undetectable[@bib3] ([Fig. 1](#fig1){ref-type="fig"}). This specific distribution of PXR in the enterohepatic system renders its crucial role as a sensor for environmental cues and inducer of xenobiotic response[@bib2].Figure 1The distribution patterns of PXR. (A) The global expression profiles of PXR in human body. High (15--30), medium or low (0--15) and not detected (below cutoff) level of TPM are indicated in red, pink and gray areas respectively[@bib3] (Image available from <https://www.ebi.ac.uk/gxa>). (B) Transcripts per million (TPM) of PXR in different organs and systems (Data available from [v18.proteinatlas.org](http://v18.proteinatlas.org){#intref0030}. <https://www.proteinatlas.org/ENSG00000144852-NR1I2/tissue>). (C) Expression of *PXR* mRNA detected by microarray and RNAseq in cell lines from different systems. (D) Expression of *PXR* mRNA in some representative cell lines from different systems (Adapted with permission from <https://portals.broadinstitute.org/ccle>).Figure 1
PXR can be activated by both endobiotic and xenobiotic chemical compounds. Besides pregnane, steroid, bile acids and other endobiotic chemicals, various clinical drugs and environmental pollutants have been demonstrated to activate PXR[@bib1], [@bib4], [@bib5], [@bib6]. Activated PXR, through direct binding to the genomic regions or indirect crosstalk with other transcriptional factors, controls many genes involved in biotransformation, transport, inflammation, cell cycle arrest, apoptosis and oxidative stress. Many of these biological signaling pathways are closely related to tumorigenesis, indicating an essential function of PXR in cancer development and progression. Indeed, PXR manifests its potential in prevention and treatment for cancers developed from toxic exposure[@bib7], [@bib8], virus infection[@bib9], and recurrent inflammation[@bib10]. In this review, we summarized the current understanding of the biological function and regulatory mechanism of PXR in the context of cancer.
2. Gene and protein of PXR {#sec2}
==========================
Human *PXR* gene locates in cytoband q13.33 of chromosome 3 with 38507 SNPs (single nucleotide polymorphisms), 2394 deletions, 1403 insertions, 18 substitutions, 8 indels, 4 genetic markers, 17 sequence alterations, 32 tandem repeats and 788 somatic sequence alterations, some of which cause monstrous change of structure and functions of PXR protein[@bib11], [@bib12], [@bib13].
PXR protein, approximately 50 kDa, consists of the N-terminal ligand-independent activation function 1 (AF-1), the DNA binding domain (DBD), the relatively short hinge region, and the ligand binding domain (LBD) which contains the ligand-dependent activation function 2 domain (AF-2)[@bib14] ([Fig. 2](#fig2){ref-type="fig"}). With the unique flexible large conformation in LBD, the capacity of binding and recognizing for a wide variety of hydrophobic ligands makes PXR a multifunctional receptor[@bib15]. It is interesting that PXR could form a homodimer unique to other nuclear receptors (NRs) by tryptophan-zipper (Trp-Zip) interaction in LBD domain, while disruption of this homodimer will significantly deprive PXR activity and its recruitment ability for transcriptional coactivator steroid receptor coactivator 1 (SRC1)[@bib16]. The sequence-specific DNA identification by the DBD of PXR is another aspect for regulating the transcriptional activation[@bib1], [@bib17]. A sub-region composing 11 sequence-specific amino acid residues called mitotic chromatin binding-determining region (MCBR) within the nuclear localization signal (NLS) mediates the binding of PXR to the DNA[@bib18]. Mechanically, PXR DBD preferentially binds to the DR (direct repeats)-3, DR-4 (the most preferred DNA-binding motif), DR-9, DR-14, DR-19[@bib17] and ER (everted repeats)-6, ER-8[@bib18], [@bib19] in the promoter region of the target genes. Genetic alterations within any domain of PXR could lead to the change of its function. For example, some splicing variants have been reported to have functional deficiency or emulative suppression due to the alteration of encoded amino acid sequences[@bib20], [@bib21]. Furthermore, other alternatively spliced isoforms of PXR, whose biological function has not been fully understood, might have some unanticipated roles.Figure 2Major domains of hPXR protein. Blue: AF-1 domain; red: the DNA binding domain; green: hinge; and yellow: the ligand binding domain (contains AF-2 domain).Figure 2
3. The regulatory mechanism of PXR activity {#sec3}
===========================================
Along with the characterization of the transcriptional activity, the multidimensional regulatory mechanism of PXR has been revealed, including the genetic and epigenetic regulation for PXR expression, transcriptional regulation, subcellular localization, ligand-dependent activation, and protein--protein interaction. The transcriptional activity of PXR also can be modulated through crosstalk with many other NRs, including farnesoid X receptor (FXR)[@bib22], constitutive androstane receptor (CAR)[@bib23], [@bib24], [@bib25], [@bib26], peroxisome proliferator-activated receptor *α* (PPAR*α*)[@bib27], liver X receptor (LXR)[@bib19], [@bib28], and androgen receptor (AR)[@bib29] ([Fig. 3](#fig3){ref-type="fig"}).Figure 3The regulatory mechanism of PXR. Part of regulatory factors for PXR could induce promotion or repression of PXR activity, or crosstalk with PXR by different modes.Figure 3
3.1. Epigenetic regulation for PXR {#sec3.1}
----------------------------------
The role of epigenetic modulation for PXR transcripts has been defined. Recently, microRNA (miR)-34a, miR-140-3p, miR-148a and miR-449a were found to downregulate the expression of PXR through the identification and interaction at the 3ʹ-untranslated region (3ʹ-UTR) of *PXR* mRNA in the hepatocellular carcinoma cell lines, which might augment the sensitivity of anti-cancer medicines[@bib30], [@bib31], [@bib32]. However, among Chinese Han population, no correlation between miR-148a and the expression of PXR or cytochrome P450 3A4 (CYP3A4) was found in livers[@bib33]. Altered 3ʹ-UTR derived from several SNPs of PXR, including rs3732360, rs1054190 and rs1054191, could change the original binding with miR-500a-3p, miR-532-3p and miR-374a-3p[@bib34]. This observation reflected how confound influence the epigenetic modulation and inter-individual variability may have on the activity of PXR. In addition to our limited understanding about miRNA-mediated silence of PXR, recently, PXR activation-mediated regulation for long non-coding RNA (lncRNA) has been shown in xenobiotic metabolism[@bib35] for the first time, indicating that it remains an open field regarding the role of non-coding RNAs (ncRNAs) for PXR activity.
Moreover, within the upstream promoter of PXR transcripts, methylation is a critical modification responsible for reduction of the expression of variant PXR isoforms[@bib21], [@bib36], while demethylation agents, such as 5-aza-2-deoxycytidine, could serve as an inducer to increase the expression of PXR isoforms[@bib21], which might be associated with biology and therapeutic outcomes of hepatocellular carcinoma (HCC)[@bib21], [@bib36]. Recently, some transcription factors have been shown to regulate PXR abundance as well. For example, transcription factor E26 transformation specific sequence 1 (ETS-1)[@bib37] and *N*-*α*-acetyltransferase 10 (NAA10)[@bib38], could interact with PXR promoter, and thus enhance the activation of downstream drug resistance related genes.
3.2. Post-translational modifications (PTMs) of PXR {#sec3.2}
---------------------------------------------------
PTMs also play a pivotal role in regulating PXR activity ([Table 1](#tbl1){ref-type="table"}[@bib32], [@bib39], [@bib40], [@bib41], [@bib42], [@bib43]). Phosphorylation[@bib44], [@bib45], acetylation[@bib40], SUMOylation[@bib42], [@bib43], poly (ADP-ribosyl)ation[@bib39], and ubiquitination[@bib32] mediated by modification enzymes could substantially cause a dynamic change of biological traits, subcellular localization, dimerization, protein stability, co-regulator interaction and degradation patterns of PXR. The activity and bioeffects of PXR are reformed as a consequence of PTMs. Of note, *O*-GlcNAcylation and other PTMs have not been defined hitherto.Table 1Post-translational modifications and sites of PXR protein.Table 1PTMSitePoly (ADP-ribosyl)ationLBD[@bib39]AcetylationK109[@bib40], K170[@bib41]SUMOylationK108, K129, K160, K170[@bib32], [@bib42], [@bib43]Phosphorylation[a](#tbl1fna){ref-type="table-fn"}S8, T57, S114, T133, T135, S167, S200, S208, T248, Y249, S256, S274, T290, S305, S350, T408, T422Ubiquitination[a](#tbl1fna){ref-type="table-fn"}K101[^1]
Protein kinase A (PKA), protein kinase C (PKC) *β*, cyclin-dependent kinase 1 (CDK1), CDK2, CDK5, cyclin A/E, casein kinase II (CK2), glycogen synthase kinase 3 (GSK3), mitogen-activated protein kinase kinase 1 (MEK1) pathway[@bib47], and 70 kDa form of ribosomal protein S6 kinase (S6K) could mediate the phosphorylation of PXR and mostly repress the activity of PXR protein by retaining the PXR protein in cytoplasm, isolating PXR from intranuclear DNA, thus diminishing the transactivation of downstream genes[@bib44], [@bib45]. Additionally, PKA activation facilitates the ubiquitination of PXR protein. E3 ubiquitin ligase ring-B-box-coiled-coil protein interacting with protein kinase C-1 (RBCK1) and several others could directly bind and ubiquitinate PXR, resulting in degradation of PXR. Suppressor for gal 1 (SUG1), a subunit of the proteasome, might contribute to the formation of proteolytic fragments of PXR as well. It is noteworthy that the increased level of ubiquitinated PXR was also observed after treatment by MG132 (the inhibitor of 26S proteasome), suggesting PXR is subjected to proteasomal degradation[@bib32].
The E1A binding protein p300 is capable of acetylating PXR at lysine 109 (K109) as the major acetylation site in the hinge, repressing PXR transcriptional activity due to loss of dimerization with RXR*α* and DNA binding. This modification can be depressed by sirtuin 1 (SIRT1)[@bib40]. Moreover, the lysine acetyltransferase, TIP60, could interact with LBD region of unliganded PXR and acetylate PXR at lysine 170 with a forfeit of ligand-dependent PXR target gene transactivation, which might promote cell migration and adhesion[@bib41]. Recent studies have focused on the involvement of acetylation of PXR in SUMO (small ubiquitin-related modifier)--acetyl switch. On account of acetylation of PXR, SUMOylation of PXR has been stimulated to repress the expression of PXR\'s target gene[@bib42]. On the contrary, through the interaction with negative charge amino acid-dependent SUMOylation motif (NDSM) in PXR, E2-conjugation enzyme UBCh9-dependent SUMOylation has been demonstrated to activate PXR. It is worth noting that the NDSM-mutated PXR (D115A) is lack of the SUMOylation event[@bib32], [@bib43]. Similarly, upon interaction with the C-terminal LBD of PXR, poly (ADP-ribose) polymerase 1 (PARP1) could directly bind and poly (ADP-ribosyl)ate PXR *via* the BRCA1 C terminus (BRCT)/automodification domain (AMD), facilitating the recruitment of PXR to the promotor of target genes and the transactivation of these target genes. Whereas, this positive regulation can be blocked by PARP1 inhibitor 3AB[@bib39].
3.3. Ligand-dependent activation for PXR {#sec3.3}
----------------------------------------
Recent studies have expanded the profile of upstream activator of PXR, including clinical drugs[@bib48], [@bib49], dietary supplements[@bib50], environmental pollutants[@bib7], endobiotics[@bib51] and other chemicals[@bib52]. Unlike many other nuclear receptors, PXR activation can be specie-specific due to the distinction of LBD[@bib53] yet produce similar transcription--regulation profiles due to the conserved DBD[@bib17]. It is worth mentioning that a number of ligands for PXR also could activate other NRs, such as CAR[@bib26] and LXR[@bib19], which impedes the development of PXR targeted therapy, and further complicate their crosstalk. PXR also responds to diverse ligands with different binding modes[@bib54], making it an abstruse target for disease therapy.
Rifampin[@bib48], [@bib49], [@bib55], Rifaximin[@bib56], St. John\'s Wort[@bib6], PCN (pregnenolone-16*α*-carbonitrile)[@bib57] and SR12813[@bib58] are classical agonists for PXR. Recently, a growing number of compounds have been established as PXR activators, basing on their binding capability to LBD and enhancement for the transactivation effect of PXR, such as nontaxane microtubule-stabilizing agents[@bib59], alismanin A[@bib60], the Chinese herbal medicine *Sophora flavescens*[@bib61], U0126[@bib47], PF-06282999[@bib62] and a series of 4-methylenesteroid derivatives isolated from *Theonella* marine sponges[@bib15]. Furthermore, some target genes of PXR, such as sphingomyelin phosphodiesterase acid-like 3A (*SMPDL 3A*)[@bib19], a hepatic nucleotide phosphodiesterase and phosphoramidase involved in purinergic metabolism and anti-inflammatory signaling pathways, is repressed by nonligand-dependent PXR while activated by PXR deficiency or ligand-dependent PXR, which manifests the influence of ligand-dependent activation for PXR on its regulation effects.
3.4. Antagonist-induced abrogation for PXR {#sec3.4}
------------------------------------------
A large number of antagonists have been reported to inactivate PXR as well, such as some active plant ingredients of sulphoraphane, coumestrol, milk thistle (silybin and isosilybin), valerian and other complementary and alternative medicines (CAM) for cancer[@bib63], the marine sulfated steroids solomonsterols A and B sourced compounds [@bib19], [@bib20], [@bib21], [@bib22], [@bib23], [@bib24]^,^[@bib64], metformin, ketoconazole, sulforaphane and SPA70[@bib52], [@bib58]. Thereinto SPA70[@bib52] is a potent selective antagonist for human PXR, suggesting the PXR targeted therapy may indeed be feasible for drug resistance in cancer. Moreover, *PXR* gene encodes some alternatively spliced isoforms, some of which exert antagonistic functions due to their competitive occupation to activators and absent interaction with target genes[@bib21]. Nevertheless, their influence on disease development remains elusive.
3.5. Cofactors of PXR {#sec3.5}
---------------------
Several transcription cofactors of PXR have been identified to either enhance or suppress the activity of PXR, depending upon the binding of coactivators' Leu-Xxx-Xxx-Leu-Leu (LXXLL) motifs and corepressors\' Ile/Leu-Xxx-Xxx-Ile/Val-Ile motifs to the AF-2 region of PXR[@bib65]. Those co-activators include SRCs[@bib32], [@bib44], [@bib58], forkhead box O 1 transcription factor (FOXO1), PPAR gamma coactivator 1*α* (PGC-1*α*)[@bib44], [@bib65], phosphatidylethanolamine binding protein (PBP), and protein arginine methyl transferase (PRMT)[@bib45], while those co-repressors include the silencing mediator for retinoid and thyroid hormone receptors/NR corepressor (SMRT/NCoRs)[@bib45], [@bib66], small heterodimer partner (SHP)[@bib47], Sterol regulatory element binding protein 1 (SREBP-1), and forkhead box A 2 transcription factor (FOXA2)[@bib65].
With the intrinsic histone acetyltransferases (HAT) activity, SRC1 could interact with PXR and further recruit secondary coactivators and histone modifying enzymes, such as CBP, p300, coactivator associated arginine methyltransferase 1 (CARM1) and PRMT1, to form a transcriptional complex, allowing the fixation of ligands and expression of target genes. This coactivation can be potentially disrupted by PXR inhibitor---metformin, which is widely used in diabetic patients to improve the metabolism of glucose and lipids[@bib58], [@bib65]. However, there are several reports suggesting that SRC2 but not SRC1 could co-activate PXR activity. In human liver cells, non-phosphorylated serum- and glucocorticoid-regulated kinase 2 (SGK2) has been demonstrated to be involved in PXR mediated co-activation for gluconeogenic genes---phosphoenolpyruvate carboxykinase (*PEPCK*) and glucose-6-phosphatase (*G6Pase*), thereby enhancing gluconeogenesis[@bib65]. Additionally, our previous data has shown hepatitis B virus (HBV) X protein (HBx) could act as a co-factor of PXR in HBV positive HCC[@bib67].
Opposing to the effect of coactivators, SMRT*α*, abundantly expressed in most human tissues and cancer cell lines, could interact with PXR through the 46-amino acid insert and the C terminal corepressor motif. This interaction is resistant to PXR ligand-induced dissociation and elicits an efficient transcriptional repression for PXR. Another major isoform of SMRT, SMRT*τ*, also inhibits PXR but with less potency than SMRT*α*[@bib66]. Previous studies also indicated that SHP is another corepressor for PXR[@bib47]. Nevertheless, this co-regulation of SHP is absent in metformin-mediated suppression for PXR--CYP3A4 pathway[@bib58]. Conversely, PXR has been proposed to attenuate *SHP* promoter activity and to repress *SHP* gene transcription[@bib65], suggesting PXR--SHP interaction might be mutual.
4. Pleiotropic regulation of PXR in cancer {#sec4}
==========================================
PXR could behave as a node of multiple signaling axes to coordinate disease progressions. The paradigm of PXR action is that, in response to the change of cellular environment, the intranuclear PXR could form heterodimer with retinoid X receptor (RXR) and bind to the specific responsive elements of genes\' regulatory domain[@bib68], and then PXR extensively manipulate the expression of downstream genes, including groups of biotransformation enzymes, transport proteins[@bib7], [@bib8], inflammatory factors, cell cycle associated proteins and anti-oxidation factors. The expression of the diverse target genes subordinates to the activity of PXR, triggering alteration in detoxification, metabolism, inflammation inhibition, cell apoptosis, cell cycle arrest, proliferation inhibition, tumor migration and anti-oxidative stress. Therefore, PXR and these targets could constitute complex cellular circuits that directly participate in various physiological and pathological progressions ([Fig. 4](#fig4){ref-type="fig"}). Due to its extensive biological regulation, the effects of PXR are noted in cancer initiation, promotion and progression, as well as in chemotherapy outcome.Figure 4Multifarious target gene-dependent biological effects of PXR. PXR and its target genes constitute complex cellular circuits to participate in cancer-related physiological and pathological progressions.Figure 4
It is generally accepted that cancer, characteristic of multistage progression and diverse etiology, is always diagnosed late and limited in therapy, which highlights the significance of prevention for it. The strategy of prevention based on the pathogenesis of cancer might lie on the intervention of toxic exposure, pathogenic infection, repeating inflammation, immune deficiency, endocrine dyscrasia and other risk factors. Accumulating evidence strongly points to the significant capacity of PXR in detoxification, defense, homeostasis maintaining and proliferation inhibition, which are antagonistic for cancer development. However, PXR and its target genes also have been reported in the association with multidrug resistance and poor chemotherapy outcome in cancer treatment[@bib69], although the mechanism of chemoresistance caused by PXR remains controversial. The pleiotropic effects of PXR in cancer are not completely explicit. It should be emphasized that the location in frontline metabolic organs, the response for internal compound imbalance, and the influence on cellular signal pathways together depict the context-specificity of PXR[@bib70], highlighting PXR might be an intrinsic central target of the huge regulatory network of cancer.
4.1. PXR in metabolism modulation {#sec4.1}
---------------------------------
Activated PXR could mediate xenobiotic detoxification, inhibit hepatic steatosis, and maintain the homeostasis of endobiotic chemicals (such as heme, bilirubin, thyroxin, bile acids, bilirubin, vitamin D, glucose and lipid)[@bib15], [@bib65], some nutrients and steroid hormones by promoting biotransformation and elimination of endobiotics and xenobiotics in normal organs. The effect of PXR on detoxification and homeostasis is exemplified by the expression of its target genes, including *CYP*s[@bib7], [@bib62], carboxylesterases[@bib8], [@bib71], *PEPCK*, *G6Pase*, estrogen sulfotransferase 1E1 (*SULT1E1*)[@bib65], glutathione S-transferases (*GST*s), glutathione peroxidase (*GPx*)[@bib72], ATP-binding cassette family proteins (*ABC*s), organic anion transporting polypeptides (*OATP*s), UDP-glucuronosyltransferases 1A1 (*UGT1A1*)[@bib71], glucose transporter 2 (*GLUT2*)[@bib57] and *MDR*s[@bib73]. Hydrolytic biotransformation and excretion by such enzymes and transport proteins[@bib53], [@bib57], [@bib62] increase water-solubility and elimination of toxic chemicals and reduce the accumulation and toxicity of substrates. Therefore, in the event of the descending expression of PXR and these target genes, detoxification capacity will be impaired[@bib74]. Meanwhile, polymorphisms of PXR\'s target genes have been documented in cancer. Some of them, such as *ABCG2* rs2231142 variant and rs6857600 minor allele, are associated with a remarkable decrease in risk of chronic lymphocytic leukemia (CLL) and B-cell lymphoma (B-NHL), respectively[@bib8], suggesting polymorphism and bi-direction of these downstream targets might concurrently contribute to PXR\'s indeterminacy.
Similar to that observed among the elderly, many reports about PXR-mediated transactivation of metabolic enzymes and transport proteins are double-edged for keeping fit in different status. The phase I biotransformation enzyme CYPs, including CYP3A4, CYP3A5, CYP3A7, CYP3A11, CYP2B6, CYP2C9, CYP2C19 and CYP24A1[@bib62], [@bib65], are executors of the hydrolytic biotransformation for many therapeutic agents and xenobiotic substance, especially the marker of activated PXR---CYP3A4. CYP3A4 has important implications for the substrate oxidation and pharmacokinetic drug--drug interactions which leads to decreased plasma levels and therapeutic efficacy of anticancer drugs[@bib73]. Therefore, treatment by PXR antagonists, which abolishes *CYP3A4* at the transcriptional level, may facilitate the therapeutic effect[@bib63], [@bib64]. However, deficiency of PXR and CYPs might be involved in arsenite[@bib7], di-ethyl-nitrosamine (DEN)[@bib74], some toxic bile acids[@bib51] and other chemicals induced pathological development to cancer. The expression of PXR, *cyp3a* and other PXR\'s target gene are also found to participate in detoxification for diclofenac (DCF)[@bib75], underlining that effects of intrinsic induction of PXR on target genes always are bidirectional for fitness. In addition to CYPs, the phase I drug metabolizing enzyme (DME) carboxylesterases (CESs)[@bib4], the typical phase II conjugation enzyme UGTs[@bib4], and another detoxification enzyme epoxide hydrolase 1 (EPHX1)[@bib62], might be responsible for the PXR-dependent metabolism of esters, amides, thioesters and/or carbamates, environmental pollutants as well as CYP-mediated oxidations on aromatic/heteroaromatic rings and/or olefinic substituents.
Concurrently, PXR is a shared master orchestrating the expression of transporters[@bib37], which mediate the acceleration of toxicant clearance and decrease of drug effectiveness resulted from "phase 0 metabolism" (reducing the entrance of harmful substances) and "phase III metabolism" (increasing the excretion of their detoxification products)[@bib8]. Through the induction for MDR1, MRP2 (multidrug resistance-associated protein 2), BCRP, UGT1A1 and SULT2A1, PXR generates an export force to remove toxins and drugs, reducing the local cellular accumulation of toxic compounds and giving the individual cell protection against toxic injuries[@bib8], [@bib58]. Due to fact that PXR mediates many metabolic enzymes and efflux transporters, the facilitation of drug metabolism and drug--drug interactions appeared inevitable in treatment of anticancer medicines. Upon PXR activation, P-glycoprotein (P-gp, also known as ABCB1 or MDR1)[@bib73], OATP[@bib4], [@bib48], MRPs[@bib69], [@bib76] and other transporter proteins are upregulated, some of which are correlated with poor prognosis of advanced cancer. Although the PXR signal pathway has drawn increasing interest in recent years for its role in the drug resistance and drug--drug interaction in cancer treatment, the intrinsic expression and activity of PXR can be inordinate given the chaotic nature of most of the cancers. Yet the capacity of xenobiotic clearance might still reflect PXR is a crucial guard in tumor initiation.
Besides the clearance of toxins and drugs, PXR modulates the metabolism of crucial nutritional compounds, such as sugars, amino acids, nucleotides and inorganic ions. This manipulation may impress the development of cancer profoundly. Infinite proliferation and rapid growth of cancer pillage excessive glucose while high acidification from this high glycolytic reaction provides cancer cells with feedback to functional polarization toward a non-inflammatory phenotype, growth and immune evasion of tumor[@bib77]. More remarkably, downregulated expression of two gluconeogenic key enzymes, PEPCK and G6Pase[@bib65], were observed after activation of PXR, which provokes an enigma for the role of PXR in shaping the microenvironment.
The double-edged effect of PXR has also been reported due to its inducibility for toxicant clearance and multidrug resistance. There is evidence showing high-level expression of PXR in stage I and low-level expression in state II and stage III in carcinoma patients. This deficient expression in advanced stages of cancer[@bib67] does not support the harmfulness of PXR activity, suggesting the precise role of PXR deserves further exploration. This inconsistency of PXR might result from its dynamical expression in the different stage of cancer with varying level heterogeneity and disorder. Whilst, false conclusion resulted from indistinguishability between some para-carcinoma tissue and cancer tissue should be noticed as well.
4.2. PXR in cell cycle arrest {#sec4.2}
-----------------------------
Other than the regulation for ADME (absorption, distribution, metabolism and elimination) of medicines, toxins, carcinogens and other substance[@bib1], PXR is also capable of suppressing growth, proliferation and migration of cancer cells by inducing cellular cycle arrest. Recently, it has been reported that the antitumor bioregulation induced by PXR impedes the tumor progression, which is achieved *via* functional interaction with the transcriptional regulation of p21 (WAF1/CIP1/CDKN1A), E2F[@bib78], [@bib79], cullin1--3, MAD2L1 (mitotic spindle assembly checkpoint protein MAD2A), ANAPC2 (anaphase-promoting complex subunit 2)[@bib79] and other PXR related signaling pathway. In addition, a recent study indicated that PXR could suppress the migration and proliferation of AsPC-1 (human metastatic pancreatic adenocarcinoma) cells, even though its target gene *CYP3A5* is related to acquired drug resistance in PDAC (pancreatic ductal adenocarcinoma)[@bib38], suggesting PXR might play an intricate role in cancer development.
In the intricate regulatory network of cancers, interplay between PXR and cell cycle regulators strikingly enhances cell cycle arrest and prevents the augmentation of cancer development, whereas the lost or greatly diminished expression of PXR might limit this effect in initial cancer[@bib78], [@bib79]. Ectopic expressed PXR was found to mediate the promotion of p21(WAF1/CIP1) and the ablation of E2F/Rb, which triggers G0/G1 cell cycle arrest with the inhibition of proliferation and tumorigenicity of colon cancer cells[@bib78]. Consistently, we have shown upon activation of cullin1-3 and MAD2L1, and suppression of ANAPC2 and CDKN1A, rifampicin-activated hPXR could attenuate the growth and proliferation of cervical carcinoma subsequently by mediating G2/M cell cycle arrest[@bib79]. Notably, PXR is involved in hepatic proliferation or inhibit apoptosis to implement liver regeneration as well[@bib2], revealing the context dependent regulatory mechanism. Considering PXR is mainly expressed in colorectum and liver, novel preventive and therapeutic strategies aimed at preventing or reversing tumorigenesis might have huge potential if the subtle regulation of these PXR associated signaling pathways is achieved in cancer initiation of these tissues.
4.3. PXR in inflammation and injury {#sec4.3}
-----------------------------------
PXR-mediated anti-inflammation, anti-oxidative stress[@bib72] and anti-apoptotic responses[@bib51] in the context of cancer[@bib80] have been documented in recent years. Considering the tumorigenic effect of inflammatory injuries and selective repopulation in contributing to the cancer pathogenesis[@bib81], PXR activation may be a promising approach for prevention from injury-induced cancer in preneoplastic stage.
PXR mitigates the inflammation injury generally through the negative regulation of NF-kappa B (NF-*κ*B)[@bib82], Toll-like receptor 4 (TLR4)[@bib80], IL-6, signal transducer and activator of transcription 3 (STAT3), tumor necrosis factor *α* (TNF-*α*)[@bib74] and other signaling pathways. Substantial evidence has proved that NF-*κ*B, one of the most central mediator of stimuli-response and immune response, could be observed with increased expression and pro-inflammation in PXR null mice and negatively interact with PXR, triggering the depression for CYP in aggravated small intestinal inflammation[@bib82]. Moreover, a study on increased burden of liver inflammation indicated that IL-6 could restrain PXR activity by inducing differentiated embryonic chondrocyte-expressed gene 1 (DEC1) to competitively bind to the dimerization partner of PXR---RXR*α*[@bib83]. Homoplastically, heightened severity of necrotizing enterocolitis (NEC) was investigated in the absence of PXR, while the secondary bile acid lithocholic acid (LCA), an agonist for PXR, could activate PXR to negatively regulate TLR4, attenuating NEC in murine intestine, suggesting promotion of PXR-dependent preventive approach might be significant for intestinal inflammation and its subsequent disease[@bib80]. More than an important origin of injure, inflammation induced by cytokines NF-*κ*B, IL-6, STAT3, TNF-*α* and other many inflammation factors also influences the expression of functional proteins, such as CYP3A11 and glutathione *S*-transferase A2 (GSTa2), which results in DEN-induced hepatic cancer of mice, while based on their negative correlation with PXR, the activation of PXR-related signaling pathway might possess a huge potential for tumorigenesis suppression[@bib74]. Recent studies shed light on PXR activation induced by ginkgolide B, PXR could mediate anti-inflammatory and anti-apoptotic effects on endothelial cells *via* suppressing TNF-*α* induced THP-1 cells (human acute monocytic leukemia cells) adhesion and expression of vascular adhesion molecule 1 (VCAM-1) and E-selectin and promoting detoxification for staurosporine and doxorubicin[@bib84].
In addition to inflammatory inhibition, PXR activation might blunt the expression of pro-apoptotic genes *TP53* and BCL2 antagonist/killer 1 (*BAK1*) to prevent toxic bile acids-induced apoptosis and subsequent selective repopulation of anti-apoptosis cells in colonic tumorigenesis[@bib51]. Furthermore, after tanshinone IIA treatment, PXR could resist ROS-induced apoptosis and oxidative stress through the inhibition for mitochondrial apoptosis pathway and the regeneration of glutathione (GSH) respectively, which might involve the PXR dependent elevation of mitochondrial membrane potential (MMP), GPx and BCL-2 and the attenuation of caspase-3, caspase-7, caspase-9, cytochrome C and BCL2 associated X (BAX)[@bib72], on this account, alleviating the oxidative injury. It is noticeable, however, genetic or pharmacological activation of PXR--CYP3A4 signaling pathway is involved in ritonavir induced hepatotoxicity[@bib85] and increased sensitization for HS (hemorrhagic shock)-induced hepatic injury[@bib86] in clinical treatment. These discoveries indicated that PXR may be a coordination center for tumorigenesis basing on its transcriptional regulation in inflammation, hepatic injury and homeostasis maintenance.
4.4. PXR in angiogenesis {#sec4.4}
------------------------
Along with the pleiotropic effects above, PXR\'s role in angiogenesis inhibition has been reported in colon cancer as well, and ligand-dependent activation might be required in this progression. Rifaximin, a gut-specific ligand for hPXR[@bib56], could significantly suppress proliferation, migration and expression of PCNA of Caco-2 cells by activating PXR to decrease release of vascular endothelial growth factor (VEGF) and nitric oxide (NO) and phosphorylation of serine/threonine-protein kinase AKT, mTOR and p38 mitogen-activated kinase (p38MAPK), as well as activity of hypoxia-inducible factor 1-*α* (HIF-1*α*), p70S6K and NF-*κ*B, while treatment by PXR\'s antagonist ketoconazole could induce the inhibition for these effects on pro-angiogenic mediators and cancer progression[@bib87]. Furthermore, upon increased survival rate and decreased tumor number, rifaximin treatment also has been shown to perform a chemoprevention for azoxymethane (AOM)/dextran sulfate sodium (DSS)-induced colon cancer dramatically *via* the PXR-dependent response and regulation[@bib88]. Another ligand for PXR---rifampin, a potent angiogenesis inhibitor targeting hepatic cancer developed from hepatitis C virus (HCV)-related liver cirrhosis, could repress human microvascular endothelial cell proliferation and migration *via* downregulating angiogenesis-associated genes[@bib55], [@bib89], while whether PXR are involved and its specific mechanism in this progression are waiting for further exploration.
5. Conclusion and prospective {#sec5}
=============================
As one of the most crucial gene-regulatory transcription factors, PXR is functioning in a wide range of cellular circuits and biological responses in different organisms. The fact that PXR is responsive to various endobiotic and xenobiotic stimulations makes it a good candidate in mediating carcinogenesis and metabolism of anticancer drugs. The regulatory network weaved by PXR and its upstream and downstream factors has drawn attention in the cancer biology. Indeed, the PXR centric signaling network has been manifested in cancer progression and a growing body of research is adding on to dissect the role of this network in the tumorigenesis. Nevertheless, the specie-specificity of diverse PXR ligands and the influence of PXR signaling on anticancer drug application need to be further defined. Furthermore, PXR targeting prevention and therapy in clinical application are hurdled by the distinctions between the different isoforms with unclear regulatory mechanisms and structurally diverse agonists/antagonists. Note that some PXR variants with antagonistic functions and their potential interaction with other NRs add more to the complexity of PXR targeting cancer treatment. Undoubtedly, the utilization of homeostasis-maintaining nature and the revelation of stage-sensitive bioeffects of PXR indicate a sally port for intervention of cancer development. Better understanding of the regulatory mechanism and biological effects of PXR in different stages of tumor development *via* the advances of research and development and the subduction of toxicity and side effects of antineoplastic will contribute to trailblazing an efficient approach for the prevention and therapy of cancer.
This study was funded by the National Natural Science Foundation of China (81772972 and 81572703); Guangdong Natural Science Foundation (2015A030313449; China); Guangdong Science and Technology Project \"Public Research and Capacity Building" Special Project Fund (2014A020212285; China); and Department of Education, Guangdong Government under the Top-tier University Development Scheme for Research and Control of Infectious Diseases (2016026, 2015060, and 2015089; China).
Peer review under responsibility of Institute of Materia Medica,Chinese Academy of Medical Sciences and Chinese Pharmaceutical Association.
[^1]: Parts of data[@bib46] are adapted with permission from <https://www.phosphosite.org>.
| {
"pile_set_name": "PubMed Central"
} |
Acute respiratory infections (ARIs) are the most common infections in humans. ARIs (nonspecific upper respiratory infections, otitis media, sinusitis, pharyngitis, bronchitis, and pneumonia) account for half of acute conditions each year [@bib1], and consistently rank among the top 10 reasons for ambulatory visits in the United States [@bib2], [@bib3], [@bib4], [@bib5], [@bib6], [@bib7], [@bib8], [@bib9], [@bib10], [@bib11], [@bib12]. Acute bronchitis episodes represent a significant portion of these illnesses. Data from the National Health Interview Survey suggest that 4% to 5% of all adults experience one or more episodes of acute bronchitis each year [@bib1]. Furthermore, over 90% of acute bronchitis episodes will come to medical attention [@bib1].
Acute bronchitis is a clinical diagnosis applied to otherwise healthy adults with acute respiratory illness of 1 to 3 weeks\' duration. Acute bronchitis usually is distinguished from other ARIs by the predominance of cough, often accompanied by other respiratory and constitutional symptoms, and the absence of findings suggestive of pneumonia. The importance placed on sputum production and wheezing when making the diagnosis of acute bronchitis varies by physician [@bib13], [@bib14], [@bib15], [@bib16]. Cough lasting longer than 3 weeks should be considered "persistent" or "chronic" cough [@bib17], [@bib18], and is not discussed here because the diagnostic considerations are significantly different than those of acute bronchitis.
This article focuses on acute bronchitis in otherwise healthy individuals, not on patients who have underlying heart or lung disease or immunosuppression, who generally have been excluded from trials evaluating etiology of and treatment for acute bronchitis. The extent to which one can generalize from the data presented herein is unknown.
Acute bronchitis: a transient form of asthma {#sec2}
============================================
Clinical features of uncomplicated acute bronchitis develop in sequential phases. Acutely, there is direct inoculation of the tracheobronchial epithelium, characterized clinically by variable constitutional symptoms, including fever, malaise, and myalgias. These symptoms usually last 1 to 5 days, depending on the infectious agent. This phase of illness is often indistinguishable from other acute upper respiratory tract infections. Most uncomplicated upper respiratory infections improve substantially within 5 to 7 days [@bib19], [@bib20]. In patients for whom the diagnosis of acute bronchitis would be appropriate, however, the acute phase is followed by a second, protracted phase characterized by persistent cough, often accompanied by phlegm production or wheezing. This second phase usually lasts 1 to 3 weeks, and has as its underlying pathophysiology the hypersensitivity of the tracheobronchial epithelium and airway receptors (reactive airway disease).
During the protracted phase, pulmonary function tests (PFTs) are frequently abnormal and do not seem to be related to either the acute cytopathic effects of the infection or the type of infection (bacterial or viral) [@bib21], [@bib22], [@bib23], [@bib24], [@bib25]. Vagal-mediated airway hyperresponsiveness has been shown to coincide with repair of the bronchial epithelium [@bib26]. Other mechanisms of bronchial hyperresponsiveness, such as adrenergic-cholinergic tone imbalance and IgE-mediated histamine release, also may be present. PFT abnormalities seem to be common in acute bronchitis, with approximately 40% of patients demonstrating significant abnormalities by forced expiratory volume (FEV1) or histamine challenge [@bib27], [@bib28]. PFT abnormalities are usually transient, typically resolving after 2 to 3 weeks, although they may last as long as 2 months [@bib27], [@bib28], [@bib29]. Recurrent episodes of "acute bronchitis" may suggest underlying asthma [@bib30], [@bib31]. Although undiagnosed asthma should be considered in patients who have acute cough illness, this diagnosis is difficult to establish because bronchial hyperresponsiveness and PFT abnormalities are frequent in patients who have acute bronchitis. Suspicion and work-up for asthma should be reserved for patients with cough lasting longer than 3 weeks [@bib17].
Microbiology of acute bronchitis {#sec3}
================================
Most acute bronchitis cases seem to have a nonbacterial etiology [@bib29], [@bib32], [@bib33]. Microbiologic study of acute bronchitis, however, similar to community-acquired pneumonia, can identify a pathogen in only 16% to 55% of cases [@bib32], [@bib34]. The significant variability in the frequency of isolation of any pathogen and the types of pathogens identified reflects the patients studied, available technology to identify certain viral and atypical pathogens, and the epidemic nature of the agents that cause acute bronchitis. Additionally, noninfectious causes of acute bronchitis also likely represent some of these cases. Occult asthma, allergic, and occupational exposures should be considered, although their prevalence in adults with acute cough illness remains unclear.
Viral bronchitis {#sec3.1}
----------------
Respiratory viruses seem to cause or serve as a copathogen in most cases of acute bronchitis in epidemiologic studies. The specific viruses most frequently associated with acute bronchitis, in order of frequency of occurrence, are influenza, parainfluenza, respiratory syncitial virus (RSV), coronavirus, adenovirus, and rhinoviruses.
Recent studies have demonstrated the importance of RSV as the etiology of ARIs in adults [@bib35], [@bib36]. The impact of RSV is greatest in the elderly, particularly those living in long-term care facilities, and those with underlying heart and lung disease and malignancy [@bib37]. Infection among exposed adults is common, with attack rates approaching 50%, particularly in households with children infected with RSV and in institutional settings [@bib24], [@bib37]. Most young and middle-aged adults develop asymptomatic or mildly symptomatic disease, often closely resembling influenza [@bib38]. RSV can be associated with more severe clinical disease and significant morbidity, even in otherwise healthy adults [@bib24]. This morbidity seems to be in part secondary to induced airway hyperreactivity.
In the elderly and institutionalized, lower respiratory illness with RSV is common, with most studies reporting rates of pneumonia and death from 10% to 20% and 2% to 5%, respectively [@bib37]. One report of an outbreak on a geriatrics ward found intense coughing and fever in 96% of patients, productive cough in 64%, and evidence of bronchopneumonia in 40% [@bib39]. In this study, it is unclear whether RSV or secondary bacterial infection caused these pneumonias.
Human metapneumovirus (hMPV), a paramyxovirus [@bib33], [@bib40], has emerged recently as an important cause of lower respiratory tract illness and acute bronchitis. Human MPV has been detected in children, adults, the elderly, and the immunocompromised in the Netherlands, Australia, North America, the United Kingdom, and Finland [@bib41], [@bib42], [@bib43], [@bib44], [@bib45]. In one study, hMPV was second only to RSV as a cause of respiratory tract illness presenting to a university hospital in the Netherlands [@bib45].
Similar to RSV, hMPV is primarily an illness of the winter months, most commonly causing significant illness in young children and immunocompromised and elderly individuals. Studies suggest that 25% to 50% of hMPV-positive patients who have significant respiratory tract illness have underlying disease [@bib46], [@bib47]. Among otherwise healthy adults, hMPV likely causes predominantly mild respiratory illness, but may cause a small but significant portion (approximately 3%) of acute respiratory illness requiring medical attention [@bib46], [@bib47], [@bib48].
Bacterial bronchitis {#sec3.2}
--------------------
When microbiologic studies are performed on select patients who have uncomplicated acute bronchitis in nonoutbreak settings, less than 10% of patients have a clear bacterial etiology [@bib29], [@bib32], [@bib33]. *Bordetella pertussis*, *Chlamydia pneumoniae*, and *Mycoplasma pneumoniae* are the only bacterial pathogens that have been established as causes of acute bronchitis. Although studies have reported the presence of *Streptococcus pneumoniae*, *Haemophilus influenzae*, and *Moraxella catarrhalis* in adults with acute bronchitis, these studies generally failed to exclude patients who had underlying lung disease, failed to distinguish between colonization and infection [@bib49], or did not differentiate adequately patients who had pneumonia from those who had acute bronchitis when determining causative agents [@bib33]. Furthermore, acute viral respiratory infections seem to increase the proliferation of these bacteria among the oropharyngeal flora [@bib50], further complicating the issue of colonization versus infection. Therefore, sputum Gram stain and culture for common bacterial pathogens have no clinical usefulness in patients who have acute bronchitis.
*Mycoplasma pneumoniae* and *C pneumoniae* have been recognized as possible causes of acute bronchitis since the 1980s [@bib51], [@bib52]. Attack rates vary highly, reflecting the seasonal, geographic, and epidemic nature of these infections [@bib21], [@bib27], [@bib33], [@bib53], [@bib54], [@bib55], [@bib56], [@bib57], [@bib58], [@bib59], [@bib60]. Studies attempting to distinguish these atypical bacterial pathogens from viral etiologies have shown that patients infected with atypical bacterial pathogens tend to present to medical attention much later than those with confirmed viral bronchitis [@bib21], [@bib60], [@bib61], and are more likely to have wheezing on clinical examination [@bib21]. In several studies, although these pathogens were present by antibody titer or gene amplification, treatment with antibiotics appropriate to atypical pathogens did not change outcome [@bib33], [@bib62], [@bib63], [@bib64], [@bib65]. This suggests that in acute bronchitis, *C pneumoniae* and *Mycoplasma pneumoniae* may reflect copathogens or inciting factors for secondary infectious processes, rather than the etiologic agent. Alternatively, because patients with atypical bacterial pathogens present late in the course of illness, the acute infectious process may have resolved with only residual reactive airway disease present at the time of presentation for medical care.
*B pertussis* causes acute bronchitis in previously immunized adults. Natural infection and vaccination with whole-cell and acellular vaccines induce protection from infection for a limited time [@bib21], [@bib66], [@bib67], [@bib68], [@bib69], [@bib70], [@bib71]. Thus, adolescents and adults gradually may become susceptible to infection again. Symptomatic adult pertussis requiring medical attention occurs at a rate of 71 to 507 per 100,000 population per year (0.1%--0.5% of the population per year) [@bib72], [@bib73], [@bib74], [@bib75], [@bib76]. This pool of frequently undiagnosed pertussis [@bib77] provides a reservoir for potentially serious infections in young infants who either are unvaccinated or whose vaccinations are not yet fully effective [@bib78].
The gradual decrease in protection against pertussis likely explains part of the wide variation in presenting symptoms in previously immunized adults. Adults with pertussis generally present with persistent cough, with a mean duration of 36 to 48 days [@bib72], [@bib75], [@bib77], [@bib79], [@bib80], [@bib81], [@bib82]. When prolonged cough (longer than 1 week) is present, a significant portion of patients will have *B pertussis* infection, with a frequency ranging from 12% to 32% [@bib71], [@bib72], [@bib74], [@bib75], [@bib79], [@bib83], [@bib84], [@bib85], [@bib86], [@bib87], [@bib88]. Cough is mostly paroxysmal, and often disturbs sleep. Choking or vomiting and whooping can be present, but less commonly than in children or previously unimmunized adults.
Some have suggested that booster pertussis immunizations for adults or adolescents may curb illness in infants [@bib89]. Whole-cell and acellular pertussis vaccines are well tolerated, with primarily local side effects [@bib90], [@bib91], [@bib92]. Only one study has assessed the efficacy of acellular (aP) vaccines in adults [@bib93]. Because of the small sample size of the trial, few pertussis cases were reported (n = 12), and no point estimate of efficacy could be given. The incidence of primary pertussis cases was decreased in the aP group (0.8 per 1000 person-years; 95% CI 0.0--2.1), however, compared with the control group (3.7 per 1000 person-years; 95% CI 1.2--6.2). An epidemiologic model has suggested that a high coverage of adults (greater than 85%) would be needed to reduce effectively the number of cases of infant pertussis [@bib94].
Antibiotic therapy does not seem to decrease duration of symptoms for pertussis unless initiated within 7 to 10 days of the onset of illness [@bib95], [@bib96], [@bib97]. Macrolide prophylaxis during outbreak situations and after intrafamilial contacts seems effective [@bib77], [@bib98], however, and decreases spread of disease [@bib96], [@bib97].
Distinguishing pneumonia from acute bronchitis {#sec4}
==============================================
In the absence of significant comorbid conditions or asthma, the primary objective when evaluating patients who have acute cough illness is excluding pneumonia. The prevalence of pneumonia in patient populations presenting with ARIs varies significantly across study populations, ranging from 3% to 10% in most studies [@bib33], [@bib99], [@bib100], [@bib101]. Cohort studies have identified clinical features useful for determining which patients do not have pneumonia [@bib99], [@bib101], [@bib102], [@bib103], [@bib104]. The absence of abnormal vital signs (heart rate greater than 100 beats/minute, respiratory rate greater than 24 breaths/minute, oral temperature above 100.5°F) and chest examination (focal consolidation; eg, rales, egophony, fremitus) reduces the likelihood of pneumonia sufficiently to render further diagnostic testing unnecessary [@bib101]. The specificity (67%--76%), but not sensitivity (62%--71%), of these clinical prediction rules for radiographic pneumonia exceeded physician judgment in a well-designed validation study of 290 adult patients who had acute cough illness [@bib100]. Notably absent from these decision rules is the presence or absence of purulent sputum because purulence (by itself) is a poor predictor of bacterial infections [@bib105], [@bib106].
Applying the pneumonia clinical prediction rules should help inform the decision about ordering a chest radiograph, but cannot substitute for clinical judgment. The pneumonia clinical prediction rules have limited application in the elderly because they may present with atypical manifestations of pneumonia (and without vital sign or examination abnormalities) [@bib107]. Conversely, during the influenza season many patients will have fever or tachycardia but not pneumonia. As a result, chest radiography often is overused in the elderly and during influenza season. In settings where chest radiography is not available readily (eg, many private office practices or rural locations), patients who have cough illness (particularly elderly) may be prescribed antibiotics to safeguard against missing a case of pneumonia.
Rapid blood tests for bacterial infections {#sec4.1}
------------------------------------------
### C-reactive protein {#sec4.1.1}
A rapid, office-based diagnostic test that improves sensitivity and specificity of detecting pneumonia could be a valuable addition to the current evaluation strategies for patients who have acute cough illness [@bib108]. European experience with an office-based, rapid c-reactive protein (CRP) test, as well as a recent study conducted in the United States, suggests considerable potential to improve diagnostic and treatment decisions for adults with cough illness [@bib109], [@bib110], [@bib111], [@bib112], [@bib113], [@bib114], [@bib115], [@bib116], [@bib117]. CRP synthesis is stimulated in response to many inflammatory conditions, and levels increase preferentially (but not exclusively) by bacterial (versus viral) infections. The serum levels of CRP associated with bacterial infections are 10 to 50 fold higher than those used to predict atherosclerotic heart disease.
Despite widespread use in Europe, and recent US Food and Drug Administration approval of a rapid CRP test in the United States, the role of rapid CRP testing in the management of adults with acute cough illness has not been defined rigorously. Most studies found a high sensitivity (80%--100%), but CRP levels may lack the specificity (60%--70%) necessary to diagnose bacterial infections in isolation. Integrating CRP testing into a clinical algorithm is one strategy to improve on its specificity while taking advantage of its sensitivity for detecting acute bacterial infections such as pneumonia. Future studies assessing the effectiveness of a CRP-based clinical algorithm are necessary.
### Procalcitonin {#sec4.1.2}
Recent studies of procalcitonin in serum also have shown levels to distinguish bacterial from viral illnesses [@bib118], [@bib119]. Early procalcitonin assays had a limited functional assay sensitivity (0.3--0.5 μg/L), and therefore were not accurate for the diagnosis of early or localized infections [@bib120], [@bib121]. A newer assay with improved functional sensitivity (0.06 μg/L) has become available in Europe, however. One recent study adopting a test-based clinical algorithm with this rapid procalcitonin testing among adults admitted to the hospital with lower respiratory tract infection demonstrated a large reduction in antibiotic use, and equivalent outcomes [@bib122].
Acute bronchitis and antibiotics {#sec5}
================================
Should antibiotics ever be used? {#sec5.1}
--------------------------------
Antibiotic prescription rates for acute bronchitis range from 50% to 80% in studies from multiple settings and countries [@bib123], [@bib124], [@bib125]. Studies have failed to show any meaningful benefit from antibiotics in the treatment of acute bronchitis, however. Systematic reviews and meta-analyses of nine randomized placebo-controlled trials conducted between 1970 and 2000 conclude that routine antibiotic treatment of acute bronchitis has no consistent effect on either the duration or severity of illness. In one meta-analysis, there was no significant impact on the duration of cough [@bib126], but two other meta-analyses reported a small but statistically significant decrease in cough duration (one third days fewer of cough after 7 days) associated with treatment with antibiotics [@bib127], [@bib128]. In all three meta-analyses, there was no significant impact on overall illness duration, activity limitation, or work loss. A recent randomized, double-blind, controlled study comparing azithromycin with vitamin C has addressed concerns that the earlier trials were performed with older antibiotics, some of which had no activity against the atypical agents implicated in acute bronchitis [@bib129]. This study found no advantage to antibiotic treatment on illness outcomes or return-to-work status.
As discussed earlier, antibiotic prescription is appropriate when the physician suspects pertussis infection. Antibiotics should be reserved for adults exposed to known pertussis infection, or to patients who have acute bronchitis in the setting of a documented pertussis epidemic. Although antibiotics do not decrease the duration of illness in this setting, they can decrease bacterial shedding and spread. Antibiotics also may be considered in the setting of a known mycoplasma or *C pneumoniae* outbreak, although data are lacking on their effectiveness in this setting.
The harm of overusing antibiotics {#sec5.2}
---------------------------------
The societal cost of inappropriate antibiotic use is the rapid emergence of antibiotic resistance among bacterial pathogens [@bib130], [@bib131], [@bib132]. Resistance is rising among common community-acquired pathogens, including *S pneumoniae* (DRSP) [@bib133], [@bib134], [@bib135]. This pathogen is a leading cause of ear and sinus infections, pneumonia, sepsis, and meningitis in the United States. At the community level, the mean increase in DRSP prevalence is directly proportional to the amount of antibiotics consumed [@bib136]. On an individual level, a person\'s risk for carriage, transmission, and invasive infection with antibiotic-resistant bacteria is associated strongly with prior antibiotic use [@bib137], [@bib138], [@bib139], [@bib140].
Finally, the sheer magnitude of antibiotic prescriptions dispensed each year for ARIs requires that excess health care costs also be considered. In 1998, 41 million antibiotic prescriptions were written for ARIs, 55% of which were likely unnecessary [@bib141]. The cost of these excess prescriptions was estimated at \$726 million. Similar high rates of inappropriate antibiotic use are seen in Europe [@bib142]. In addition, the result of antibiotic resistance on antibiotic selection and clinical outcomes further increases health care costs [@bib143].
If they don\'t work, why are antibiotics so frequently prescribed for acute bronchitis? {#sec5.3}
---------------------------------------------------------------------------------------
Physician education likely reflects a small component of inappropriate antibiotic use. Evidence suggests that physicians and patients are more likely to believe that antibiotics are appropriate if purulent secretions are present [@bib144], [@bib145], despite significant evidence to the contrary. Physician specialty and level of training also are associated with antibiotic prescriptions for ARIs. Family medicine physicians are more likely to prescribe antibiotics to children with ARIs than pediatricians [@bib146]. Also, providers that are further from medical school graduation and practicing in rural areas are more likely to prescribe antibiotics [@bib147].
Antibiotic prescribing behavior is associated poorly with clinicians\' subjective norms and intentions, which suggests that external forces such as patient-specific beliefs and health plan factors play a greater role [@bib148] than physician knowledge. Patients frequently expect to receive antibiotics for uncomplicated acute bronchitis [@bib149], [@bib150] and patients or parents who expect antibiotics are more likely to receive them [@bib150], [@bib151]. Communication elements associated with antibiotic prescriptions for ARIs include patient appeals to specific life circumstances (eg, a pressing social engagement), identification of a previous positive experience with antibiotic use [@bib81], or being labeled as having "acute bronchitis" rather than a "chest cold" [@bib149].
Not surprisingly, clinicians with greater patient workloads prescribe antibiotics for ARIs more frequently, likely reflecting the perceived time it would take to discuss the inappropriateness of antibiotic use in ARIs [@bib152]. Other health plan factors that may contribute to prescribing behavior include restricting formularies and practice characteristics such as payment structure. A recent survey of physicians\' attitudes regarding the role of societal risks in making antibiotic treatment decisions for individual patients found that societal concerns about promoting antibiotic resistance ranked below patient-centered factors such as ease of use and cost to the patient [@bib153].
Despite physician concerns about patient expectations, most studies find that satisfaction with care for ARIs is tied more closely to how much time the physician spent explaining the illness, rather than receipt of antibiotics [@bib150], [@bib151], [@bib154]. Communication elements associated with high patient satisfaction include positive responses to the following statements: "the doctor spent enough time with me"; "the doctor explained the illness to me"; and "the doctor treated me with respect" [@bib147]. An intervention strategy consisting of patient and clinician education reduced antibiotic prescription rates for acute bronchitis in adults [@bib155], but did not decrease patient satisfaction [@bib147]. Furthermore, antibiotic prescribing does not seem to reduce additional care seeking in adult patients [@bib156].
Nonantibiotic treatment of acute bronchitis {#sec6}
===========================================
Anti-influenzal therapy {#sec6.1}
-----------------------
Influenza is the most common cause of acute bronchitis, and influenza vaccination is the most effective strategy for preventing influenzal illness. Treatment for high-risk exposed individuals and those who present within 48 hours of symptom onset is also possible. Amantadine, rimantidine, zanamivir, and oseltamivir decrease illness duration by approximately 1 day and lead to a 0.5-day quicker return to normal activities [@bib157]. The primary difference between the agents is that the neuraminidase inhibitors are effective against influenza A and B, whereas amantadine and rimantidine are effective only against influenza A. The relative proportion of cases caused by each type of influenza virus varies from year to year, and is determined best through consultation with local public health agencies. Adverse events are modestly more common with rimantidine (32% of patients, most commonly central nervous system) than with the neuraminidase inhibitors (24% of patients, mostly gastrointestinal) [@bib157]. Because each of these therapies is only effective if initiated within the first 48 hours, and preferably 30 hours, of symptom onset, rapid diagnosis is key. During documented influenza outbreaks, the positive predictive value of clinical diagnosis based on clinician judgment is good (correct approximately 70% of the time) [@bib158], and compares favorably with rapid diagnostic tests for influenza (sensitivities of 63%--81%) [@bib158], [@bib159], [@bib160]. Diagnosis of influenza in a nonoutbreak period is more difficult and diagnostic testing should be considered.
Antiviral treatment for other viral illness either have been studied inadequately, carry inappropriately high side-effect profiles, or are ineffective in otherwise healthy individuals [@bib161]. Ribavirin is indicated in bone marrow transplant patients who have RSV, and in this population reduces morbidity and mortality [@bib162].
Bronchodilator therapy {#sec6.2}
----------------------
Three randomized, controlled trials have demonstrated a consistent benefit to bronchodilator treatment [@bib163], [@bib164], [@bib165]. Approximately 50% fewer patients report the presence of cough after 7 days of treatment. This benefit seems to be greatest in the subset of patients who had bronchial hyperresponsiveness. A large trail of patients who had URI-associated cough, but not clearly acute bronchitis, reported no benefit of bronchodilator treatment [@bib166]. A meta-analysis of these studies showed no significant benefit from b2-agonists [@bib167], but is limited by the addition of the Littenberg study, which enrolled patients who had acute, nonspecific cough.
Whether anticholinergic bronchodilator therapy is effective in patients who have uncomplicated acute bronchitis is not known. Similarly, no studies have examined the effect of inhaled corticosteroid therapy, although the delay in onset of action for this type of therapy (1--2 weeks) may preclude finding a major benefit.
Antitussive therapy {#sec6.3}
-------------------
The effectiveness of antitussive therapy seems to depend on the cause of cough illness. Acute or early cough caused by colds and other upper respiratory tract infections does not seem to respond to dextromethorphan or codeine. Cough of greater than 3 weeks\' duration, cough associated with underlying lung disease, and experimentally induced cough seem to respond to these agents. Given that the cough of acute bronchitis often lasts for 2 to 3 weeks, these agents likely have a modest impact on cough severity and duration.
Immunomodulating therapies {#sec6.4}
--------------------------
Most trials of immunomodulatory (alternative) therapies have been conducted on patients who have early symptoms of colds and nonspecific ARIs. As a result, these data are difficult to extrapolate to patients who have acute bronchitis, who generally present later and with more severe illness. Vitamin C at doses exceeding 1 g/d seems to offer small but significant reduction in illness duration of about 0.5 day per cold episode [@bib168]. Well-performed clinical trials comprising mostly small studies of zinc gluconate and zinc acetate lozenges have had mixed results [@bib169] and their benefit is unclear. Echinacea seems to be of benefit in some preparations [@bib170], but there is significant heterogeneity of study design, as well as preparations tested. Also, quality control of echinacea preparations sold to the community is poor, with one study demonstrating that 10% of single-herb echinacea preparations in one metropolitan area had no active ingredient, and less than half met the quality standards described on the label [@bib171].
A recent randomized, double-blind, placebo-controlled trial has shown the benefit of an extract of *Pelargonium sidoides* roots in acute bronchitis [@bib172]. This plant extract is used commonly in Europe and Mexico. Its mechanism of action is poorly understood, but is believed to be immunomodulatory in nature, having been used first in the early 1900s as a treatment for tuberculosis. In the recent study, adult patients who had acute bronchitis of greater than 48 hours\' duration and a bronchial severity score (BSS) of at least 5 points were enrolled. Patients were excluded if they were to receive or recently had received antibiotics or had other serious illnesses. Patients were randomized to receive active ingredient or color-, smell-, viscosity-, and taste-matched placebo. Among patients receiving pelargonium, decrease in BSS on day 7 was 5.9 points compared with 3.2 points for placebo (*P* \< .0001). Duration of illness (*P* \< .001) and inability to work (16% versus 43%, *P* \< .0001) were significantly less in the pelargonium group compared with placebo. Further studies are necessary to confirm these interesting results. In the United States, *Pelargonium sidioides* is marketed under the trade name Umcka (Nature\'s Way, Springville, Utah).
Approach to the patient with acute bronchitis {#sec7}
=============================================
The approach to the otherwise healthy patient with acute cough illness first should be to assess his or her likelihood of pneumonia. In the nonelderly patient without abnormal vital signs or consolidative lung findings, the likelihood of pneumonia is less than 1% in the ambulatory care setting [@bib99], [@bib101]. When these abnormalities are present, a chest radiograph should be considered, depending on overall clinical impression and likelihood of influenza. For patients who present with prolonged cough (longer than 1 week), pertussis should be considered, along with bronchial hyperresponsiveness.
Once a diagnosis of acute bronchitis has been made, providers should address symptomatic treatment and patient expectations for the visit. Physicians should validate the severity of the patient\'s illness (because it has affected the patient\'s activities enough to seek care and acute bronchitis significantly decreases quality of life) [@bib173]. Treatment discussions should focus on alleviating symptoms and providing realistic expectations for the duration of symptoms. Patients should be informed that they should expect their cough to last 10 to 14 days after the office visit. Providers should also inform patients of which symptoms should prompt a return to the clinic or office.
For patients who request antibiotics for clear viral infections, providers should discuss the lack of benefit and the risks of inappropriate antibiotic use. These risks should be personalized as much as possible, informing them that previous antibiotic use increases their personal risk of carriage and infection with antibiotic-resistant infections. In addition, antibiotics cause frequent side effects, especially of the gastrointestinal tract.
Symptomatic treatment will depend on severity of illness and time at presentation. Alternative and over-the-counter preparations may be most effective in the early stages of illness. For those with prolonged or severe cough or clear bronchial hyperresponsiveness, bronchodilator treatment and antitussives should be considered. Further studies are necessary on the plant extract *Pelargonium sidoides* to assess further its benefit in this setting.
| {
"pile_set_name": "PubMed Central"
} |
{#sp1 .201}
{#sp2 .202}
{#sp3 .203}
{#f1 .202}
{#f2 .203}
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#ss1}
============
Prolonged coughing is a frequent symptom in children and is often associated with respiratory tract infection \[[1](#b1){ref-type="ref"}\]. Although prolonged coughing is a prominent feature of infection with *Bordetella pertussis*, infections with other respiratory tract pathogens may also cause prolonged coughing \[[2](#b2){ref-type="ref"}\]. Mixed infections with a combination of two or more pathogens occur, but data on clinical manifestations of combined infections derived from retrospective studies are often difficult to interpret \[[2](#b2){ref-type="ref"}, [3](#b3){ref-type="ref"}, [4](#b4){ref-type="ref"}, [5](#b5){ref-type="ref"}\]; nevertheless, it has been suggested that mixed infections may result in more severe illness, especially in younger children \[[3](#b3){ref-type="ref"}, [4](#b4){ref-type="ref"}, [6](#b6){ref-type="ref"}\]. In an earlier retrospective observational study (unpublished data) involving 81 children with serologically proven *B. pertussis* infection, evidence was found in 28% of the children for concomitant infections with other respiratory tract pathogens. It is known that many such pathogens may cause prolonged coughing in children, but prolonged coughing is not always caused by a respiratory tract infection \[[1](#b1){ref-type="ref"}, [7](#b7){ref-type="ref"}, [8](#b8){ref-type="ref"}\]. The present study investigated the role of different respiratory pathogens in prolonged coughing in children, as well as the frequency of occurrence of possible mixed infections. In addition, the severity of disease in patients with one or more pathogens was compared with that in patients for whom no pathogens were detected.
Materials and methods {#ss2}
=====================
Patients {#ss3}
--------
Participating patients (aged ≤ 18 years) were those referred with persistent coughing to the outpatient clinic of the Department of Pediatrics at Groene Hart hospital, a 500‐bed general hospital in Gouda, The Netherlands, between September 2001 and September 2003. Disease duration was defined as the time between onset of symptoms and first visit to the hospital. The definition of prolonged, chronic or persistent cough has varied from 5 days to \> 1 month \[[2](#b2){ref-type="ref"}, [7](#b7){ref-type="ref"}, [9](#b9){ref-type="ref"}, [10](#b10){ref-type="ref"}, [11](#b11){ref-type="ref"}, [12](#b12){ref-type="ref"}\]. Patient selection in the present study was based on a persistent cough lasting 1--6 weeks. Patients with aspiration of a foreign body or patients known to have cystic fibrosis were excluded. Blood samples were analysed at the first visit for erythrocyte sedimentation rate, C‐reactive protein level, leukocyte count and differentiation, and serological evidence of respiratory pathogens (see below). Oropharyngeal, nasal and nasopharyngeal swabs were taken for culture and PCR analysis for respiratory pathogens (see below). When the disease duration on the first visit was \< 14 days, serological tests were repeated after 2 weeks, except for *B. pertussis*. When serology and PCR results at the first visit were negative for *B. pertussis*, the serological testing was repeated 4 weeks later (after 6 weeks for children aged \< 1 year). The patients, or their parents, were asked to complete a questionnaire regarding symptoms, previous diseases and vaccination status. All gave their written informed consent to participate in the study. A follow‐up telephone interview was conducted *c*. 4 weeks after enrolment. The study was approved by the Medical Ethics Committee of the hospital. A healthy control group was not included in this study.
Serology {#ss4}
--------
Positive serology for *B. pertussis* was defined as at least a four‐fold increase in IgG to pertussis toxin (IgG‐PT) in paired sera to a level of ≥ 20 U/mL, or a high IgG‐PT concentration in a single serum, i.e., \> 100 U/mL, as measured by the in‐house IgG‐PT ELISA of the National Institute of Public Health and the Environment (RIVM), Bilthoven, The Netherlands \[[13](#b13){ref-type="ref"}\]. The interpretation criteria of the IgG‐PT ELISA have been shown to have a sensitivity and specificity of 90% and 97%, respectively \[[14](#b14){ref-type="ref"}\].
All sera were tested for antibodies to respiratory syncytial virus (RSV), influenza viruses A and B, adenovirus, parainfluenza viruses 1, 2 and 3, *Mycoplasma pneumoniae*, *Chlamydia* spp. and *Coxiella burnetii* by the Regionaal Medisch Microbiologisch Laboratorium (Rotterdam, The Netherlands) with the complement fixation method (Serion Immunodiagnostica, Würzburg, Germany). Two‐point serology against these pathogens was considered proof of a recent infection when there was a four‐fold or more increase or decrease in titre. In one‐point serology, a titre ≥ 128 was considered to indicate recent infection, while a titre ≥ 64 was considered to be equivocal, except for *C. burnetii* (≥ 4 equivocal) and *Chlamydia* (≥ 8 equivocal) \[[15](#b15){ref-type="ref"}\]. An indirect IgM immunofluorescence assay (Serion Immunodiagnostica) was performed when the complement fixation method was equivocal. A positive IgM result, together with an equivocal complement fixation test result, was also considered to be proof of a recent infection. As there are serological cross‐reactions between parainfluenza virus types 1, 2 and 3, these results were reported together.
Pcr {#ss5}
---
Detection of *B. pertussis* and *Bordetella parapertussis* was by PCR, using a MagNa Pure LC Total Nucleic Acid Isolation Kit (Roche Diagnostics, Indianapolis, IN, USA) and primer pairs based on IS*481* and IS*1001*. The final PCR product was analysed by gel electrophoresis and by dot‐blot hybridisation \[[16](#b16){ref-type="ref"}, [17](#b17){ref-type="ref"}\].
PCRs for *Chlamydia pneumoniae*, *M. pneumoniae*, influenza viruses A and B, coronaviruses 229E and OC43, RSV, human metapneumovirus, rhinovirus and enterovirus were performed on a nasal and a throat swab. A negative control (virus transport medium) was included for every four clinical samples. A positive control containing each pathogen was included in each DNA and RNA extraction and PCR run. DNA was isolated with a proteinase K and sodium dodecyl sulphate extraction, essentially as described previously \[[18](#b18){ref-type="ref"}\]. RNA was isolated with a High Pure RNA isolation kit (Roche), with the addition of poly(A) RNA as carrier, according to the manufacturer\'s instructions. An aliquot of the isolated DNA was used in a PCR detecting either *M. pneumoniae*\[[18](#b18){ref-type="ref"}\] or *Chlamydia pneumoniae*\[[19](#b19){ref-type="ref"}\]. An aliquot of the eluted RNA preparation was used for the detection of rhino/enterovirus \[[20](#b20){ref-type="ref"}\], influenza viruses A \[[21](#b21){ref-type="ref"}\] and B \[[22](#b22){ref-type="ref"}\], and human metapneumovirus \[[23](#b23){ref-type="ref"}\] in separate single‐tube RT‐PCRs. RSV \[[24](#b24){ref-type="ref"}\] and coronaviruses OC43 and 229E \[[25](#b25){ref-type="ref"}\] were amplified in a multiplex single‐tube nested RT‐PCR.
Culture for respiratory bacterial pathogens {#ss6}
-------------------------------------------
An oropharyngeal swab was inoculated on to a blood agar plate, a blood agar plate containing oxolinic acid (10 mg/L), and a chocolate agar plate. β‐Haemolytic group A streptococci, *Haemophilus influenzae*, *Streptococcus pneumoniae* and *Moraxella catarrhalis* were considered to be potential pathogens. *Staphylococcus aureus*, *Neisseria meningitidis*, non‐group A β‐haemolytic streptococci and *Candida albicans* were also reported, but were not considered to be pathogens. Only cultures showing significant growth were considered to be positive \[[26](#b26){ref-type="ref"}\]. Non‐pathogenic oropharyngeal flora (e.g., viridans streptococci, diptheroids, coagulase‐negative staphylococci) were not reported.
Statistical analysis {#ss7}
--------------------
For the description of patient characteristics, median and interquartile ranges were calculated. A statistical comparison was made between three groups of children, i.e., children in whom no pathogen was detected, children with only one pathogen, and children with two or more pathogens. The chi‐square test or, for continuous variables, analysis of variance (ANOVA) was used.
Results {#ss8}
=======
During the 2‐year period, 152 patients were referred for prolonged coughing, of whom 11 refused to participate in the study. Five patients were not included because of protocol violation. The remaining 136 episodes related to 135 patients, since one patient had two episodes of coughing with an interval of 5.2 months, which were considered as two independent events. Patient characteristics are summarised in [Table 1](#t1){ref-type="table"}. One or more pathogens were found in 91 (67%) patients. Of these patients, 49 (36%) had one pathogen, 35 (26%) had two pathogens, six (4%) patients had three pathogens, and one (1%) had four pathogens. The presence of more than one pathogen in patients was detected throughout the entire year of the study in children of all ages. The different pathogens detected are listed in [Table 2](#t2){ref-type="table"}.
######
Clinical data for all patients, grouped according to the number of potential pathogens isolated
All patients No pathogen detected One pathogen detected More than one pathogen detected p value
------------------------------------------------------------------------------------ ---------------- ---------------------- ----------------------- --------------------------------- ---------
*n* 136 45 (33%) 49 (36%) 42 (31%)
Age, years (median and interquartile range) 2.7 (0.6--6.4) 1.3 (0.6--5.8) 1.0 (0.4--5.0) 4.4 (1.8--9.0) 0.005
Male gender 80 (59%) 27 (60%) 31 (63%) 22 (52%) 0.56
Disease duration, days (median and interquartile range)[^a^](#t1n9){ref-type="fn"} 14 (10--20) 15 (13--21) 14 (9--17) 14 (9--20) 0.33
Vaccination status[^b^](#t1n8){ref-type="fn"} 124 (92%) 41 (91%) 44 (90%) 39 (95%) 0.64
Admission to hospital 51 (38%) 16 (36%) 22 (45%) 13 (31%) 0.37
Length of stay following admission, days (median and interquartile range) 5 (3--7) 3.5 (2--6) 5 (3--7) 5 (4--7) 0.49
Antibiotics prescribed 68 (50%) 22 (49%) 23 (47%) 23 (55%) 0.75
Disease duration: time between first day of illness and first presentation in the hospital.
Information on vaccination status (regular vaccinations appropriate for age) was missing for one patient.
######
Respiratory pathogens detected in the study and means of detection
Children with pathogen *n* = 91 (67%)[^a^](#t2n7){ref-type="fn"} Diagnostic test
--------------------------------------- ------------------------------------------------------------------ ----------------- ----- ----- -----
Pathogens
*Bordetella pertussis* 23 (25%) 3 8 12 ND
*Bordetella parapertussis* 1 (1%) ND 1 ND ND
Influenza virus A or B 5 (5%) 5 0 0 ND
Adenovirus 6 (7%) 6 ND ND ND
*Mycoplasma pneumoniae* 6 (7%) 4 1 1 ND
Parainfluenza virus 9 (10%) 9 ND ND ND
Respiratory syncytial virus 15 (16%) 14 1 0 ND
Rhinovirus 43 (47%) ND 43 ND ND
Human metapneumovirus 1 (1%) ND 1 ND ND
Enterovirus 4 (4%) ND 4 ND ND
*Chlamydia pneumoniae* 1 (1%) 0 0 1 ND
Carriage organisms
*Streptococcus pneumoniae* 1 (1%) ND ND ND 1
*Haemophilus parainfluenzae* 7 (8%) ND ND ND 7
*Staphylococcus aureus* 1 (1%) ND ND ND 1
*Haemophilus influenzae* non‐type b 12 (13%) ND ND ND 12
*Haemophilus influenzae* type b 1 (1%) ND ND ND 1
β‐Haemolytic streptococci group C 1 (1%) ND ND ND 1
β‐Haemolytic streptococci group G 1 (1%) ND ND ND 1
β‐Haemolytic streptococci group B 1 (1%) ND ND ND 1
*Candida albicans* 2 (2%) ND ND ND 2
Total 141
Two or more pathogens were found in 42 patients.
All respiratory pathogens found in the present study were considered to be possible causes of prolonged coughing, even though patients may also be colonised by some of the pathogens found. The pathogens detected most frequently were rhinovirus (43 patients; 47%), *B. pertussis* (23 patients; 25%) and RSV (15 patients; 16%) ([Table 2](#t2){ref-type="table"}). In the 42 patients with a mixed infection ([Table 3](#t3){ref-type="table"}), the most frequent combination was *B. pertussis* and rhinovirus (*n* = 14).
######
Combinations of respiratory pathogens detected in 42 patients with more than one pathogen
*Bordetella* *pertussis* *Bordetella* *parapertussis* Influenza virus A or B Adenovirus *Mycoplasma* *pneumoniae* Parainfluenza virus Respiratory syncytial virus Rhinovirus
--------------------------------------- ------------------------------- ------------------------------ ------------------------ ------------ ------------------------------ ------------------------------ ------------------------------ ------------------------------
Pathogens
*Bordetella pertussis*
*Bordetella parapertussis*
Influenza virus A or B
Adenovirus 2[^a^](#t3n6){ref-type="fn"} 1[^b^](#t3n5){ref-type="fn"} 1
*Mycoplasma pneumoniae*
Parainfluenza virus
Respiratory syncytial virus 1 1 1
Rhinovirus 14[^c^](#t3n4){ref-type="fn"} 2 2[^d^](#t3n3){ref-type="fn"} 2[^a^](#t3n6){ref-type="fn"}
Human metapneumovirus
Enterovirus 1 1 1
*Chlamydia pneumoniae*
Carriage organisms
*Streptococcus pneumoniae* 1
*Haemophilus parainfluenzae* 1 1
*Staphylococcus aureus*
*Haemophilus influenzae* non‐type b 1 3 1
*Haemophilus influenzae* type b 1
β‐Haemolytic streptococci group C
β‐Haemolytic streptococci group G 1[^e^](#t3n2){ref-type="fn"}
β‐Haemolytic streptococci group B 1[^c^](#t3n4){ref-type="fn"}
*Candida albicans* 1
Three pathogens were detected in six patients, and four pathogens in one patient.
One patient also had *Haemophilus parainfluenzae.*
Also respiratory syncytial virus.
One patient also had *Mycoplasma pneumoniae.*
One patient also had respiratory syncytial virus and *Haemophilus parainfluenzae.*
One patient also had *Haemophilus influenzae* non‐type b.
Fifteen of the 27 positive bacterial cultures were found in patients with more than one pathogen ([Table 3](#t3){ref-type="table"}). There were 11 patients with a positive oropharyngeal culture and one other pathogen. In only one of these patients (with a C‐reactive protein level of 109 mg/L) was the bacterial infection considered to influence the course of the disease, while the other ten patients showed no clinical sign of a bacterial infection. It was difficult to assess whether the results for these ten patients represented a mixed or a consecutive infection, or whether the bacteria found simply reflected carriage. Therefore, these cases were not considered to be possible mixed infections. On the basis of serology, PCR and culture, 21 (24%) of the remaining 32 patients with more than one pathogen detected were considered to have a possible mixed infection (14 of these had a positive PCR for two or three pathogens), nine had a consecutive infection, and two had both.
There were no significant differences between patients with or without pathogens with respect to disease duration before presentation, disease severity (antibiotics, hospitalisation, length of stay, coughing and other symptoms, such as headache, sore throat or pain) or vaccination status ([1](#t1){ref-type="table"}, [4](#t4){ref-type="table"}. However, children with more than one pathogen were, on average, 3 years older than children with one or no pathogens ([Table 1](#t1){ref-type="table"}; p 0.005). Although there were no significant differences in coughing during different parts of the day, there seemed to be a tendency for children with more than one pathogen to cough more frequently ([Table 4](#t4){ref-type="table"}). There were no significant differences in coughing or other symptoms for the three pathogens found most frequently, except that patients with *B. pertussis* had less fever. All patients recovered, although the time to recovery was not recorded accurately. Data regarding coughing in the family or other potential contaminators were not recorded.
######
Number (%) of patients with coughing and other symptoms
No pathogen One pathogen More than one pathogen *Bordetella* *pertussis* Rhinovirus Respiratory syncytial virus
-------------------------------------------- ------------- -------------- ------------------------------------ -------------------------- ------------ -----------------------------
All (*n* = 135)[^a^](#t4n1){ref-type="fn"} 45 (33) 49 (36) 41 (31)[^a^](#t4n1){ref-type="fn"} 23 (17) 43 (32) 15 (11)
Coughing
Morning (*n* = 82) 25 (30) 26 (32) 31 (38) 16 (20) 30 (37) 11 (13)
Afternoon (*n* = 70) 22 (31) 21 (30) 27 (39) 17 (24) 26 (37) 9 (13)
Evening (*n* = 88) 25 (28) 32 (36) 31 (35) 20 (23) 31 (35) 9 (10)
Night (*n* = 89) 25 (28) 32 (36) 32 (36) 17 (19) 29 (33) 13 (15)
Symptoms
Fever (*n* = 47) 18 (38) 14 (30) 15 (32) 1 (2) 13 (28) 7 (15)
Tachypnoea (*n* = 51) 16 (31) 20 (39) 15 (29) 6 (12) 15 (29) 9 (18)
Dyspnoea (*n* = 63) 22 (35) 21 (33) 20 (32) 9 (14) 14 (22) 7 (11)
There were no significant differences, except that there was less fever associated with *Bordetella pertussis* infection.
Data missing for one patient.
Discussion {#ss9}
==========
To our knowledge, this is the first study in which a broad spectrum of respiratory pathogens has been studied prospectively in children selected by persistent coughing. As found in the present study, rhinovirus is associated frequently with respiratory disease in children of all ages \[[27](#b27){ref-type="ref"}, [28](#b28){ref-type="ref"}, [29](#b29){ref-type="ref"}, [30](#b30){ref-type="ref"}\], with incidences ranging from 21% to 40% in children with overt infection, but with an incidence of 20% in one study of children without nasal symptoms \[[29](#b29){ref-type="ref"}\]. Rhinovirus can cause mild upper airway disease, and can also play a role in more severe respiratory tract infection. It is difficult to compare the results regarding the frequency of various pathogens obtained in the present study with those of other studies on respiratory tract infections among children, since such studies have been mainly retrospective \[[2](#b2){ref-type="ref"}, [4](#b4){ref-type="ref"}, [5](#b5){ref-type="ref"}, [6](#b6){ref-type="ref"}, [8](#b8){ref-type="ref"}, [9](#b9){ref-type="ref"}, [10](#b10){ref-type="ref"}, [30](#b30){ref-type="ref"}, [31](#b31){ref-type="ref"}, [32](#b32){ref-type="ref"}, [33](#b33){ref-type="ref"}, [34](#b34){ref-type="ref"}, [35](#b35){ref-type="ref"}, [36](#b36){ref-type="ref"}, [37](#b37){ref-type="ref"}, [38](#b38){ref-type="ref"}, [39](#b39){ref-type="ref"}, [40](#b40){ref-type="ref"}, [41](#b41){ref-type="ref"}, [42](#b42){ref-type="ref"}, [43](#b43){ref-type="ref"}, [44](#b44){ref-type="ref"}\]. The incidence of pathogenic agents responsible for respiratory infection in the present study was 67%, which is comparable with the results of some studies (incidences of 56--64%) \[[2](#b2){ref-type="ref"}, [9](#b9){ref-type="ref"}, [33](#b33){ref-type="ref"}, [40](#b40){ref-type="ref"}\], but different from those of others (incidences of 80--90%) \[[9](#b9){ref-type="ref"}, [30](#b30){ref-type="ref"}, [37](#b37){ref-type="ref"}\].
A relatively high incidence (10% of all patients) of co‐infection with rhinovirus and *B. pertussis* was observed. Among children infected with *B. pertussis*, co‐infection with rhinovirus was found in 61% of cases. The role of *B. pertussis* in rhinovirus infection is complicated; indeed it has been reported that rhinovirus‐induced changes in airway smooth muscle responsiveness in isolated rabbit and human airway smooth muscle tissue, as well as cultured airway smooth muscle cells, were largely prevented by pretreating the tissues with pertussis toxin or with a monoclonal blocking antibody to intercellular adhesion molecule‐1, the principal endogenous receptor for most rhinoviruses \[[45](#b45){ref-type="ref"}\]. This might indicate that co‐infection with rhinovirus and *B. pertussis* is not as severe as might be expected. This hypothesis was supported by the observation in the present study that coughing and other clinical symptoms did not differ significantly from those in patients infected with either rhinovirus or *B. pertussis*, and the fact that, although there was a co‐infection incidence of 61% with *B. pertussis* and rhinovirus, there was less fever in the co‐infected group, as would be expected with *B. pertussis* infection alone. In agreement with other studies \[[2](#b2){ref-type="ref"}, [9](#b9){ref-type="ref"}, [31](#b31){ref-type="ref"}, [44](#b44){ref-type="ref"}, [46](#b46){ref-type="ref"}\], no pathogens or positive serological results were obtained for 33% of the patients in the present study, although pathogens as yet unidentified may be associated with prolonged coughing in this group, as exemplified by the recent description of human metapneumovirus and human coronavirus NL63 \[[23](#b23){ref-type="ref"}, [47](#b47){ref-type="ref"}, [48](#b48){ref-type="ref"}\].
Asymptomatic children were not recruited as a control group, since this was considered to be too distressing for children. However, the occurrence of pathogens in subjects without coughing would provide data on the incidence of carriage of potential pathogens without disease. Some studies have observed fewer viruses in non‐symptomatic controls compared with patients \[[32](#b32){ref-type="ref"}, [33](#b33){ref-type="ref"}, [34](#b34){ref-type="ref"}, [35](#b35){ref-type="ref"}\]. Gunnarsson *et al.*\[[31](#b31){ref-type="ref"}\] isolated *S*. *pneumoniae*, *H. influenzae* and *Moraxella catarrhalis* more frequently from patients with long‐standing cough than from healthy individuals of the same age. In a case‐control study involving general practitioner patients of all ages with acute respiratory tract infections, significantly fewer viruses were detected in controls than in patients (19% vs. 54%) \[[36](#b36){ref-type="ref"}\].
Most patients with a positive culture for pathogenic bacteria will not have a bacterial respiratory tract infection. Since carriage rates for potential pathogenic bacteria in healthy young children may vary between 11% and 48%\[[31](#b31){ref-type="ref"}, [49](#b49){ref-type="ref"}\], it is difficult to establish whether these bacteria may have a clinical influence. In a logistic regression analysis, although not statistically significant, there was a greater association with coughing in the morning for pathogens diagnosed by PCR or serology than for pathogens diagnosed by oropharyngeal bacterial culture (OR, 2.0 vs. 0.8). Similar associations were found when coughing patterns at other times of the day were analysed. Therefore, it can be argued that a positive bacterial oropharyngeal culture is of little value in the search for respiratory pathogens causing prolonged cough.
More than one pathogen was detected in 31% of all children. The presence of one (viral) respiratory pathogen may predispose to a second or third (bacterial) infection \[[39](#b39){ref-type="ref"}, [42](#b42){ref-type="ref"}, [46](#b46){ref-type="ref"}\]. The proportion of mixed infections reported in other studies shows large variations (10--74%) \[[9](#b9){ref-type="ref"}, [33](#b33){ref-type="ref"}, [34](#b34){ref-type="ref"}, [38](#b38){ref-type="ref"}, [39](#b39){ref-type="ref"}, [46](#b46){ref-type="ref"}\]. There is a significant seasonal influence on disease, so the observation period of the present study was 2 years. In the second year, a low incidence of prolonged coughing was observed, with few referrals, perhaps associated with the mild winter of 2002--2003 and the warm summer of 2003 in The Netherlands. However, the frequency of isolation of two or more pathogens during both years of the study in every season remained the same. Previous studies have associated mixed infections with more severe disease \[[3](#b3){ref-type="ref"}, [4](#b4){ref-type="ref"}, [5](#b5){ref-type="ref"}, [6](#b6){ref-type="ref"}, [43](#b43){ref-type="ref"}, [44](#b44){ref-type="ref"}, [48](#b48){ref-type="ref"}\], but this was not the case in the present study ([Table 1](#t1){ref-type="table"}). Generally, it seemed that there was one dominant pathogen, with no cumulative or synergic effect of a second pathogen. Patients with more than one pathogen were significantly older than those with one or no pathogens, which might indicate that they were more seriously ill than children of similar age with one or no detectable pathogen. This is supported by the fact that there was a slight tendency for coughing to be more frequent in patients with two or more pathogens ([Table 4](#t4){ref-type="table"}).
In summary, one pathogen was detected in 36% of children with prolonged coughing, and more than one pathogen in 31%. There was a strong seasonal influence on the number of cases, but not on the pathogens found or on the percentage of multiple infections. Clinical data did not distinguish between pathogens, or whether or not pathogens were found. There was no increase in disease severity, with respect to symptoms and hospitalisation, for cases of mixed or consecutive infection, although this group was significantly older, suggesting that they had a more severe disease than their peer groups with one or no pathogens. Obviously, not all possible pathogens found will cause illness, although it is still not clear whether or how the presence of a colonising potential respiratory pathogen influences or facilitates infection caused by other possible pathogens. It is therefore important to extend studies on respiratory tract infections associated with more than one pathogen in order to evaluate their influence on disease severity.
Acknowledgements {#ss10}
================
We thank J. C. M. Hoekx, E. G. H. van Leer, J. S. Starreveld, C. V. Tjon Pian Gi and all the interns of the Department of Pediatrics at Groene Hart Hospital for their participation in this study. We thank B. McClements for revising the text, and J. Buitenwerf and R. Woudenberg (Regionaal Medisch‐Microbiologisch Laboratorium, Rotterdam, The Netherlands) for helping to interpret the serological data. We thank Abbott and the Groene Hart Hospital for their support, which made this study possible. This study was supported by an unrestricted grant from Abbott.
| {
"pile_set_name": "PubMed Central"
} |
Xie H., Hu L., Wu F., Chen M., Wu L. (2016). Self‐Templated Synthesis of Ultrathin Nanosheets Constructed TiO~2~ Hollow Spheres with High Electrochemical Properties. Adv. Sci., 3: 1600162. doi: 10.1002/advs.201600162
1. Introduction {#advs194-sec-0010}
===============
Hollow hierarchical nanostructures with properly engineered building blocks have been devoted considerable interest for the fundamental research and practical applications in various fields, such as catalysis, biomedicine, water purification, energy storage and conversion, etc.[1](#advs194-bib-0001){ref-type="ref"}, [2](#advs194-bib-0002){ref-type="ref"}, [3](#advs194-bib-0003){ref-type="ref"}, [4](#advs194-bib-0004){ref-type="ref"}, [5](#advs194-bib-0005){ref-type="ref"}, [6](#advs194-bib-0006){ref-type="ref"}, [7](#advs194-bib-0007){ref-type="ref"}, [8](#advs194-bib-0008){ref-type="ref"}, [9](#advs194-bib-0009){ref-type="ref"}
Generally, two main strategies, e.g., template‐based (hard‐ and soft‐templates) and template‐free (Kirkendall effect, galvanic replacement, and Ostwald ripening), are used to obtain well‐defined hollow spheres with different compositions and shapes.[5](#advs194-bib-0005){ref-type="ref"}, [10](#advs194-bib-0010){ref-type="ref"}, [11](#advs194-bib-0011){ref-type="ref"}, [12](#advs194-bib-0012){ref-type="ref"}, [13](#advs194-bib-0013){ref-type="ref"}, [14](#advs194-bib-0014){ref-type="ref"}, [15](#advs194-bib-0015){ref-type="ref"}, [16](#advs194-bib-0016){ref-type="ref"}, [17](#advs194-bib-0017){ref-type="ref"} Recent researches have demonstrated that hollow spheres constructed with 1D nanorods/nanoneedles[18](#advs194-bib-0018){ref-type="ref"} and 2D nanosheets[19](#advs194-bib-0019){ref-type="ref"}, [20](#advs194-bib-0020){ref-type="ref"}, [21](#advs194-bib-0021){ref-type="ref"}, [22](#advs194-bib-0022){ref-type="ref"} even exhibit faster surface reaction rates and better physicochemical stability because they occupy much larger specific surface areas than the most hollow spheres composed of 0D nanoparticles,[23](#advs194-bib-0023){ref-type="ref"} and simultaneously inherit the characteristics of the constituent units. However, synthesis of this kind of hierarchical hollow structures is generally challenging due to the difference of stability between core and shell materials.[24](#advs194-bib-0024){ref-type="ref"}, [25](#advs194-bib-0025){ref-type="ref"} For example, TiO~2~ is a kind of well‐known nanomaterials as photocatalysts,[26](#advs194-bib-0026){ref-type="ref"}, [27](#advs194-bib-0027){ref-type="ref"} and anode material of Li‐ion batteries,[28](#advs194-bib-0028){ref-type="ref"}, [29](#advs194-bib-0029){ref-type="ref"}, [30](#advs194-bib-0030){ref-type="ref"}, [31](#advs194-bib-0031){ref-type="ref"}, [32](#advs194-bib-0032){ref-type="ref"}, [33](#advs194-bib-0033){ref-type="ref"}, [34](#advs194-bib-0034){ref-type="ref"}, [35](#advs194-bib-0035){ref-type="ref"}, [36](#advs194-bib-0036){ref-type="ref"}, [37](#advs194-bib-0037){ref-type="ref"}, [38](#advs194-bib-0038){ref-type="ref"}, [39](#advs194-bib-0039){ref-type="ref"}, [40](#advs194-bib-0040){ref-type="ref"}, [41](#advs194-bib-0041){ref-type="ref"} owing to its low cost, environmental benignity, physicochemical stability, and safety, but most of them are nanoparticles or primary nanoparticle‐constructed hollow spheres, very few researches involve the synthesis of 1D‐ or 2D‐constructed hollow spheres, through template methods.[42](#advs194-bib-0042){ref-type="ref"}, [43](#advs194-bib-0043){ref-type="ref"}, [44](#advs194-bib-0044){ref-type="ref"}
In this study, we report a self‐templated synthesis of ultrathin nanosheets constructed titania hierarchical hollow spheres (UNTHS) by hydrothermally treating the titania--silicone composite particles and then calcination. Because the nanosheets favor the short Li^+^ ion diffusion pathway and fast electrochemical reactions, and the hierarchical hollow architectures can still accommodate the volume change and avoid stacking of nanosheets during cycling process, the uniquely structured hollow spheres as the anode material of Li‐ion battery exhibit excellent rate capability and very good cycle stability even at current density of ≈10 C.
2. Results and Discussion {#advs194-sec-0020}
=========================
2.1. Synthesis of Hierarchical TiO~2~ Hollow Spheres {#advs194-sec-0030}
----------------------------------------------------
The synthesis of UNTHS was started from the pre‐synthesized hybrid particles via the sol‐gel reactions of titania precursor, titanium tetraisopropoxide (Ti(OiPr)~4~), and a small amount of long hydrophobic alkyl silane hexadecylsilanetriols (C~16~TS) as a hydrolysis inhibitor and nucleation agents,[45](#advs194-bib-0045){ref-type="ref"} in ethanol under the catalysis of concentrated ammonia. Without C~16~TS, only irregular and nonuniform particles were produced (Figure S1, Supporting Information). The composite particles display monodisperse and spherical morphology with a mean diameter of 380 nm (**Figure** [**1**](#advs194-fig-0001){ref-type="fig"}a). The higher magnification SEM image (Figure [1](#advs194-fig-0001){ref-type="fig"}b) indicates that the composite particles with rough surfaces are composed of numerous primary particles. The energy‐dispersive X‐ray spectrometry (EDS) further reveals that 35.2 wt% of Ti and 3.30 wt% of Si are uniformly dispersed in the particles as shown in Figure [1](#advs194-fig-0001){ref-type="fig"}c and Table S1 of the Supporting Information.
{#advs194-fig-0001}
After hydrothermally treated in NaOH solution followed by the H^+^‐exchange process, these titanate--silicone composite particles were successfully converted to yolk--shell composite spheres as shown in **Figure** [**2**](#advs194-fig-0002){ref-type="fig"}a. Their X‐ray diffraction (XRD) pattern shows broad diffraction peaks with low intensity (Figure S2a, Supporting Information), which can be indexed to H~2~Ti~5~O~11~·3H~2~O,[46](#advs194-bib-0046){ref-type="ref"} suggesting its poor crystallization. Meanwhile, an obvious low angle diffraction peak at around 8.7° indicating the layered structures.[46](#advs194-bib-0046){ref-type="ref"}, [47](#advs194-bib-0047){ref-type="ref"} The peak marked by black rhombus at around 21.3° could be attributed to polycondensed C~16~TS, like some other organosilica,[48](#advs194-bib-0048){ref-type="ref"} which is absent after calcination. Raman spectra show typical titanate phase with weak and broad peaks (Figure S2b, Supporting Information), further confirming the poor crystallized titanate phase.[49](#advs194-bib-0049){ref-type="ref"} The yolk is mainly composed of Si and C elements indicated by the EDS mapping images in Figure [2](#advs194-fig-0002){ref-type="fig"}e. And the neglectable amount of Na within the samples reveals that it has been almost entirely removed by the H^+^‐exchange. The inner diameter of the yolk--shell structure is ≈380 nm. The nanosheets constructing the hierarchical microspheres are semitransparent with a thickness of ≈3 nm, and a lamellar structure with a spacing of ≈1.0 nm is observed in high‐resolution transmission electron microscopy (HRTEM) image (Figure [2](#advs194-fig-0002){ref-type="fig"}b), which is corresponding to the XRD results. After calcination at 350 °C, the yolk has been burned away to form a complete ultrathin nanosheets constructed hollow spheres (Figures [2](#advs194-fig-0002){ref-type="fig"}c,f, Supporting Information), and titanate has been transformed into anatase titania with a small content of TiO~2~‐B, which can be confirmed by the XRD (Figure S2a, Supporting Information), and the Raman spectrum (Figure S2b, Supporting Information).[50](#advs194-bib-0050){ref-type="ref"} Meanwhile, no obvious peaks of silica are observed due to its small content and amorphous phase, and the low‐angle diffraction peak disappears and a broad peak at ≈12° appears due to the collapse of the interlayer spacing by dehydration and recrystallization.[51](#advs194-bib-0051){ref-type="ref"}, [52](#advs194-bib-0052){ref-type="ref"} The obtained hierarchical UNTHS still have good morphology, whose Brunauer--Emmett--Teller (BET) surface area and total pore volume are calculated to be ≈285.3 m^2^ g^−1^ and ≈0.88 cm^3^ g^−1^, respectively, based on the N~2~ sorption isotherms (Figure S3, Supporting Information). The thermogravimetric analysis indicates a main weight loss of ≈19.7% (Figure S4a, Supporting Information), which is consistent with the decomposition of titanate to form TiO~2~ and organic components. The HRTEM images (Figure [2](#advs194-fig-0002){ref-type="fig"}d, inset) show nanocrystals within nanosheets with a interspacing of ≈0.35 nm corresponding to the (101) planes of anatase TiO~2~. The selected area electron diffraction (SAED) pattern of the same region reveals well defined diffraction pattern, which can also be assigned to the anatase TiO~2~.[31](#advs194-bib-0031){ref-type="ref"} The EDS mapping images of the UNTHS further display that the Ti is absolutely located in the nanosheet‐based shell, only residual silica is dispersed in the inner TiO~2~ shell (Figure [2](#advs194-fig-0002){ref-type="fig"}f).
{#advs194-fig-0002}
To illustrate the formation mechanism of the yolk--shell spheres and UNTHS, a time‐dependent hydrothermal treatment experiment in NaOH aqueous solution (1 [m]{.smallcaps}) at 150 °C was performed. Products were collected at different stages of the heating process, and their morphologies were subjected to TEM investigation. As shown in **Figure** [**3**](#advs194-fig-0003){ref-type="fig"}, only after hydrothermal treatment for 20 min under NaOH solution, some irregular and whisker‐like nanosheets can be seen from the smooth surfaces of the composite particles (Figures [3](#advs194-fig-0003){ref-type="fig"}b vs [3](#advs194-fig-0003){ref-type="fig"}a), forming pompons‐like particles with porous outer layer. The XRD pattern does not show any broad diffraction peak which can be ascribed to amorphous titanate--silicone (Figure [3](#advs194-fig-0003){ref-type="fig"}f), except for only several weak and broad peaks.[25](#advs194-bib-0025){ref-type="ref"}, [53](#advs194-bib-0053){ref-type="ref"} After continuous hydrothermal etching for 40 min, the core of the pompons‐like particle is shrinking, producing yolk--shell nanostructures with a portion of titanate compositions in the yolk as shown in Figure [3](#advs194-fig-0003){ref-type="fig"}c and Figure S5 of the Supporting Information. Simultaneously, the nanosheets are growing up and an obvious peak at ≈8.5° appears, indicating the formation of lamellar structure. When the reaction time is increased to 60 min, the broad and weak peak at 21.3° become much clearer and sharper, indicating an increasing degree of polycondensation of C~16~TS. And the size of the yolk is further decreased with well defined hierarchical shell (Figure [3](#advs194-fig-0003){ref-type="fig"}d). After 2 h reaction time, the hierarchical yolk--shell structure was kept unchanged (Figure [3](#advs194-fig-0003){ref-type="fig"}e). Meanwhile, the intensity of the low‐angle diffraction peak is obviously enhanced as shown in Figure [3](#advs194-fig-0003){ref-type="fig"}f, suggesting a more pronounced layered structure of the titanate nanosheets.
{#advs194-fig-0003}
Based on the above experiments and discussion, the hierarchical microspheres could be formed through an inward etching and outward recrystallization process in a short time scale as shown by the cartoons in Figure [3](#advs194-fig-0003){ref-type="fig"}. With the concentrated NaOH solution is used to assist etching, titanate nanosheets form on the surfaces of the composite particles owing to heterogeneous nucleation, which is energetically favored relative to the homogeneous growth in solution.[54](#advs194-bib-0054){ref-type="ref"} And the outer surface is becoming porous, which facilitates the alkali solution penetrating inward for further etching. Meanwhile, the long hydrophobic alkyl C~16~TS molecules distributing uniformly in the composite particles mostly shrink toward core and gradually separate from titanate‐rich phase due to their polycondensation with hydrophobic effect. As a result, the dissolution rate of the amorphous composite particles is gone well with the continuously outward epitaxial growth of titanate nanosheets around the surfaces of the particles to form cavity, which is similar to the Ostwald‐ripening process.[42](#advs194-bib-0042){ref-type="ref"}, [55](#advs194-bib-0055){ref-type="ref"} And the polycondensed C~16~TS molecules form yolk. The whole etching and recrystallization process come to an end within 60 min, showing high efficiency of the hydrothermal reaction. When these yolk--shell spheres were calcinated, the TiO~2~ hollow spheres composed of ultrathin titania nanosheets were obtained as discussed above.
2.2. Electrochemical Properties of the Hierarchical TiO~2~ Hollow Spheres {#advs194-sec-0040}
-------------------------------------------------------------------------
Considering this unique hollow structure constructed with thin nanosheets, we further investigated their charge storage performance in nonaqueous electrolyte including Li^+^ ions in half‐cells using Li metal as counter and reference electrodes. The principal electrochemical Li^+^ insertion/extraction reaction in a TiO~2~/Li half‐cell can be described as $$\left. {TiO}_{2} + x{Li}^{+} + xe^{-}\leftrightarrow{Li}_{x}{TiO}_{2} \right.$$where the coefficient *x* is the amount of inserted Li^+^ in TiO~2~ and the maximum value of it is ≈0.5, which leads to a theoretical storage capacity of 167.5 mAh g^−1^.[56](#advs194-bib-0056){ref-type="ref"}
As shown in **Figure** [**4**](#advs194-fig-0004){ref-type="fig"}a, at a sweep rate of 0.1 mV s^−1^ and in the potential range between 1.0 and 3.0 V (vs Li/Li^+^), the dominated pairs of redox peaks at ≈1.7 and 2.0 V denoted as A of the cyclic voltammetry (CV) are corresponding to the diffusion controlled process of Li^+^ ion into anatase TiO~2~,[57](#advs194-bib-0057){ref-type="ref"} while the two minor peaks at around 1.54 to 1.68 V denoted as S1 and S2 are ascribed to the surface Faradaic pseudocapacitive reactions of TiO~2~‐B.[58](#advs194-bib-0058){ref-type="ref"}, [59](#advs194-bib-0059){ref-type="ref"} Figure [4](#advs194-fig-0004){ref-type="fig"}b shows the CV curves of UNTHS based electrodes obtained at different scan rates ranging from 0.1 to 2 mV s^−1^. As the scan rate increases, all redox peaks are becoming broader and broader, and the peaks S1 and S2 are becoming larger and larger, namely increasing currents as the rate of surface confined faradaic reaction is faster than the diffusion controlled process, indicating that faradaic pseudocapacitance is gradually dominating the electrochemical capacitance at high scan rates, which is favor for high rate capacitance.[60](#advs194-bib-0060){ref-type="ref"}, [61](#advs194-bib-0061){ref-type="ref"}
{#advs194-fig-0004}
The galvanostatic charge--discharge voltage scans at current density of 0.1 A g^−1^ over a voltage range of 1.0--3.0 V clearly exhibit three distinct domains (Figure [4](#advs194-fig-0004){ref-type="fig"}c). The first domain is a monotonous voltage drop from 3.0 to ≈1.7 V induced by Li^+^ ion diffusion into the bulk, providing a limit of capacity. The second domain is a characteristic plateau around 1.7 V showing reversible insertion of Li^+^ ion into the anatase TiO~2~, the final domain is a slow decay of the voltage associated with intercalation pseudocapacitance and surface faradaic reaction.[60](#advs194-bib-0060){ref-type="ref"}, [62](#advs194-bib-0062){ref-type="ref"} At a current rate of 0.1 A g^−1^ (Figure [4](#advs194-fig-0004){ref-type="fig"}c), the discharge capacity at the first cycle is 333.3 mAh g^−1^, while the first charge process releases only 249.8 mAh g^−1^, corresponding to a initial 25.1% capacity loss. From the TGA analysis of the UNTHS (Figure S4b, Supporting Information), the sample calcined during temperature region below 100 °C shows ≈3.4 wt% weight loss which could be attributed to the evorporation of adsorbed moisture due to this highly porous structure, and ≈1.3 wt% weight loss as calcined to 450 °C indicating small amount of carbonaceous residues within the UNTHS structure, which is also found in the EDS results (Figure [2](#advs194-fig-0002){ref-type="fig"}f). These residues and water may cause irreversible reaction of Li^+^ ion and electrolyte decomposition.[63](#advs194-bib-0063){ref-type="ref"} It should be noted that a small amount of silica within the UNTHS (12.5 wt% determined from the elemental analysis by inductively coupled plasma‐atomic emission spectrometry (ICP‐AES)) devotes negligible capacity at voltage range of 1.0--3.0 V.[64](#advs194-bib-0064){ref-type="ref"}, [65](#advs194-bib-0065){ref-type="ref"}, [66](#advs194-bib-0066){ref-type="ref"} However, after further cycles, the coulombic efficiency gradually increases to almost 100%. And a reversible capacity of 232.7 mAh ^−1^ during the initial 10 cycles, corresponding to *x* = 0.64 (higher than the theoretical value of 0.5), furthur suggests abundant surface area for Faradaic pseudocapacitive reactions owing to the hierarchical hollow structure.[21](#advs194-bib-0021){ref-type="ref"}, [60](#advs194-bib-0060){ref-type="ref"}
Figure [4](#advs194-fig-0004){ref-type="fig"}d also displays the rate performance of UNTHS at discharge rates from 0.1 up to 5 A g^−1^, then back to 0.1 A g^−1^, each step lasts for 10 cycles. The UNTHS exhibits high specific capacity of ≈105 mAh g^−1^ even at 5 A g^−1^, indicating that this material with ultrathin titania nanosheets assembled hollow structure has excellent rate capability. Moreover, when the current density returns to 0.1 A g^−1^, the capacity reverts to values above 210 mA h g^−1^ in spite of some capacity decay compared to the first ten cycles, which may be due to the volume swing of the TiO~2~ crystal structure.[19](#advs194-bib-0019){ref-type="ref"}, [22](#advs194-bib-0022){ref-type="ref"} The UNTHS also demonstrates nice cycling performane. As shown in **Figure** [**5**](#advs194-fig-0005){ref-type="fig"}a, despite initial capacity decay, the UNTHS could still deliver discharge capacity as high as ≈123 mAh g^−1^ at a current density of 2 A g^−1^ (≈10 C) after 200 cycles, which is almost double that of the commercial TiO~2~ (C‐TiO~2~, 5--10 nm). As far as we know, these electrochemical properties of the UNTHS‐based anode are better than those of most of the previously reported TiO~2~‐ based electrodes (Table S2, Supporting Information). To understand the improved electrochemical performance of the UNTHS, both the Nyquist plots of UNTHS and C‐TiO~2~ nanocrystals electrodes are shown in Figure [5](#advs194-fig-0005){ref-type="fig"}b, which exhibit depressed semicircle during high‐ and midium‐frequency region, and a straight line in the low‐frequency rigion. The high frequency semicircle is associated with the internal resistance including the interface resistance, the separator, and the charge transfer resistance. The medium‐frequency linear region which are relevant to Li^+^ ion interfacial diffusion resitance, coupled with a double layer capacitance at the interface. It can be seen that the UNTHS is provided with lower ionic resistance and enhanced kinetics of electrolyte peneatration than the C‐TiO~2~. This could be ascribed to the special hierarchical hollow structure assembled with titania nanosheets. The unique hollow titania spheres provide sustainable electrode/electrolyte contact area and reduce diffusion pathways of Li^+^ ions, promoting the electrochemial reactions within the surface to achieve high specific capacitane and good rate performance. Furthermore, the TEM images (Figure S6, Supporting Information) clearly reveal that the UNTHS still maintains the original hierarchical hollow spheres with nanosheets subunits, although the detailed nanosheet structures have somewhat been diminished. Intelligibly, such ultrathin nanosheets are not robust enough to withstand the fast Li^+^ insertion/extraction process under the extended high rate cycling process. The XRD analysis of the UNTHS electrode slurry before and after 200 cycles (Figure S7, Supporting Information) suggested that the anatase crytal structure can still be maintained with no apparent intensity change of the dominate (101) peak. This observation indicates that the hollow structure with interweaving ultrathin nanosheets could efficiently keep them stable rather than aggregation, leading to outstanding cycling stability.
{#advs194-fig-0005}
3. Conclusion {#advs194-sec-0050}
=============
We have demonstrated the simple and novel synthesis of the ultrathin nanosheets constructed titania hollow spheres through a solvothermal treatment of the titanate--silicone composite particles combined with calcination. When such unique hierarchical TiO~2~ hollow spheres with a high specific area used as the anode of the Li‐ion battery, they can effectively shorten the diffusion path of Li^+^ ions and enlarge contact surface area of electrodes/electrolyte, promoting the electrochemical ractions. Accordingly, they show very good cycle stability even at current density of ≈10 C and excellent rate capacilities.
4. Experimental Section {#advs194-sec-0060}
=======================
*Synthesis of Composite Titanate--Silicone Paticles*: Ti(O^i^Pr)~4~ (1.42 g, ≈5.0 mmol) was dissolved in hexane (7.0 mL) and heated to 60 °C, and added by C~16~TS (0.25 g, ≈1.0 mmol) in 2.5 mL THF under stirring. This mixture was held at that temperature for 1 h, during which the solvent and resultant isopropyl alcohol were fractionated off continuously under vaccum. When no isopropyl alcohol came out, the composite precursors were obtained and cooled down and kept sealed.
The composite precursors (0.5 g) were dissolved in hexane (1.89 mL, 40 wt%), then added into a flask charged with absolute ethanol (50 mL) and concentrated ammonia solution (0.782 mL, 28 wt%), and the reaction was allowed to proceed at 60 °C for 24 h under continuously vigrous stirring. The product was centrifuged in 5000 r min^−1^ and washed with ethanol for three times, redispersed in deionized water for use.
*Fabrication of Hierarchical TiO~2~ Hollow Spheres*: The dispersion of the above abtained composite paticles (10 mL) was mixed uniformly with NaOH solution (10 mL, 2 [m]{.smallcaps}), and then added into a Teflon‐lined stainless‐steel autoclave (30 mL in capacity), and heated to 150 °C for 2 h, and then allowed to cool down to room temperature naturally. The product was washed with HCl aqueous solution (0.1 [m]{.smallcaps}) for three times until pH value being close to 7, and immersed in HCl for 2 h, then washed with deionized water and ethanol for several times, and dried at 60 °C in vacuum over night. The obtained products were finally annealed at 350 °C for 6 h at a heating rate of 2 °C min^−1^.
*Characterization*: The surface morphology of the as‐prepared samples was examined by field‐emission scanning electron microscopy (FESEM; Zeiss, Ultra 55, 3 kV), and their microstructures was observed by TEM (FEI, Tecnai G2 20 TWIN, 200 kV) and HRTEM (JEOL JEM‐2010). Crytallographic information was collected using powder XRD (Bruker, D8 Advance X‐ray diffractometer, Cu ΚR radiation (λ = 1.5406 Å)). The Raman spectra were measured on a micro‐Raman spectroscope (Horiba Jobin Yvon, XploRA, the excitation wavelength of 532 nm). The elemental compositions of the samples were analyzed by EDS attached to the FESEM instrument and inductively coupled plasma‐atomic emission spectrometry (ICP‐AES; Hitachi, P4010). The surface area of the samples was measured using a Quantachrome Instruments Quadrasorb evo instrument. Thermogravimetric analysis (TGA) was carried out under a flow of air with a temperature ramp of 10 °C min^−1^.
*Electrochemical Properties*: The electrodes for half‐cell tests were prepared by homogeneously mixing the active materials (e.g., the UNTHS, 70 wt%) with super‐P carbon (20 wt%) and polyvinylidene fluoride (10 wt%) in N‐methyl‐2‐pyrrolidone (NMP). And the mass loading of the electrodes was carefully controlled within the range of 0.4--1.0 mg cm^2^. The specific capacity was caculated based on the mass of TiO~2~ within the UNTHS. The resulting slurries were applied to the Cu foil. After heated at 120 °C for 12 h, the sheet was punched into 15 mm diameter electrodes. The cells were assembled in an argon‐filled glove box with the concentrations of moisture and oxygen below 1 ppm. For half‐cell tests, the anode was tested using 2032‐type coin cells, where Li metal was used as as both the counter and reference electrode, and 1.0 M LiPF~6~ in ethylene carbonate/dimethyl carbonate (EC/DMC, 1:1 volume ratio). The cyclic voltammetry and galvanostatic charge--discharge measurements were conducted on a CHI 660B electrochemical workstation (Shanghai CH Instrument Company, China). Cycle‐life measurements for the half‐cell were carried out on a battery test system (Land CT2001A model, Wuhan Land Electronics. Ltd.).
Supporting information
======================
As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.
######
Supplementary
######
Click here for additional data file.
Financial supports of this research from the National Natural Science Foundation of China (51133001, 21374018 and 51372040), Science and Technology Foundation of Ministry of Education of China (20110071130002), the Shanghai Rising‐Star Program (16QA1400700), Science and Technology Foundation of Shanghai (13JC1407800) are appreciated.
| {
"pile_set_name": "PubMed Central"
} |
Background {#Sec1}
==========
When an implant is inserted into bone the healing process starts by creating new bone that interlocks with the implant surface, so called osseointegration. It has been found that an implant with a rough surface topography enhances osseointegration and load bearing capacity \[[@CR1], [@CR2]\]. A significant amount of research has been undertaken in order to identify the biomechanical response to a structured surface, which is characterized by a set of statistical surface roughness parameters \[[@CR3]--[@CR7]\]. The general result of *in vivo* experiments is that increased implant surface roughness (S~a~ value) of cylindrical implants results in increased interfacial shear strength \[[@CR8]\]. Whether this empirical correlation is an effect of enhanced mechanical strength of the bone caused by the biological response to the surface and/or enhanced interlocking capacity is unclear. In addition, it is unclear whether surface roughness parameters truly reflect the load bearing capacity \[[@CR9]\]. Hansson and Norton \[[@CR10]\] developed artificial surfaces characterized by various pit size, pit shape and pit density which were used to simulate the theoretical interfacial load bearing capacity. They found that the topographical structure affects the interfacial load bearing capacity \[[@CR10]\]. However, the theoretical surfaces considered by Hansson and Hansson \[[@CR9]\] do not represent the surface topography of implants. Besides the implant surface topographical structure, the interfacial load bearing capacity depends on the mechanical properties of the bone during healing. In brief, bone formation in a gap proceeds, regardless of the presence of an implant, in two steps: (1) woven bone formation (2) remodeling of woven bone to lamellar/Harversian bone \[[@CR11], [@CR12]\]. Woven bone is characterized by a random arrangement of collagen fibers with poor mechanical properties \[[@CR13]\]. Woven bone is thereafter remodeled to Haversian bone over time by basic multicellular units (BMU) \[[@CR14]\]. The mineralization process of a new Haversian system can be divided into two stages; primary mineralization and secondary mineralization \[[@CR15]\]. Primary mineralization is characterized by a rapid, constant rate of mineralization that proceeds until 50--60% of the mineralization maximum has been reached \[[@CR15], [@CR16]\]. Following primary mineralization, a decreased rate of mineralization progressively continues and typically stabilizes at around 90--95% of the maximum level \[[@CR15], [@CR16]\]. The current knowledge of the mechanical properties of bone during healing is scarce. However, it is known that the degree of mineralization greatly affects the mechanical properties of bone \[[@CR17]--[@CR20]\]. It is evident that newly formed bone has different mechanical properties compared to mature bone due to different degrees of mineralization \[[@CR21], [@CR22]\]. Nano indentation has been used to investigate the mechanical properties of newly formed bone. Leong and Morgan \[[@CR13]\] measured the mechanical properties of fractured rat bones after 24 days of healing and found that Young's modulus of woven and cortical bone was 36.2 MPa and 7.2 GPa respectively. In a study by Ishimoto et al. \[[@CR23]\] the mechanical properties of regenerated (2 weeks) and mature rabbit bone were measured. They obtained an elastic modulus of 17.0 and 27.6 GPa respectively which is higher than the maximum Young's modulus for mature rabbit bone (\~8 GPa) found by Isaksson et al. \[[@CR21]\]. Few studies have measured the mechanical properties of bone in the vicinity of implants. Chang et al. \[[@CR24]\] obtained a value of Young's modulus of pig alveolar bone close to an implant surface, after one month of healing, of 6.17 GPa that gradually increased to 7.9 GPa at a distance of 150 µm from the implant surface. Finally, it reached 10.1 GPa at a distance of 1,500 µm from the implant surface. This is slightly lower than what Vayron et al. \[[@CR25]\] found in rabbit bone after longer healing times (7 and 13 weeks). An alternative way to determine the mechanical properties during healing was used in a study by Warzen et al. \[[@CR26]\]. They measured the stiffness of the implant bone interface *in vivo* of mice hind legs after 0--6 days of healing and by reverse engineering, obtained values of Young's modulus ranging from 3 to 17 MPa. The mechanical properties of the bone and the surface topography affect the bone implant interfacial shear strength \[[@CR10]\]. The objective of the present study was to estimate the theoretical bone-to-implant interfacial shear strength for different surface topographies during healing by means of finite element analysis (FEA).
Methods {#Sec2}
=======
Geometrical representation of a surface topography {#Sec3}
--------------------------------------------------
The coordinates (x, y and z) from a representative patch of each of the three different (A, B and C) surfaces used in the study by Halldin et al. \[[@CR27]\] were extracted from interferometry measurements. Group A were blasted with coarse titanium particles thereafter treated in oxalic acid and hydrofluoric acid sequentially (CB-AT-I). Group B were blasted with fine particles titanium oxide particles and thereafter treated in oxalic acid and hydrofluoric acid sequentially (FB-AT-I). Group C implants were blasted with coarse titanium particles thereafter treated in hydrofluoric acid (CB-HF). The representative patches were selected to have surface average height values (S~a~) close to the mean S~a~ values of the implant in Halldin et al. \[[@CR27]\]. The x, y and z coordinates were imported into MATLAB (MATLAB 2013b, Mathworks Inc., USA) to calculate the average height values (S~a~ \[μm\]), root-mean-square (S~q~ \[μm\]), skewness (S~sk~ \[μm/μm\]) and the root-mean-square of the surface slope (S~dq~ \[μm/μm\]) of the selected patches \[[@CR8]\]. Thereafter, the surfaces were exported to STL file format followed by conversion with high accuracy, to IGES format using Geomagic studio® (Geomagic Solutions, Cary, NC, USA). The IGES surfaces were imported into Pro Engineer (PTC, Needham, MA, USA) for creation of a 3D geometry (Figure [1](#Fig1){ref-type="fig"}). Finally the geometries were imported to ANSYS® 14.5 (ANSYS INC, Canonsburg, PA, USA) for simulation of the interfacial load bearing capacity.Figure 1Surface geometry. Geometrical representation of the selected patch of the implant surfaces (CB-AT-1, FB-AT-1 and CB-HF) derived from interferometry measurements in the study by Halldin et al. \[[@CR27]\]. Surface characteristics of the patches are presented in Table [3](#Tab3){ref-type="table"}.
Finite element model {#Sec4}
--------------------
The bone-to-implant interface was modeled according to Figure [2](#Fig2){ref-type="fig"}. The implant surface was in contact with interfacial bone which in turn was adjacent to the fictitious bone. The fictitious bone represents the structural stiffness of the surrounding bone and was modeled as an isotropic material with a thickness (*l*~0~) of 10 µm and a Young's modulus represented by E~s~. Substituting the surrounding bone by a fictitious bone results in a less computationally heavy model. The implant was locked in y direction and was moved in x direction by a constant acceleration (a~x~) during the simulation time. The interface between interfacial bone and fictitious bone was fixed in x direction and free to move in y direction. The outer sides of the X--Y plane of the 3D model were restricted to move in z direction. The implant was modeled as a rigid component and the interface bone-to-implant interface was modeled as a frictional contact with an assumed coefficient of friction of 0.4 \[[@CR28]\].Figure 2Illustration of the 3D FEA model with boundary conditions.
Mechanical properties {#Sec5}
---------------------
The Young's modulus (E~s~) of the fictitious bone was determined by simulating the deformation (*d*) at a predefined internal pressure (*P*) of 100 MPa in a 2D plane strain model representing a disc of mature bone. The disc had an inner diameter of 3.5 mm similar to the outer diameter of the implant in the study by Halldin et al. \[[@CR27]\]. The structural stiffness was computed for different outer diameters (D~o~) of the disc ranging from 2.5 to 12.5 mm. The fictitious Young's modulus E~s~, for each outer diameter (D~o~) with corresponding deformation (d), was calculated using Eq. ([1](#Equ1){ref-type=""}).$$\documentclass[12pt]{minimal}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$$E_{s} = \frac{{P \cdot l_{0} }}{d}$$\end{document}$$Since the current knowledge of the mechanical properties of bone in the vicinity of an implant during healing is limited, the assumed mechanical properties of bone during healing were derived from the correlation between mechanical properties and mineral content \[[@CR17]\] and from the correlation between mineral content and healing time \[[@CR15]\]. Currey \[[@CR17]\] investigated the mineral content, expressed as milligrams of calcium (Ca) per gram dried defatted bone, for various species with corresponding Young's modulus (E), yield strain (ε~y~) and post yield strain (ε~py~). Using Currey´s data the correlation between Ca and E, ε~y~ and ε~py~ can be identified (Figure [3](#Fig3){ref-type="fig"}a). Furthermore, by assuming 100% mineralization after 350 days of healing \[[@CR15]\], and a corresponding Young's modulus of 7,950 MPa \[[@CR21]\], the Young's modulus (E), yield strain (ε~y~) and ultimate strain (ε~u~ = ε~y~ + ε~py~) at different healing times can be approximated (Figure [3](#Fig3){ref-type="fig"}b). In the present study it was assumed that the implant interfacial bone exhibits a bilinear isotropic material behavior without hardening. Thus the stress strain curves and the mechanical properties at different healing times can be estimated (Figure [4](#Fig4){ref-type="fig"}; Table [1](#Tab1){ref-type="table"}). The strength and Young's modulus of titanium are an order of magnitude higher than those of the bone. Therefore, to reduce computing time the implant was assumed to be rigid. In the present study the shear strength of a bone-to-implant interface was simulated with explicit finite element analysis (Figure [2](#Fig2){ref-type="fig"}) using the ultimate strain as a failure criterion. According to the simulations the maximum shear force occurred at implant displacements of less than 0.75 µm. Therefore, the simulations were discontinued after an implant displacement of 0.75 µm was reached. The displacement (0.75 µm) was achieved by a constant acceleration of the implant during the simulation time. Using the representative patch area the maximum reaction force between implant interfacial bone and the fictitious bone was converted, to interfacial shear strength (τ~p~). In explicit FE simulation the computing time depends on the choice of simulation time, minimum element size and material density \[[@CR29]\]. Reduced computing time is obtained by decreased simulation time (increased acceleration), increased material density and increased minimum element size \[[@CR29]\]. Prior to the 3D simulation a parameter study (convergence test) with several combinations of different values of the acceleration (P1), material density increase (P2) and minimum element size (P3) (Table [2](#Tab2){ref-type="table"}) was performed to obtain reasonable computing time without compromising the accuracy. The parameter study was performed by simulating the reaction force on a 2D profile. The 2D profile was extracted from an x--y plane of the 3D model (Figure [2](#Fig2){ref-type="fig"}). The meshes were auto generated in ANSYS® with different values of the minimum element size at the bone-to-implant interface. Otherwise the simulations were performed with the same settings. The simulated shear strength for the three surfaces were converted to removal torque value (RTQ) of an implant with the same design as in the Halldin et al. \[[@CR27]\] study according to Eq. ([2](#Equ2){ref-type=""}).$$\documentclass[12pt]{minimal}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$$RTQ = \tau_{p} \cdot A_{i} \cdot r \cdot BIC$$\end{document}$$where *A*~*i*~ represents the outer threaded area of the implant which was 78 mm^2^, *r* represents the radius to the implant surface, which was 1.75 mm, and *BIC* represents the bone to implant contact obtained in the *in vivo* study \[[@CR27]\], which was set to 40% for all three surfaces.Figure 3Estimation of mechanical properties during healing: **a** Derived functions (*solid line*) for correlation between the mineral content and the mechanical properties derived from the data (o) published by Currey \[[@CR17]\]. **b** Mechanical properties during healing derived from the data presented by Currey \[[@CR17]\] and Fuchs et al. \[[@CR15]\].Figure 4Stress--strain curves at different healing times.Table 1Mechanical properties during healingHealing time (weeks)E (GPa)σ~y~ (MPa)ε~u~ (%)12,20029.311.2243,16437.69.2124,00944.38.0507,95071.35.1The properties were derived from the data by Currey \[[@CR17]\] and Fuchs et al. \[[@CR15]\].Table 2Convergence analysisP1P2P3Acceleration m/s^2^Material density increase factorMinimum element size (µm)610.566.71000.6256001,0000.7560,00010,0000.8751,000,0001.0Values used in the 2D analysis to identify appropriate parameter settings to obtain reasonable computing times for the following 3D simulations.
Results {#Sec6}
=======
The surface roughness (S~a~ value) of the selected representative patch (Figure [1](#Fig1){ref-type="fig"}) and the S~a~ value with the standard deviation (SD) of the implant in the study by Halldin et al. \[[@CR27]\] are presented in Table [3](#Tab3){ref-type="table"}. The fictitious Young's modulus E~s~ of the bone surrounding a 3.5 mm (D~i~) implant with different values of the outer diameter (D~0~) is presented in Figure [5](#Fig5){ref-type="fig"}. The fictitious Young's modulus was chosen to 33 MPa. A reasonable computing time was obtained with an acceleration of 600 m/s^2^ (resulting in a simulation time of 5 × 10^−5^ s), material density increase by a factor of 1,000 and a minimum element size of 1 µm. The results of the 2D analysis indicate that these settings do not seem to significantly affect the reaction force compared to simulation results using lower parameters values (P1, P2 and P3) (Figure [6](#Fig6){ref-type="fig"}). Hence these settings were used in the 3D simulations which resulted in a computing time of 24 h for each simulation. The 3D models were meshed with ANSYS^®^ default settings and a minimum element size of 1 µm resulting in meshes according to Table [4](#Tab4){ref-type="table"} and Figure [7](#Fig7){ref-type="fig"}. The interfacial shear strength of the three surfaces is presented in Figure [8](#Fig8){ref-type="fig"}a. The theoretical removal torques, according to Eq. ([1](#Equ1){ref-type=""}), and the corresponding *in vivo* removal torques of Halldin et al. \[[@CR27]\] are presented in Figure [8](#Fig8){ref-type="fig"}b. The ratio (*test/control*) of the simulated shear strength and the mean of the corresponding ratios of the *in vivo* study by Halldin et al. \[[@CR27]\] for the surfaces are presented in Figure [8](#Fig8){ref-type="fig"}c.Table 3Surface characterizationParameterCB-HFCB-AT-1FB-AT-1ImplantS~a~1.52 (SD 0.23)1.49 (SD) 0.230.95 (SD 0.19)Representative patchS~a~ \[μm\]1.481.630.91S~q~ \[μm\]1.932.081.21Ssk \[μm/μm\]0.18−0.07−0.44Sdq \[μm/μm\]1.001.020.69The surface roughness (S~a~ value) of the selected representative patch and the S~a~ value of the implant in the study by Halldin et al. \[[@CR27]\].Figure 5Calculation of fictitious Young's modulus. Young's modulus of mature bone surrounding a 3.5 mm implant with different values of the outer diameter (D~0~).Figure 6Convergence test analysis. The results of the 2D simulation indicate that a parameter setting of 600 m/s^2^, material density increase of 1,000 and minimum element size of 1 µm (*green circle*) do not significantly affect the reaction force compared to lower parameter values.Table 4Mesh descriptionSurfaceNumber of nodesNumber of elementsElement typeCB-HF378,1312,025,934Four-node tetrahedral elementCB-AT-1356,0301,888,789Four-node tetrahedral elementFB-AT-1314,1641,662,405Four-node tetrahedral elementResults of the mesh when using ANSYS® predefined meshing algorithm with a minimum element size of 1 µm and a defeaturing tolerance of 0.75 µm.Figure 7Illustration of mesh. Mesh used in the present simulation with a minimum element size of 1 µm.Figure 8Load bearing capacity. **a** Maximum interfacial shear strength of the three surfaces simulated in the 3D FEA model. **b** Theoretical removal torque calculated according to Eq. ([2](#Equ2){ref-type=""}) compared to the mean removal torque value found by Halldin et al. \[[@CR27]\]. **c** The ratio (test/control) of the simulated interfacial shear strength and the mean of the corresponding removal torque ratios of the *in vivo* study by Halldin et al. \[[@CR27]\].
Discussion {#Sec7}
==========
In the present study the theoretical interlocking capacity of three different implant surfaces with bone after different healing times was evaluated. The simulated ratios (test/control) seem to agree with those of the Halldin et al. \[[@CR27]\] study for longer healing times (Figure [8](#Fig8){ref-type="fig"}). However, the simulated absolute removal torque is underestimated compared to the biological removal torque as found in Halldin et al. \[[@CR27]\] (Figure [8](#Fig8){ref-type="fig"}b). There are many factors that may influence the simulated results: (1) model setup (2) representative mechanical properties of the interfacial bone during healing (3) selection of the representative patch (4) pressure at the bone-to-implant interface (5) bone-to-implant contact. In simulation it is essential to use a model that captures the essence of what is intended to be analyzed. However, a model will always be a simplification of reality. In this study a simplified model of the interface was developed to analyze the interlocking capacity of a rough surface with bone.
In the present 3D-model the circumferential structural displacement was neglected. The circumferential structural displacement would affect the angle of rotation but is not assumed to affect the interfacial shear strength. This simplification reduces the size of the model and thus the computing time. The 2D analyses were performed to investigate how the reaction force is affected by different simulation settings. A reduced element size was assumed to provide a higher accuracy but would have implied a longer computing time \[[@CR29]\]. The results of the 2D analysis showed that a minimum element size of 1 µm could be used without negatively affecting the accuracy of the results. Furthermore to reduce computing time the simulation time was decreased by the use of increased acceleration. An increased simulated acceleration might affect the interfacial shear force. To reduce the inertia effects the simulation was performed with a constant acceleration of 600 m/s^2^, which ramped up the velocity from 0 to 30 mm/s during 5 × 10^−5^ s. To reduce the computing time even further the material density was increased in turn affecting the inertia and consequently the results. In the 2D analysis it was found that the material density could be increased by a factor of 1,000 without substantially affecting the interfacial shear strength.
The magnitude of the interfacial shear strength is affected by the mechanical properties of interfacial bone as well as the interface properties (i.e. friction) To the authors' knowledge the coefficient of friction on micro level is unknown and the assumed coefficient of friction was based on a coefficient of friction between bone and titanium \[[@CR28]\]. The coefficient of friction is assumed to affect the magnitude of the interfacial shear strength to the same extent between the simulations.
The knowledge about the mechanical properties of bone, including failure behavior, in the vicinity of an implant surface during healing is limited. Therefore, an estimation of the mechanical properties of bone was derived from the mineral content measurements of osteoid under mineralization in an osteon during healing by Fuchs \[[@CR15]\] and the relationship between mineral content and material properties found by Currey \[[@CR17]\]. Furthermore, it was assumed that the mineral content was 100% after 350 days of healing which corresponds to a Young's modulus of 7.95 GPa \[[@CR21]\]. To estimate the degree of mineralization during the first weeks of healing, extrapolation was made of the experimental data obtained by Currey \[[@CR17]\]. Thus in the present study, Young's modulus during healing was set in the range of 2--10 GPa. Other studies have found a variation in Young's modulus during healing which could be caused by differences in experimental setup and choice of animals used \[[@CR13], [@CR23]--[@CR25], [@CR30]\]. The assumed mechanical properties used in this study might not reflect the true material properties during healing which in turn affect the absolute value of the removal torque. However, when the relative interfacial load bearing capacity of two surfaces is considered the actual material property during healing is assumed to have limited impact. Furthermore, the same bone material properties were used for all surfaces at coinciding healing times and a potential biological change in the mechanical properties caused by a specific surface topography was not considered. This might be an explanation of the consistency in ratios for longer healing times and the deviations during the early healing phase. This indicates that, for longer healing times the surface topography might be the main parameter influencing the interlocking capacity of an osseointegrated implant surface.
In this simulation the ultimate strain was used as a failure criterion based on available ultimate strain data. Principal strain failure can be used to represent brittle or ductile failure in materials \[[@CR29]\]. Other material failure criteria, such as Mohr--Coulomb, Tsai--Wu and Hill, have been discussed for bone \[[@CR31]\]. However to our knowledge the anisotropy and a suitable yield and failure criterion for the bone material during healing in the vicinity of the implant surface are unknown. When the failure criterion was reached in an element the element was removed and did no longer contribute to the interfacial shear strength. This might result in an underestimation of the simulated interfacial shear strength value. The magnitude of the reaction force depends on the surface topography of the selected patch. The standard deviation of the S~a~ value (Table [3](#Tab3){ref-type="table"}) of the implants in the Halldin et al. \[[@CR27]\] study, from which the representative patches were selected, indicates a variation. Even though the S~a~ value of the representative patch was selected to be in the range of the S~a~ value of the implant, the selected patch might not fully represent the surface topography with respect to load bearing capacity. When an implant is inserted it induces static strains in the bone that gradually decrease during healing \[[@CR32], [@CR33]\]. These initial residual stresses increase the removal torque value during the early healing period. The current simulations did not include the change of residual stresses during healing time. Therefore, the absolute removal torque value might be underestimated during the early healing phase. During osseintegration the biological processes start to integrate the implant with bone and thus increase the bone-to-implant contact and the removal torque over time. In this simulation, 40% bone-to-implant contact was assumed, for coinciding healing times, which might lead to an over- or underestimated simulated absolute removal torque, but is assumed to have a limited impact when analyzing the ratios.
Conclusion {#Sec8}
==========
In this simulation the theoretical interfacial shear strength of the three surface topographies was simulated and compared to the results in the study by Halldin et al. \[[@CR27]\]. The simulated ratios (test/control) seem to agree with those of the Halldin et al. \[[@CR27]\] for longer healing times (Figure [8](#Fig8){ref-type="fig"}). Despite differences in the absolute removal torque values between simulations and the *in vivo* data, finite element analysis is a promising method, that can be used prior to *in vivo* study, to compare the theoretical load bearing capacities of the bone-to-implant interfaces especially for longer healing times. It was also concluded that the mechanical properties and the material model of bone in the vicinity of an implant affect the theoretical load bearing capacity and requires more research.
BIC
: bone to implant contact
RTQ
: removal torque.
AH developed appropriate material model performed finite element simulations and drafted the article. MA participated in the design and development of the model and drafting the article. SH performed the research of the mechanical behavior of bone and developed appropriate material model and drafting the article. MJ contributed to develop biological interfacial model and drafting the article and contributed for intellectual content. All authors read and approved the final manuscript.
Acknowledgements {#d30e1336}
================
This work was supported by the Swedish Research Council (621-2010-4760). We would like to thank Klas Johansson and Mikael Lauth at EDR&Medeso for the valuable support during simulation.
Compliance with ethical guidelines {#d30e1341}
==================================
**Competing interests** The authors declare that they have no competing interests.
| {
"pile_set_name": "PubMed Central"
} |
Introduction
============
Parkinson's disease (PD) is the second most common neurodegenerative disease that affects almost 1% of individuals over 60 years of age. First described in 1920,^[@B1]^ the ketogenic diet (KD) comprises fat (80-90%), carbohydrate, and protein and metabolically induces a fasting-like condition.^[@B2],[@B3]^ Some studies have shown the usefulness of this diet in PD.^[@B2],[@B4],[@B5]^
Studies on experimental models of Modeling Parkinson's disease in primates (MPTP)-induced Parkinsonism have shown that a restriction in glucose intake bolsters resistance of the cells located in the substantia nigra against neurotoxic effects of MPTP and prevents the progression of symptoms associated with PD 2.^[@B5]^
Another study, by Vanitallie et al.^[@B6]^ on PD patients, shown that the KD for almost 1 month could significantly lessen symptoms and consequently the unified PD rating scale score. Although there is not a consensus on the mechanism underlying the effect of the KD on cerebral pathologies, it seems that the efficacy of ketone bodies in this regard stem from an enhancement of mitochondrial functions and a decrease in oxidative stress^[@B7],[@B8]^ the prevention of the excitotoxicity due to a neurotransmission escalation of excitatory amino acids,^[@B8]^ and fending off inflammatory processes and apoptosis.^[@B9]^
Pramipexole is a non-ergo dopamine receptor agonist with high affinity toward D2 and D3 dopamine receptors, which has been used for symptomatic treatment of PD in recent years. Noting the mechanism of PD in which there is a decrease in cerebral dopamine levels after degradation of neurons in the substantia nigra, this medication can ameliorate PD motor symptoms through exerting dopamine agonist effects and binding to its receptor.^[@B10]^
Because available treatments in PD are usually along with diminished symptoms without affecting progression of the disease and at the same time the potential of causing motor fluctuations, it is essential to use new therapeutic strategies in this regard.^[@B1]^
Because there is no available study regarding the effects of the KD on motor symptoms in PD using rat model, this study seeks to examine the effect of this regimen on motor symptoms in rats with 6-hydroxydopamine (6-OHDA)-induced PD and to evaluate the efficacy of this regimen alone or in combination with pramipexole.
Materials and Methods
=====================
In this experimental study, a total of 56 Wistar rats weighing 200-240 g and aged 12-14 weeks were randomized in seven 8-rat groups including controls on a regular diet, sham surgical group, controls on the KD, rats with 6-OHDA-induced PD on a regular diet (negative controls), rats with 6-OHDA-induced PD on the KD for 25 days, rats with 6-OHDA-induced PD on a regular diet and pramipexole for 14 days, and rats with 6-OHDA-induced PD on the KD combined with pramipexole for 14 days.
This study was performed at Tabriz Neuroscience Laboratory (Iran) from July 2014 to September 2015. The exclusion criteria were an occurrence of any disease during the study period; the expiration of the rats due to intracerebral injections, and no documentation of ketonemia in rats after being on consecutive days of the KD.
This research was conducted in accordance with the latest ethical regulations for laboratory animal studies issued by Tabriz University of Medical Sciences in terms of providing an appropriate environment for animals, free access to water, and using painless stereotaxic injections.
Rats were first underwent a period of training for a standard bar test, beam traversal task test, and cylinder task test and then randomized in the aforementioned groups. To induce ketonemia, a KD containing medium chain triglycerides (MCTs) oils accounting for a total of 50% of the required calories plus a regular diet for the remaining calories was administered. The MCTs oil was administered orally through standard gavage feeding tubes.
To ensure induced ketonemia, serum levels of beta-hydroxybutyrate (BHB), as an index for the production of ketone bodies, were measured in both groups including rats on the KD and regular diets. In case of documenting a statistically significant difference between the two groups, an induced ketonemia was confirmed in rats on the KD.^[@B11],[@B12]^
In the present experiment, serum levels of ketone bodies were measured using biochemical kits of BHB at baseline and on the day 14 after being on the KD.
To induce experimental Parkinsonism, 11 days after the commencement of the KD and regular diets an intranigral injection of 6-OHDA (12 µg in 2 µl normal saline and 0.2% ascorbic acid) was carried out. 30 minutes before, this injection desipramine (25 mg/kg) was injected into the peritoneal cavity to preclude a possible reabsorption of 6-OHDA back into the noradrenergic neurons and subsequent injuries. In the sham surgical group, 6-OHDA was replaced with normal saline.
Finally, 14 days after induction of Parkinsonism the rats underwent a standard bar test, beam traversal task test, and cylinder task test.
***Bar test***
As illustrated in [figure 1](#F1){ref-type="fig"} to perform a standard bar test, the forearms of the animal were placed on a bar (9 mm in diameter) fixed at the height of 9 cm away from the platform of the testing device. The duration of maintaining this position (catalepsy time) was documented. In case of any exploratory head movements or displacement of one or both forearms, the test was considered terminated.
***Beam traversal task test***
As illustrated in [figure 2](#F2){ref-type="fig"}, [5](#F5){ref-type="fig"} cm wide and 1 m long wooden bridge fixed 50 cm above the ground was used. The animal was placed on one end of the bridge and released. The time needed for total crossing the bridge was documented.
***Cylinder task test***
As illustrated in [figure 3](#F3){ref-type="fig"}, an animal was placed inside a seven-through glass cylinder, and the total number of forearm contacts with the wall was documented within a 10-minute period. Accordingly, the final score was calculated using the following formula.^[@B13]^
{#F1}
{#F2}
{#F3}
Score = 100 × (Number of forearm contacts on the lesion side + 1/2 of total forearm contacts)/total forearm contacts.
The KD was started from the 1^st^ day of experiment and pramipexole (0.3 mg/kg once a day) was administered on the day 12 for 14 consecutive days in relevant groups.
The data were presented as mean ± standard error (SE) of the mean. The SPSS software (version 16, SPSS Inc., Chicago, IL, USA) was used. The one-way ANOVA test coupled with the Tukey post-hoc analysis was employed for comparisons. Levels of serum ketone bodies were compared between groups using the Wilcoxon test. A P \< 0.050 was considered statistically significant.
Results
=======
The mean level of serum ketone bodies in the group on the KD was 0.25 ± 0.03 mmol/L (range: 0.21-0.28) at baseline and 1.83 ± 0.17 mmol/L (range: 1.60-2.10) on the day 14. According to the results of Wilcoxon test, the mean serum level of ketone bodies increased significantly on the day 14 compared to that at baseline (P = 0.010).
***Results of the bar test (catalepsy time)***
Results of the bar test in different groups were as follows ([Figure 4](#F4){ref-type="fig"}): The control group on a regular diet: 15.00 ± 1.18 seconds (range: 10-20); the surgical sham group: 18.75 ± 1.15 seconds (range: 15-25); the control group on the KD: 15.88 ± 1.30 seconds (range: 10-21); the PD group on a regular diet: 1 03.13 ± 2.65 seconds (range: 95-115); the PD group on the KD: 80.63 ± 1.21 seconds (range: 75-85); the PD group on a diet with pramipexole: 73.63 ± 1.87 seconds (range: 67-80); and the PD group on the KD with pramipexole: 72.50 ± 2.17 seconds (range: 65-80).
Comparing the controls on a regular diet with the sham surgical group showed no significant difference in terms of the mean catalepsy time (P = 0.730). A similar finding was found in the comparison between the controls on a regular diet and the controls on the KD (P = 0.990). Comparing unaffected groups with PD groups revealed that the mean catalepsy time was significantly shorter in the former (P \< 0.001 for all paired comparisons). Similarly, the mean catalepsy time was significantly longer in the PD group on the KD compared to that in the remaining three PD groups (P \< 0.001 for all paired comparisons). Comparing the PD group on the KD with the PD group on a regular diet with pramipexole did not show a statistically significant difference between the two groups in terms of the mean catalepsy time (P = 0.090). The PD group on the KD, however, had significantly longer mean catalepsy time then the PD group on the KD with pramipexole (P = 0.030). Finally, comparing the PD group on a regular diet with pramipexole with the PD group on the KD with pramipexole did not show a statistically significant difference between the two groups in terms of the mean catalepsy time (P = 0.990).
{#F4}
***Results of the beam traversal task test***
Results of the beam traversal task test in different groups were as follows ([Figure 5](#F5){ref-type="fig"}): The control group on a regular diet: 3.61 ± 0.17 seconds (range: 3-4); the sham surgical group: 3.76 ± 0.13 seconds (range: 3-4); the control group on the KD: 3.64 ± 0.12 seconds (range: 3-4); the PD group on a regular diet: 6.25 ± 0.26 seconds (range: 5-8); the PD group on the KD: 5.41 ± 0.32 seconds (range: 4-7); the PD group on a diet with pramipexole: 5.08 ± 0.10 seconds (range: 5-6); and the PD group on the KD with pramipexole: 4.83 ± 0.06 seconds (range: 4.5-5).
There was not a significant difference between the controls on a regular diet and the sham surgical group in terms of the mean results of the beam traversal task test (P = 0.990); nor in the comparison between the controls on a regular diet and the controls on the KD (P = 0.990). Comparing unaffected and PD groups, however, showed that the mean results of the beam traversal task test were significantly lower in the former (P \< 0.001 for all paired comparisons). Comparing the PD group on a regular diet with the three remaining PD groups showed that the mean results of the beam traversal task test were significantly higher in the former (P = 0.040 in comparison with the PD group on the KD and P \< 0.001 for the other two comparisons). Comparing the PD group on the KD and the PD group on a regular diet with pramipexole showed no significant difference between the two groups as to the results of the beam traversal task test (P = 0.860). A similar finding was found in comparisons between the PD group on the KD and the PD group on the KD with pramipexole (P = 0.300), and between the PD group on a regular diet with pramipexole and the PD group on the KD with pramipexole (P = 0.960).
{#F5}
{#F6}
***Results of the cylinder task test***
Results of the cylinder task test in different groups are as follows ([Figure 6](#F6){ref-type="fig"}): the control group on a regular diet: 48.38 ± 1.50% (range: 42-55); the sham surgical group: 50.38 ± 1.48% (range: 45-56); the control group on the KD: 49.88 ± 1.68% (range: 42-56); the PD group on a regular diet: 71.63 ± 1.86% (range: 62-80); the PD group on the KD: 64.13 ± 1.53% (range: 58-70); the PD group on a diet with pramipexole: 60.63 ± 1.10% (range: 57-65); and the PD group on the KD with pramipexole: 58.00 ± 1.48% (range: 52-65). Comparing the controls on a regular diet and the sham surgical group showed no significant difference between the two groups in terms of the mean results of the cylinder task test (P = 0.970). Similarly, no significant difference was found in the comparison between the controls on a regular diet and the controls on the KD (P = 0.990). Comparing the unaffected and PD groups showed significantly lower mean results of the cylinder task test in the former (P \< 0.001 for all paired comparisons). In addition, comparing the PD group on a regular diet with the remaining three PD groups showed significantly lower mean results of the cylinder task test in the former (P = 0.020 in comparison with the PD group on the KD and P \< 0.001 for the other two comparisons). There was no significant difference between the PD group on the KD and the PD group on a regular diet and pramipexole in terms of the results of the cylinder task test (P = 0.670). There was also no significant difference between the PD group on the KD and the PD group on the KD and pramipexole in this regard (P = 0.090). Finally, there was no significant difference between the PD group on a regular diet with pramipexole and the PD group on the KD with pramipexole regarding the results of the cylinder task test (P = 0.890).
Discussion
==========
Nutritional and metabolic therapeutic strategies have been tried in a wide range of neurologic diseases such as epilepsy, headache, neurotrauma, Alzheimer's disease, sleep disorders, brain cancer, autism, pain, multiple sclerosis (MS), and PD. The related incentive possibly originates from the ineffectiveness of available pharmacologic treatments in many of such diseases, as well as a general inclination toward more natural products.^[@B14]^
The present experimental study on rats showed that the KD is significantly effective in promoting motor functions in PD animals and at the same time, it may enhance the therapeutic effects of concomitantly administered pramipexole, albeit in an insignificant fashion.
In accordance with these findings, using a standard scale, Vanitallie et al.^[@B6]^ showed that the KD was able to exert beneficial effects in 5 out of 7 PD patients.
One of the major limitations in the current study was using a small sample size that was unable to exclude the placebo effect. In a study by Tieu et al.,^[@B5]^ the injection of BHB acid also effectively prevented MTPT-induced neurodegenerative and aging processes in dopaminergic cells located in rat brain.
In another animal model used by Cheng et al.^[@B15]^ the employment of ketone bodies protected dopaminergic neurons located in the substantia nigra against 6-OHDA-induced neurotoxicity. In a study by Yang and Cheng,^[@B16]^ the anti-inflammatory effects of ketone bodies were confirmed using a model of MPTP-induced neurotoxicity.
In a similar experiment by Beckett et al.^[@B17]^ using a rat model of Alzheimer's disease, in conformity with our findings the KD managed to significantly improve some aspects of motor functions. In this study, all aspects of motor functions of the examined animals were not influenced equally. Although the underlying cause of this finding is not clear, it seems that various muscular groups are affected unequally by the KD (for example the quadriceps muscles vs. the paw flexors).
In another series by Brownlow et al.,^[@B18]^ there was also a considerable improvement in motor functions in Alzheimeric rats using the KD. Ruskin et al.^[@B19]^ also showed that the administration of the KD could increase motor capabilities of animals with induced Huntington's disease (HD).
Although the exact neuroprotective property of the KD is not known, some hypotheses have been suggested. The two major aspects in treating with the KD is an increase in the production rate of ketone bodies in the liver and a decrease in serum glucose levels. An increased level of ketone bodies is assumed to be related to fatty acid oxidation. Certain polyunsaturated fatty acids such as arachidonic acid (AA), docosahexaenoic acid (DHA), and eicosapentaenoic acid (EPA) may regulate the stimulatory properties of the neural sheaths through inhibiting voltage-dependent calcium and sodium channels, decreasing inflammation and inducing mitochondrial uncoupling proteins that could lead to production of reactive oxygen species (ROS).^[@B14]^ Ketone bodies themselves may also bear neuroprotective properties.^[@B2]^
This effect develops through increasing the levels of adenosine triphosphate and decreasing ROS production via enhancing nicotinamide adenine dinucleotide hydrogen (NADH) oxidation and preventing mitochondrial permeability change. Along with other enhanced bioenergetics pathways, ketone bodies are able to stimulate mitochondrial biogenesis and stabilize synaptic functions. The second major biochemical feature of ketone bodies is diminishing glycolytic flow. This condition is the main feature in calorie intake limitation that could elongate the life of various species including primates. It seems that the observed neuroprotective effect is due to a diminished incidence of the brain-derived neurotrophic factor and its principle receptor the tyrosine kinase B, an improved mitochondrial function, diminished oxidative stress, a compromised activity of pro-apoptotic factors, and the prevention of inflammatory mediators such as interleukins and tumor necrosis factor-alpha.^[@B14]^
Possible mechanisms involved in the effectiveness of the KD in relieving PD symptoms could be summarized as follows:
- Providing an efficacious source of energy, which is capable of preventing local hypometabolism in the brain
- Diminished oxidative injury
- Increasing mitochondrial biogenesis pathways
- Exploiting ketone capacities to bypass a failure in the Complex I activity.^[@B20]^
Noh et al.^[@B21]^ showed that one of principle mechanisms of the KD in preventing neurodegeneration was via impeding neural apoptosis by caspase-3.
The early pathophysiology of PD is an excitotoxic degeneration of the dopaminergic neurons located in the substantia nigra. On the basis of some findings, ketone bodies may bypass defects in mitochondrial Complex I activities, which are possibly involved in the pathogenesis of PD.^[@B6]^ Therefore, it seems that mitochondrial abnormalities play a pivotal role in this regard.^[@B22]^
Kashiwaya et al. used an analog of heroin to destroy dopaminergic cells. The suggested mechanism involved in this process was the blockade of a mitochondrial NADH dehydrogenase multi-enzyme complex.^[@B4]^ Accordingly, much of previous studies on the role of the KD in PD have been focused on this aspect of the disease (i.e., mitochondrial abnormalities).^[@B23]-[@B27]^ For example, in a study by Kashiwaya et al.^[@B4]^ on an animal model of PD induced by MPTP, the administration of BHB reduced mitochondrial respiration cycle lesions, which are usually induced by the used toxin.
In a study by Kim et al. the protective effects of ketone bodies against mitochondrial respiratory complex lesions developed by inhibitors of complex I (rotenone) and II (exogenous 3-nitropropionic acid) were examined. They suggested their findings as potential neuroprotective mechanisms of ketone bodies in PD.^[@B7]^
Nevertheless, the current study is one of the rare experiments that examine the effect of the KD on one aspect of PD in a rat model. The observed results are promising and could pave the way for further human studies. For instance, recently some commercial treatments have been available which are based on development and exacerbation of ketonemia such as a formulation using medium-chain triglycerides. It is not clear whether such treatments are effective against PD as much as they are against Alzheimer's disease. Further studies are needed in this regard.^[@B28]^
Conclusion
==========
According to the findings of the present study, since the KD is effective in improving motor function in PD rats further human studied are recommended in this regard.
The authors would like to sincerely thanks to the staff of laboratory of NSRC and Research Council in Tabriz University of Medical Sciences for providing financial and emotional assistance in this project.
Conflict of Interests
=====================
The authors declare no conflict of interest in this study.
Notes:
======
**How to cite this article:** Shaafi Sh, Najmi S, Aliasgharpour H, Mahmoudi J, Sadigh-Etemad S, Farhoudi M, et al. The efficacy of the ketogenic diet on motor functions in Parkinson's disease: A rat model. Iran J Neurol 2016; 15(2): 63-9**.**
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#s1}
============
Stroke remains the leading cause of adult disability worldwide. Recovery following stroke is variable but only a third of patients regain some dexterity in the first 6 months (Kwakkel et al., [@B11]) and the majority of stroke survivors are left with residual motor deficits (Duncan et al., [@B7]). Upper limb deficits limit the ability to perform activities of daily living, primarily due to muscle weakness. This loss of independence contributes to increased healthcare costs and burden of stroke on the individual and the community (Deloitte Access Economics, [@B6]).
Transcranial magnetic stimulation (TMS) has been used to investigate properties of the corticomotoneuronal projection and the potential for upper limb recovery post-stroke. When TMS is performed in the early stages post-stroke, the size of the evoked response in the weak hand is predictive of recovery of finger control (Turton et al., [@B30]). However, most patients have small responses to TMS over the affected motor cortex, which is indicative of reduced excitability of cortical projections and/or a decreased cortical representation of the affected hand following stroke (Rossini and Dal Forno, [@B24]).
The representation of muscles within the motor cortex can be investigated with TMS by moving the stimulating coil over successive positions of the scalp and recording the size of the evoked response in the target muscle (Cramer and Bastings, [@B5]). The resultant map is relatively stable over time in healthy individuals (Uy et al., [@B31]) but can change as a result of rehabilitation following stroke (Traversa et al., [@B28], [@B29]; Liepert et al., [@B13]) or specific training interventions (Liepert et al., [@B14]). The resultant change in the map is thought to represent reorganization of the motor cortex.
Reorganization of motor cortex can occur following repeated, meaningful sensory input in healthy adults. Combined stimulation of sensory receptors in the hand and the hand representation of the motor cortex, repeated daily, results in expansion of the hand motor map (McKay et al., [@B17]). Interestingly, the motor map expansion persists for at least 2 days post-stimulation (McKay et al., [@B17]). The current study sought to investigate if continuous passive movement (CPM, used regularly following orthopedic surgery) is also capable of evoking expansion of the motor cortical representation of the muscles involved. Passive movements are associated with acute changes in corticomotoneuronal excitability in healthy adults (Chye et al., [@B3]), largely due to increased muscle spindle firing rates during muscle lengthening. Furthermore, use of blood oxygen level-dependent magnetic resonance imaging reveals substantial reorganization of sensorimotor cortex in healthy adults following daily passive movement of the wrist (duration 20 min) for 1 month (Carel et al., [@B2]). A single session of passive movement of the wrist can also alter the corticomotor representation of a forearm muscle, with increased volume of the motor map of the flexor carpi radialis muscle in healthy adults, but not participants following stroke (Lewis and Byblow, [@B12]). However, the after-effects of repeated passive movement of the thumb on corticomotoneuronal excitability has not been investigated.
The aim of the current study was to determine if three consecutive days of passive thumb movement (30 min per day), increases the representation of thumb muscles in the motor cortex in healthy adults following the intervention. We also investigated if the change in the representation of the thumb muscles was still evident 5 days after cessation of the passive movement paradigm. The thumb was chosen because opposition of the thumb is an essential requirement for functional use of the hand, which is often impaired after stroke. It was hypothesized that repeated afferent input, with attention directed to the thumb, could transiently enlarge the cortical representation of the thumb. If the results of the study support the hypothesis, then use of CPM on the weak hand following stroke could prevent a reduction in the cortical representation of the weak hand muscles and thus aid rehabilitation.
Materials and Methods {#s2}
=====================
Participants {#s2-1}
------------
Thirteen healthy adults (7 men, 6 women) aged 20--33 years participated in the study. Inclusion criteria were right-hand dominant (according to the Edinburgh Handedness Inventory (Oldfield, [@B20])), no neurological or orthopedic conditions affecting the hand, and no contra-indications to TMS (Rossi et al., [@B23]). All participants provided written, informed consent in accordance with the Declaration of Helsinki and the study was approved by the local institutional ethics committee.
Testing Procedures {#s2-2}
------------------
A close-fitting flexible cap was placed on the participants head. The cap was marked with a grid of 1 cm spacing. TMS was delivered over points on the grid using a Magstim 200^2^ stimulator and a figure-of-eight coil (part number: 3281--00, 90 mm external diameter of wings, Magstim, Whitland, UK). The coil was initially placed over the motor cortex in the left hemisphere, over the abductor pollicis brevis (APB) motor area. The handle of the coil pointed posteriorly at approximately 45 degrees to the midline and tangentially to the skull. This coil position induces a posterior-to-anterior current in the brain and is optimal for stimulating the hand region of the motor cortex. Single stimuli were delivered at a rate of \~0.2 Hz.
The response to TMS (motor evoked potential or "MEP") in the target muscle (right APB) was recorded with electromyography (EMG). Two surface EMG electrodes (Ag-AgCl, 10 mm diameter) were placed over the muscle belly and tendon. EMG signals were sampled at 2000 Hz, amplified (1000×), and filtered (20--1000 Hz) using a data acquisition system (1902 with Power 1401 interface and Signal software, Cambridge Electronic Design, Cambridge, UK).
The experiment began with determination of the optimal scalp site for eliciting a MEP for APB. The optimal site was marked and resting motor threshold (RMT) was determined. Determination of RMT involved initially setting the intensity of stimulation well above threshold and then reducing the intensity in steps of 1--2% of stimulator output until it was below threshold. RMT was defined as the stimulus intensity that produced an MEP of amplitude greater than 0.05 mV in 5 out of 10 consecutive stimuli. Fifteen stimuli were then delivered during complete muscle relaxation at an intensity of 120% RMT. Visual feedback was provided via a computer screen to ensure complete muscle relaxation and any trials with pre-stimulus voluntary EMG activity were discarded from the analysis. After collection of MEPs evoked at a stimulus intensity of 120% RMT, participants underwent mapping of the cortical representation of the right APB muscle during relaxation, similar to the method described by Uy et al. ([@B31]). Five stimuli were delivered at each scalp site at an intensity of 110% RMT. The first scalp site was the optimal site (APB hot spot) and selection of subsequent sites involved moving the coil outwards in 1 cm steps until no MEPs were elicited. The number of scalp sites that exhibited a MEP amplitude ≥ 0.05 mV is an index of the area of the cortical representation of APB. The sum of the averaged MEP amplitude evoked at all active scalp sites is an index of the volume of the cortical representation of APB (Cicinelli et al., [@B4]; Meesen et al., [@B18]). The center of gravity for the map was also calculated according to the method described by Wassermann et al. ([@B32]), using the following formula: $$\text{CoG} = {{\sum{v_{i}x_{i}}}/{\sum{v_{i},}}}{{\sum{v_{i}y_{i}}}/{\sum{v_{i},}}}\text{ for scalp sites }x_{i},y_{i}\text{ and amplitudes }v_{i}$$
The between-session difference in the CoG was also calculated.
The above TMS procedure was performed at baseline, following each day of CPM (days 1--3), and 5 days after the final session of CPM (day 8). The distance from the vertex to the nasion, inion, and bilateral pre-tragus was recorded during the baseline session to enable consistent cap (and thus coil) placement on subsequent days.
Continuous Passive Motion {#s2-3}
-------------------------
The CPM intervention was applied to each participant for 30 min per day over three consecutive days using a custom made device (see Figure [1](#F1){ref-type="fig"}). Participants sat with their right shoulder in slight abduction (10°), elbow flexed to 90° and forearm restrained in a supinated position. The thumb was supported with a flexible sling around the proximal phalanx. An AC servomotor (model YM2760, Jaycar Electronics, Adelaide, Australia) located above the device enabled variable amplitude and frequency movements to be generated in the thumb support, moving the thumb in an arc of flexion/abduction and extension/adduction. The speed (0--240° s^−1^) and amplitude (0--60°) of the movement was varied in a random fashion to prevent habituation to the passive movement with repetitive, identical stimuli. The investigator was able to alter the pattern of the passive motion from random to regular, and participants were asked to pay attention to their hand and to alert the investigator when the movement changed from random to regular. The investigator covertly altered these settings once every approximately 5 min and recorded the time interval between changing the setting and the participant reporting the change (called "detection" time). At the end of the CPM session, participants voluntarily removed their thumb from the device and replaced it on a supportive pillow for the remainder of the TMS mapping.
{#F1}
Statistical Analysis {#s2-4}
--------------------
Data are presented as means ± standard deviations (SD). All parameters were analyzed for normality using Kolmogorov-Smirnov tests and log transformation was applied to the map volume data. The acute effect of CPM on TMS parameters were analyzed with Student's paired *t*-tests (pre- vs. post-CPM on day 1). Any long-lasting effect of CPM on TMS parameters and attention were analyzed with one-way repeated measures analysis of variance (ANOVA) (time points: post CPM day 1, post-CPM day 2, post-CPM day 3, and day 8) or the non-parametric Friedman's test for the percent change from baseline. The effect of the initial area of the motor map and attention were considered as covariates in an ANCOVA to investigate whether they influenced MEP amplitude, COG or map area. All statistical tests were performed with IBM SPSS Statistics Version 21 and the level of significance was set at *P* \< 0.05.
Results {#s3}
=======
The average age of participants was 24.3 ± 4.3 years and the average score on the Edinburgh Handedness Inventory was 87 ± 15. Group data for TMS parameters obtained before and/or after CPM are shown in Table [1](#T1){ref-type="table"}. There was no difference in RMT between baseline and the end of the first CPM session (paired *t*-test, *P* = 0.47) or between days 1, 2, 3 and 8 (ANOVA, *F*~4,12~ = 0.43, *P* = 0.78).
######
**Group data for TMS parameters and the attention task**.
Day 1 Pre (Baseline) Day 1 Post Day 2 Post Day 3 Post Day 8
------------------------------------ ---------------------- ------------ ------------ ------------ ------------
**RMT SI (%MSO)** 40.1 ± 6.7 40.4 ± 6.0 39.8 ± 6.3 39.8 ± 6.4 40.2 ± 6.7
**MEP amplitude at hot spot (mV)** 1.2 ± 0.8 0.9 ± 0.5 0.9 ± 0.6 1.0 ± 0.5 0.9 ± 0.7
**Map area (sites)** 19.1 ± 9.3 19.5 ± 9.1 18.8 ± 9.6 16.9 ± 7.6 19.9 ± 9.5
**Map volume (mV)** 4.8 ± 2.9 6.0 ± 6.4 5.1 ± 4.2 5.2 ± 5.0 6.2 ± 6.0
**Detection task time (s)** -- 59 ± 23 65 ± 24 75 ± 50 --
*RMT = resting motor threshold, SI = stimulus intensity, MSO = maximal stimulus output, MEP = motor evoked potential. The detection task was only performed during the three CPM sessions on days 1--3*.
MEP Amplitude at the Hotspot {#s3-1}
----------------------------
There was no difference between the MEP amplitude recorded at an intensity of 120% RMT at the hotspot before and after the first session of CPM on day 1 (see Table [1](#T1){ref-type="table"}, paired *t*-test, *P* = 0.13). Raw (ANOVA, *F*~4,10~ = 0.67, *P* = 0.62) and normalized (percent change from baseline, Friedman's test, Chi-square = 1.338, *p* = 0.720) MEP amplitude also did not differ across testing sessions.
Map Area and Volume {#s3-2}
-------------------
The area of the cortical representation of APB did not differ between pre and post CPM on day 1 (see Table [1](#T1){ref-type="table"}, paired *t*-test, *P* = 0.77) or across testing sessions (raw data: ANOVA, *F*~4,9~ = 0.76, *P* = 0.58; normalized data (percent change from baseline): Friedman's test, Chi-square = 2.504, *p* = 0.475).
The APB map volume varied greatly between participants (1.1--24.5 mV) but did not significantly differ between pre- (4.8 ± 3.0 mV) and post-CPM (6.0 ± 6.4 mV) on day 1 or across testing sessions (raw data: ANOVA, *F*~4,9~ = 1.16, *P* = 0.34; normalized data (percent change from baseline): Friedman's test, Chi-square = 3.092, *p* = 0.378).
Center of Gravity {#s3-3}
-----------------
The COG of motor maps changed between sessions for all individuals in an inconsistent manner. The distance (mm) from the COG at baseline on day 1 was calculated, and there were no significant differences in the distance moved after the first CPM session (average COG distance moved 1.49 ± 1.42 mm) or across the remainder of the testing days (ANOVA, *F*~3,36~ = 1.047, *p* = 0.359).
Attention {#s3-4}
---------
During the CPM intervention on days 1--3, the average time taken to detect a change in the pattern of CPM (i.e., index of attention) was 57.8 ± 37.6 s. Detection time did not significantly change across the six epochs within a session (ANOVA, *F*~5,14~ = 1.7, *P* = 0.16) or across testing sessions (ANOVA, *F*~2,8~ = 0.02, *P* = 0.98) and no epoch \* day interaction was observed (ANOVA, *F*~10,54~ = 0.93, *P* = 0.51).
Influence of Confounding Variables {#s3-5}
----------------------------------
In order to determine whether baseline map characteristics or attention influenced the response to CPM, a further analysis of variance was conducted with these variables as covariates. The initial area of the motor map had no influence on the map area (ANCOVA, *F*~4,44~ = 0.40, *P* = 0.81) or volume after CPM (ANCOVA, *F*~4,9~ = 0.63, *P* = 0.65). Likewise, attention had no influence on the post-CPM map area (ANCOVA, *F*~4,44~ = 1.88, *P* = 0.20) or map volume (ANCOVA, *F*~4,44~ = 0.48, *P* = 0.75).
Discussion {#s4}
==========
The results of the current study suggest that single and repeated sessions of CPM of the thumb are insufficient to induce a lasting change in the representation of the APB muscle in the motor cortex of healthy adults. The lack of a change in the APB representation was evidenced by an unaltered motor map after a single session of CPM or daily sessions repeated over three consecutive days. The size of the response evoked by a specified intensity of stimulation over the APB hotspot was also unaltered suggesting unchanged excitability in this part of the motor cortex and in the APB corticomotoneuronal projection.
The lack of an acute change in the APB motor map and excitability following a single session of thumb CPM is surprising for three reasons. First, Lewis and Byblow ([@B12]) observed a significant increase in the volume of the motor map of another muscle, the flexor carpi radialis, after a single 30-min session of passive wrist movement in healthy adults. Second, fMRI studies show that passive movement of the wrist activates primary motor cortex, again in healthy adults (Lotze et al., [@B15]). Lastly, two hours of tactile stimulation increases the size of the hand representation in sensory cortex in animals and improves tactile discrimination in healthy adults (Godde et al., [@B10]). Our results suggest that the after-effects of passive movement on the motor cortex may differ between muscles and that the mechanisms that underlie plasticity may differ for tactile stimulation and passive movement.
Our study is the first to investigate the effect of repeated sessions of passive thumb movement on the representation and excitability of thumb muscles involved in the task. However, we observed no effect of repeated sessions of thumb CPM on the representation and excitability of APB. We speculate that passive movement-induced reorganization may be better suited to muscles that have a smaller representation in the human motor cortex, like the flexor carpi radialis, or to movements that involve many muscles, like flexion and extension of the wrist. This could be investigated by comparison of modulation of APB corticomotor excitability during passive thumb movement with modulation of FCR corticomotor excitability during passive wrist flexion and extension. Alternatively, a greater number of sessions may be required to effect a change in the motor map. The observation of increased activity in motor cortices of healthy adults following 20 min of wrist movement repeated daily for 1 month (Carel et al., [@B2]) and enlargement of the cortical motor map following 1 h of transcutaneous electrical stimulation of the thumb daily, for 3 weeks in healthy adults (Meesen et al., [@B18]) suggests that this may be the case.
Factors that are known to affect experimental and use-dependent plasticity within motor cortex include attention (Stefan et al., [@B27]), circadian rhythm, sex, central nervous system active drugs and physical activity (see Ridding and Ziemann, [@B22] for review). The current experimental design attempted to minimize any between-subject differences in attention by including an attention task and quantifying attention within- and across-sessions. No significant changes in attention were observed within- or across-sessions and lack of attention to the hand is unlikely to explain the absence of an effect on motor maps in the current study. The effect of circadian rhythm and cortisol levels (Sale et al., [@B25]) were minimized by scheduling experiments at the same time of day.
There are several limitations to our study. First, the average age of our participants (24.3 ± 4.3 years) was considerably younger than most people following stroke. We recruited healthy young participants because we anticipated that any effect of a motor task on cortical reorganization was most likely to be seen in younger compared with older adults (Rogasch et al., [@B33]). The lack of an effect in this younger sample suggests that an effect is unlikely to be present in a sample of older adults. Second, the study had a small sample size. However, the absence of even a trend towards an increase in map size is unlikely to translate into a significant effect in a larger group. Third, tactile input arising from contact between the sling and the skin on the thumb during CPM is unlike the tactile input that would normally be associated with voluntary thumb movement during for example, manipulation of objects. Finally, we did not control for or assess thumb use before the study and thus are unable to comment on the role of activities such as playing a musical instrument or prolonged video gaming on the results of the current study.
Identifying a robust and reproducible method for experimentally inducing reorganization of the sensorimotor cortex could benefit a wide range of patients. Reorganization of the sensorimotor cortex is already known to occur in natural situations such as reading Braille (Elbert and Rockstroh, [@B9]), learning to play a musical instrument (Elbert et al., [@B8]) and in response to pathological situations such as amputation (Ridding and Rothwell, [@B21]). A common feature is that use of a body part is associated with increased representation, while non-use, or experimentally depriving the cortex of afferent information, is associated with a decrease in the representation. It appears that change in sensory (afferent) input is a critical component of this so-called "use-dependent" plasticity (Bütefisch, [@B1]). Recently, some novel rehabilitation paradigms, which modify afferent input, have been developed to manipulate use-dependent plasticity in a functionally beneficial way following stroke (McDonnell et al., [@B16]; Schabrun and Hillier, [@B26]). We proposed that CPM could provide an avenue for tactile and proprioceptive input, which may help to reorganize the sensorimotor cortex and ultimately improve recovery of the hemiplegic arm post-stroke. The lack of an effect following CPM in our healthy sample may indicate that either passive movements are not functionally-meaningful enough to induce reorganization, as has been shown with some motor tasks in healthy adults (Ngomo et al., [@B19]), or that it is difficult to show a change in an intact, healthy motor representation. While we expect that the shortening and lengthening of the APB would have activated muscle spindles and cutaneous mechanoreceptors, the passive movement did not take the carpometacarpal joint of the thumb to the end of range and thus did not activate joint receptors. This may have contributed to the inability of the CPM to induce a lasting change.
Conclusion {#s5}
==========
Single and repeated sessions of thumb CPM do not alter the motor cortical representation of the APB in healthy adults. Thus, use of CPM on the weak hand following stroke may not reverse the reduction in the cortical representation of the weak hand muscles or aid rehabilitation.
Conflict of Interest Statement {#s6}
==============================
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
**Acknowledgments**
We wish to thank Stan Flavel for his invaluable technical assistance with the passive movement device. GT received support from the National Health and Medical Research Council of Australia (Career Development Award, ID 627003).
[^1]: Edited by: A. M. Barrett, Kessler Foundation, USA
[^2]: Reviewed by: Filippo Brighina, University of Palermo, Italy; Winston D. Byblow, University of Auckland, New Zealand; Suzy Ngomo, Université du Québec à Chicoutimi, Canada
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#s1}
============
Psychiatric symptoms, such as anxiety, depression, insomnia, and somatization, are the clinical manifestations of common psychiatric and neurological diseases. Many neurological disorders show these indistinguishable psychological symptoms in the early stages, especially when symptoms similar to Parkinson's disease (e.g., difficulty walking, stiff limbs, and tremors) are present. These patients are likely to be misdiagnosed with Parkinson's disease and are treated with dopamine replacement therapy; however, in rare cases, increasing the dosage of dopamine can elicit restless leg syndrome (RLS). Extended durations of these psychiatric symptoms can be detrimental to the patient's physical and mental health.
The present study assessed a group of patients who were misdiagnosed with Parkinson's disease and were administered large doses of Madopar. All the patients exhibited rare RLS-like symptoms, such as difficulty in walking, stiff limbs, and tremors, which were accompanied by anxiety, depression, and other psychiatric symptoms. Clinical data of all patients were collected, and strategies of treatment and prognosis were analyzed.
Materials and Methods {#s2}
=====================
Patients {#s2_1}
--------
The present study was approved by the Ethics Committee of the First People's Hospital of Huainan, and written informed consent was provided. Twelve patients demonstrating symptoms of anxiety, depression, and somatization due to misdiagnosis of Parkinson's disease and taking a large dose of Madopar were identified and recruited from January 2010 to December 2017. Two Parkinson's disease patients did not meet the inclusion criteria, and one patient declined to follow up. Therefore, nine patients (47--78 years old) were enrolled. All nine patients were hospitalized, and after a detailed evaluation of medical history, rigorous neurological physical examination, and related auxiliary tests, they were determined to not meet the criteria for Parkinson's disease according to the British Parkinson's Disease Society ([@B1]). Anxiety, depression, insomnia, and somatization symptoms were diagnosed as depression and anxiety disorders according to the *Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition* (DSM-5) ([@B2]). All patients underwent general routine examinations as well as biochemical and imaging examinations. No abnormalities were noted except for the primary disease. The diagnosis for RLS was based on clinical criteria ([@B3]) and included an urge to move the legs, usually associated with unpleasant sensations; symptoms occurring during periods of rest, such as sitting or lying down; symptoms relieved by movement; and worsened symptoms in the evening or night.
The education level of all patients was above primary school, and they could independently complete the questionnaire without communication barriers. All patients agreed to follow up.
Laboratory and Imaging Examinations {#s2_2}
-----------------------------------
Routine blood, urine, fecal, serum glucose level, liver and kidney function, thyroxine, and electrolyte laboratory and physical examinations were conducted. Electroencephalogram and brain magnetic resonance imaging were performed in all patients.
Clinical Evaluation and Follow-Up {#s2_3}
---------------------------------
Severity of RLS was evaluated on the basis of the International RLS Rating Scale (IRLS-RS) ([@B4]). The diagnosis and severity of insomnia, anxiety, and depression in all patients were assessed by two neurologists and a psychiatrist according to the Insomnia Severity Index (ISI) ([@B5]), Hamilton Anxiety Rating Scale (Hamilton) ([@B6]), Hamilton Depression Rating Scale (HDRS) ([@B7]), and DSM-5 diagnostic criteria ([@B2]) combined with clinical symptoms and signs. Follow-up data for all patients with RLS were obtained during face-to-face or telephone interviews.
### Clinical Research Flow {#s2_3_1}
The clinical study flow is shown in [**Figure 1**](#f1){ref-type="fig"}.
{#f1}
Statistical Analysis {#s2_4}
--------------------
All statistical analyses were performed using Statistical Product and Service Solutions (SPSS) version 19.0 (SPSS Inc., Chicago, IL, USA). The normality of the distribution was assessed using the Kolmogorov--Smirnov test. Normally distributed quantitative data were presented as "mean ± standard deviation (SD)." The international RLS scores of patients before and after treatment were compared by *t*-test. The anxiety, depression, and insomnia scores of patients before and after taking a large dose of Madopar were compared using the Student's *t*-test. *P* values \<0.05 were considered significant.
Results {#s3}
=======
Nine patients took Madopar orally due to being misdiagnosed with Parkinson's disease, and the starting dosage ranged from 1/2 to 1 tablet (0.25 g/tablet). All patients gradually increased the amount of medication administered. Some were under the guidance of a doctor, but then to achieve the "curative effect," patients increased the amount of medication themselves. Some patients increased their doses by themselves from the beginning (i.e., without the doctor's assistance). The amount of medication in most patients was 2--3 tablets per dose, 3--4 times per day, which was at maximum 5 tablets per dose, 3--5 times a day in one case. When the average dosage reached 6--8 tablets per day and the duration of administration lasted 2--4 weeks, the onset of bilateral lower limb discomfort appeared. Initially, the symptoms were minimal, which did not alert the attention of the patients. As the medication dosage and duration increased, so did the symptoms, which appeared as unexplained abnormal sensations in both lower extremities to varying degrees, such as numbness, swelling, crawling, burning, and traction pain at night. The symptoms could be temporarily reduced by activity, which forced patients to stay out of bed for exercises, which affected their sleep. As a result, patients typically increased the dose of Madopar, which could reduce the symptoms, especially when the symptoms were unbearable. The increasing dosage of Madopar could effectively improve the symptoms, and thus forced patients to increase the amount of medication.
During this cycle, when symptoms appeared during the daytime, the upper limbs and occasionally the entire body displayed varying degrees of involvement. As shown in [**Table 1**](#T1){ref-type="table"}, the average time from the use of the Madopar to the onset of RLS symptoms was 2.61 ± 0.60 months. For symptoms to appear, the minimum of the average daily dose of Madopar was 1.44 ± 0.21 g; moreover, the average duration for the nine patients with RLS from the time of high-dose Madopar administration to the time of hospital admissions was 18.17 ± 9.40 months. The original symptoms of anxiety, depression, insomnia, and general discomfort worsened in seven patients before onset of the disease. The other two cases displayed anxiety, depression, and insomnia, as well as whole-body burning-like and mobile pain, accompanied by the gradually aggravated discomfort of the bilateral lower limbs. As shown in [**Table 2**](#T2){ref-type="table"}, the symptoms of anxiety, depression, and insomnia were significantly worse in all nine patients after taking a large dose of Madopar (*P* \< 0.0001, *P* \< 0.05, *P* \< 0.0001).
######
Demographic and clinical characteristics of the nine patients.
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Case Gender Age (years) Diagnosis Madopar dosage\ Time1\ Minimum\ Time2\ Clinical manifestations
(tablets/d) (months) dosage (g/d) (months)
------ -------- ------------- ------------------------------------------------------------------------------ ----------------- ---------- -------------- ---------- ------------------------- ----- ---- ----- ----- ----- -----
1 F 78 Panic disorder 10--18 3 1.25 25 B 0 0 0 ++ \+ \+
A +++ ++ +++ +++ ++ +++
2 F 76 Sleep-associated leg spasms, liver cirrhosis,upper gastrointestinal bleeding \ \ \ \ \ \ \ \ \ \ \
6--12 2.5 1.5 20 B\ 0\ 0\ 0\ 0\ 0\ 0\
A +++ + + ++ + ++
3 F 74 Cerebral infarction, poststroke depression 6--10 3 1.25 24 B\ 0\ 0\ 0\ \+\ ++\ \+\
A +++ ++ ++ +++ +++ ++
4 F 47 Somatization disorder 8--12 2.5 1.25 13 B\ 0\ 0\ 0\ ++\ \+\ 0\
A ++ 0 + +++ ++ ++
5 M 56 Postencephalitis 6--15 3.5 1.5 4.5 B 0 0 0 0 0 0
A +++ \+ ++ +++ ++ ++
6 F 52 Anxiety and depressive disorder 9--12 2 1.75 5 B\ 0\ 0\ \+\ ++\ ++\ ++\
A +++ + ++ +++ ++ +++
7 F 67 Anxiety and depressive disorder 6--9 3 1.5 31 B\ 0\ 0\ 0\ \+\ ++\ \+\
A +++ 0 + ++ ++ +++
8 F 71 Somatization disorder 6--12 2.5 1.25 15 B\ 0\ 0\ 0\ ++\ \+\ 0\
A +++ + ++ +++ ++ +++
9 M 68 Cerebral infarction, poststroke depression 6--13 1.5 1.75 26 B\ 0\ 0\ 0\ 0\ ++\ \+\
A ++ + 0 ++ ++ ++
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
M, male; F, female; LE, lower extremity; UL, upper limb; GD, general discomfort; 0, no symptoms; +, mild symptoms; ++, moderate symptoms; +++, severe symptoms. B, before treatment; A, after treatment. Time 1, time of onset of restless leg syndrome (RLS) after taking Madopar; Minimum dosage, minimum dose of Madopar at the onset of RLS; Time 2, time of onset of RLS to visit. Clinical manifestations, clinical manifestations before and after taking Madopar.
######
Comparison of anxiety, depression, and insomnia scores of the nine patients.
1 2 3 4 5 6 7 8 9 $\overline{x}$ ± *s* *P*
------ ---- ---- ---- ---- ---- ---- ---- ---- ---- -------------- ---------------------- ----------
HARS B 25 4 18 25 7 22 16 23 10 16.67 ± 2.66 \<0.0001
A 52 26 50 42 38 45 27 47 26 39.22 ± 3.50
HDRS B 22 12 33 22 6 30 30 22 30 23.00 ± 3.03 \<0.05
A 33 25 62 31 32 33 33 32 32 34.78 ± 3.50
ISI B 10 7 10 5 3 17 11 5 10 8.667 ± 1.40 \<0.0001
A 26 13 19 21 17 26 25 25 19 21.22 ± 1.54
B, before taking Madopar; A, after taking large-dose Madopar; HARS, Hamilton Anxiety Rating Scale; HDRS, 24-item Hamilton Depression Rating Scale; ISI, Insomnia Severity Index; p-value, Student's t-test.
All nine cases were asked to gradually reduce their dose of Madopar. Low doses of long-acting dopamine agents, dopamine receptor agonists, α2δ calcium channel ligands, clonazepam, and other drug treatments were administered. All psychiatric symptoms were greatly alleviated but did not fully disappear and lasted for approximately 2 years. The severity of symptoms in seven patients with more than 6 months of disease course was significantly improved, but after 3 months of treatment, there was no obvious further improvement of symptoms and fluctuations were present.
As shown in [**Table 3**](#T3){ref-type="table"}, there was a significant difference in the IRLS scores 1 month before and after treatment. During the first month of follow-up, IRLS scores of all patients were significantly lower than the initial assessment \[21.22 + 2.05 points (indicative of severe symptoms) compared to 35.33 ± 2.40 points; *P* \< 0.0001\]. At the 3-month follow-up, the IRLS scores of patients were significantly lower than the first month of follow-up (*P* = 0.001). Finally, at the 6-month follow-up, the IRLS scores were 13.89 ± 5.06 points, indicating moderate severity. Before treatment and at the 1-month follow-up, there was statistically significant difference between IRLS scores (*P* \< 0.0001); when the 3-month follow-up was compared to this, although the symptoms were improved, it was not found to be statistically significant (*P* = 0.33). And, at the 12-, 18-, and 21-month follow-ups, when compared with the 6-month follow-up, two cases at the 21-month follow-up demonstrated that the IRLS score continued to decrease, reaching a mild level of severity; however, the subsequent treatment did not provide additional benefits to the remaining seven cases. The RLS symptoms showed no obvious improvements compared to the 6-month follow-up and were still classified as moderately severe symptoms. Another patient died of primary disease (cirrhosis and hemorrhage of upper digestive tract) during the 12 months of follow-up. This demonstrated that improvements were no longer obvious after 3 months, which suggested that early diagnosis and treatment might be the key factor to improving prognosis.
######
Treatment and outcome of treatment of the nine patients and follow-up.
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Case RLS score before\ Length of treatment\ IRLS-RS at follow-up (months) Treatment strategy
treatment at follow-up (months)
------ ------------------- ----------------------- ------------------------------- -------------------- ---- ----- ---- ---- ----- ----- --------- --------- ---- ----- -------- -------- --------
1 39 38 25 20 16 18 25 20 3 -- 50 150 -- 0.5 -- 60 5
2 31 10 18 12 10 die -- -- 0.5 0.5 50 75--100 -- 0.5 -- -- --
3 35 13 22 20 23 18 -- -- 1 0.5 50 -- -- 1 20 -- 2.5
4 34 19 21 18 10 12 15 -- -- 1 50--100 150 2 -- -- 60 5
5 37 32 20 12 10 12 13 8 -- 1 50 150 -- 0.5 20 -- 2.5
6 36 25 23 16 10 12 10 9 1 -- 100 150 -- -- -- 60 2.5--5
7 33 23 20 16 16 17 16 -- -- 0.5 50 100 2 -- -- 60 2.5
8 36 18 22 18 20 16 18 -- -- 1 50 100 -- -- -- 60--90 5
9 37 36 20 16 10 15 14 12 -- 1 50 200 -- 0.5 20--30 -- 2.5
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
IRLS-RS, International Restless Legs Syndrome Rating Scale; MAD, Madopar; CLSRT, carbidopa and levodopa sustained-release tablets; PSRT, piribedil SR tablets; PRE, pregabalin; EST, estazolam; CLO, clonazepam; PAR, paroxetine; DUT, dutoxetine; OLA, olanzapine.
Discussion {#s4}
==========
RLS is a common nervous system sensory dyskinesia disease, and the clinical manifestations are extreme discomfort during rest and nocturnal sleep. Symptoms can be remitted through movement of the lower extremities, which forces patients to continue to move their limbs, therefore disturbing sleep and rest. According to the etiology, RLS can be divided into two subtypes: primary and secondary. The former etiology is unclear and may be heredity, while the latter is often due to iron deficiency, pregnancy, chronic renal failure, and other causes. In the present study, we reported for the first time that nine patients who had no family history of primary or secondary RLS showed RLS-like symptoms accompanied by anxiety and depressive symptoms, as induced by high-dose Madopar.
Previous reports and case studies have suggested that certain medications may cause or exacerbate RLS. These medications include several classes of antidepressants, including tricyclic antidepressants ([@B8]) such as imipramine ([@B9]); selective serotonin or norepinephrine reuptake inhibitors ([@B10]), such as citalopram ([@B11]), escitalopram ([@B12]), fluoxetine ([@B13]), sertraline ([@B14]), paroxetine ([@B15]), trazodone ([@B16]), venlafaxine ([@B17]), dutoxetine ([@B18]), and mirtazapine ([@B19]); and neuroleptics that have significant dopaminergic blockade ([@B20]), such as olanzapine ([@B21]), risperidone ([@B22]), and quetiapine ([@B23]). In addition, antihistamines operating on the H1 receptor ([@B24]) and selected antiemetics with dopamine antagonism such as metoclopramide ([@B25]) and prochlorperazine ([@B26]) have also been associated with RLS. However, as none of the nine patients we observed took the aforementioned drugs, RLS-like symptoms caused by these drugs were excluded.
The clinical manifestations of anxiety, depression, and somatoform disorders are complex and diverse. Other than the emotional aspects, these illnesses can manifest as different forms of somatic symptoms, such as dizziness, headache, limb weakness (especially in the lower extremities), and difficulty walking, which is likely to be misdiagnosed as a primary disease. Among the nine patients in the present study, seven had varying degrees of anxiety and depression. For example, Case 5 was misdiagnosed with Parkinson's disease and treated with Madopar due to the spasm gait of the double lower limb after encephalitis. Since the effect of treatment was not obvious, the patient increased the dosage by themselves. RLS-like symptoms and severe anxiety, insomnia, and other psychiatric symptoms occurred with an increasing amount of medication. Case 2 was diagnosed with a sleep-related leg spasm according to the predisease clinical manifestation, and then the patient was misdiagnosed with Parkinson's disease and treated with a high dose of Madopar, which induced RLS-like symptoms and anxiety. As one of the effective drugs to treat RLS, Madopar has been widely used in clinical practice; however, the onset of RLS is very rare and its pathogenesis should be further discussed.
In 2016, single-photon emission computed tomography imaging was used to study the pathophysiological mechanisms of RLS at the Tri-Service General Hospital, National Defense Medical Center ([@B27]). The results showed a significantly reduced uptake in striatal dopamine transporter (DAT) density and activity in RLS patients ([@B27]). This study supported that symptoms of RLS resulted from the striatum due to dopaminergic system dysfunction ([@B27]).
To date, many studies have shown that using drugs such as levodopa and other dopamine agonists can significantly improve the symptoms of RLS; therefore, the central dopaminergic nervous system (particularly the nigra-striatum system or intermediate cortical system) has been considered to be associated with the onset of RLS. In the present group, patients displayed RLS symptoms after the use of long-term high doses of Madopar (from the use of Madopar to the onset of RLS-like symptoms), suggesting that the cause was the dysfunction of the central dopamine system. A number of studies have reported that the use of dopamine drugs in the treatment of RLS may promote symptom deterioration, reverse jump, and other adverse reactions. This is especially true after the long-term use of levodopa, as the proportion of deteriorated symptoms is 18--80% ([@B28]). Since these adverse reactions are more common in patients with long-term high doses of levodopa treatment, it suggests that the mechanism of symptom deterioration may be associated with dopamine overdose in the central nervous system ([@B29]). Therefore, we speculate that the mechanism of RLS-like symptoms in the nine patients of the present study may be caused by excessive dopamine in the central nervous system after the use of long-term high doses of dopaminergic agents.
High concentrations of dopamine can excite D1 receptors and cause D1 receptor-related pain, which results in periodic limb movements. Several studies have shown that certain concentrations of external toxic substances \[such as levodopa, dopamine (DA)\] may damage dopamine transporters (DAT), therefore significantly reducing their abundance. Additionally, compared to the mitochondria, DAT is more sensitive to injury stimulation from external toxic substances. Before the cells' mass death, the number of DAT on the cell membranes is significantly reduced. The remaining DAT functions display compensatory hyperfunction and are therefore eliminated due to the reciprocal inhibition, which allows them to ingest more dopamine and its metabolites into the cells. This results in a large number of free radicals and the inhibition of the mitochondrial respiratory chain, and eventually causes cell death. Therefore, it is speculated that the long-term use of dopamine in patients of the present study may lead to a reduction in the number of dopamine receptors in the brain and spinal cord or a decrease in DAT function, finally resulting in dopaminergic systemic dysfunction and the occurrence of RLS symptoms.
In the present study, nine patients with long-term high doses of dopaminergic drugs displayed RLS symptoms, while the original symptoms of anxiety, depression, insomnia, and somatization appeared or were aggravated. Current epidemiological studies report that RLS is a common cause of insomnia, and the rate of comorbidity with depression and anxiety is high ([@B30]). Winkelmann et al. ([@B31]) assessed 238 RLS patients with a standardized diagnostic interview \[Munich-Composite International Diagnostic Interview for DSM-IV (*Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition*)\]. Rates of anxiety and depressive disorders were compared between them and 2,265 community respondents from a nationally representative sample. RLS patients revealed an increased risk of having anxiety and depressive disorders with particularly strong associations with panic disorder, generalized anxiety disorder, and major depression ([@B31]). Moreover, the Baltimore Epidemiologic Catchment Area follow-up study suggested a strong association between RLS and major depressive disorder and/or panic disorder ([@B32]). An anonymous survey study in Appalachia suggested that those with RLS were significantly more likely to indicate a history of depression and anxiety and report sleep impairments both 4 and 7 days/week, with a mean sleep duration \<5 h/night ([@B33]). These associations increased in both strength and magnitude with increasing symptom frequency ([@B33]). More recently, a study on the clinical characteristics of RLS in adult patients from Peking Union Medical College Hospital demonstrated that primary RLS patients suffer from poor sleep and are more susceptible to anxiety and depression ([@B34]). The scores of Hospital Anxiety and Depression Scale for depression and anxiety were significantly correlated with those of the Pittsburgh Sleep Quality Index and IRLS ([@B34]).
The underlying cause of the high incidence of RLS and anxiety and depression is unclear. It is possible that these illnesses may share a basic pathophysiological mechanism leading to their development. Pan et al. ([@B35]) explored the regional gray matter (GM) density in depressed drug-naïve RLS patients using voxel-based morphometry, which showed that GM density of the bilateral anterior cingulate cortex (ACC) was significantly reduced in RLS patients with depressive symptoms (RLS-D) compared to RLS patients without depressive symptoms or healthy controls. Additionally, a significant negative correlation between right ACC density and HDRS scores and duration of depressive symptoms in patients with RLS-D was found ([@B35]). It was speculated that depressive symptoms are associated with GM abnormalities in the ACC of patients with RLS.
In the present study, anxiety was more obvious than the depressive symptoms in our patients; however, there is no evidence to suggest its mechanism to date, thereby requiring further discussion. By administering small doses of long-acting dopamine agents, dopamine receptor agonists, α2δ calcium channel ligands, clonazepam, and other treatment to this group of patients, the IRLS score gradually declined and symptoms improved, but RLS symptoms did not completely disappear. Here, the prognosis of RLS was significantly different from secondary RLS patients, as they typically demonstrated complete disappearance of symptoms after cause elimination. It is suggested that long-term high doses of Madopar cause excessive accumulation of dopamine in the central nervous system, thereby irreversibly decreasing the number of dopamine receptors or DAT function, resulting in the persistence of clinical RLS-like symptoms.
The present study reports drug-induced (high-dose levodopa) RLS, which is different from the idiopathic RLS in treatment; however, the method treatment is identical. First, the dose of levodopa was gradually reduced, but as the clinical symptoms of the patients were severe, measures needed to be taken accordingly, depending on the patient's condition and their accompanying anxiety, depression, and insomnia. We referenced the European guidelines on management of restless legs syndrome:report of a joint task force by the European Federation of Neurological Societies, the European Neurological Society and the European Sleep Research Society ([@B36]). According to the recommendations and precautions for drug treatment, the principle of individualization was followed, simultaneously supplemented by physical therapy (hot water bath before sleep, limb massage, etc). Thus, the clinical symptoms of the nine patients were relieved to varying degrees. The specific method recommends the minimization and withdrawal of the use of Madopar. The therapy was changed to a small dose of a long-acting dopamine agent (Xining 100--200 mg/day) to reduce the risk of dosage increase. Madopar was finally discontinued in five of the nine patients; however, Case 1 was found to have difficulty when the Madopar dose was reduced to 3 pills/day. A small dose of a dopamine receptor agonist and a small-to-moderate dose of a α2δ calcium channel ligand (pregabalin) were added depending on the severity of the patient's symptoms. To improve the symptoms of anxiety, depression, and insomnia, the therapy was supplemented by small-dose paroxetine, duloxetine, and olanzapine; an obvious curative effect was achieved by all of them.
In summary, clinicians should be mindful of differential diagnoses when patients present with walking difficulties and limb stiffness. In addition to the original diseases, anxiety, depression, and somatization disorders should be considered. In particular, clinicians should strengthen the management of patients who use dopaminergic agents to reduce the great physical and mental adverse events due to misdiagnosis and mistreatment.
Consent for Publication {#s5}
=======================
The nine patients gave written consent for both participation and publishing the data in a scientific journal. They understood that the information will be published without their names attached, but that full anonymity cannot be guaranteed. They understood that the material may be published and placed on worldwide website and journals. Both the printed version and the website are seen and read by doctors, journalists, and members of the public. The material will not be used for advertising or packaging.
Ethics Statement {#s6}
================
This study was approved by the Ethics Committee of the First People Hospital of Huainan, and written informed consent was obtained from the patient for publication of this case report.
Author Contributions {#s7}
====================
LZ, MZ, WZ: study design and critical revision of the manuscript. LZ, MZ: collection and interpretation of data. LZ: analysis data and drafting of the manuscript. JL, CR, MX, CY: collection data. All authors approved the final version for publication.
Funding {#s8}
=======
This work was supported by the Chinese Academy of Medical Sciences Innovation Fund for Medical Sciences (2016-I2M-1-006).
Conflict of Interest Statement {#s9}
==============================
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
[^1]: Edited by: Chunxue Wang, Beijing Tiantan Hospital, China
[^2]: Reviewed by: Li Cao, Shanghai Jiao Tong University, China; Chun-Feng Liu, Second Affiliated Hospital of Soochow University, China
[^3]: This article was submitted to Mood and Anxiety Disorders, a section of the journal Frontiers in Psychiatry
| {
"pile_set_name": "PubMed Central"
} |
The authors confirm that all data underlying the findings are fully available without restriction. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium (<http://www.proteomexchange.org>) via the PRIDE partner repository with the dataset identifier PXD000778, DOI:10.6019/PXD001269.
Introduction {#s1}
============
Proper temporal organization of behavioral, physiological and biochemical processes and their synchronization with the environmental light/dark cycle are fundamental features of most organisms [@pgen.1004695-Reppert1]. In mammals, the central pacemaker that coordinates this adaptive response to temporal cues resides in the suprachiasmatic nucleus (SCN) of the anterior hypothalamus [@pgen.1004695-Moore1], [@pgen.1004695-Stephan1]. The SCN is uniquely positioned to receive light signals from the retina and to relay the time information to peripheral clocks via synaptic and humoral mechanisms. Both central and peripheral clocks use a series of autoregulatory transcription-translation feedback loops to drive cell-autonomous, circadian (∼24 h) rhythms of gene expression of core clock components as well as tissue-specific, clock-controlled outputs [@pgen.1004695-Reppert1].
In order to gain a bird\'s eye view of circadian regulation, a number of gene expression profiling studies using microarrays have been done to examine the circadian transcriptome of the murine SCN and liver [@pgen.1004695-Panda1], [@pgen.1004695-Ueda1]. However, numerous studies have shown that transcript levels are not necessarily reliable predictors of protein abundance, and thus of functional outcome [@pgen.1004695-Greenbaum1]. A previous attempt at elucidating the circadian proteome, using two-dimensional differential gel electrophoresis (2D-DIGE) combined with mass spectrometry (MS), identified 34 rhythmically expressed proteins within the SCN [@pgen.1004695-Deery1]. Until now, the small size of the SCN, and the limited amounts of protein that can be extracted from it, has posed a significant challenge to acquiring accurate and comprehensive quantitative proteomics data. However, recent technological advances in MS-based quantitation, and our growing awareness of the importance of post-transcriptional regulation of circadian rhythms [@pgen.1004695-Lim1], encouraged us to re-evaluate SCN functions from a proteomic perspective. In our previous study, we employed the AutoProteome system in conjunction with spectral counting to identify 2131 unique proteins in the SCN, of which 387 were acutely up- or down-regulated following nocturnal light exposure [@pgen.1004695-Tian1].
In this report, we performed an unbiased interrogation of the SCN proteome over a 24 h cycle using an alternative MS approach: a centrifugal proteomic reactor (CPR) coupled with stable isotope labeling by amino acids in cell culture (SILAC)-based quantitation. Neuro2A murine neuroblastoma cells were used as the SILAC-labeled internal reference standard for murine SCN tissues, based on reports that numerous core clock genes are expressed in this cell line, and that serum shock can induce robust circadian oscillations of their transcripts [@pgen.1004695-Chang1], [@pgen.1004695-Musiek1]. Furthermore, a previous study indicated \>97% overlap between the proteomes of mouse whole brain and Neuro2A cells [@pgen.1004695-Ishihama1]. Our proteomics screen identified a total of 3275 unique proteins in the murine SCN, 421 of which displayed time-of-day-dependent expression profiles. Within this smaller subset, 48 proteins fluctuated in a circadian manner. Bioinformatics analyses of these 421 proteins highlight the potentially important role that post-transcriptional mechanisms such as miRNAs may play in shaping the final profile of the time-of-day proteome, and the orchestrated expression of multiple proteins involved in neurosecretory processes and mitochondrial oxidative phosphorylation within the SCN.
Results {#s2}
=======
Interrogation of the SCN Proteome {#s2a}
---------------------------------
To examine the murine SCN proteome, we stably entrained male C57Bl/6J mice to a 12 h light∶12 h dark (LD) cycle, and transferred them to constant darkness (DD) for two days. Starting at circadian time (CT) 2 on the third day of DD, we collected SCN tissues from four mice at 4 h intervals for a full circadian cycle ([Figure 1A](#pgen-1004695-g001){ref-type="fig"}). SCN samples were processed individually to yield four independent biological replicates for each time point (4 mice per CT, n = 24 total mice). SCN protein lysates (30 µg) were mixed with equal quantities of lysates prepared from Neuro2A cells that had previously been cultured for \>10 passages in heavy SILAC medium. Under these conditions, the Neuro2A proteome is estimated to be \>98% heavy SILAC-labeled and thus useful as a spike-in reference standard. The CPR, with its superior recovery of hydrophobic membrane proteins compared with other approaches [@pgen.1004695-Zhou1], was used for rapid protein preconcentration, derivatization, enzymatic digestion, and fractionation of the samples. Peptides were eluted from the CPR in ten fractions and analyzed by high performance liquid chromatography-electrospray tandem mass spectrometry (HPLC-ESI-MS/MS) in a total of 240 runs. From the raw mass spectrometric data, Maxquant and Andromeda identified 3275 protein groups with a false discovery rate (FDR) of 1% ([Table S1](#pgen.1004695.s005){ref-type="supplementary-material"}). Out of these 3275 proteins, only 7 lacked a corresponding SILAC-labeled peak, indicating that these proteins are expressed in the SCN but not in Neuro2A cells ([Table S1](#pgen.1004695.s005){ref-type="supplementary-material"}). As expected from recent proteomic studies [@pgen.1004695-Robles1], [@pgen.1004695-Mauvoisin1], our MS screen failed to detect any core clock proteins, likely due to their low abundance relative to the many cytoplasmic proteins which were detected. The raw dataset was further filtered for proteins that were identified by a minimum of two peptide ratio counts and where accurate quantification values were obtained in a minimum of 12 out of 24 independent samples. Downstream bioinformatics and statistical analyses were performed on this stringently filtered dataset of 2112 proteins (64.5% of total identified proteins), hereafter referred to as the SCN proteome ([Table S2](#pgen.1004695.s006){ref-type="supplementary-material"}). High r values (between 0.83 and 0.97, [Table S3](#pgen.1004695.s007){ref-type="supplementary-material"}) were obtained for the pairwise Pearson\'s correlation analysis of 24 measurements, indicating excellent reproducibility within our SCN proteome data.
{#pgen-1004695-g001}
To identify proteins whose expression significantly fluctuated as a function of time-of-day, we subjected the SCN proteome to an analysis of variance. ANOVA revealed that 421 proteins (i.e., 20% of the SCN proteome) exhibited statistically significant (p\<0.05) alterations in abundance across the 24 h cycle. This significantly altered proteome, hereafter referred to as the time-of-day proteome, was evaluated for reproducibility using pairwise Pearson\'s correlation analysis of the 24 samples ([Table S4](#pgen.1004695.s008){ref-type="supplementary-material"}). The r values between biological replicates at a specific CT were extremely high ([Figures 1B and 1C](#pgen-1004695-g001){ref-type="fig"}) compared with the lower r values observed between samples of different CTs ([Figure 1D](#pgen-1004695-g001){ref-type="fig"}), indicating a high degree of reproducibility within biological replicates. As a second independent measure of variability, we calculated the relative standard deviation (RSD) for all proteins at every CT. The median RSD for the time-of-day proteome was 16%, compared with 11% from a previous study of the circadian hepatic proteome [@pgen.1004695-Robles1]. Considering that their study [@pgen.1004695-Robles1] utilized single pooled tissue samples per CT per 24 h cycle (2 cycles were analyzed), while ours uses unpooled samples from 4 animals per CT, an RSD of 16% reflects an acceptable level of variability within our dataset. Lastly, the temporal profiles of 12 proteins ([Figure 2](#pgen-1004695-g002){ref-type="fig"}) with quantification values in all 24 samples also showed a higher degree of correlation within a CT than between CTs. The expression of five of these proteins---endophilin A1 (SH3GL2), synaptobrevin 2 (VAMP2), serine/threonine-protein kinase PAK 1 (PAK1), synaptotagmin 1 (SYT1), and synaptic vesicle glycoprotein 2A (SV2A)---was evaluated by Western blot (WB) analysis using independent batches of SCN tissues ([Figure 2](#pgen-1004695-g002){ref-type="fig"}). In all cases, the WB results correlated well with our MS-based quantification (Pearson\'s coefficients = 0.78 \[SH3GL2\], 0.69 \[VAMP2\], 0.75 \[PAK1\], 0.77 \[SYT1\], 0.66 \[SV2A\]).
{#pgen-1004695-g002}
Next, the time-of-day proteome was subjected to Gene Ontology (GO) analysis by DAVID [@pgen.1004695-Jiao1] in order to investigate its biological relevance. Relative to both the SCN proteome and the entire mouse genome (by DAVID), the time-of-day proteome was significantly enriched for GO total cellular components that were classified as mitochondrion or mitochondrial membrane (Fisher\'s exact test, p\<0.05) ([Figure 1E](#pgen-1004695-g001){ref-type="fig"}). Additionally, several metabolic pathways including generation of precursor metabolites and energy, oxidation reduction, energy derivation by oxidation of organic compounds, and cellular respiration were significantly enriched in this dataset based on GO total biological process analysis ([Figure 1F](#pgen-1004695-g001){ref-type="fig"}). A more in-depth examination using GO FAT analysis, which filters out very broad GO terms based on a measured specificity of each term, confirmed that a significant portion of the time-of-day proteome was associated with the mitochondrion, energy generation and consumption, and hydrogen ion transmembrane transporter activity ([Figure S1](#pgen.1004695.s001){ref-type="supplementary-material"}).
Hierarchical clustering of the time-of-day proteome revealed segregation of proteins into six different expression clusters ([Figure 3A](#pgen-1004695-g003){ref-type="fig"}, [Table S5](#pgen.1004695.s009){ref-type="supplementary-material"}). Two dominant clusters emerged from the analysis: cluster D (118 proteins) and cluster E (189 proteins). Closer examination of the protein levels within each cluster ([Figure 3B](#pgen-1004695-g003){ref-type="fig"}) revealed that clusters D and E are mirror images of one another. A second interesting feature of clusters D and E is their bimodal expression profile. For instance, cluster E is characterized by an increase in expression from CT 2 to CT 10, a sharp decrease from CT 10 to CT 14 (the anticipated light-to-dark transition), a gradual increase through the night (CT 14 to CT 22), followed by an abrupt decrease from CT 22 to CT 2 (the anticipated dark-to-light transition). Interestingly, when compared to the time-of-day proteome, cluster E was selectively enriched for several GO biological processes such as generation of precursor metabolites and energy, cellular respiration, and energy derivation by oxidation of organic compounds ([Figure 3C](#pgen-1004695-g003){ref-type="fig"}). Collectively, the data reveal that a substantial portion of the SCN proteome (20%) exhibits significant changes in abundance as a function of time-of-day.
{#pgen-1004695-g003}
The Murine SCN Proteome: Existence of Ultradian Components and Relationship with the SCN Transcriptome {#s2b}
------------------------------------------------------------------------------------------------------
Next, we sought to identify proteins that exhibited a strictly circadian pattern of expression, as well as those that were ultradian. To this end, we employed the JTK_CYCLE algorithm [@pgen.1004695-Hughes1], [@pgen.1004695-Hughes2] to identify the subsets of proteins within the time-of-day proteome that oscillated with periods of 8, 12 and 24 h. JTK_CYCLE was recently developed to detect rhythmic components within large genomic datasets, and is superior to other similar algorithms in its sensitivity, specificity and efficiency [@pgen.1004695-Hughes1]. Recent studies have also used JTK_CYCLE to analyze the circadian acetylome and the diurnal transcriptome of the murine liver and heart, respectively [@pgen.1004695-Masri1], [@pgen.1004695-Podobed1]. Given that the observed free-running period of C57BL6 mice is ∼23.6 to 23.8 h, we used the nearest integer value (24) to approximate a circadian cycle using JTK_CYCLE. Based on a p-value cutoff of 0.05 [@pgen.1004695-Hughes2], 48 proteins were deemed to be circadian, with phase of peak expression distributed across the entire 24 h cycle ([Figure S2C](#pgen.1004695.s002){ref-type="supplementary-material"}, [Table S6](#pgen.1004695.s010){ref-type="supplementary-material"}). Surprisingly, a relatively large proportion of the time-of-day proteome exhibited ultradian periods of 8 h (25 proteins) and 12 h (59 proteins) (p\<0.05, JTK_CYCLE, [Figures S2A and S2B](#pgen.1004695.s002){ref-type="supplementary-material"}, [Table S7](#pgen.1004695.s011){ref-type="supplementary-material"}). Those 12 h rhythmic proteins tended to peak in expression at either the early day and early night, or late day and late night ([Figure S2B](#pgen.1004695.s002){ref-type="supplementary-material"}), mirroring the profiles of clusters D and E ([Figure 3B](#pgen-1004695-g003){ref-type="fig"}), respectively. Moreover, subjecting the larger dataset of the SCN proteome to JTK_CYCLE analysis resulted in the assignment of an additional 11, 41 and 43 proteins as rhythmic with periods of 8, 12 and 24 h, respectively. The fact that these proteins were identified as rhythmic by JTK_CYCLE but were not significantly altered based on ANOVA suggests that they might exhibit weak fluctuations that are mistaken as rhythmic. Thus, we focused subsequent downstream analyses on those circadian and ultradian proteins that were identified within the time-of-day proteome rather than the SCN proteome. Our results are somewhat reminiscent of the findings of Hughes et al. [@pgen.1004695-Hughes2], which identified clusters of transcripts that cycled at the second and third harmonics of circadian rhythmicity in the murine liver; however, in that study the transcripts exhibiting these subharmonics accounted for only a small fraction of the entire rhythmic transcriptome.
To investigate the relationship between transcript levels and protein abundance, we compared our time-of-day, circadian and ultradian (8 h and 12 h) proteomes with mRNA data extracted from two published microarray studies (MAS4 Panda et al. and gcrma Panda et al.) of the mouse SCN transcriptome from the CIRCA database (<http://bioinf.itmat.upenn.edu/circa/>) using identical JTK_CYCLE filtering criteria (p\<0.05, 0 to 40 h period rhythmicity) ([Figures 4A--D](#pgen-1004695-g004){ref-type="fig"}). Notably, \>40% of each proteome was encoded by non-rhythmic transcripts. Circadian transcripts were found to encode a smaller subset of proteins within the time-of-day, circadian and 12-h ultradian proteomes. Each proteome also consisted of a subset of proteins that were encoded by rhythmic (non-24 h) transcripts cycling at intervals of 16, 20, 28 or 32 h. Notably, none of the 8 h and 12 h rhythmic proteins were encoded by transcripts that oscillated with the same period. Moreover, there were only 9 genes that exhibited circadian rhythms at both the transcript and protein level, and even amongst most of these there was a significant time lag (mean ∼8 h) between the peak in expression of the mRNA and the protein ([Figure 4E](#pgen-1004695-g004){ref-type="fig"}).
{#pgen-1004695-g004}
Collectively, our data indicate that transcript levels are a generally poor predictor of protein abundance in the murine SCN. By extension, this suggests that post-transcriptional mechanisms play a dominant role in shaping the ultimate landscape of the SCN proteome. Another key observation from our study is that, for a substantial portion of the time-of-day proteome, the anticipated light-to-dark and dark-to-light transitions trigger robust changes in protein abundance that are similar in direction (either up- or down-regulated).
The Potential Role of microRNAs in the Time-of-Day-Dependent SCN Proteome {#s2c}
-------------------------------------------------------------------------
microRNAs (miRNAs) are small (∼22--24 nt), noncoding RNAs that act as potent post-transcriptional modulators of gene expression. Various miRNAs have been implicated in the regulation of circadian rhythms in multiple model systems [@pgen.1004695-Mehta1], [@pgen.1004695-Cheng1]. To examine a possible involvement of miRNAs in shaping the SCN proteome, we first asked whether predicted targets of known miRNAs were enriched within particular hierarchical clusters of our time-of-day proteome. To this end, we compared the time-of-day proteome with the predicted targets of 86 broadly conserved miRNA families extracted from the TargetScanMouse version 6.2 database (<http://www.targetscan.org/mmu_61/>). Out of 86 broadly conserved miRNA families examined, only miR-133ab showed a significant enrichment of its respective targets in at least one hierarchical cluster. [Figure 5A](#pgen-1004695-g005){ref-type="fig"} illustrates the number of predicted miR-133ab target genes within each hierarchical cluster based on the presence of at least one conserved site within the annotated 3\' untranslated region (UTR) of the transcript. Compared to other hierarchical clusters, cluster E had the largest number of, and was statistically enriched for, predicted miR-133ab targets (Fisher\'s exact test, *p*\<0.05; [Figure 5A](#pgen-1004695-g005){ref-type="fig"}).
{#pgen-1004695-g005}
As miR-133ab levels have been reported to be low in the brain [@pgen.1004695-Heyer1], we used an ultra-sensitive qRT-PCR approach to quantify levels of mature miR-133a and -133b in the murine SCN ([Figure 5B](#pgen-1004695-g005){ref-type="fig"}). miR-133b abundance was elevated throughout the subjective night, whereas miR-133a levels peaked sharply at CT 14 ([Figure 5B](#pgen-1004695-g005){ref-type="fig"}). Notably, the expression profile of miR-133ab showed an inverse trend when compared with the MS-quantified expression of their predicted target genes in cluster E from CT 10 to CT 22 ([Figure 5C](#pgen-1004695-g005){ref-type="fig"}).
To provide functional evidence that these are authentic targets of miR-133ab, we selected three predicted targets within cluster E and examined their expression in Neuro2A cells in which levels of miR-133a or miR-133b have been enhanced using microRNA mimics, or suppressed by microRNA inhibitors. Either one or both mimics of miR-133a and miR-133b strongly suppressed the levels of SH3GL2, SYT1 and SV2A proteins in transfected Neuro2A cells compared with controls ([Figures 5D--F](#pgen-1004695-g005){ref-type="fig"}). On the other hand, silencing of miR-133a or miR-133b robustly elevated the expression of SH3GL2 but not SYT1 or SV2A ([Figures 5D--F](#pgen-1004695-g005){ref-type="fig"}). These data are consistent with SH3GL2, SYT1 and SV2A being authentic targets of miR-133ab.
Collectively, our results raise the possibility that miRNAs, such as miR-133ab, are orchestrating the temporal profiles of multiple target genes as suggested previously [@pgen.1004695-AlvarezSaavedra1]. Given the fact that predicted miR-133ab targets are not solely restricted to a single hierarchical cluster with a common temporal expression pattern, other miRNAs are likely to work in concert with miR-133ab to fine-tune the temporal expression profile of its targets. More generally, our data suggest that miRNAs may be key post-transcriptional regulators of time-of-day-dependent protein expression within the SCN.
Pathway Analysis of the SCN Proteome {#s2d}
------------------------------------
The concerted expression of a large number of proteins within each hierarchical cluster suggests that many of these proteins might interact directly with one another to modulate a shared set of biological responses. To test this hypothesis, we used the IPA software to perform a very restrictive interaction analysis using only interactions from public repositories and limiting the network to direct interactions between proteins within a single cluster (i.e., no neighbors). Proteins in cluster E exhibited a high degree of connectivity (35 out of 189 proteins, or 18.5%) through direct protein-protein interactions ([Figure 6](#pgen-1004695-g006){ref-type="fig"}). The top functions within this network were neurological disease and psychological disorders. Notably, this network included a relatively large number of proteins that are involved in neurotransmitter release (i.e., NSF, SV2A, SYT1 and VAMP2) and synaptic transmission (i.e., CNP, NSF, SV2A, SH3GL2 and SYT1) ([Figure 6](#pgen-1004695-g006){ref-type="fig"}). A more comprehensive functional protein interaction network analysis of the entire time-of-day proteome (421 proteins) revealed that one of the largest networks is driven from the interactions among proteins involved in protein trafficking and carbohydrate metabolism ([Figure S3](#pgen.1004695.s003){ref-type="supplementary-material"}). This network cluster included 9 proteins that oscillated with 12-h rhythms. The larger protein interaction network also included proteins involved in more general cellular functions, such as cellular assembly and organization, cellular function and maintenance, and cell morphology ([Figure S3](#pgen.1004695.s003){ref-type="supplementary-material"}).
{#pgen-1004695-g006}
We further mapped the time-of-day proteome onto 192 known canonical pathways using IPA to identify pathways that might be significantly impacted ([Figure S4](#pgen.1004695.s004){ref-type="supplementary-material"}). Three canonical pathways that had previously been implicated in the regulation of the SCN clock [@pgen.1004695-Obrietan1]--[@pgen.1004695-Ginty1] were identified in our IPA analysis: Ca^2+^/cAMP response element binding protein (CREB) signaling in neurons (p = 0.0031); extracellular signal-regulated kinase (ERK)/mitogen- activated protein kinase (MAPK) signaling pathway (p = 0.0019); and synaptic long-term potentiation pathway (p = 0.0035).
In agreement with our previous GO analysis, the top-ranked canonical pathway was the mitochondrial dysfunction pathway, with 29 out of 421 proteins mapped. To delve further into the potential implications of an apparent temporal regulation of the mitochondria, we performed a KEGG pathway enrichment analysis by DAVID to compare our time-of-day proteome against all 2112 quantified proteins in the SCN proteome. Our data reveal that the KEGG pathways for Huntington\'s disease, Parkinson\'s disease, oxidative phosphorylation, pyruvate metabolism, and arginine and proline metabolism were significantly enriched with at least 8 proteins mapped in each pathway ([Figure 7A](#pgen-1004695-g007){ref-type="fig"}). Indeed, many of the proteins within the Huntington\'s disease, Parkinson\'s disease, and oxidative phosphorylation (OxPhos) pathways overlapped and were mitochondrial in their localization. Particularly noteworthy was the OxPhos pathway, which accounted for 25 proteins within the time-of-day proteome. Of these 25 OxPhos-related proteins, 22 were present in cluster E and thus exhibited a similar trend in expression profile. Results from KEGG analysis mirrored those from canonical pathway analysis by IPA, which identified 19 OxPhos-related proteins (a subset of the 25), with 16 of these belonging in cluster E ([Figure 7B](#pgen-1004695-g007){ref-type="fig"}). One of these OxPhos proteins, NDUFA10, a subunit of NADH∶ubiquinone oxidoreductase (complex I), was selected for validation by IF. NDUFA10 immunoreactivity within the SCN exhibited a pronounced increase at the CT10-CT14 transition, in keeping with the MS results at these two time points ([Figure 7C](#pgen-1004695-g007){ref-type="fig"}).
{#pgen-1004695-g007}
The time-of-day proteome is significantly enriched for mitochondrial proteins (111/421, or 26.4%) relative to the SCN proteome (2112 proteins). Further analysis revealed that 51.4% of these mitochondrial proteins belonged within Cluster E, although this "cluster bias" did not reach statistical significance ([Figure 7D](#pgen-1004695-g007){ref-type="fig"}). However, when we restricted our analysis to mitochondrial proteins that were involved in the OxPhos pathway, we found significant enrichment within cluster E. Our collective data point to a hitherto unappreciated temporal regulation of mitochondrial functions, particularly oxidative phosphorylation, within the central pacemaker.
Discussion {#s3}
==========
Despite recent advances in quantitative proteomics and their application to the study of clock-controlled processes in the liver [@pgen.1004695-Robles1], [@pgen.1004695-Mauvoisin1], [@pgen.1004695-Masri1], the SCN proteome has been challenging to characterize in a comprehensive manner due to its inherently low sample availability. A previous attempt, using 2-dimensional difference gel electrophoresis (2D-DIGE) coupled with MS for protein identification, uncovered 115 proteins with time-of-day-dependent expression, of which 34 were circadian, out of 871 protein spots detected [@pgen.1004695-Deery1]. In our present study, we took advantage of the quantitative accuracy of SILAC, and combined it with the enhanced detection sensitivity that is achieved using the CPR, to provide a large-scale interrogation of the SCN proteome ([Figure 1A](#pgen-1004695-g001){ref-type="fig"}). The outcome was the identification of 421 and 48 proteins whose expression profiles were time-of-day-dependent and circadian, respectively, from a stringently quantified dataset of 2112 proteins. In contrast with recently published liver circadian proteome studies [@pgen.1004695-Robles1], [@pgen.1004695-Mauvoisin1], we used 4 independent biological replicates to represent each CT across one 24 h cycle, rather than a single (pooled) sample for each time point across two cycles. Despite these procedural differences, the percentage of the detected proteome that exhibited circadian (24 h) rhythms was reasonably similar in all three studies (2.2%, our study; 6.0% Robles et al. [@pgen.1004695-Robles1]; 4.8% Mauvoisin et al. [@pgen.1004695-Mauvoisin1]). Furthermore, based on our proteomic analysis of the SCN, we were able to conclude that transcript expression is a relatively poor predictor of protein abundance, that ultradian rhythms in protein expression are prevalent in the SCN, and that the mitochondria, in particular oxidative phosphorylation, is a major target of temporal control in the central pacemaker. Additionally, our findings support the argument that post-transcriptional mechanisms, including miRNAs, may play a prominent role in shaping the ultimate landscape of the SCN proteome.
Comparisons with the Deery et al. report revealed that our SCN proteome (2112 proteins) included 30 of their 115 time-of-day-dependent proteins. Eleven out of these 30 were found within our time-of-day proteome and two belonged in our circadian proteome (p\<0.05, JTK analysis). The higher frequency in sampling intervals in our study compared to theirs (4 h vs. 6 h) along with different statistical algorithms employed may partially account for the differences between the two datasets. Nevertheless, both studies identified vesicular trafficking and neurosecretory processes as novel points of temporal regulation. Interestingly, a number of proteins involved in neurotransmitter release (i.e., NSF, SV2A, SYT1 and VAMP2) and synaptic transmission (i.e., CNP, NSF, SV2A, SH3GL2 and SYT1) shared similar expression trends (cluster E), exemplifying how the SCN might coordinate the expression of proteins that act within the same biological pathway.
One unexpected observation was the sharp, and parallel, changes in protein expression at the anticipated transitions from light-to-dark and dark-to-light. This pattern was prevalent not only in the 12 h ultradian proteome ([Figure S2B](#pgen.1004695.s002){ref-type="supplementary-material"}) but also in the time-of-day proteome (clusters D and E, [Figure 3B](#pgen-1004695-g003){ref-type="fig"}). Hughes et al. [@pgen.1004695-Hughes2] made similar observations in analyzing the 12 h cycling transcripts of the liver. They speculated that these dawn- and dusk-peaking transcripts took part in cellular mechanisms that anticipated the stress of daily transitions into light and darkness. Such an explanation might also be pertinent to our study. In particular, changes in levels of oxidative phosphorylation might constitute one such cellular response to light-dark/dark-light (or rest-wake/wake-rest) state transitions ([Figure 7B](#pgen-1004695-g007){ref-type="fig"}). The mechanisms that drive these bimodal expression profiles within the SCN are unknown, but could conceivably involve interactions between the molecular clock and systemic cues. Alternatively, as shown by Westermark and colleagues [@pgen.1004695-Westermark1], 12 h rhythms may arise theoretically through the actions of pairs of circadianly expressed transcription factors with defined phase relationships relative to each other.
Our study adds to the growing body of evidence that post-transcriptional regulation is a key feature of central and peripheral clocks. Less than 60% of the circadian and ultradian proteomes were encoded by transcripts that were rhythmic. In the case of the 8 h and 12 h proteomes, none of those rhythmic transcripts exhibited the same period as the encoded protein. Furthermore, for the 9 genes that were circadian at both the mRNA and protein level, there was a great disparity between the phase of peak expression of the transcript vs. that of the protein (mean time difference of ∼8 h). From these collective data, one can infer that post-transcriptional (and post-translational) regulation has a major influence on protein expression in the SCN. Specific post-transcriptional events, such as miRNA regulation, may intersect with circadian transcriptional rhythms (or even constitutive gene transcription) to establish the final time-course and profile of protein expression. Along these lines, we provide evidence that a microRNA, miR-133ab, has the ability to regulate the expression of various target proteins belonging to cluster E. Furthermore, we noted an inverse trend between miR-133ab levels and expression of their *in silico* targets in the SCN.
Another interesting finding is the prevalence of ultradian proteins in the SCN. Hughes et al. [@pgen.1004695-Hughes2] also discovered circadian harmonics within his rhythmic transcriptome, but they occurred with a much lower frequency relative to the circadian transcripts. The mechanisms that are driving a larger subset of proteins towards ultradian expression are unclear, but could conceivably involve one or several post-transcriptional mechanisms.
Finally, our results strongly support the notion that mitochondrial energetics within the SCN is under strict temporal control ([Figure 7A](#pgen-1004695-g007){ref-type="fig"}). While recent studies emphasize the circadian control of mitochondrial metabolism in the liver [@pgen.1004695-Masri1], [@pgen.1004695-Peek1], a key organ for energy storage and mobilization, almost none have directly studied mitochondrial function in the SCN. However, recent observations that the NAD^+^-dependent deacetylase SiRT1 positively modulates the expression of CLOCK and BMAL1 in the SCN [@pgen.1004695-Chang1], and that resveratrol activates SiRT1 through an increase in mitochondrial complex I-dependent NADH oxidation [@pgen.1004695-DesquiretDumas1], raise the possibility that mitochondrial metabolism, in particular oxidative phosphorylation, may play a prominent role in maintaining robustness of the SCN clock. Interestingly, others have noted that hepatic NAD^+^ levels exhibit a bimodal rhythm and attributed this entirely to its biosynthesis by the enzyme nicotinamide phosphoribosyltransferase [@pgen.1004695-Ramsey1]. Our proteomics data, which highlight the bimodal expression of a large number of OxPhos-related proteins, provide an additional mechanism by which NAD^+^ levels may be shaped in the SCN.
In conclusion, our study provides a broader perspective on the temporal control of the SCN proteome. Our results underscore the significance of post-transcriptional regulation, the surprising prevalence of ultradian protein expression, and the functional implications on mitochondrial energy metabolism. Future investigations should help to clarify how each of these aspects contributes to the central pacemaker function of the SCN.
Materials and Methods {#s4}
=====================
Ethics Statement {#s4a}
----------------
All animal handling and experimental procedures were conducted at the University of Toronto Mississauga animal facility, and were approved by the local animal care committee in compliance with institutional guidelines and the Canadian Council on Animal Care.
Reagents {#s4b}
--------
Ammonium bicarbonate (NH~4~HCO~3~), dithiothreitol (DTT), iodoacetamide (IAA), citric acid, and urea were obtained from Sigma-Aldrich (Saint Louis, MO). Acetonitrile, with 0.1% formic acid, and water, with 0.1% formic acid, were purchased from J.T. Baker (Phillipsburg, NJ). Trypsin was purchased from Promega (Madison, WI). Strong cation exchange (SCX) beads were obtained from Polymer Laboratories, Varian, Inc. (Palo Alto, CA). CHAPS (3-\[(3-cholamidopropyl)dimethylammonio\]-1- propanesulfonate, BP 571), ammonium hydroxide (NH~4~OH), and methanol were purchased from Fisher Scientific (Hampton, NH).
Animals {#s4c}
-------
Eight- to 12-week-old male C57BL6/J mice that were obtained from in-house breeding or purchase from The Jackson Laboratory (Bar Harbor, ME) were used for all experiments. Mice were group-housed in polycarbonate cages and given *ad libitum* access to rodent chow and water throughout the study.
Tissue Harvest {#s4d}
--------------
Mice were stably entrained for a minimum of 2 weeks to a 12 h light∶12 h dark (LD) schedule (light intensity during the L phase was ∼200 lux) prior to transfer to complete darkness (DD) for 2 full cycles. Dark adaptation was achieved by placing cages into light-tight ventilated cabinets. On day 3 of DD, mice were killed at 4-h intervals at circadian times (CT) 2, 6, 10, 14, 18, and 22, where CT corresponds to the Zeitgeber time (ZT) of the previous LD cycle. Mice were killed by cervical dislocation and decapitated, and eyes were covered with black electrical tape under dim red light. Brains were dissected and cut into 800-µm thick coronal sections containing the SCN in cooled oxygenated media using an oscillating tissue slicer [@pgen.1004695-AlvarezSaavedra1]. SCN was isolated from the tissue slice using a razor blade, and frozen immediately on dry ice. For IF, the entire coronal slice was fixed in 4% (w/v) paraformaldehyde in phosphate-buffered saline (PBS; 6 h, room temperature), cryoprotected in 30% (w/v) sucrose (overnight, 4°C), and cut into thin sections (30-µm) using a freezing microtome.
Cell Culture {#s4e}
------------
Neuro2A cells (American Type Culture Collection \[ATCC\], Manassas, VA) were grown in customized DMEM (AthenaES, Baltimore, MD, USA) and supplemented with \[^13^C~6~,^15^N~4~\]-L-Arginine (Arg-10), \[^13^C~6~,^15^N~2~\]-L-Lysine (Lys-8) at Arg 42 mg/L, Lys 146 mg/L, Met 30 mg/L and supplemented with 10% (v/v) dialyzed FBS (GIBCO-Invitrogen; Burlington, ON, Canada), 1 mM sodium pyruvate (Gibco-Invitrogen) and 28 µg/mL gentamicin (Gibco-Invitrogen). Labeled amino acids are purchased from Sigma-Aldrich (Oakville, ON, Canada). Cells were maintained in culture for at least 10 doubling times to allow for complete (\>98%) incorporation of the isotope-labeled amino acids into the cells.
SCN Proteomic Analysis by Centrifugal Proteomic Reactor {#s4f}
-------------------------------------------------------
SCN tissues of individual mice were homogenized in 80 µL lysis buffer (8M urea, 4% CHAPS, 100 mM NH~4~HCO~3~ with fresh proteinase inhibitor mixture) in a 1.5 mL Pellet pestle, and sonicated 3 times for 10 s each with \>30 s on ice between each pulse. Protein concentration was determined by the Bradford method. Proteins were processed in a centrifugal proteomic reactor device as previously described with some modifications [@pgen.1004695-Zhou1]. Briefly, lysates from SCN tissues and heavy SILAC-labeled Neuro2A cells were mixed 1∶1 (30 µg protein each), and vortexed (1 min) vigorously in the presence of 30 µL of SCX slurry and 1.2 mL of 5% formic acid. Samples were centrifuged (16,100*×g*, 3 min), and SCX bead pellets were washed twice with 1.2 mL 0.1% formic acid. Proteins were reduced by incubating samples in the presence of 20 µL of 150 mM NH~4~HCO~3~, 20 mM DTT (1200 rpm, 56°C, 15 min), and were subsequently alkylated by the addition of 20 µL of 150 mM NH~4~HCO~3~, 100 mM IAA (room temperature, 15 min in darkness). The reaction was stopped by adding 1.2 mL of 0.1% formic acid supplemented with 6 µg trypsin. Following centrifugation, the SCX bead pellet was resuspended in 40 µL of 1 M NH~4~HCO~3~ and trypsin digested for 4 h (37°C, 1200 rpm). Finally, pH step elution of the peptides from the SCX beads was performed by adding: 1.2 ml of 0.1% formic acid pH 2.5, followed by additional nine pH fractions (pH 3.0, 3.5, 4.0, 4.5, 5.0, 5.5 6.0, 8.0, 12) by subsequent additions of 400 µL 10 mM citric acid/NH~4~OH pH buffers. The fractionated samples were desalted using in-house-made C~18~ desalting cartridges, and dessicated using a Speed-Vac prior to LC-MS measurement.
LC-MS Measurement {#s4g}
-----------------
All resulting peptide fractions were analyzed by HPLC-ESI-MS/MS, which consists of an automated Agilent 1100 micro-HPLC system (Agilent Technologies, Santa Clara, CA) coupled with an LTQ-Orbitrap mass spectrometer (ThermoFisher Scientific, San Jose, CA) equipped with a nano-electrospray interface operated in positive ion mode. Each peptide mixture was reconstituted in 20 µL of 0.5% (v/v) formic acid, and 10 µl was loaded on a 200 µm×50 mm fritted fused silica pre-column packed in-house with reverse phase Magic C18AQ resins (5 µm; 200-Å pore size; Dr. Maisch GmbH, Ammerbuch, Germany). The separation of peptides was performed on an analytical column (75 µm×10 cm) packed with reverse phase beads (3 µm; 120-Å pore size; Dr. Maisch GmbH, Ammerbuch, Germany). Gradient elution was performed over 75 min from 5--30% acetonitrile (v/v) containing 0.1% formic acid (v/v) at an eluent flow rate of 200 nL/min after in-line flow splitting. The spray voltage was set to +1.8 kV and the temperature of the heated capillary was 200°C. The instrument method consisted of one full MS scan from 400 to 2000 m/z followed by data-dependent MS/MS scan of the 10 most intense ions, a dynamic exclusion repeat count of 2, and a repeat exclusion duration of 90 s. The full mass was scanned in the Orbitrap analyzer with R = 60,000 (defined at *m/z* = 400), and the subsequent MS/MS analysis was performed in the LTQ analyzer. To improve the mass accuracy, all measurements in the Orbitrap mass analyzer were performed with on-the-fly internal recalibration ("Lock Mass"). The charge state rejection function was enabled, and single charge and unassigned charge ions were rejected. All data were recorded with the Xcalibur software (ThermoFisher Scientific, San Jose, CA).
Database Search and Bioinformatic Analysis {#s4h}
------------------------------------------
Raw files were processed and analyzed by MaxQuant, Version 1.3.0.5 against the mouse International Protein Index protein sequence database (IPI Mouse, version 3.75), including commonly observed contaminants. The following parameters were used: cysteine carbamidomethylation was selected as a fixed modification; and the methionine oxidation and protein N-terminal acetylation were set variable modifications. Enzyme specificity was set to trypsin. Up to two missing cleavages of trypsin were allowed. SILAC double labeling (light: K0R0; heavy: K8R10) was set as the search parameter in order to assess the conversion efficiency. The precursor ion mass tolerances were 6 ppm, and fragment ion mass tolerance was 0.8 Da for MS/MS spectra. The false discovery rate (FDR) for peptide and protein was set at 1% and a minimum length of six amino acids was used for peptide identification.
The proteingroup file was imported into Perseus (version 1.3.0.4) for statistical analysis of the data. The raw dataset (3275 proteins) was filtered to include only proteins with a minimum peptide ratio count of 2 and with quantification values in a minimum of 12 of 24 MS measurements (or 24 independent SCN samples), resulting in a stringently quantified dataset of 2112 proteins. One-way ANOVA was used to analyze this stringent dataset for temporal regulation, with p-values\<0.05 indicating statistical significance. For the hierarchical clustering analysis, median value of logarithmized values for the normalized L/H ratio of each protein profile was performed after z-score normalization of the data within Euclidean distances.
To identify the subset of 24-h rhythmic proteins, JTK_CYCLE algorithm [@pgen.1004695-Hughes1] was used on the SCN proteomic (2112 proteins) or the time-of-day proteomic (421 proteins) dataset under R language. The JTK_CYCLE algorithm allows the user to input integer values when defining the (non-statistical) parameters of a search. The values for the ratio L/H normalized of each protein profile from 24 mice were used. Any missing values (i.e., not detected by MS) were replaced with zero prior to JTK_CYCLE analysis, as per expert recommendation (Dr. Michael Hughes, personal communication). Using a similar strategy as reported previously [@pgen.1004695-Dallmann1] to deal with missing values (ie., by replacing them with the minimum values observed for any given peptide) yielded the same results as replacing the missing values with zero, affirming the validity of our approach. p-values (ADJ.P) less than 0.05 were considered significant, and the corresponding proteins were classified as displaying a circadian rhythm [@pgen.1004695-Robles1]. To find the 8-h or 12-h rhythmic proteins within the 421 protein dataset, another JTK_CYCLE analysis was separately performed with period lengths set at 8 and 12-h, respectively.
Canonical pathways analyses and protein network of the time-of-day proteomic (421 proteins) dataset were mapped and summarized by Ingenuity Pathways Analysis (IPA), version 8.5 (Ingenuity Systems, Redwood City, CA). Canonical pathways analyses were performed with p value of 0.05 and networks were displayed with minimum significant score of 16. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis was achieved using the DAVID Bioinformatics Resources (<http://david.abcc.ncifcrf.gov>). DAVID statistical analyses are performed against the whole genome [@pgen.1004695-Ishihama1]. Proteomics has a tendency to oversample proteins from the cytosol while undersampling nuclear and membrane-associated proteins. To calculate exactly significant enrichment of time-of-day proteome in each GO term, we first calculated the amounts of matched proteins enriched either in the time-of-day proteome (421 proteins) or the total SCN proteome (2112 proteins). Fisher\'s exact test was used to check that the GO results were significantly enriched in the time-of-day proteome relative to the total SCN proteome. This was done to avoid any pathway/GO enrichment biases that would result from comparing our time-of-day proteome against the whole mouse database.
The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium (<http://www.proteomexchange.org>) via the PRIDE partner repository [@pgen.1004695-Vizcaino1] with the dataset identifier PXD000778.
Real-Time PCR {#s4i}
-------------
Total RNA was extracted from individual SCN tissues using the Trizol Reagent according to manufacturer\'s instructions. RNA concentration and purity were determined using the NanoPhotometer P-Class (Implen GmbH, Germany), and RNA integrity was confirmed by agarose gel electrophoresis. cDNA synthesis was performed using the Universal cDNA Synthesis Kit II (Exiqon) and 20 ng of total RNA (for miR-133b) or 100 ng of total RNA (for miR-133a). cDNA was diluted 1∶40 and real-time PCR was performed using the ExiLENT SYBR Green Master Mix (Exiqon) and hsa-miR-133a and hsa-miR-133b LNA™ PCR primer sets (Exiqon), on the Stratagene Mx3000P qPCR System. Values were normalized to 18S ribosomal RNA abundance.
*In Vitro* miRNA Mimic and Inhibitor Experiments {#s4j}
------------------------------------------------
Neuro2A cells were grown on 6-well plates in DMEM containing 5% FBS and 1% penicillin-streptomycin at 37°C and 5% CO~2~ until they reached 75--80% confluence. Cells were then transfected in duplicate using Lipofectamine 2000 (Invitrogen) according to manufacturer\'s instructions. Cells were transfected with one of the following: miRCURY LNA Power Inhibitor (Exiqon) targeted towards either miR-133a or miR-133b; miRCURY LNA microRNA Mimic for either miR-133a or miR-133b; or miRCURY LNA microRNA inhibitor negative control. Protein lysates were harvested 24 h post-transfection for Western blot analysis.
Western Blot Analysis {#s4k}
---------------------
SCN containing tissues were homogenized on ice in RIPA buffer containing protease inhibitors. Homogenized tissues were incubated on ice for 20 min and centrifuged at 4°C at 17,000*× g* for 20 min. The supernatant was collected and stored at −80°C for downstream analysis. Protein concentration was measured using the Bradford assay. Protein lysates were mixed with SDS loading buffer to 1× concentration, heated at 95°C for 5 min, and centrifuged for 1 min at 17,000× g. Lysates (20 µg/well) were electrophoresed in a SDS polyacrylamide gel for approximately 2 h at 100 V at room temperature (RT) and electroblotted onto polyvinylidene fluoride (Immobilon P; Millipore, Bedford, MA) membrane for either 1 h at RT at 85 V, or overnight (O/N) at 4°C at 30 V. Protein transfer was confirmed using Ponceau S, followed by 3 washes for 5 min each in Tris Buffered Saline with 0.1% Triton X-100 (TBS-T). Membranes were blocked in 5% skim milk in TBS-T for 1 h at RT, followed by O/N incubation at 4°C with one of the following primary antibodies in blocking solution: rabbit anti-SYT1 (1∶500; cat \#3347; Cell Signaling Technologies); rabbit anti-SV2A (1∶500; cat \#ab32942; Abcam); rabbit anti-VAMP2 (1∶2000; cat\#13508; Cell Signaling Technologies); rabbit anti-SH3GL2 (1∶2000; cat\#ab169762; Abcam); rabbit anti-PAK1 (1∶500; cat\#40852; Abcam); and rabbit anti-actin (1∶10,000; Sigma-Aldrich). The next day, membranes were washed in TBS-T and incubated for 2 h at RT with goat anti-rabbit horseradish peroxidase (HRP) conjugated secondary antibody (1∶250,000; ThermoFisher Scientific) in blocking solution. Chemiluminescent signal was developed using the SuperSignal West Femto Maximum Sensitivity Substrate reagent (ThermoFisher Scientific). Quantitation of western blots performed using the "measure" function in ImageJ (<http://rsbweb.nih.gov/ij/>) yielded a "mean gray" value for each protein band, which were normalized to background "mean gray" values. Values are presented as median relative abundance of the protein examined normalized to relative abundance of actin from 3 mice per time point.
Immunofluorescence {#s4l}
------------------
Tissue sections were washed 5 times for 5 min each in Phosphate Buffered Saline with 0.1% Triton X-100 (PBS-T). Sections were blocked in 10% horse serum in PBS-T for 1 h at RT and incubated O/N with one of the following primary antibodies in blocking solution: rabbit anti-NDUFA10 (1∶1000; cat \#ab103026; Abcam). The next day tissues were washed 5×5 min in PBS-T, and incubated for 2 h at RT in the dark with Alexa Fluor 488 donkey anti-rabbit secondary antibodies (1∶1000; Invitrogen) in blocking solution. Sections were washed 5×5 min in PBS-T, incubated with DAPI for 5 minutes, and washed with PBS. Sections were mounted on microscope slides, cover-slipped with fluorescence DAKO mounting medium and sealed with nail-polish. Slides were stored at 4°C.
Imaging {#s4m}
-------
IF images were captured using a Zeiss Axio Observer Z1 inverted microscope equipped with a Laser Scanning Microscope (LSM) 700 module along with the ZEN 2010 software (Zeiss, Oberkochen, Germany). Individual fluorochrome signals were collected sequentially using the multitrack setting along with appropriate barrier filters using the "smart set-up" option. IF images were acquired from a central focal plane of 2.3 µm optical thickness using the 10×, 20× or 40× objectives, or 1.0 um optical thickness using the 63× objective. Identical settings for gain, pinhole size, and brightness were used to acquire all images of the same magnification within each experiment. Adjustments to brightness and contrast were applied equally to all images within an experiment using Adobe Photoshop CS.
For quantitative analysis, the bilateral SCN from the micrographs were outlined using the polygon tool in ImageJ. The "measure" function yielded a "mean gray" value for each of the two bilateral SCN, which were normalized to background "mean gray" values obtained from surrounding non-immunoreactive hypothalamalic regions. Values are presented as median relative abundance of the protein examined from 3 mice per time point.
Supporting Information {#s5}
======================
######
FAT Gene Ontology enrichment analysis of the time-of-day proteome. (A--C) GO FAT analysis of 421 time-of-day-dependent proteins (significantly altered: blue bars) and 48 circadian proteins (red bars) for (A) cellular component, (B) biological process, and (C) molecular function by DAVID. Black bars represent the total SCN proteome of 2112 proteins. GO FAT analysis confirmed that a significant proportion of the time-of-day-dependent proteins were associated with the mitochondrion, energy generation and consumption, and hydrogen ion transmembrane transporter activity. \*p\<0.05, \*\*p\<0.01, and \*\*\*p\<0.001.
(TIF)
######
Click here for additional data file.
######
Heat maps of ultradian and circadian proteins within the time-of-day proteome. (A--C) Hierarchical clustering of (A) 8-h rhythmic, (B) 12-h rhythmic and (C) 24-h rhythmic proteins (p\<0.05, JTK_cycle analysis) within the time-of-day proteome. The color of spots corresponds to the value for each protein (in rows) for each of 24 SCN samples (in columns) based on the logarithmized values of L/H normalized ratios after z-score normalization. (Larger than zero: red, green: smaller than zero, grey: NaN, not detected).
(TIF)
######
Click here for additional data file.
######
Protein interaction networks of the time-of-day proteome. A comprehensive functional protein interaction network analysis of the time-of-day proteome revealed that several diseases and functions were connected and enriched in the SCN proteome by using IPA software. Proteins labeled in green, orange, or pink indicate those which belong to the 8 h, 12 h or 24 h proteome group, respectively.
(TIF)
######
Click here for additional data file.
######
Canonical pathway analysis by IPA of the time-of-day proteome. Enriched canonical pathways observed in the time-of-day proteome (*p*\<0.05). and three canonical pathways previously implicated in the regulation of the SCN clock were identified. In addition, the top-second ranked canonical pathway was the oxidative phosphorylation pathway, with 19 out of 421 proteins mapped, in agreement with our previous KEGG analysis by DAVID. Interestingly, the "breast regulation by stathmin pathway" was also identified, due to the presence of 15 proteins (CAMK1, GNAQ, TUBA4A, HRAS, KRAS, CDK1, GNB1, STMN1, PAK1, CAMK2A, RRAS2, TUBA1A, RHOA, TUBB4A, and MAP2K1) that mapped to this pathway but that are also involved in more general functions of the cell.
(TIF)
######
Click here for additional data file.
######
Total identified protein dataset in the murine SCN. The list of 3275 identified proteins resulting from the proteinGroups of MaxQuant analysis of mass spectrometry RAW files. Following analysis using the Perseus software, values for the ratio L/H normalized (where L and H represent the normalized intensities of SCN and N2A lysates for each identified protein) for each biological sample (24 total, n = 4 per CT, 6 different CT) are presented in columns A to X. Column Y shows the order of proteins which exhibit statistically significant (1--421), or non-significant (422--3275) expression within a 24-h cycle. A total of 7 proteins (3269--3275) lacked a corresponding SILAC-labeled peak, indicating that these proteins are expressed in the SCN but not in Neuro2A cells.
(XLSX)
######
Click here for additional data file.
######
The SCN proteome. The list of 2112 accurately quantified proteins representing the SCN proteome. This list was generated by filtering the raw protein dataset (3275 identified proteins) for only those proteins which were identified by a minimum of two peptide ratio counts and for which accurate quantification values were obtained in a minimum of 12 out of 24 independent samples.
(XLSX)
######
Click here for additional data file.
######
Pearson r correlation values of the 2112 proteins in the SCN proteome. Pairwise Pearson\'s correlation analysis of 24 measurements using the SCN proteome dataset (2112 proteins).
(XLSX)
######
Click here for additional data file.
######
Pearson r correlation values of the 421 proteins in the time-of-day proteome. Pairwise Pearson\'s correlation analysis of 24 measurements using the time-of-day proteome dataset (421 proteins) revealed that the r values of biological replicates at a specific CT were higher than between samples of different CTs.
(XLSX)
######
Click here for additional data file.
######
The time-of-day proteome. The time-of-day proteome is comprised of 421 significantly altered proteins, segregated into six different expression clusters. Columns AK to AP indicate the median value of logarithmized values for the normalized L/H ratio of each protein profile in six CTs after z-score normalization of the data within Euclidean distances. Column AQ indicates the results of hierarchical clustering.
(XLSX)
######
Click here for additional data file.
######
The circadian proteome of the murine SCN. Within the time-of-day proteome, a total of 48 proteins were considered to be circadian (24-h rhythmic) based on a JTK p-value cutoff of 0.05. Column AL shows the p-value (ADJ.P). Columns AM, AN and AO indicate the period (PER), optimal phase (LAG), and amplitude (AMP) estimates, respectively, for each protein according to JTK_CYCLE algorithm analysis.
(XLSX)
######
Click here for additional data file.
######
List of 8 h and 12 h rhythmic proteins in the murine SCN. Within the time-of-day proteome, 25 and 59 proteins were characterized as either 8-h or 12-h rhythmic, respectively, based on a JTK p-value cutoff of 0.05.
(XLSX)
######
Click here for additional data file.
We thank Dr. Michael Hughes for helpful advice in analyzing our data using JTK_CYCLE. We also thank the PRIDE team for proteomics data deposition and accessibility.
[^1]: The authors have declared that no competing interests exist.
[^2]: Conceived and designed the experiments: HYMC DF CKC. Performed the experiments: CKC NM PZ AP JM. Analyzed the data: CKC NM ZN WYLS HYMC DF. Contributed reagents/materials/analysis tools: CKC NM PZ HYMC DF. Wrote the paper: CKC NM HYMC DF.
| {
"pile_set_name": "PubMed Central"
} |
Background {#Sec1}
==========
There is an ever-growing number of patients requiring aortic valve replacement (AVR) \[[@CR1]\]. The two principal reasons for AVR are aortic regurgitation (AR) and aortic stenosis (AS), the latter being the most common indication. Although the majority of patients is older \[[@CR2]\], younger patients are of particular concern as they have a longer lifespan with their replaced valve and are dependent on properly functioning, non-deteriorated valves over this much longer term.
In general, a wide spectrum of therapies can be offered to younger patients such as sparing valve techniques and mechanical valve replacement. Homografts are possible but less popular due to inferior longevity. The Ross procedure gains in popularity in selected expert centres. Mechanical valves have been preferred over bioprosthetic valves in younger patients, but this is not equivocal. While some studies have shown a survival benefit of mechanical valves in younger patients \[[@CR3]--[@CR7]\], large retrospective observational studies \[[@CR8]--[@CR12]\] and one randomized controlled trial \[[@CR13]\] have shown similar long-term survival in patients 50 to 69 years of age undergoing mechanical versus bioprosthetic valve replacement.
Based on these data, the 2017 American Heart Association (AHA)/American College of Cardiology (ACC) guidelines on valvular heart disease \[[@CR14]\] recommend mechanical over bioprosthetic valves in patients below the age of 50 years (class IIa, lowered from 60 years in the 2014 version) and suggest an individualised choice (so called grey-zone) of either a mechanical or bioprosthetic valve in patients between 50 and 70 years. Conversely, the 2017 European guidelines recommend the use of mechanical valves in patients under the age of 60 years unless good quality anticoagulation is unlikely and a grey zone between 60 and 65 years \[[@CR15]\]. Both guidelines emphasize the need to consider the desire of an informed patient when it comes to the choice of the valve treatment.
The INSPIRIS RESILIA aortic Valve™ (Edwards Lifesciences, Irvine, USA) is a stented bioprosthetic, tri-leaflet valve comprised of bovine pericardial tissue. Specific new tissue preservation techniques result in a stable capping process, which blocks residual aldehyde groups known to bind calcium, in addition to a phospholipid removal process. A final tissue glycerolisation step allows valve storage without further tissue exposure to glutaraldehyde. Finally, INSPIRIS RESILIA features an expansion feature, called VFit, intended for future potential valve in valve procedures.
RESILIA tissue demonstrated, in a large pre-clinical randomized control trial conducted in juvenile sheep in mitral position, to significantly reduce tissue calcification (− 72%) and even to improve haemodynamic performance compared with the Perimount valve \[[@CR16]\]. The RESILIA tissue also has been studied in two clinical trials to date \[[@CR17], [@CR18]\]. Bartus et al. found, in a single-arm observational study of 133 patients, that the RESILIA tissue provided excellent performance and safety without structural valve deterioration (SVD) \[[@CR18]--[@CR20]\]. In the COMMENCE trial \[[@CR17], [@CR20]\], 679 patients underwent Carpentier-Edwards PERIMOUNT Magna Ease™ aortic valve replacement with RESILIA™ tissue (Model 11000A) and similar excellent safety and effectiveness were demonstrated for up to 4 years without SVD. Both of these trials have included relatively young patients, with a mean age between 65 and 67 years and up to 26% of patients aged less than 60 years.
These trials generated useful insights on the safety and effectiveness of the RESILIA tissue, but they were not specifically designed to assess durability in younger patients which received the INSPIRIS RESILIA aortic valve and data on this topic in general is scarce \[[@CR21], [@CR22]\].
It is for this reason that we designed a prospective long-term registry around the INSPIRIS RESILIA aortic Valve™. With 400 patients under the age of 60 years included and a follow-up of 5 years, we will collect data on the short-term clinical effectiveness of the valve's implantation, as well as pivotal data on the long-term haemodynamic and structural performance of the valve.
Methods/design {#Sec2}
==============
The INDURE registry is a prospective, open-label, multicentre, international registry with a follow-up of 5 years to assess the clinical outcomes of patients younger than 60 years of age who undergo surgical AVR with the INSPIRIS RESILIA aortic valve (Edwards Lifesciences). The registry is conducted according to ISO 14155:2011. Approximately 400 patients will be enrolled across 20--22 European sites (including Austria, Belgium, France, Germany, Italy, the Netherlands, Spain, and UK) and Canada, resulting in about 20 patients per site. It was estimated, from the COMMENCE Trial dataset \[[@CR17], [@CR20]\], that freedom from time-related valve safety events at 1-year (composite endpoint according to VARC-2 criteria) is around 91.5%, suggesting that 400 patients will arrive at a 95% confidence interval (CI) of ±2.14%. Lower rates (80%) will broaden the 95% CI to ±3.92% and higher rates (99%) narrow it down to ±1%.
Patients {#Sec3}
--------
Patients under the age of 60 years undergoing SAVR and receiving the INSPIRIS RESILIA aortic valve prosthesis will be enrolled on a consecutive basis. In addition to the applicable criteria of the device Instructions for Use (IFU), the registry inclusion criteria stipulate that patients require a planned replacement of their native valve as indicated in a preoperative evaluation, are scheduled to undergo planned AVR with or without concomitant root replacement and/or coronary bypass surgery. The latter is understood as isolated AVR with or without CABG and ascending aortic replacement. Also allowed is pulmonary vein isolation if it is not a full cox-maze procedure. Patients with a Bentall procedure or any surgery on other valves are not allowed in this registry. Patients need to be available to attend yearly follow-up visits at the registry centre for up to 5 years and all patients are required to provide written informed consent.
Patients will be excluded from the study if 1) they have active endocarditis/myocarditis at the time of surgery or have had it within the last 3 months of the scheduled SAVR, 2) have had previous AVR, 3) valve implantation is not possible in accordance with the device IFU, or 4) they have an estimated life expectancy of less than 12 months for any reason. The intraoperative exclusion criterion is that valve implantation is not possible in accordance with the device IFU.
Objectives {#Sec4}
----------
The primary objectives (Table [1](#Tab1){ref-type="table"}) are to 1) determine the time-related valve safety at 1-year (composite endpoint according to the VARC-2 criteria) depicted as freedom from events \[[@CR23]\] and 2) determine freedom from stage 3 SVD following the Salaun definition at 5 years \[[@CR23], [@CR24]\]. Events include SVD (either valve-related dysfunction, defined by haemodynamic parameters or the need for repeat procedure), prosthetic valve endocarditis, prosthetic valve thrombosis, thromboembolic events (e.g., stroke) and valve-related bleeding. Table 1Registry objectives**Primary objectives**Time-related valve safety at 1 year (composite endpoint according to VARC-2 \[[@CR23]\] depicted as freedom from the following events: • SVD (valve-related dysfunction \[MPG ≥20 mmHg, EOA ≤0.9--1.1 cm^2^ and/or DVI \< 0.35 m/s, AND/OR moderate or severe prosthetic valve regurgitation\], requiring repeat procedure \[TAVI or SAVR\]) • Valve-related dysfunction • Requirement of repeat procedures (Re-intervention) • Prosthetic valve endocarditis • Prosthetic valve thrombosis • Thromboembolic events (e.g., stroke) • Valve-related VARC bleedingTo determine freedom from stage 3 SVD following \[[@CR24]\] at 5 years**Secondary objectives**Haemodynamics and durability: • Haemodynamic performance of the INSPIRIS RESILIA Aortic Valve™ including PPM • SVD following Salaun \[[@CR24]\] • Description of potential ViV procedures and clinical outcome including follow-upClinical outcomes: • NYHA functional class compared to baseline • Freedom from valve-related rehospitalisationQuality-of-life: • 3--6-month, 1-year and 3-year change from baseline in quality-of-life assessed by the KCCQ and SF-12 Health Survey**Exploratory: Rehospitalisation and costs**Length of hospital stayLength of time in intensive care unitTime to return to workRate of valve-related rehospitalisation and associated costs (average costs per country)Rate of transfusion for bleeding and associated costs (average costs per country)Costs of a major bleeding eventCosts of daily anticoagulationRate of re-intervention for valve degeneration and associated costs (average costs per country)**Exploratory: Safety**SVDNon-SVDThromboembolic events (e.g., stroke)Valve thrombosisAll bleeding/haemorrhageMajor bleeding/haemorrhageAll paravalvular leakMajor paravalvular leakEndocarditisAll-cause mortalityCardiac-related mortalityValve-related mortalityValve-related re-interventionConduction disturbancesMyocardial infarctionDeep sternal wound infectionAcute kidney injury*DVI* Doppler velocity index, *EOA* Effective orifice area, *KCCQ* Kansas City Cardiomyopathy Questionnaire, *MPG* Mean pressure gradient, *NYHA* New York Heart Association, *PPM* Patient prosthesis mismatch, *SAVR* Surgical aortic valve replacement, *SF-12* Short Form-12, *SVD* Structural valve degeneration, *TAVI* Transcatheter aortic valve implantation, *VARC* Valve Academic Research Consortium, *ViV* Valve-in-valve
The secondary objectives are designed to assess haemodynamic performance and further durability parameters, clinical outcomes and quality-of-life (QoL). The first group of objectives is further defined as the haemodynamic performance of the INSPIRIS RESILIA aortic valve including patient prosthesis mismatch (PPM); SVD following the Salaun definition; and the description of potential valve-in-valve procedures and clinical outcomes. Clinical outcomes of interest are NYHA functional class compared to baseline and freedom from valve-related hospitalisation. Quality-of-life will be assessed using the Kansas City Cardiomyopathy Questionnaire (KCCQ) and Short Form-12 Health Survey (SF-12) \[[@CR24]\]. Various additional exploratory analyses regarding rehospitalisation, costs and safety will also be performed.
Data collection {#Sec5}
---------------
The clinical outcome data collected will be based on the site's standard-of-care for surgical AVR. Data will be collected prospectively, according to the timetable set out in Table [2](#Tab2){ref-type="table"}, and include medical history, physical assessments, electrocardiogram (ECG), laboratory results, computerised tomography (CT) scans (if performed as a standard-of-care), transthoracic/transoesophageal echocardiography and QoL measures. Anti-thrombotic therapy and medications are at the discretion of each investigator. Data will be captured on an electronic case report form (eCRF) by either a study nurse or physician, and data will be checked automatically for plausibility and completeness. Table 2Data collection scheduleBaseline/screeningSurgeryDischarge3--6 months^a^Years 1--5Signed informed consentXMedical history^b^XPhysical examination^c^XXXECG (12-lead)XXXEchocardiogram (TTE)XXXCore Lab echoX^e^MSCT (if available)X^f^NYHA class/CCS angina classXXXQoL questionnaire (SF-12 and KCCQ)XXX^g^Anti-thromboembolic therapy and medicationsXXXXProcedural informationXAetiologyXSAE reportingXXXXDischarge dataXRehospitalisation data^d^XXReturn to workXX*CCS* Canadian Cardiovascular Society, *ECG* Electrocardiogram, *KCCQ* Kansas City Cardiomyopathy Questionnaire, *MSCT* Multi-slice computed tomography, *NYHA* New York Heart Association, *QoL* Quality of life, *SAE* Serious adverse event(s), *SF-12* Short Form-12 Health Survey, *TTE* Transthoracic echocardiogram^a^Data captured over a telephone call^b^Includes cardiovascular and non-cardiovascular conditions, prior cardiac interventions and surgeries^c^Physical examination, includes height, weight and vital signs (blood pressure and heart rate)^d^Includes re-interventions, potential valve-to-valve procedures, associated costs^e^Solely performed at 1-year and 5-years of follow-up^f^Solely documented at 5-years of follow-up^g^Solely performed at baseline, 3--6 months and years 1 and 3 of follow-up
Echocardiography core lab {#Sec6}
-------------------------
Digital imaging and communication in medicine (DICOM) files of echocardiograms generated at years 1 and 5 will be collected for analysis by the Echo Core Laboratory to ensure unbiased and consistent analysis of the diagnostic data and, with the use of serial echocardiographic studies conducted on the same patient, for evaluating patient status over the course of 5 years.
Statistical analysis {#Sec7}
--------------------
Continuous variables will be presented as mean ± standard deviation (SD) or as median with interquartile range (IQR), and categorical variables (e.g., gender) will be reported as frequencies and percentages. The Kolmogorov-Smirnov test will be used to test for normal distribution. Accordingly, the Student t-test or Mann-Whitney U test will be used to test for statistically significant differences. The Chi-Square or Fisher Exact test will be used for statistical distribution analysis of categorical variables. Kaplan-Meier analyses will be performed for survival and safety outcomes. Linearized rates and actuarial probability statistics will be used where appropriate for adverse event reporting. A *P*-value of \< 0.05 will be considered statistically significant. Statistical analysis will be performed using SPSS Version 24.0 (Armonk, NY, IBM Corp.).
Discussion {#Sec8}
==========
The INDURE registry has been designed to provide prospectively collected data that can be used to elucidate the benefits and risks of the surgical implantation of INSPIRIS RESILIA in patients with AVR who are younger than 60 years of age, as well as the long-term haemodynamic and structural performance of the valve in patients in this age group. Analysis of the data may also provide additional support for the earlier use (e.g., at a younger age) of bioprosthetic valves in patients undergoing AVR.
Bioprosthetic vs. mechanical valves {#Sec9}
-----------------------------------
Mechanical valves are generally preferred over bioprosthetic valves for younger patients undergoing AVR because of their perceived greater durability with the 15-year rates of redo-surgery being 6.9% for mechanical and 12.1% for biological heart valves \[[@CR10]\]. It is suggested that mechanical valves will last throughout the remainder of the patient's lifetime \[[@CR25]\]. Mechanical valves do, however, require daily treatment with anticoagulants, which will increase the risk of bleeding. Lifelong anticoagulation can be difficult for patients with a history of bleeding issues or an increased risk of injury related to an active lifestyle. There may also be dietary restrictions, including reducing the intake of foods rich in vitamin K when taking vitamin K antagonists \[[@CR26]\]. Newer (or non-vitamin K) oral anticoagulants (NOACs) are strictly contraindicated in patients with any mechanical prostheses \[[@CR15], [@CR27], [@CR28]\]. Next to all the anticoagulation-related problems, reoperations can be needed even in mechanical valves in case of pannus overgrowth. Bioprosthetic valves do not require long-term daily anticoagulants but are at risk of SVD requiring reoperation \[[@CR26]\]. The risk/benefit profile of mechanical versus bioprosthetic valves has led to both American and European guidelines on valvular heart disease recommending the use of mechanical valves in patients younger than 50 years \[[@CR14], [@CR15]\] with the European version extending this recommendation to patients up to 60 years (class IIa, level C) and the American guidelines considering both mechanical and bioprosthetic valves in patients between 50 and 70 years of age (class IIa, level B, no RCT data). Despite these recommendations, the use of bioprosthetic valves has significantly increased over the last few decades across all age groups \[[@CR26]\]. Currently bioprosthetic valves are being developed that avoid the risk of valve required anticoagulation while reducing the reoperation rates seen with earlier generation bioprosthetic valves.
Determinants and surrogates of valve failure {#Sec10}
--------------------------------------------
The ultimate goal when developing durable bioprosthetic valves, which are particularly required for younger patients, is to ascertain an uncompromised haemodynamic function over the very long term with no structural degeneration that would otherwise lead to a requirement for valve replacement or valve-in-valve (ViV) interventions or death \[[@CR29], [@CR30]\]. The data required, however, would take 10, 15 or even 20 years to be collected and assessed and, as such, shorter-term surrogates of valve degeneration have been developed which facilitate shorter valve development cycles. The criteria are plenty, but have been recently reviewed by different author groups including Capodanno et al. \[[@CR31]\], Dvir et al. \[[@CR32]\] and Salaun et al. \[[@CR24]\], partly in an attempt to provide standardised definitions of SVD for bioprosthetic aortic valves. The definition of SVD by Salaun \[[@CR24]\] (Table [3](#Tab3){ref-type="table"}) has been adopted in the current project as it incorporates terminology proposed by both Dvir \[[@CR32]\] and Capodanno \[[@CR31]\] and was compatible with the definition used by Pibarot et al. \[[@CR34]\] (see below). We will, however, capture the components of the other definitions as well aiming to explore and compare these as well. Table 3Definition of structural valve deterioration of aortic bioprostheses following Salaun \[[@CR24]\]Stage 0^a^\
No SVDStage 1^a^\
SVD with no HVD\
'Morphological SVD'^b^Stage 2^a^\
SVD with moderate HVD\
'moderate hemodynamic SVD'^b^Stage 3^a^\
SVD with severe HVD\
'Severe hemodynamic SVD'^b^Valve leaflet morphology and motion by TTE, TEE or MDCT Leaflet morphology Valve leaflet thickening: At least one leaflet with thickness ≥ 2 mmAbsentPresentPresentPresent Valve leaflet fibrocalcific remodelling: Hyper echogenicity (TTE/TEE) or hyper density (MDCT) Detectable leaflet calcification at MDCTAbsentPresentPresentPresent Leaflet motion Reduced mobilityAbsentAbsent or mildMild to moderateModerate to severe Leaflet tear/avulsionAbsentAbsentMay be presentMay be presentValve haemodynamics by TTE Mean transprosthetic gradient Absolute increase from baseline^c^\< 10 mmHg\< 10 mmHg10--19 mmHg≥20 mmHg Mean gradient at post-AVR echo^d^\< 20 mmHg\< 20 mmHg20--39 mmHg≥40 mmHg Valve effective orifice area Absolute decrease from baseline\< 0.30 cm^2^\< 0.30 cm^2^0.30--0.59 cm^2^≥0.60 cm^2^ Doppler velocity index Absolute decrease from baseline\< 0.1\< 0.10.1--0.2≥0.2 Transprosthetic valve regurgitation^c^ Worsening compared with baselineAbsentNone≥1 Grade≥2 Grades Grade of regurgitationNone, trace or mildNone, trace or mildModerateSevereClinical statusSubclinicalOften subclinicalGenerally clinically expressive New onset or worsening symptomsAbsentOften absentGenerally present New onset or worsening LV dilation/hypertrophy/dysfunctionAbsentGenerally absentOften present New onset or worsening pulmonary hypertensionAbsentGenerally absentOften presentThe classification and criteria presented in this table are based on recommendations and standardised definitions of medical societies or group of experts \[[@CR31]--[@CR33]\]Stage 0 refers to a normal valve. Stage 1 consists in the presence of morphological abnormalities of valve leaflets but with no evidence of HVD. At echocardiography, the leaflets are thickened (\> 2 mm), often irregular and hyperechogenic. MDCT without contrast may be used to detect and quantitate macroscopic valve leaflet calcification by the modified Agatston method or the volumetric method. Stage 2 consists in SVD with moderate HVD defined as: (1) an increase in mean transprosthetic gradient ≥10 mmHg since early post-SAVR or TAVI echocardiography with concomitant decrease in valve EOA and DVI; and/or (2) a new onset or worsening of transprosthetic regurgitation by at least one grade with a final grade of moderate. An increase in transprosthetic velocity and gradients with concomitant increase in valve EOA and DVI is actually related to an increase in flow during follow-up and should not be mistaken for an HVD. Stage 3 consists in SVD with severe HVD characterised by: (1) an increase in mean transprosthetic gradient ≥20 mmHg since SAVR or TAVI with concomitant marked decrease in valve EOA and DVI and/or (2) new onset or worsening of transprosthetic regurgitation by at least two grades with final grade of severe regurgitation*AVR* Aortic valve replacement, *DVI* Doppler velocity index, *EOA* Effective orifice area, *HVD* Haemodynamic valve deterioration, *LV* Left ventricle, *MDCT* Multidetector CT, *SVD* Structural valve deterioration, *SAVR* Surgical aortic valve replacement, *TAVI* Transcatheter aortic valve implantation, *TEE* Transesophageal echocardiography, *TTE* Transthoracic echocardiographyClassification terminology proposed by: ^a^Dvir et al. \[[@CR32]\] and ^b^Capodanno et al. \[[@CR31]\]^c^The most important criteria to define haemodynamic HVD is a significant increase in mean transprosthetic gradient with concomitant decrease in valve EOA and DVI; and/or a new onset or a worsening of transprosthetic valve regurgitation^d^This criterium is corroborative but should not be used in isolation to define HVD
INDURE in perspective {#Sec11}
---------------------
The RESILIA tissue has been studied in two trials to date \[[@CR17], [@CR18]\] with a total of 812 patients showing that its use results in excellent haemodynamic performance and safety up to 2 years. Both the RESILIENCE trial \[[@CR34]\] and the INDURE registry were set up in order to assess the long-term performance and structural integrity of bioprosthetic valves using the RESILIA tissue in younger patients. RESILIA tissue incorporates an anti-calcification process, by permanently blocking the residual aldehyde groups that are known to bind calcium. Calcification is known to occur more commonly on bioprosthetic valves than mechanical valves \[[@CR35]\]. It, therefore, has the potential to increase valve longevity and consequently reduce re-intervention rates.
Both INDURE and RESILIENCE (Table [4](#Tab4){ref-type="table"}) are prospective studies, including patients with either the INSPIRIS RESILIA valve (INDURE) or any RESILIA tissue bearing valve (RESILIENCE). While INDURE will follow patients from the time of surgery for 5 years, RESILIENCE pursues retrospective inclusion of patients with the first visit being 5-years after surgical intervention and a prospective follow-up (up to year 11 after the implant). On the one hand, INDURE puts emphasis on a combination of time-related valve safety at 1-year, SVD defined according to Salaun \[[@CR24]\] using a CoreLab and clinical outcomes, while on the other hand RESILIENCE focuses on the multi-slice computed tomography (MSCT) and echo-based (both CoreLab) prediction of re-intervention or valve-related death. Projected completion dates are 2025 (INDURE) and 2027 (RESILIENCE), respectively. Table 4INDURE vs. RESILIENCEINDURE (NCT03666741)RESILIENCE (NCT03680040)Valve usedINSPIRIS valvesRESILIA tissue valvesDesignProspectiveRetrospective inclusion, prospective follow-upStudy Start date26 April 20195 November 2018BaselineImplantation5 yearsFollow-up5 years -- projected completion 20256 years (from year 5 to year 11) -- projected completion 2027Subjects/centres400 subjects, 20--25 centres (EU and Canada) under the age of 60 years at the time of their SAVR220 subjects, up to 15 centres (US and EU) under the age of 65 years at the time of their SAVRObjectiveAssess clinical outcomesTime to valve failure due to valve degeneration requiring re-intervention & early potential predictors of valve durabilityPrimary endpointsTime-related valve safety at 1 year (VARC-2)Time to BVF due to SVD, defined as requiring re-intervention (redo surgery or ViV), or confirmed valve related death, according to Akin criteria \[[@CR29]\]Rate of severe SVD (stage 3 following Salaun \[[@CR24]\]) at 5 years (Echo CoreLab)Secondary endpointsHaemodynamics and durability (Echo CoreLab)Clinical outcomes (NYHA and freedom from rehospitalisation)Quality-of-life (KCCQ & SF-12)Early possible predictors of valve failure including leaflet calcification and morphological/haemodynamic valve degeneration: -Valve leaflet calcification via CoreLab evaluated MSCT (no contrast) -Haemodynamic performance (Echo CoreLab)*BVF* Bioprosthetic valve failure, *EU* European Union, *KCCQ* Kansas City Cardiomyopathy Questionnaire, *MSCT* Multi-slice computed tomography, *NYHA* New York Heart Association, *SF-12* Short Form-12, *SVD* Structural valve degeneration, *US* United States, *VARC* Valve Academic Research Consortium, *ViV* Valve-in-valve
Up and beyond INDURE and RESILIENCE there is a third long-term data collection ongoing (IMPACT; NCT04053088) using the INSPIRIS RESILIA valve. It is being conducted in Germany, Austria, Switzerland and The Netherlands and will follow up to 500 patients for 5 years. The principal objective of IMPACT is the assessment of the impact of comorbidities such as chronic kidney disease (CKD), diabetes, hypertension, metabolic syndrome and inflammation on all-cause mortality. Among the secondary objectives there is, again, assessment of SVD, which will complement the data derived from INDURE and RESILIENCE.
Appreciation of the study design {#Sec12}
--------------------------------
The INDURE registry is a prospective, open-label, multicentre, international registry. The multinational nature of this registry increases the applicability of findings to clinical practice all over Europe and Canada. However, it has no control group making a comparison of different bioprosthetic valves or valve generations impossible. Furthermore, there is no comparison of the bioprosthetic valve data with the outcomes of mechanical valve implantation, which would be desirable, but goes beyond the possibilities of such a project. Because of the multicentre design, an Echo CoreLab has been established to have a uniform assessment of SVD over the time course of the 5-year follow-up. We considered establishing an MSCT CoreLab, as has been incorporated in the RESILIENCE trial, but it would have violated the non-interventional nature of the INDURE registry as most sites reported that an MSCT is not standard-of-care at their institution, which may be considered as a limitation. Nonetheless these data can be documented in case they are available from routine practice. Finally, the same INSPIRIS RESILIA valve will be used in all patients in the INDURE registry which will abolish any bias introduced by the use of different bioprosthetic valves.
Conclusions {#Sec13}
===========
INDURE is a prospective, multicentre registry in Europe and Canada that will provide much needed data on the long-term durability of bioprosthetic valves in general and the INSPIRIS RESILIA valve in particular. The data may help to gather a deeper understanding of the longevity of bioprosthetic valves and may expand the use of bioprosthetic valves in patients under the age of 50 and 60 years.
ACC
: American College of Cardiology
AHA
: American Heart Association
AS
: Aortic stenosis
AVR
: Aortic valve replacement
CT
: Computerised tomography
DICOM
: Digital imaging and communication in medicine
EACTS
: European Association for Cardio-Thoracic Surgery
ECG
: Electrocardiogram
eCRF
: Electronic case report form
ESC
: European Society of Cardiology
IFU
: Instructions for use
KCCQ
: Kansas City Cardiomyopathy Questionnaire
MCST
: Multi-slice computed tomography
NYHA
: New York Heart Association
PPM
: Patient prosthesis mismatch
QoL
: Quality of life
SD
: Standard deviation
SF-12
: Short Form-12 Health Survey
SVD
: Structural valve degeneration
**Publisher's Note**
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Data were captured using the s4trials Software provided by Software for Trials Europe GmbH, Berlin, Germany.
BM, RdP, TB, BB, PB and MB were involved in the conception and design of the study. PB and BB drafted the manuscript and all other authors revised the article for important intellectual content. All authors gave approval of the final version.
This work was supported with a research grant provided by Edwards Lifesciences (Nyon, Switzerland) to the sponsor Institute for Pharmacology and Preventive Medicine, Cloppenburg, Germany.
Not applicable.
Ethic committee approval has been sought at all participating centres prior to patient enrolment. All patients will be required to provide signed informed consent.
Not applicable.
BM, RdP, TB, PB and MB have received lecture fees and/or research support from Edwards Lifesciences. The institutions of these and those of the remaining authors representing study centres have received patient inclusion-based funding. VG is an employee of Edwards Lifesciences. BB has no conflict of interest to disclose.
| {
"pile_set_name": "PubMed Central"
} |
Introduction
============
The lack of integration within the public health system is one of the most important problems of dental treatment for patients with special needs. Moreover, the peculiarity of decentralization of Public Health Services in Spain, makes public health patients with special needs coverage, very different within the same country ([@B1]).
This creates a lackof care that makes patients to keep looking for solutions to their dental problems in the field of private assistance ([@B2]).
We will present two examples of dental treatment integration in patients with special needs in a private hospital setting.
To evaluate the service provided to this type of patients, not only in Spain, a systematic review of the literature was performed using database PubMed-Medline using the keywords dentistry and handicapped patient hospital.
With the keywords used 151 articles were found, and then we proceeded to do a manual search to select the suitable items, select 22 to perform the review ([Table 1](#T1){ref-type="table"}).
Table 1Most prominent research works in oral treatment of handicapped patients.
Private hospital dental practice in adults
==========================================
Private Hospital Dentistry, as a professional activity, only differs from the one made out in publics hospitals in the kind of patients who come or may come, and in the characteristics related to the way of work or the way in which assistance is provided.
On the other hand, indications, protocols, techniques, and ultimately, how patient is treated should be no different from a public or private one. But actually they have many distinguishing features: from how to capture the patient, the use of the cabinet and the operating room, the relationship with family and guardians, loyalty of patient and monitoring of treatment.
Considerations
==============
1\. Public and private.
Understanding both , public and private assistance have the same purpose, oral and dental health. It is clear that the only difference between both assistance services is that in the public assistance, patient does not pay directly and he does it in the private one, with differences that it entails at all levels (from the management, training, technology, treatment options, human resources).
2\. Hospital option.
Dental Clinic understood, as usual, (private, not big clinic with only 1 or 2 dentists, open to everybody) has its essence in business logic: It is private medicine that is practiced as close as possible to patients.
Hospital Dentistry is often only associated with the use of the operating room as a procedure room. It is a very simplistic and logical, derived from the tradition in the practice of dentistry.
Reality is very different.
Hospital Dentistry can not be only a pure Dentistry activity, because it allows dentistry to expand to large spaces of Stomatology and Oral Surgery. It could open dentistry to a prominent place in medicine, resulting from interaction with other hospital acts and medical specialities. It Can include all "Dental Clinics" features and may offer more.
Moreover, hospitals are those who are reluctant to provide space for dentistry because of its characteristics, that make dentistry not very related to other medical specialities.
Privates hospitals are usually interested in renting rotating spaces to different specialists better than one that also does not generate other benefits added to hospital.
Private hospital dental model.
After consider options where we can work, and once decided by Hospital dentistry, it is appropriate that we take into account several factors that will determine the success of this model that includes two types of patients.
• Patients without underlying systemic diseases.
• Handicapped Patients or patients with systemic diseases.
a\) Patients without underlying systemic diseases:
Dentistry practiced in a Hospital competes with multitude of common dental clinics full of arguments for them: the next customer location, integration into the commercial life of the neighborhoods etc.
Then there is the assumed tradition by patients in concept of Dentist (something more aesthetic and associated with repairing teeth) and in concept of Hospital (associated with diseases).
Hospitals are in most cases associated to sophisticated treatments, technological advances and specialization these are important arguments that we should offer to conquer a reference position, position that will give us the famous "word of mouth" (patients talking with other people about our services) and can be reinforced with simple advertising campaigns.
Another essential aspect for the proper positioning within the Hospital of our dental service is to be able to choose the location close to fields related to our workspace: ENT, Dermatology, Plastic Surgery, Pediatrics, Psychiatry/Psychology ([@B3]).
Looking to our hospital, like any other specialty, you need to be more than a tenant: you should generate traffic of patients that can consume in other departments (Analytical, Rx, interdepartmental, operating rooms). This, which could be required from Hospital direction, comes up by activity that generates working in a hospital enviroment. In fact, from the beginning, we will find care demands not only in teeth areas and also we could find complex patients who have not been successful treated in their regular dentists, and who need dental attention in other environment ([@B4]).
b\) Handicapped Patients.
Probably the most important differential area compared to Dental Clinics.
The special care in dentistry.
Those people who for physical or mental characteristics can only be treated in hospital environment, either for monitoring and controlling risks or because the only way to access to their mouth, is using general anesthesia or other form of sedation.
It's easy, even today, associate this operation to the public health service but since transfers were initiated to differents Spanish health services, we find many contradictions and discrimination for being resident in one or another part of the country ([@B5]).
Additional cost that Dental performance in these kind of patients means, does, in many cases, impossible to treat them under the private health system when they need General Anesthesia, and they only have access to dental health through mutilation (extractions multiple) or the bounty of the Administration.
At this point is very important that hospital direction could get involved to promote patients referral with some agreements that can range from the "total" (completely covered dental care for patients with disabilities) to "partial"promoting access to the hospital (General Anes-thesia and Operating Room) and giving only economic coverage overall performance (extractions and surgery).
This simple formula, a combination of public and private health services, allows to do more just and equitable dental practice: handicapped patient can be treated at all levels at the same price as everyone else in the field of dentistry, and has the same right to Oral Health than any other person without pathologies because the state could provide additional medical facilities (hospital, operating room and anesthesia) for dental care.
Probably this health service model could be the final format that can be applied by the state because of economic rationality and social justice .
But from the point of view of independent dental professional, it is possible that this patient loads is not enough itself to maintain Dental Service.
Again is necessary to apply business judgment in this kind of professional dedication to find profits and for Monitoring evolution and loyalty of patients, promote oral health in those centers which care of disabled patients, collaborate in monitoring and maintaining hygiene and also promote a familiar motivation.
These actions, carried out systematically and notarized, have enabled successful experiences in creating Dentistry Services at Private Hospitals. Being also an obvious fact that morbidity and dependence in chronic patients in our society have increased. The incidence of cerebral palsy with a neonatal origin have been maintained and even decreased but dependencies related to age and senile dementias have gradually increased ([@B6]).
All of them have teeth as positive / negative factor in systemic health and they need to be treated.
Guidelines for special attention to handicapped patients in a private hospital.
Some private hospitals have technical resources to perform under sedation or general anesthesia oral treatments in both, normal patients and in handicap pedpatients. But not all of them have a Dental Service that could provide dental care in different specialties (as General Dentistry, Preventive Dentistry, Orthodontics, Speech, Maxillofacial Surgery, Special care in Dentistry) integrated within the hospital structure and related with other medical areas that can provide a comprehensive, multidisciplinary solution to health problems with joined protocols appropriate to patients needs.
Handicapped Patients in Pediatric Dentistry
Pediatric patients with special needs include a large number of patients with many different diseases that can be classified into 4 groups:
\- Group 1 would be medically compromised patients.
\- Group 2 include patients with motor deficits to a greater or lesser degree.
\- Group 3 would consist of those with some sensory deficit (DEAF, BLIND people).
\- Group 4 would bring together all patients with pathologies with some degree of intellectual disability.
Most of these patients, due to wide range of general pathology that may present, requires specially trained professionals, and individualized treatment plans that cover dental treatment needs.
In all cases we should always perform a complete clinical history, we should also request medical reports and study the underlying pathology. We should also know the medical treatments that are being performed as the same time of dental attention.
On the other hand, we must perform a dental history aimed to obtain an individual and objective assessment that will show us the most appropriate oral therapeutic needs in each case and will help us to decide what would be the best way to perform dental attention in clinic or under sedation or general anesthesia ([@B7]-[@B9]).
All of this makes necessary an interrelation and intercommunication with other medical and dental specialties, developing integrated protocols in many cases.
If all these actions could be coordinated from a hospital dental service that can organize referrals to other services (not only related to mouth) integrated in the same hospital, it will make everything easier, because not only number of movements of patient get reduced, because visits and tests could be jointly programmed, but, thanks to the computerization of services, it faci-litates access from any specialty to all tests that have been done to that patient.
Clinical dental treatment in handicapped and madically compromised patients
===========================================================================
Within this diverse spectrum of individual needs in dental treatment of handicapped patients, we must consider on one hand the underlying disorder which will mark the most appropriate treatment decisions, the different medical specialists who must intervene to schedule referrals and which protocols must be followed in each case.
In case of medically compromised pediatric patients, we should always perform a complete clinical record and we should also ask their doctors for a full report of their illness, treatment and an update prognosis. When patients come for the first time to our clinic, we always ask them from 5 years of age onwards, about an orthopantomography which can be performed the day before our appointment. Oral treatment may be performed in normal clinical following special protocols for each specific type of pathology and in some cases it will require coordination with others specialists who could directly control overall patient´s illness.
Patients from this group, usually attend a Hospital Dental Service because they feel safer getting dental care in this environment.
Pathologies that medically compromised patients could present can be very large and diverse and always require an overall assessment of patient status, to design individually dental treatment plan which suits better, each case needs.
Diabetic patients have no specific oral manifestations. For dental practice we must always control diabetes. Poorly controlled patients should be referred to specialist some days before dental treatment to control blood sugar levels.
It is recommended to perform dental treatment within two hours after the insulin injection and we should not modificate patient´s usual breakfast, not change medication schedules and especially food intake, both, before and after treatment.
Local anesthesia can be performed normally. We should remember that in these patients healing is slower and therefore, if extractions are performed, we recommend coverage with high spectrum antibiotics ([@B10],[@B11]).
The most common complication in this kind of patients is hypoglycemia, which usually can be solved with administration of orally quick absorption carbohydrates (or parenterally), but we can avoid it easily if we adapt schedules appointments to patient´s intake.
Another group of patients are those who suffer from cardiac diseases. In these cases, patients who come more often are those with congenital cardiac abnormalities and those with functional murmurs. In first group we must always request specialist report (where should appear pathology´s current situation) , because is specialist who makes pathology following. We should always apply prophylaxis recommendation for bacterial endocarditis prevention in cases where it is indicated ([@B12],[@B13]). In the second group, treatment does not require any special preparation.
In pediatric patients, breathing problems we see more frequently are those associated with bronchial asthma. Most common oral abnormalities in this kind of patients are an increased predisposition to tooth decay, associated with prolonged use of steroids and other drugs in suspension form for inhalation and dispensed daily ([@B14]).
It is recommended to perform Dental treatment in these patients in periods in which child is asymptomatic, appointments should be done at morning hours because we can monitor better patient´s situation, anesthetic with vasoconstrictor can be used, always performing aspiration during injection and should be avoided anesthetics with sulphites.
After inhaler use it is recommended to these kind of patients to always rinse their mouth with water, to avoid decay risk ([@B14]).
In epileptic patients without intellectual disabilities who take regular medication, are well controlled and who usually have no crisis, dental treatment can be normally performed at clinic. It is recommended to treat these kind of patient within 2 hours after they have consumed their medication.
In these patients we should avoid triggers factors for epilepsy, such as stress and anxiety. If during treatment, suddenly, a crisis appears, we should immediately remove all tools and materials from mouth, put patient in a supineposition, wemust tilt patients head sideways, and we should avoid mouth closing for avoid tongue biting ([@B15]).
For patients with mild motor deficits involving only arms and legs and not accompanied by intellectual disabilities, there won´t be any problem for normal dental treatment. In cases which shortfall affects upper limbs it is important to involve their family in proper daily dental hygiene.
Patients with sensory deficits, which more often come to clinic are those with impaired hearing ability or deaf patients, who in earlyonset cases it will probably be associated with speech deficits. In these cases, patient communication is the most important problem, which will make treatment exclusive and personalized: You need a substantial visual communication and specialized collaboration in sign language by their family or by a professional ([@B16]).
Dental tratment of handicapped and medically compromised patients in an operating room
======================================================================================
Within resources that can be provided in a hospital, one of the most important and requested by handicapped patients is dental treatment under general anesthesia or sedation ([@B17],[@B18]).
Within this group of patients we can find medically compromised patients (congenital cardiac abnormalities, blood dyscrasias, allergic reactions to local anesthetics,. uncontrollable epilepsy, etc.) On the other hand, we have all the patients with motor deficits that don´t allow proper treatment in clinics and all patients who have a mild or severe intellectual disability, whose condition or treatment inhibit a dental treatment in clinics ([@B8],[@B16]).
This group of patients (intellectual disabled) generally presents big problems, with many different oral pathologies, because they themselves are not able to seek medical care and disability also involves a failure to perform a proper daily oral hygiene and proper maintenance and it is also the group of patients which generates a greater request for hospital treatment under general anesthesia or sedation.
Inside this group we can find genetic intellectual disabilities such as Down syndrome, fragile X syndrome, Angelman syndrome etc. Others would be patients with severe neuromuscular disorders such as cerebral palsy and spina bifida we can also include in this group patients with autism spectrum disorders, affected by Asperger Syndrome and Rett Syndrome.
To assess these patients, we must make a general record of underlying disease with complete laboratory blood analysis and electrocardiogram , and dental history (orthopantomography etc.). With these data we can make an inquiry with the anesthesiologist to determine the health risk presented by the patient.
For this purpose the ASA group classification is used, which is a 6-degree scale created by the American Society of Anesthesiologists that relates the degree of surgical risk in the patient with his main pathology and with that, he will determine the type of anesthetic technique, general anesthesia or sedation, to perform dental treatment plan with the most appropriate option ([@B17]-[@B19]).
To achieve sedation treatments patients can be only candidates if they are included in both ASA groups I and II ([Table 2](#T2){ref-type="table"}).
Table 2ASA patients classification.
The anaesthesiologist will establish the preoperative protocol with minimum 6 fasting hours and suppression or not, of patient's underlying medications, depending on patient´s disease and type of drug. On this visit, an informed consent, for the type of anesthetic technique will be performed.
If it´s necessary to administer some treatment to reduce anxiety and fear, at the time of admission before intervention, premedication and administration way( nasal, rectal or sublingual) will be scheduled ([@B20]-[@B22]).
With all this we make a second patient reassessment where informed consent is signed to carry out the proposed dental treatment. You give to patient´s family, written preoperative and postoperative recommendations to prevent oversights and errors.
In those cases in which correct exploration is impossible and therefore we can not perform a preoperative treatment plan, parents will be informed and we will proceed to evaluate patient needs and establish a treatment plan when patient is slept, forcing us to make changes in predicted treatment, make quick decisions taking into account needs of patient and degree of cooperation in the later maintenance of treatment .
Patient will enter into health service as an ambulatory surgery proceed,a couple of hours before surgery. In these cases treatments are usually performed along morning so patient could remain as shortest as possible in hospital, and they can return sooner to their usual environment.
At the moment of admission at hospital, as a medical order, premedication ordered by anesthesiologist should appear, and it will be administered by nursing staff and, during surgery, certain drugs could be administered parenterally according to patient's requirements if it ´s needed, as in case of endocarditis prophylaxis in patients with heart disease, analgesics to reduce postoperative pain etc.
After intervention, patient remains with an injecting dropper and under hospital supervision for about four hours and if liquid tolerance is adequate, no vomiting and no complications appear we will proceed to discharge patient with drug regimen to be followed at home depending on each case of dental treatment performed and the patient's underlying disease.
With this we will try to interfere as little as possible with patient's general environment and allows that time at hospital is as minimum as possible. Patient should return for a control treatment in the recommended period and in that time, we will proceed to establish the inspection and maintenance protocol to be followed by the patient in order to minimize the emergence of new diseases. On this visit is we will suggest referrals to other services if it is necessary ([@B8]).
In conclusion we have described a sequence of performances for outpatient treatment in handicapped patients with different pathologies whether children or adults, both in private practice at dental clinics and in hospitals, in order not to changetheir routine as much as possible and create a favorable environment to face treatment, prescribing sedation or general anesthesia as a last resource in extreme cases only if patient´s pathology, medication or irregular collaboration, require it.
[^1]: **Conflict of interest statement:** The authors have declared that no conflict of interest exist.
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#S1}
============
With advancement in minimally invasive surgical interventions, arthroscopic surgery is now one of the most common procedures performed in modern orthopaedics \[[@R1]\]. Unlike open surgery, arthroscopic procedures involve 2-D image projection of a three-dimensional operative field requiring more technical dexterity and visuo-spatial coordination \[[@R2]--[@R4]\]. Proficiency in arthroscopy has a steep learning curve and before embarking on clinical arthroscopy, surgeons should be competent in handling instruments and learn the basic skills of arthroscopy \[[@R3]--[@R5]\]. There is a need for out of operating room practice of these skills so as to decrease errors in instrument handling, risks of iatrogenic injury, financial burden and operative time during the initial phase of practice. This is further limited by introduction of restricted working hours for residents across Europe and North America affecting *hands-on* training \[[@R2], [@R3], [@R6]--[@R10]\]. Arthroscopic simulators, including cadaver and bench models. More recently, computerized virtual reality simulators have evolved over time to teach psychomotor skills of arthroscopic surgery \[[@R2]--[@R5], [@R7]--[@R9], [@R11]--[@R24]\]. Cadaver models and virtual reality simulators are resource demanding whereas the bench models are easy to set up, simulate realistic environment and have shown development of motor skills and technical training \[[@R2], [@R7]--[@R9], [@R11], [@R20], [@R22]--[@R24]\]. Simulated tasks involving *image tracking*, *triangulation and probing* and *handling of basic tools* like shaver, punch and grasper can replicate basic skills of arthroscopic surgery. Performance in these tasks has been shown to match clinical experience however, there is need of *structured modules* that can be applied in training curriculum to establish the proficiency of trainees \[[@R2], [@R3], [@R25]--[@R31]\]. Collaborative efforts of American orthopaedic and arthroscopy associations have proposed one such module for arthroscopic skills training; the "*FAST* program" \[[@R29]\].
The module uses a simple box-trainer type arthroscopic simulator and we hypothesized that more advance surgeons would perform better in *tasks* based on experience. Hence, this study was designed and conducted to test the usefulness of this model to assess performance based on clinical experience and obtain data to design guidelines for future arthroscopic training.
Methodology {#S2}
===========
Study design and subjects {#S3}
-------------------------
This prospective cohort study was conducted in the sports medicine and arthroscopy unit of a University Hospital during the period of September to October 2014. The study included 20 orthopaedic surgeons of various levels of arthroscopic experience including professors, lecturers, assistant lecturer grade surgeons, trainees including year 1--3 orthopaedic residents. None of the subjects had previous exposure to practice on the test workstation. They were given written instruction and video demonstration of the tasks and consent to participate in the study was obtained.
The 20 orthopaedic surgeons in the study consisted of 12 participants nonproficient in arthroscopic surgery (7 residents, 3 clinical fellows, 2 trauma surgeons) and 8 arthroscopy specialists (5 assistant lecturers or lecturers, 3 professors) and their demographic details were obtained as presented in [Table 1](#T1){ref-type="table"}.
Table 1.Participants profile and arthroscopic experience of subgroups.Group characteristicsNoviceBeginnerIntermediateAdvanceIndependent proceduresNone\<5050--100\>200Number (% of total)9 (45%)4 (20%)3 (15%)4 (20%)Mean age (years, ±*SD*)32.3 (±6.1)33.5 (±3.3)32 (±2.6)44 (±6.9)Mean years of orthopaedic practice38719.5Clinical arthroscopy practice (years)None\<22--4\>4Knee arthroscopy exposure[\*](#TFN1){ref-type="table-fn"}NoneAllAllAllShoulder arthroscopy exposure[\*](#TFN1){ref-type="table-fn"}None1/42/3AllIndependent surgeon[\*](#TFN1){ref-type="table-fn"}None2/4AllAll[^1]
Arthroscopy experience was quantified as the number of independently performed procedures involving basic knee arthroscopic surgery. The subjects were divided into four groups as follows: (1) Novice -- no exposure to arthroscopy, (2) Beginner -- less than 50 cases independently performed, (3) Intermediate -- 50--100 cases performed independently and (4) Advance -- more than 200 cases as independent surgeon.
Equipment and task specification {#S4}
--------------------------------
The study was conducted using a bench-type workstation with geometrical objects placed in a three-dimensional environment for the use of basic arthroscopy instruments. The *box-trainer* was developed by "SAWBONES" (Pacific Research Laboratories) in collaboration with the "FAST" (Fundamentals of Arthroscopic Surgery Training) program. The "FAST" program consists of modules for sequential proficiency in basic and advanced motor skills proposed by combined efforts of AANA (Arthroscopy Association of North America), AAOS (American Academy of Orthopaedic Surgeons) and ABOS (American Board of Orthopaedic Surgery) to become a part of structured curriculum and training in arthroscopic surgery \[[@R30]\]. The workstation consists of multiple detachable rotating platforms like maze navigation, randomly arranged number probing station, horizontal and vertical pillar stands for object placement and extraction and meniscus platform for simulating basic and interventional knee arthroscopy skills ([Figure 1](#F1){ref-type="fig"}).
Figure 1.Modules for "FAST" bench workstation. (A) Maze navigation, (B) number probing, (C) object handling-vertical pillars and (D) meniscus stations.
Competency in scope navigation, triangulation, probing and handling of objects within three-dimensional space and partial menisectomy were chosen from the modules to assess basic arthroscopy skills. The tasks used as surrogate for these skills were: (1) maze navigation, (2) number probing, (3) object extraction and insertion and (4) partial menisectomy of medial and lateral meniscus ([Table 2](#T2){ref-type="table"}, [Figure 2](#F2){ref-type="fig"}).
Figure 2.Opaque dome (OD) modules for "FAST" bench workstation. (A) Maze navigation, (B) number probing, (C) object handling-vertical pillars and (D) meniscus stations. Table 2.Description of basic arthroscopy skills and surrogate tasks on the "FAST" workstation.SkillDescription of skillTask surrogateDiagnostic arthroscopyScope navigation (telescoping, periscoping), Triangulation, ProbingMaze navigation task, Number probing taskTissue handlingProbing, Simultaneous image tracking and instrument handling, Bimanual dexterity, Grasp and releaseLoose body extraction and insertion task -- vertical and horizontal pillarsPartial menisectomyMeniscus resection and balancing, Use of instrument -- punchMedial and Lateral meniscus partial menisectomy task
After the instructional video on task performance, participants were asked to perform all four tasks under a transparent dome (TD) with direct vision and then under the opaque dome (OD) with the arthroscope using the portals representative for antero-medial and antero-lateral arthroscopy portals for the right knee. The equipment was standardized for all in terms of position of platform and instruments used which consisted of a standard 30° arthroscope with light source and camera, LCD monitor, 5 mm probe and straight grasper for object handling and a straight cutting punch for menisectomy.
Individual task protocols ([Figure 2](#F2){ref-type="fig"}) {#S5}
-----------------------------------------------------------
Maze navigation: Subjects completed the task of probing a steel ball (3 mm diameter) through the maze without skipping channels or dropping the ball which was counted as an error. The platform was fixed at 0° rotation.Number probing: In this task the participants were asked to sequentially probe numbers 1--21 randomly arranged on a predesigned perforated platform which fixed at 270° front facing and the time to complete was computed under both TD and OD.Object handling: This task was considered representative of tissue handling skill and ability to use instruments with bimanual dexterity. The task required placement and retrieval of 10 tubular cylinders of each 12 mm size into two different spatial arrangements of horizontal and vertical pillars. Dropping the object was counted as an error and the time to complete all 10 objects in the task was taken as the end point.Partial menisectomy: The specially designed platform for representation of meniscus was rotated to 45° on either side to represent medial (MM) and lateral meniscus (LM) for the right knee. A 1 mm standard density foam material was tested and agreed upon by three arthroscopy experts to represent the feel of cutting meniscus tissue by punch. A semicircular meniscus was fashioned out of the foam and a mark was made to represent red-white zone with a radial tear. The participants were then asked to trim the meniscus to within the marked zone with precision first under the transparent then under opaque the dome. Time to complete task and the precision of "menisectomy" were noted.
Additional platforms for shoulder arthroscopy tasks like anchor placement, suture passage and knot tying were not used in this study.
Assessment parameters and analysis {#S6}
----------------------------------
Time for task completion and number of errors were used as objective assessor of proficiency and discriminant validity of construct. For each error the participants were penalized 5 seconds on the clock but allowed to complete the task. Subjective assessment of steadiness of scope, simultaneous image tracking and instrument handling, bimanual dexterity and face validity was done by feedback questionnaire from the participant and assessment by an unbiased observer of performance using a Likert scale from 1 to 5 (1-very easy or very good/2-easy or good/3-not so difficult or not so good/4-difficult or poor/5-very difficult or very poor). Performance measured as time for task completion under transparent and opaque dome comparisons within and between groups was estimated using the non-parametric statistical tests (Mann-Whitney U, Kruskal-Wallis H test). All analysis was done using commercial statistical package SPSS (Version 16, SPSS Inc, Chicago, IL) for MS Windows. A *p-value ≤ 0.05* was considered statistically significant during the analysis.
Results {#S7}
=======
The mean time taken for completion of tasks performed under the transparent dome and opaque dome is presented in [Table 3](#T3){ref-type="table"} and the comparison of performance between groups with improvement of performance based on experience is presented in [Table 4](#T4){ref-type="table"}.
Table 3.Mean times (±*SD*) for completion of task for different sub-groups under transparent dome (TD) and opaque (OD) domes.Maze navigationNumber probingObject retrievalObject placementPartial menisectomyTDODTDODTDODTDODTDODNovice66.5 (±10.4)179.7 (±71.4)128.1 (±32.7)657.0 (±195.7)108.2 (±23.4)388.3 (±165.8)230.6 (±46.8)626.2 (±150.3)120.4 (±32.3)364.1 (±156.0)Beginner56.0 (±8.9)130.0 (±32.0)134.0 (±29.5)333.0 (±166.9)90.7 (±11.8)247.7 (±45.3)162.7 (±30.8)433.7 (±116.2)100.7 (±22.9)208.7 (±29.6)Intermediate53.6 (±8.1)104.3 (±18.8)141.0 (±19.0)267.0 (±119.3)79.3 (±2.5)162.3 (±28.9)141.3 (±30.0)269.0 (±46.5)110.3 (±48.3)142.6 (±110.3)Advance57.2 (±10.3)100.7 (±23.8)127.2 (±36.7)258.2 (±71.3)87.2 (±11.2)146.7 (±36.1)159.2 (±43.3)276.2 (±49.2)56.75 (±10.2)85.7 (±39.2)*p value* (*ANOVA*) -- *between groups*0.1520.0270.9240.0010.0840.0110.090\<0.0010.0520.005 Table 4.Difference in performance under the opaque dome (OD) as measured by percent improvement between subsequent groups with higher level of expertise of subjects.Task in ODNovice time (min)Beginner time (min)Intermediate time (min)Advance time (min)Novice to Beginner differenceBeginner to Intermediate differenceIntermediate to Advance differenceMedian (range)Median (range)Median (range)Median (range)% improv*p* value% improv*p* value% improv*p* valueMaze166 (124--360)135.5 (86--163)100 (88--125)97.5 (78--130)750.273670.593330.285Probing684 (325--960)254.5 (240--583)228 (172--401)260 (187--326)1000.068671.00671.00Object extraction319 (227--679)240 (202--309)171 (130--186)143.5 (106--194)750.1441000.109670.285Object insertion630 (456--935)392 (351--600)250 (235--322)270 (225--340)750.1441000.102671.00Meniscus resection364 (203--640)204 (178--249)82 (76--270)94 (37--118)1000.068670.285670.285
No difference was noted in terms of time for task completion in the clear dome but under the opaque dome lesser experienced groups showed greater times. Significant difference was seen on comparison of performance between the groups only under the opaque dome where use of the scope was required ([Table 4](#T4){ref-type="table"}).
Individual task results ([Tables 3](#T3){ref-type="table"}, 4 and [Figures 2](#F2){ref-type="fig"}, 3) {#S8}
------------------------------------------------------------------------------------------------------
Maze navigation: The mean time to completion under the opaque dome was sequentially less as experience level increased with Novice = 179.7 s (±71.4), Beginner = 130 s (±32.0), Intermediate = 104.3 s (±18.8) and Advance = 100.7 s (±23.8), difference between groups being significant with *p* = 0.027. Although oncomparing the improvement in performance ([Table 4](#T4){ref-type="table"}) between the groups, we found that there was no statistical significance (*p* \> 0.05) between subsequent groups in terms of improvement of performance. However, it was notable that improvement was least between intermediate and advance groups (33% positive performance compared to 75% between novice and beginner and 67% between beginner and intermediate).Number probing: As for the maze navigation task no significant difference was found under the TD but the increased experience groups were noted to have faster times under the OD. The mean times for each group under opaque dome were Novice = 657 s (±195.7), Beginner = 333 s (±166.9), Intermediate = 267 s (±119.3) and Advance = 258.2 s (±71.3) with *p* = 0.001, although the trend of improved performance was best between novice and beginner groups (100% positive performance).Object handling: Similar results were noted as the other tasks, with Novice and Beginners being significantly slower than Intermediate and Advance under the OD with *p* = 0.011 for object retrieval and *p* \< 0.001 for object insertion. There was 67% positive performance between intermediate and advance on paired comparison with actually an increased mean time for advance group (276.2 ± 49.2 s) versus intermediate group (269 ± 46.5 s) in the object placement task.Partial menisectomy: [Figure 4](#F4){ref-type="fig"} shows the individual times for each group taken for MM and LM partial menisectomy under TD and OD. The novice group took almost three times the time under opaque dome than under transparent dome for either MM or LM with a *p* = 0.008 whereas the intermediate and advance had hardly any time difference for either the dome or the site of meniscus (*p* = 0.197).
Error rate, face validity and subjective assessment {#S9}
---------------------------------------------------
The errors among the groups were, as expected, higher for lesser experienced participants. There were no errors observed for any group under the transparent dome. Under the opaque dome, the median number of errors in maze navigation was novice-2, beginner-2, intermediate-0 and advance-0 and in object handling it was novice-7, beginner-5, intermediate-4 and advance-2. There were four out of nine novices with inappropriate meniscus resection as compared to one in beginner group and none in intermediate or advance group. The median rating on a scale of 1--5 for smooth scope navigation, image tracking, triangulation, instrument handling and bimanual dexterity was 4 (poor) for novice, 2.5 (between not so good and good) for beginners and 1 (very good) for both intermediate and advance groups. The same was the response to difficulty experienced. All the participants agreed that this model represented basic skills of scope movement, simultaneous image tracking and instrument use, tissue handling and basic meniscus resection skills required for arthroscopy. Irrespective of experience levels, 85% (17) felt there was improvement in performance with subsequent tasks and 90% felt this model would be useful in arthroscopic skills development. Except the advance, all other participants (16 of 20) expressed desire to train on this model to improve their skills.
Discussion {#S10}
==========
We observed that task performance under the transparent dome was not related to experience of the surgeon ([Table 3](#T3){ref-type="table"}, [Figure 3](#F3){ref-type="fig"}) unlike the opaque dome which highlighted the importance of hand-eye coordination required in arthroscopy. Transparent dome tasks require isolated motor skills with direct visualization whereas the opaque dome requires visuo-spatial coordination which is a skill that develops with practice as shown by the difference in performance between the different groups of surgeons based on their experience. The times for maze navigation and number probing show no difference in performance between the groups although tasks for object handling and partial menisectomy show slight improvement in performance with increase in arthroscopy experience but it is insignificant (*p* \> 0.05). When comparing the performance under the opaque dome we found that all the tasks are clearly able to distinguish the skills of surgeons based on experience (*p* \< 0.05), suggesting the use of arthroscope and bimanual dexterity becomes more proficient as the experience increases. This is best observed on comparing the performance for partial menisectomy ([Figure 4](#F4){ref-type="fig"}). We also observed that there is significant difference in time for novice and beginners between transparent and opaque domes (*p* \< 0.005) but intermediate and advance level surgeons hardly had any difference (*p* = 0.197). This means that this model has the ability to distinguish between pure motor skills and the skills required while using an arthroscope. This establishes the construct validity, as it can effectively differentiate the surgeon\'s skills of using the arthroscope and instruments is different and require repeated practice for improving dexterity and visuo-spatial orientation.
Figure 3.Graphical representation of mean time taken by groups to complete tasks under transparent dome (TD) and opaque dome (OD). Figure 4.Individual times (s) for medial meniscus (MM) and lateral meniscus (LL) taken by groups under transparent dome (TD) and opaque dome (OD).
When we compared arthroscopic skills between the groups, we focused on the time for task completion under the opaque dome as the surrogate for competency. There was a consistent trend for lesser time for completion with higher experience level ([Table 3](#T3){ref-type="table"}, [Figure 3](#F3){ref-type="fig"}), except one aberration for the task of object placement where the advance took a mean of 276.2 s compared to 269 s when compared to intermediates. This difference between groups was statistically significant when comparing novice to all groups and beginner to intermediate and advance but not between intermediate and advance with highly significant difference (*p* \< 0.005) for tasks like number probing, object placement and partial menisectomy. This establishes that fine difference in performance after a certain degree of experience requires more than just basic psychomotor skills.
These observations confirm our hypothesis that higher experienced arthroscopic surgeons will perform better in tasks requiring specific arthroscopic skills as seen in many previous studies \[[@R2], [@R7], [@R12], [@R15]--[@R17], [@R19], [@R22], [@R25]--[@R31]\]. We can thus suggest that this model has good construct validity. The differences between groups are not simply in mean time for task performance but also in performance change. The jump in improvement ([Table 4](#T4){ref-type="table"}) is seen most for novice to beginner (75--100%) followed by beginner to intermediate (67--100%) and least for intermediate to advance (33--67%). This improvement reflects that development of comprehensive technical skills in the early stages of arthroscopic training may be faster and more efficient. This kind of assessment would also allow identification of trainees who may pick up skills better than other. This is also shown in other studies where certain students acclimate to arthroscopy earlier than others \[[@R31], [@R32]\].
The face validity of this construct should be established before we can suggest its use as a surrogate for arthroscopic skills training. Using grading from very good to very poor to assess performance of different groups in terms of arthroscope navigation, image tracking, triangulation, object and instrument handling and bimanual dexterity we found that there was a progressive improvement in median grade from 4 (poor) to 1 (very good) from novice to advance. This was coupled by similar feedback response by participants as grade 4 (difficult) for novice to 1 (very easy) for advance group. Similar feedback assessment for face validity was used by Braman et al. \[[@R30]\]. We acknowledge that our analysis is limited by the fact that the number of participants in each of the groups was small and statistical comparison could not be made. Nonetheless, the objective construct validity was reinforced by the sequential improvement in performance and response-based face validity making the model applicable for training surrogate. The entire advance group with more than 15 years or arthroscopic experience between them agreed that these four tasks adequately represent the basic skills in arthroscopy.
Establishing benchmark criteria for objective assessment of skills and for a stepwise training is an important criterion for any training module. Our results show that, for novice to beginners and beginners to intermediates and/or advance, the time for task completion is significantly disparate and shows that the differences are in a gradual slope. We suggest the possibility of using the median time for task performance of beginners and intermediates as guideline to assess the improvement during training and for progression of skills in a structured program (from [Table 4](#T4){ref-type="table"}). If we take the task of meniscus resection as an example, a novice that begins training on day one and trains for a fixed period of time e.g. two weeks, then, to consider him competent with the technical skill, he should perform the task within 205 seconds (beginner level). At this stage the training continues and the next goal is to be able to do the same task within 80--95 seconds (intermediate-advance level). This would allow an objective assessment of skills and help trainers to recognize adequacy of skills in trainees to progress to clinical setting.
Cadavers and computerized high fidelity simulators which incorporate three-dimensional anatomy, virtual reality, haptic feedback, trajectory and force data analysis are useful in providing a training atmosphere which attempts to recreate anatomy, tissue response and clinical scenarios \[[@R2], [@R4], [@R7]--[@R9], [@R12]--[@R17], [@R21], [@R22], [@R26], [@R27], [@R33], [@R34]\]. Recent evidence also suggests that there is considerable ability to transfer skills acquired on simulators to the operating room \[[@R34]\]. However, both cadaveric and high technology computerized simulators are limited by availability, expense and resources \[[@R8], [@R9], [@R11], [@R19], [@R20], [@R35]\]. Low fidelity simulation allows opportunity to learn and practise basic skills with goal-directed modules; like the incentive to reach training benchmarks to allow trainees to progress \[[@R23], [@R24], [@R31], [@R32], [@R36]\]. We have used a low fidelity *box-type* bench model and assessed its utility in distinguishing arthroscopic skills and possibility of using the results to develop a structured training program.
Limitations and future directions {#S11}
---------------------------------
This study is limited by small sample size and insufficient variation in tasks like more repetitions, changing orientation of platforms which would have provided more data for better comparison. However, we take this study as an opportunity to differentiate surgeons on basic skills and allow us to generate guidelines for training. We suggest using this or a similar module for assessing the baseline skills and then using the performance based on experience to develop a structured program. An example from this study that can provide a baseline to develop benchmark scores is timing of intermediate level surgeons. For the future, we aim to use these parameters for trainees on this model and assess their skills over a period of time (an on-going study). However long the time taken to reach the target score, the trainees need to stop when they reach intermediate level timings. Although further recommendations will not only depend on the time spent on the model but are influenced by other factors as well.
Conclusions {#S12}
-----------
Pure motor skills with direct visualization tasks are inherent skills of surgeon but the use of arthroscope needs visuo-spatial coordination; a skill that develops with practice as shown by the performance of different groups of surgeons based on their experience. From this study we conclude that this model has adequate construct validity for distinguishing level of basic arthroscopic skills among surgeons and provides us with guidelines for further research (which we are doing currently) to see progression of trainees as they spend more time training on this model. This model is especially useful in institutions where resources to a develop surgical skills laboratory are limited.
Conflict of interest {#S13}
====================
The author(s) declare no conflict of interest in relation with this paper.
**Cite this article as:** Goyal S, Radi MA, Ramadan IK & Said HG (2016) Arthroscopic skills assessment and use of box model for training in Arthroscopic surgery using Sawbones -- 'FAST' Workstation. SICOT J, **2**, 37
[^1]: Out of the total number in the group.
| {
"pile_set_name": "PubMed Central"
} |
Plain english summary {#Sec1}
=====================
The effects of a 14-week DMT intervention on body image and alexithymia were tested in 12 patients with EDs. The DMT group (*n*=7) improved in various aspects of body image after the treatment, while no change occurred in the control group (*n*=5). Both groups continued to receive the standard treatment offered by the Clinic. Neither of the groups improved significantly in alexithymia scores. The qualitative analysis revealed that participants were satisfied with this intervention. According to this analysis, the DMT sessions increased their self-awareness, improved their mood states, and promoted the building of meaningful relationships. Taken together, the findings suggest that DMT could contribute to the treatment of EDs on various levels.
Background {#Sec2}
==========
The prevalence of Eating Disorders (EDs) is increasing worldwide \[[@CR1]\]. Dropouts \[[@CR2]\] from eating disorder therapy and full relapses are common \[[@CR3]\]. Comorbidity with other psychological disorders and health issues pose a challenge for professionals and require multidisciplinary treatments \[[@CR4]\]. Widely used specialized treatments for EDs mostly focus on cognitive-behavioral aspects and weight restoration \[[@CR5]\]. However, randomized controlled trials have generated contradictory results regarding the effectiveness of those treatments \[[@CR6]\]. Furthermore, patients with EDs are often not content with the treatment they receive according to a review of qualitative research \[[@CR7]\]. In light of these findings, many authors have underlined the need for novel approaches to treat EDs \[[@CR8], [@CR9]\].
One of the core characteristics and a risk factor of EDs is body image disturbance, which has implications on cognitive, emotional and behavioral levels \[[@CR10]\]. A person's body image is formed since birth. The first experience of the self is through proprioception, i.e. sensations that arise from within the body, which develops by a person's perceptions while relating to others and influences the emotions, attitude, psychological functioning, and self-value \[[@CR11]\]. Cash describes it as an "inner view" which refers to the multidimensional embodied experience related mostly to one's physical appearance \[[@CR12]\]. Patients with EDs, especially those with anorexia nervosa, tend to focus on shape, weight, and food perhaps as a way of ignoring thoughts and emotions that provoke anxiety \[[@CR13]\]. Thus, they consider their bodies with disdain and often fail to recognize their bodily needs \[[@CR14]\].
Another personality construct that is common in individuals with EDs is alexithymia \[[@CR15]\], which literally means "no words for emotions". Alexithymia is characterized by an externally focused cognitive style and a difficulty in differentiating between feelings and bodily sensations \[[@CR16]\]. This personality construct is also related to poor self-regulation \[[@CR17]\]. As people with alexithymia face difficulties in mentalizing the emotional state of others and their own, they can be perceived as non-empathetic and experience difficulties in social relations \[[@CR18]\]. Alexithymia is also associated with poor creative capacity and a limited ability of symbolization and play \[[@CR19]\]. The absence of consciousness of the internal experience and the difficulty of trusting the perception of one's emotions, sensations, and reflections are related to the sense of poor self-efficacy that patients with EDs face, as a result of not being able to resolve the conflict between the illusion of perfectionism, their actual potentials, and the expectations of others \[[@CR20]\].
The development of body image and socioemotional functioning starts on a nonverbal level in infancy \[[@CR21]\]. The infant first experiences its body through the tactile experience provided by the caregiver \[[@CR11]\]. The empathetically attuned interactions with the caregiver provide a secure holding which defines the infant's inner state by offering psychological and physical soothing, something that is the critical basis for secure attachment \[[@CR22]\]. The term of psychesoma as defined by Winnicott \[[@CR23]\] has its origins in the holding phase. When the handling is good enough the psyche can indwell in the body (soma) providing a sense of internal reality and structure of the self \[[@CR23]\]. On the contrary, misattunement with the needs of the infant can arouse an inner confusion and a detachment between the subjective and physical elements of emotions, which is at the core of ED's pathology \[[@CR20]\]. The foundation of emotion-regulation and self-regulation is based on the secure attachment with the caregiver \[[@CR24]\]. It is well established that attachment insecurity is an important risk factor for the development of EDs and a predictor for poor outcomes \[[@CR25]\]. One study has found that improvement in attachment insecurity led to a decrease in interpersonal problems in patients with EDs \[[@CR26]\]. Moreover, a review of three meta-analytic procedures found that secure attachment style is a crucial predictor for beneficial psychotherapeutic outcomes \[[@CR27]\]. These findings indicate that treatment models for EDs should address attachment related disorders. Considering that the development of attachment is rooted in a pre-verbal phase, such a treatment model can make use of a bodily, non-verbal component.
The importance of bodily experiences and body memory for the psychological functioning has been shown by cognitive neuroscience studies \[[@CR28]\]. Recent research in interpersonal neurobiology considers the brain in the whole nervous system, existing in the whole body \[[@CR29]\]. These findings have encouraged psychotherapists to investigate the role of the body in verbal psychotherapies \[[@CR30]\]. One bodily based approach is Dance Movement Therapy (DMT), which explores and uses movement in a therapeutic way with a focus on self-awareness, in order to promote the psychophysical integration of the individual \[[@CR31]\]. This psycho-dynamic model, forms part of the Creative Art Therapies, in which music, art, and drama therapy are also included \[[@CR32]\]. Some of the core characteristics among others for this group of therapies are the possibility of active participation, the self-expression, the emphasis in imagination, and the mind-body connections \[[@CR33]\]. In particular, DMT was formed by dancers and psychotherapists in the 1940s, who realized the therapeutic impact of the artistic expression in mental health patients \[[@CR33]\]. It is a form of psychotherapy based on embodiment and enaction \[[@CR34]\]. The core basis of DMT is the innate connection between body and mind \[[@CR34]\]. The unintentional expression through movement is fostered, which permits the enactment of unconscious ideas and emotions that often cannot be put into words \[[@CR35]\]. This not only discharges psychological or corporal tensions but, simultaneously, can help the individual become aware of his/her emotional resources and strengths \[[@CR36]\]. This therapeutic approach is both non-verbal and verbal \[[@CR37], [@CR38]\]. The non-verbal interactions can promote insights regarding the participants' behavior, beliefs, relationship patterns, and emotional apparatus \[[@CR39]\]. The verbal communication of thoughts and emotions has a fundamental role in the therapeutic process, and gives meaning to the symbolic aspects of the movement \[[@CR33]\]. The therapist-patient relationship is a central component in DMT approximation. The DMT therapist uses specific tools of movement analysis in order to observe and interpret the participants' body language. Building upon these observations, techniques such as attunement and mirroring are applied, which provide a link between attachment and non-verbal behavior \[[@CR40]\]. The creative play and the transitional object concepts are also included in the therapeutic process \[[@CR41]\]. Those means help the improvement of embodied relationships and give meaning to the movement \[[@CR42]\]. The emotional state of the patients, as well as their early misattuned experiences with the caregiver, can be experienced, explored, felt and understood by the patient \[[@CR34]\] through a deep attunement with the therapist on an emotional and bodily level \[[@CR43]\]. These moments of meeting serve as a healing emotional experience \[[@CR44]\]. The union between body and mind, as opposed to the mind-body dualism, assumes the principle that "the body reflects the mind and the mind reflects the body" \[[@CR43]\]. Following up on this principle, our expectation is that changes in movement patterns can cause a modification in the internal comprehension of the individual and vice versa \[[@CR45]\]. Other authors like Jung have mentioned the importance of the symbolic part in the therapeutic process \[[@CR46]\]. The symbolic part of movement and the communication via movement metaphors, which is a core feature of DMT, can be helpful for alexithymic patients as it can promote the exploration of one's emotional state and facilitate the therapeutic process \[[@CR35]\]. The bodily and sensorimotor components of empathy are indicated by studies in neuroscience \[[@CR47]\] so that, the psychotherapeutic use of body movement in DMT can enhance empathy among alexithymic individuals. In the same way, DMT can help people who suffer from EDs to experience themselves in a more complete and embodied way, since they are encouraged to make connections between the cause factors of EDs and the corporal manifestations of them \[[@CR48]\], gaining body awareness \[[@CR34]\].
Significant improvements have been associated with the use of DMT as an adjunct in the treatment of a wide variety of disorders, including anxiety, depression, schizophrenia, dementia \[[@CR49], [@CR50]\]. Improvements included an increase in quality of life ratings, as well as interpersonal and cognitive skills \[[@CR50]\]. Such outcomes were comparable to pharmacological interventions and verbal psychotherapies \[[@CR49], [@CR50]\]. Some studies have found DMT interventions to improve the body image in patients with obesity \[[@CR51]\], cancer \[[@CR52]\] and fibromyalgia \[[@CR53]\]. Apart from those findings, a major amount of research regarding DMT is focused on case studies and qualitative research which could be explained by the characteristics of DMT as the emphasis is put on the subjective experience, therapeutic relationship, personal development and change of the person as well as on the creative and aesthetic part of the therapeutic process \[[@CR54]\]. Considering the alexithymic aspects and body image issues in EDs, treatment should promote the physical and psychic integration of the individual, which is why DMT is a potentially adequate approach for patients with EDs.
The proposed therapeutic intervention was designed for the needs of a group, which consisted of patients with EDs. As it was mentioned in previous studies, difficulties in accessing self and body awareness as well as the resistance towards movement exploration \[[@CR55]\] and negative body experiences were identified \[[@CR56]\]. The guided imagery by Reddemann \[[@CR57]\] was used as a mediator to facilitate the connection of the participants with the body and the self in movement and to enrich one's movement vocabulary.
The current research is a pilot study. It aims to explore a DMT intervention that could complement the already existing ED treatment in an attempt to provide a more holistic and effective approach. Due to the theoretical considerations, research findings and the mixed nature of the sample, a transdiagnostic perspective was followed \[[@CR58]\]. Firstly, changes in the body image and alexithymic traits as a result of the proposed intervention were explored. Secondly, as the DMT intervention on the ED population is rarely investigated, a qualitative analysis was included in the research, to develop a better understanding of the patients' experience. To the authors' opinion, a mixed study that integrates quantitative and qualitative analysis can provide a more holistic view.
Method {#Sec3}
======
Participants {#Sec4}
------------
The participants were 14 young women who attended a private day clinic in Spain, specializing in the treatment of EDs. Participants' ages ranged from 14 to 32 (M = 20.28, SD = 5.91). BMI ranged from 14.03 to 24.69 (M = 18.94, SD = 3.56). Criteria of inclusion were: receiving standard treatment from the Clinic and having been diagnosed with an ED according to the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) \[[@CR59]\]. Exclusion criteria were: comorbidity with personality disorders and/or the administration of medications, as those could modify the movement qualities of the participants. These criteria were aiming at creating a more homogeneous group in order for the obtained results to be as representative as possible for the ED population. One participant was excluded from the study as she did not complete the predetermined criteria. The final sample size was determined by the number of the patients in the Clinic who fit these criteria.
Setting and procedure {#Sec5}
---------------------
The Clinic where the study was conducted offered mainly a cognitive behavioral treatment approach. The weekly schedule consisted of three group therapies, six psychoeducational sessions, and one individual therapy session for each patient. The number of sessions that each patient received was adapted according to her needs. The Clinic was informed of and approved the structure, content, and procedures related to the DMT intervention, as well as the possibility of publication of the obtained research data. Written approval was provided to the authors by the ethical committee of the Clinic. All participants signed a consent form, agreeing to the anonymized use of self-report questionnaires in the current study. The DMT group signed a second consent form regarding the anonymized use of reflective diaries that were used in this investigation. As part of the consent forms, they were provided with detailed information about the framework of the investigation and the research questions. The groups were quasi-randomized, as a professional from the Clinic assigned the participants to the DMT group or the control group depending on their timetable. The DMT group was semi-open; new participants could join the sessions once they entered the day clinic and they had to leave the group when they were released from the Clinic. Twelve sessions were conducted over a 14-week period. Participation was optional. Both, intervention and control group, continued their treatment in the Clinic as usual. Self-report questionnaires were first completed prior to the participants' entry to the DMT sessions, and again within four weeks from the last session. The control group followed the same procedure. The research, as well as the interventions, were part of the Master's training in Dance Movement Therapy, which runs under the Department of Psychology of the Autonomous University of Barcelona. The sessions were facilitated by a dance movement therapist in training, who received clinical supervision from the University during the entire process. The research was supervised by a researcher from a different university, who specialized in DMT and EDs. Another dance movement therapist in training participated in the data analysis.
Intervention {#Sec6}
------------
The DMT sessions were based on Chace's model \[[@CR60]\], and guided imagery according to Reddemann \[[@CR57]\]. Each session was divided into 6 parts: check-in, warm-up, guided imagery, exploration in movement, writing, check out. During check-in, the participants were invited to share their current emotional state. Based on that, the therapist guided the warm-up in order to activate the body and prepare the participants for the main part of the session. During the guided imagery part, the participants were provided with self-regulatory techniques in order to remain in the "here and now" and to reduce stressful thoughts. The aim was to experience feelings of safety and comfort as well as to explore one's own emotional resources. During the exploration in movement, themes that have emerged in the check-in, warm-up or guided imagery part could be accessed in depth. The creative process aimed to transform participants' inner experiences into external realities. The objectives and dynamics proposed in the movement exploration part are shown in Table [1](#Tab1){ref-type="table"}. Afterwards, the participants had to write down their reflections upon the process. In addition, to the investigation purposes, this part also aimed at reintegrating the lived experience. During the check-out, the participants were given the opportunity to verbalize and share their experiences. Table 1Dynamics of the DMT sessions**ObjectiveInterventionExamples**Connection with one'sBreathing exercises--Participants observe the impact of deep breathing in their bodies.self-- Breathing in circle: A touches her stomach and the back of B,B touches her stomach and the back of C, and so on.Use of light and sustained movements.-- Participants observe their own breathing and the one of theother's.Body awarenessExploration with--Participants move an object through different parts of the bodyobjectsand observe the impact of this contact and the difference betweenthe parts of the body that have not been touched.Use of light, free, direct, and sustained movements.Interaction among theMovement in pairs-- Participants work in pairs:participants-- A touches B and B responds by accepting or rejecting the touch.--A touches B and B moves that part of the body. Light, indirect movements.-- A touches B and B resists by putting force in that part of the body. Strong,direct movements.Exploration of limitsExploring the space-- Participants use objects (e.g., wool, elastics, fabrics) to visualizetheir personal space.- Exploration with elastics: First each participant uses an elasticband to integrate her muscles in movement and explore her limits.As a next step, they continue this exploration in pairs, sharingone elastic band. Use of different movement qualities.Group cohesionGroup games-- First in a circle, participants pass each other the ball in a certainorder, saying the name of the person that the ball is directed to.They continue in the same order without saying the names.Movement and music can be introduced and the pace can be varied.-- Use of fabrics: After experimenting individually with a fabric,participants get together in small groups and create a group dance.EmpathyMovement exploration-One participant chooses a place in the room and closes her eyes(optional). Other group members slowly approach her and offer thetouch they think she needs (e.g., holding a hand, gently touching thehead). This activity can be proposed once there is sufficientconfidence in the group. Participants are advised to say no if theyfeel uncomfortable. Use of light, direct, and sustained movements.ReflectionGroup circle-- After experimenting with different movement qualities (e.g. fluid,robotic) participants are invited to share their associations withthese, and possibly relate them to their personal lives and EDsymptoms.-- Participants use different objects to transmit to the group whatthey felt during the session.Note: The references in movement qualities, derived from Laban movement analysis that is used in DMT. They are related to the indulgence or the resistance of the participant and may indicate the way of relating with the other.
The sessions lasted 90 minutes. For the check-in and check-out (oral verbalization) 10' was given for each. The warm-up took 10'. For the guided imagery part the participants had 20', followed by the movement exploration (30'). Finally, the writing part (personal reflections) took 10'. The time could be varied according to the general group status.
The intervention was divided into the following phases: in the first four sessions (1-4), the focus was on creating an atmosphere of safety and trust which played a crucial part throughout the whole therapeutic process. The connection of the participants with their own self was also promoted from the very first encounters. From the 5^th^ to the 10^th^ session the focus was shifted to a facilitation of the group cohesion. Participants were invited to explore their movements in pairs which led to empathetic interactions among the group members. The main aim of this phase was to promote self- and body-awareness as well as creating a safe atmosphere where the participants could share their experiences. During the last two sessions (11-12) the focus was shifted in preparing the participants for the upcoming closure of the DMT intervention. The therapist brought the participants' attention back to the connection with their own self and with others, as a way to support their internal reorganization.
The sessions were conducted in a spacious room of the Clinic, where participants could move freely. Objects played a significant part in the therapeutic process. They were used in the majority of the sessions in order to facilitate the contact of the individuals with their bodies, as well as their interaction with the other participants. The objects used were: elastics, different forms of fabrics, small hard (tennis) balls and soft balls, balloons, ribbons, scarves, paper and crayons.
Qualitative measures {#Sec7}
--------------------
Participants of the DMT group were invited to keep reflective diaries in order to obtain a better understanding of their processes. Each participant was given a notebook and wrote in it freely at the end of each session. Due to the exploratory nature of this study, no specific instructions were given about the content of these texts. The therapist told the participants to write about whatever they wanted, in order to minimize possible biases. Reflective diaries were transcripted, and codified. Successive readings of the material were made looking for themes. This process was repeated until saturation was reached, following a classical content analysis \[[@CR61]\]. Investigator triangulation was applied by including a second researcher who independently and blind to the intervention process, carried out the content analysis. The whole analysis was made once the therapeutic process was over. Nvivo 12 software was chosen for the analysis \[[@CR62]\]. The extracts from the reflective diaries presented in the current research were translated to English by one of the authors. The trainee therapist's observation diary was also included as another source of data collection. Information from the therapist's observation diary was used to illustrate some specific moments during the sessions.
Quantitative measures {#Sec8}
---------------------
Two self-report questionnaires were used for the quantitative part of this study.
Toronto Alexithymia Scale (TAS-20) \[[@CR63], [@CR64]\]. The Spanish version \[[@CR65]\] of the TAS-20 was used to measure alexithymia. Items are rated on a 5-point Likert scale. It consists of three subscales: Difficulty Identifying Feelings, Difficulty Describing Feelings to others, Externally Oriented Thinking. The scales have adequate reliability (*α*=.794, *α*=.732, *α*=.613 respectively) and adequate convergent validity.
Multidimensional Body-Self Relations Questionnaire (MBSRQ) \[[@CR66]\]. The 45-item Spanish adaptation of the MBSRQ questionnaire was used to measure body image \[[@CR67]\]. The subscales Appearance Evaluation (satisfaction or dissatisfaction with one's physical appearance), Appearance Orientation (importance placed on one's physical appearance and effort put in maintaining one's looks), Body Areas Satisfaction (satisfaction or dissatisfaction with specific body areas e.g., stomach, legs, etc.) and Overweight Preoccupation (ideas regarding fat anxiety, weight vigilance, etc.) were used in this pilot study. Cronbach's *α* were calculated for the MBSRQ subscales used in the current study. All the subscales were found to have adequate internal consistency (*α*=0.937, *α*=0.867, *α*=0.929, *α*=0.900).
Data analysis {#Sec9}
-------------
Following the confirmation of the parametric distribution of the quantitative variables, paired-samples t-tests were conducted to compare differences pre- to post-intervention in each group. Cohen's d effect sizes were calculated. For not normally distributed data, Wilcoxon Signed Rank test was applied to compare pre- and post-intervention scores in each group and Rosenthal's r effect sizes were calculated. Values of 0.2, 0.5, and 0.8 are interpreted as small, medium, and large effect sizes respectively \[[@CR68]\]. Given the small sample size, both statistically and clinically significant results were taken into account. Statistical Package for the Social Sciences (SPSS), version 20.0 for Windows \[[@CR69]\] was used for statistical analysis.
Results {#Sec10}
=======
Demographic summary {#Sec11}
-------------------
Of the 14 young women who were recruited for the study, 7 were assigned to the DMT group and 7 to the control group. They were all high school or university students of Spanish nationality. The DMT group consisted of 4 EDNOS and 3 AN patients aged between 14 and 32 (M = 20.1, SD = 5.9). The BMI of the DMT group ranged from 14.03 to 24.69 (M = 19.82, SD = 4.37). The length of time since their admission to the Clinic at the beginning of the intervention ranged from 1 to 23 weeks (M = 13.43, SD = 9.11). Two participants from the control group did not complete the follow-up questionnaires, so that they had to be excluded from the sample. The final control group (*n* = 5) consisted of 2 AN, 2 EDNOS and 1 BN patients aged between 17 and 23 (M = 20.3, SD = 2.5). The BMI of the control group ranged from 17.33 to 22.9 (M = 19.07, SD = 2.28). The length of time since their admission to the Clinic at the time of the first data collection ranged from 24 to 59 weeks (M = 42, SD = 15.7). T-tests showed that the two groups were homogeneous concerning their age (t=0.21; *p*=0.84) and BMI (t=0.19; p=0.83), however not regarding the length of time at the Clinic. The participants of the control group had been in the Clinic for significantly longer period at the time of the first data collection, t=4.0, p=0.02. The implications of this difference for the current study are considered in Limitations.
The 7 participants in the intervention group completed a minimum of 4 and a maximum of 11 sessions (M=7.71, SD=2.93). One participant completed 11, two completed 10, one completed 9, two completed 5 and one completed 4 sessions. As a result, an average of 4 to 5 participants were involved in each DMT session.
Quantitative outcomes {#Sec12}
---------------------
The mean scores for the MBSRQ subscales are presented in Table [2](#Tab2){ref-type="table"}. Prior to intervention, for Appearance Evaluation participants in the DMT group were more than 1 SD below the norm (M = 3.36, SD = 0.87). For Appearance Orientation they were within 1 SD from the norm (M = 3.91, SD = 0.60). For Body Areas Satisfaction they were more than 1 SD below the norm (M = 3.23, SD = 0.74). For Overweight Preoccupation they were more than 1 SD above the norm (M = 3.03, SD = 0.96) \[[@CR66]\]. As all the MBSRQ subscales were normally distributed in both groups, paired samples t-tests were conducted to compare the pre- and post-intervention scores in each group. As can be seen in Table [2](#Tab2){ref-type="table"}, the DMT group increased significantly in Appearance Evaluation and Body Areas Satisfaction, and decreased significantly in Appearance Orientation pre- to post-test. Large Cohen's d effect sizes were observed for each of these variables. There was a tendency to decrease in Overweight Preoccupation which was not statistically significant, however a medium to large effect size was observed. Post-test mean scores of the DMT group for Appearance Evaluation, Body Areas Satisfaction, and Overweight Preoccupation were within 1 SD from the norm. No significant differences were found in the control group between pre- and post-test in any of the subscales. Table 2Results of paired samples T-Test for pre- and post-intervention for MBSRQ subscales in DMT and Control groups**Descriptive statisticsT-testMBSRQGroupPretestPosttestt (df)pCohen's dSubscalesMean (SD)Mean (SD)**AppearanceDMT2.34 (1.09)2.94 (0.81)-2.51 (6).046-.95EvaluationControl2.80 (0.97)3.04 (0.89)-1.04 (4).358-.46AppearanceDMT3.92 (0.79)3.41 (0.73)3.44 (6).0141.30OrientationControl3.43 (0.88)3.46 (0.73)-.23 (4).831-1.02Body AreasDMT2.24 (0.80)2.76 (0.71)-2.91 (6).027-1.10SatisfactionControl3.07 (1.27)2.93 (1.12).75 (4).494.34OverweightDMT4.29 (0.76)3.64 (0.56)2.00 (6).093.75PreoccupationControl2.70 (1.92)2.80 (1.82)-.41 (4).704-.18
When a mixed-effects ANOVA model was applied, results for time effect were consistent with previous results for Appearance Evaluation (F=5.93; $\documentclass[12pt]{minimal}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$\eta ^{2}_{p}=0.37$\end{document}$; *α*=0.04) and Appearance Orientation (F=5.54; $\documentclass[12pt]{minimal}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$\eta ^{2}_{p}=0.36$\end{document}$; *α*=0.04). For Body Areas Satisfaction time effect was not significant (F=2.23; $\documentclass[12pt]{minimal}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$\eta ^{2}_{p}=0.18$\end{document}$; *α*=0.17). Regarding the interaction effect of time and group, significant changes appeared in Appearance Orientation (F=6.90; $\documentclass[12pt]{minimal}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$\eta ^{2}_{p}=0.41$\end{document}$; *α*=0.03) and Body Areas Satisfaction (F=6.32; $\documentclass[12pt]{minimal}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$\eta ^{2}_{p}$\end{document}$=0.39; *α*=0.03). The interaction effect for Appearance Evaluation was not statistically significant (F=1.09; $\documentclass[12pt]{minimal}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$\eta ^{2}_{p}=0.09$\end{document}$; *α*=0.32). Finally, for the Overweight Preoccupation the ANOVA model was consistent with the previous analysis. Neither time (F=1.55; $\documentclass[12pt]{minimal}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$\eta ^{2}_{p}= 0.13$\end{document}$; *α*=0.24) nor interaction effects (F=2.90; $\documentclass[12pt]{minimal}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$\eta ^{2}_{p}=0.23$\end{document}$; *α*=0.12) were significant.
Prior to the intervention, alexithymia scores ranged between 47 and 71 in the DMT group and 27 to 60 in the control group. According to the cutoff scoring, in the DMT group 4 participants did not display alexithymia, 1 displayed possible alexithymia and 2 displayed alexithymia. In the control group, 4 people did not display alexithymia, and 1 displayed possible alexithymia. As the data was not normally distributed, Wilcoxon Signed Ranks tests were undertaken to compare pre- and post-intervention results of the DMT and control groups (see Table [3](#Tab3){ref-type="table"}). There was a slight tendency of decrease in the total score and the subscales of alexithymia in the DMT group, which was not statistically significant. Small Rosenthal effect sizes were observed in Difficulty Identifying Feelings and in Difficulty Describing Feelings to Others. On the contrary, the control group had a tendency to increase in the total score and the subscales of alexithymia. This increase was not statistically significant, however moderate effect sizes were observed in the Alexithymia total score, in Difficulty Identifying Feelings, and in Difficulty Describing Feelings to other, and a small effect size was observed in Externally oriented thinking. Table 3Results of paired samples T-Test for pre- and post-intervention for Alexithymia subscales in DMT and Control groups**Descriptive statisticsWilcoxon signed rank testGroupPre MdnPost MdnzpRosenthal's r(IQR)(IQR)**AlexithymiaDMT49 (48-70)46 (42-61)-0.593.553-.16Total ScoreControl35 (30-52.5)50 (40-53)-1.753.080-.55DifficultyDMT21 (17-26)20 (12-26)-0.734.463-.20Identifying FeelingsControl9 (9 - 21.5)13 (10-23)-1.841.066-.58Difficulty DescribingDMT14 (13-20)11 (10-19)-1.156.248-.31Feelings to OthersControl10 (6.5-13.5)13 (11.5-14.5)-1.604.109-.51Externally OrientedDMT18 (15-23)17 (15-21)-0.171.865-.05ThinkingControl14 (11-22)19 (12.5-24.5)-0.816.414-.26Note: Cut-off scoring for Alexithymia total score goes as follows: \<51 = non-alexithymia, 51-60 = possible alexithymia, \>60 = alexithymia
Qualitative outcomes {#Sec13}
--------------------
There were five domains that emerged from the analysis of the qualitative data. Within each domain, a number of themes are identified (see Table [4](#Tab4){ref-type="table"}). Each theme is presented, followed by a selection of related extracts from the data. Due to space limitations only some of the evidences are presented to illustrate each category. The code of each quote indicates the code given to the participant and the number of the session when the text was written (e.g., P1S01 refers to Participant 1, Session 1). Table 4Qualitative analysis. Domains, themes, and sub-themes**DomainThemeSub-theme**1. Emotion, mood,1. Emotional states (7)1.1.1. Patterns of mood changesand alexithymiaduring the sessions (6)1.1.2. Getting in touch / dealingwith strong feelings (3)1.2. Alexithymia, connection, and disconnection (6)2. Body and movement2.1. Joys of movement (7)2.2. Difficulties related to DMT activities (7)2.3. A pattern of activating during the session (6)3. Interpersonal aspects3.1. Enjoyment of the interactions in the group (6)3.2. Difficulties in relationships (5)4. Metaphors4.1. "Safe place" (6)4.2. The use of metaphor to describe a feeling (2)5. Reflections5.1. Reflections on one's process (3)5.2. Looking back on the DMT sessions (4)Numbers in parentheses indicate how many participants shared this category
### Domain 1: Emotion, mood, and alexithymia {#Sec14}
**Emotional and mood states**
All the participants made attempts to identify and describe how they felt at the beginning, during or at the end of the session, related to both the present moment and the larger context of their lives. The level of detail of their descriptions varied, from "fine" to naming specific emotions to describe how they felt (the use of metaphors to describe a feeling is considered later, under the theme of metaphors).
Extracts 1 "*"Today I was emotionally less calm and peaceful because of circumstances and decisions to take in the near future." - P1S04*"
"*"Today I came in feeling indifferent and tired, I didn't feel like doing anything after feeling not understood by my father yesterday." - P5S08*"
These participants described their moods briefly, comparing it to the previous days or weeks, or relating it to a situation that triggered these feelings. They mostly referred to anxiety vs. calmness, happiness vs. sadness and a lack of motivation.
**A pattern of mood shift during the session**
Most participants reported changes in their moods during the sessions.
Extracts 2 "*"Today I've felt very calm. I've been through a couple of days with a lot of anxiety and the session helped me to lower it." -P7S12*"
"*"I'm leaving the session more energetic and happy, feeling better than I was before. Motivated to do many things and endure whatever may come. Stronger." -P5S10*"
"*"I am leaving the session a bit calmer and more relaxed (even though I don't think it will last long, but I'm happy that even though for only a while I wasn't stuck in my negative thoughts.)" -P6S10*"
In these instances, the participants considered that the DMT sessions helped them reduce their anxiety, or offered a temporary relief from it. Moreover, they reported feeling more motivated after the sessions.
**Getting in touch with strong feelings**
Some participants described getting in touch with strong feelings that they probably were unaware of.
Extract 3a "*"I was surprised that I cried so much, so deeply and so intensely. I wasn't aware that I was feeling all that so intensely." - P6S07*"
P6 expresses her surprise as she becomes aware of the strong feelings that were unknown to her before the session.
Extracts 3b "*"When A.*[1](#Fn1){ref-type="fn"}*encouraged me to work through my anger, I started feeling it very strongly (even crying), first I thought it was anger towards myself, and then I realized it was anger towards the disease." P2S08*"
"*"I felt moved by P2's thing. Oddly enough my head was very lucid there, I was left with the urge to deal with my own anger because I feel that I have a lot and I quite identify the images, where they come from. But again, if I don't have a clear secure place to turn to, perhaps it's not a good idea to go to this anger because I might get stuck there \[...\] I feel confused, my head is going to explode." P6S08*"
These extracts illustrate the attempts of these participants to accept and work through their anger, as well as their difficulties to stay with these feelings. P6 expresses how P2's process inspires her to work through difficult feelings herself.
**Alexithymia, connection, and disconnection**
In addition to the quality of a feeling, a theme of recognition and awareness of the feelings emerged. Participants often included a dimension of being "connected" versus "disconnected" when talking about their mood states, which in clinical terms seem to describe their varying experiences of dissociation. The term dissociation is used loosely here in order to encompass the wide range of experiences of being more or less present in the body as described by the participants.
Extracts 4a "*"I don't know exactly how I felt." - P2S05*"
"*"Initial feeling: I've been emotionally shattered the whole week, without knowing exactly how I feel. Physically I feel well, but inside I'm nervous (because of the exams), tired of everything, sad..." - P5S03*"
These participants recognized their difficulties in identifying feelings. P5 described an apathetic mood that had been going on for several days.
Extracts 4b "*"I felt very comfortable during the session. That's for sure, but sometimes it was easy for me to disconnect and think about other things." - P3S03*"
"*"At the beginning of the session, I found it very difficult to connect with myself, to stop connecting with my anxieties and unease, and to focus on the moment, on living in the moment." - P5S11*"
"*"Today I feel absent, my head is somewhere else. I realize that everything is heavy. My body, my head, my soul, everything..." - P6S08*"
These extracts show the variety of meanings that connection and disconnection have for the participants. Both P3 and P5 mention having difficulties to live in the present moment and an urge to think of other things. For P3, this seems to enable her to escape from current stimuli *"it was easy for me to disconnect"*, while for P5 getting into contact with the present is overwhelming. In general, these extracts present a tendency of these individuals to live more in their minds rather than in their bodies. In the last quote, as P6 acknowledges her feelings of absence, her level of self-awareness increases paradoxically as she directs her attention to the other aspects of her being.
Extract 4c "*"After the breathing exercises, I felt very RELAXED. I was so focused on the breathing and on A.'s voice in my head that I even forgot about the existence of my body and it wasn't until when A. told us to delicately start moving it \[our bodies\] that I remembered it again and I started feeling it and connecting with it again. It was a strange feeling but very pleasant and I felt an absolute physical and mental relaxation." - P6S12*"
P6 describes a subjective disconnection from the body during relaxation that she experiences as pleasant. Her narrative suggests a powerful experience *"forgot about the existence of my body"* that is positive to her, probably because she takes a break from the negative thoughts that were mentioned in the previous section.
### Domain 2: Body and movement {#Sec15}
This domain includes participants' responses to DMT activities that involve the body and its movement. These activities can vary from breathing techniques, becoming aware of the body and exploring different movements with it, individually, in pairs or as a group, with or without materials. (Interpersonal aspects of these activities will be discussed in Domain 3).
**The joys of movement** Participants enjoyed different aspects of the movements in the session.
Extracts 5 "*"\[\...\]with the ball, I was able to identify which parts of my body were more receptive." - P6S05*"
"A vignette of S05 from the therapist's diary is presented:*" They had their eyes closed and seemed very concentrated on themselves. Their movements were free, direct, sustained and light. I was observing them while they were passing the ball across their body. There was no expression in their faces and I was trying to understand what they were experiencing, what the contact with the object made them feel. I noticed that they would only take care of, touch with the ball, the front part of their body, while ignoring the rear part."*"
"*"Later when we had to walk, my fun and "playful" side got out with the ball and the big red cloth (like power)." - P5S10*"
"A vignette of S10 from the therapist's diary is presented:*" They were all trying the movement proposals. I was very surprised that everyone not only permitted the contact with the other group members but also seemed to enjoy it, laughing and searching different body parts to propose. Seeing that this was well accepted I proposed them to follow the touch of their partners with all of their body. Immediately light, indirect and free movements emerged while the participants were experimenting with the quality of time. When I proposed them to do exactly the opposite, resist to the movement, something completely different happened. They had to put force and use the weight of their body as well as they had to find a minimum grounding. P5 seemed very comfortable and she was enjoying challenging P4. She surprised everyone with what her body was able to do."*"
"*"As I was using the rope, at first the space was smaller and more closed, even though with a slit so someone could come near. I liked interacting with the balls, I felt comfortable, integrated and at the same time with the space for myself. I also liked observing how others interacted with each other." - P2S04*"
"A vignette of S04 from the therapist's diary is presented:*" Each participant created her own personal space using a woollen thread. They started exploring that space having in general a low tone. I observed light, sustained, indirect and bound movements, while there was no interaction among the participants. In an attempt to facilitate the group coherence I added some balls in the available materials and put some upbeat music. The participants started to interact with each other. At first everyone was lying down and was throwing the balls without using any force. I was mirroring their movements making them progressively bigger and bigger. Suddenly, the participants started to put force in their movements and they were throwing the balls far away, something that made them change from the low to the medium-high level. Their movements were strong, sudden, free and direct. Their bodies seemed bigger, more open and they transmitted a sense of grounding*[2](#Fn2){ref-type="fn"}.""
In the first quote, P6 describes how her sense of body awareness increased by getting in physical contact with an object. P5 describes how she connects with her "playful" side, perhaps a side that she has not been in touch with much since she developed the ED, or even before. P2's comments experience is an example of how the use of different objects (e.g. ropes, balls) can facilitate the interaction between the participants and also give them a chance to symbolically explore their personal space and limits.
**Difficulties** Several difficulties were mentioned by the participants with regard to the DMT activities.
Extracts 6 "*"Sometimes I felt that my body couldn't move as much as I wanted, that the movements were smaller than I wished." - P2S05*"
"*"The part with pushing and using strength, UNCOMFORTABLE. I felt awkward. I couldn't simply enjoy \[the activity\] because I didn't fully let myself go. The cloth was great but since we were holding it between us all, I didn't feel completely free, to move as I actually needed, to flow." - P6S06*"
"*"Today we started with breathing. There I felt nervous." - P7S10*"
"*"At the beginning of the session I didn't feel like doing movements because I felt embarrassed, but as time passed, I started feeling more and more comfortable. Today I listened more to my body and less to my fears." - P4S06*"
"*"Later they (my companions) touched me where they believed I needed it, and I touched them. It was very difficult for me to think "where they need it" or "what they need". And when they touched me, I felt that they didn't touch me where I needed it, even though I liked it, actually I didn't know what I needed either." - P7S12*"
"*"I missed some music in the movement part. It would have made it easier for me." -P3S03*"
A variety of difficulties are mentioned in these extracts. P2 recognizes her difficulty using her body fully when moving. In the same line, P4 expresses an initial resistance to moving due to feelings of embarrassment that she eventually, at least partially, overcomes. Certain activities and movement qualities tended to be more challenging than others and triggered ambiguous responses, especially the use of strength and breathing exercises. P7 acknowledged her difficulty in recognizing her bodily needs as well as guessing others'. The lack of music was a challenging factor for P3.
**A pattern of activating during the session**
Some of the difficulties mentioned above were overcome during the course of the session as their energy level increased. These extracts were examined in order to identify the helpful elements in the session that triggered a change in the participants' attitude.
Extracts 7 "*"I had two opposite feelings during the session. On the one hand, in the beginning, I was angry and anxious and didn't want to do anything. It was very difficult to make the movements and the only way for me to be comfortable was on the floor with my face covered. On the other hand, when we started passing the ball I cheered up much more and I even enjoyed it at some point. I think the music cheered me up a lot and helped me." - P3S04*"
"*"I couldn't find a safe place and the blanket that was covering me (imaginary) was bothering me. As I opened my eyes, the feeling diminished. Then we made an activity of using strength, which soothed me very much. I felt liberated, feeling less burdened. The movement made me feel lighter. The tiredness of this week has left me, I feel more active. Now I feel content and motivated to talk and move." - P7S10*"
Analyzing these extracts, it is observed that often the participants who were burdened with negative feelings at the beginning of the session (as presented in theme 1.1.) and without any motivation or energy to move, struggled during the initial self-awareness exercises that required them to be alone with themselves. The key moments that shifted the energy level were identified to happen in pair or group interactions. Moreover, the use of objects like cloths and balls, and of music further facilitated such playful interaction. P7 self-regulates her feelings of anxiety by opening her eyes, and later she finds the activity that requires the use of strength soothing, even though such activities were received more skeptically by some of the participants.
### **Domain 3: Interpersonal aspects** {#Sec16}
Interpersonal aspects were identified as a domain by itself, due to the abundant evidence that the participants presented around this theme.
**Enjoyment of the interactions in the group**
The interactions in the group were identified as an important resource for the participants, in order to raise their energy level. Here they are considered in more detail.
Extracts 8 "*"I liked having P7 as my partner. I was surprised positively and it has been a new and different way to get to know her and get a little closer to her." - P6S10*"
"*"Later, as P6 was touching me and I had to move, I felt very comfortable and as time passed, we were finding each other more. Finally, we were moving freely but looking at each other, something that made me nervous at first, but in the end, I felt very comfortable and confident." P5S05*"
"*"I really liked the next exercise as well. Someone touching my chin, holding it from below. It made me feel especially good, it was like never doubt or never forget to hold your head up high." - P6S12*"
In addition to the increase in the energy level that was mentioned in the previous section, these extracts show the impact of the non-verbal interactions on the participants. P5 describes an intimate moment she found with another participant, as they encountered each other through their movements. P6 describes her enjoyment as she gets in physical contact with the other participants in an exercise focused on giving and receiving care from one another. For her, this encounter has a symbolic meaning of being supported by her companions.
**Difficulties in interpersonal relationships and interaction**
Several challenges and difficulties were mentioned with respect to interactions and relationships, within the group and outside. These issues can be summarized as fear of intimacy, difficulty expressing one's needs, and preoccupations about what others might think of oneself.
Extracts 9a "*"The thing with eye contact was very intense for me. It is like letting someone in. It wasn't exactly uncomfortable, but maybe strange, a little violent perhaps." - P6S05*"
"*"I also liked interacting with the other person, especially when I was the one being touched, feeling the movement of the other person conveyed more uncertainty because you don't know where she will take you. I felt more uncomfortable having to move while looking at the other, the movements were less fluent as if they were conditioned." - P2S05*"
"*"I didn't want to let people enter my life just to leave later. I went back to dancing on the terrain, finding my companions without any kind of tie that would bind me to someone. I felt free, liberated." - P7S09*"
Touching another group member was mostly received positively, yet maintaining eye contact with a partner, especially while moving, was challenging for many participants. P6 expresses her discomfort and her words suggest that it is linked to a fear of intimacy "like letting someone in". P2 also describes the uncertainty she felt concerning "where \[the other person\] will take \[her\]". Similarly, P7, who was new in the group at this point, expresses her distrust of others and reluctance in building meaningful relationships with people "just to leave later", suggesting an underlying fear of abandonment.
Extracts 9b "*"I didn't have the strength or energy to look for contact but when it came to me (when the others came near to interact) it made me feel good, even though also a little sad at the same time for not being able to do the same, for having so little energy. - P6S09*"
"*"\[...\] I only wanted to feel human contact (\...) She \[the therapist\] told us to get in pairs and it was the last thing I wanted to do, and when I could, I started playing with the balloon alone. I connected with the feeling of loneliness that I have been carrying for a while, usually, it comes up with losses. The best of the session: When P4 came to give me a hug." - P5S11*"
While the participants were eager to give care and received it thankfully, it was difficult for them to ask for it. When P6 receives the contact that she needed, she is not able to enjoy it purely as she is concerned that she cannot offer the same to the others. P5 recognizes her need for contact but she is ambivalent about it and tries to avoid others by playing alone. It is noteworthy that even though these participants did not express their needs openly, in some cases the group sensed them and responded to them.
Extracts 9c "*"Today I found it harder to connect than usual, I don't think I could really connect, because I was thinking about what others would think of me (that's why I don't think the movement part helps me much since I hold myself back." - P4S03*"
"*"I felt very comfortable with her, even though I was worried that she didn't feel the same way" (referring to an exercise in pairs) - P5S10*"
"*"I am sad, sad because I don't have P. now \[she is referring to her long-distance relationship\] and I look for any tiny detail to support my distortion that he doesn't like me and he will end up getting tired of me (...). I feel sorry that today was the last session and I've been so distant \[\...\]." - P5S12*"
The first two quotes express the participants' worries about what others might think about them, whether they will be liked or accepted. This seems to impair their capacity to get the most out of the movement experience. In the last quote, P5 expresses a deep fear of not being loved by her significant other, and while rationally she regards this as her "distortion", emotionally she seems to be heavily burdened to the point that she is not able to be present in the session.
### **Domain 4: Metaphors** {#Sec17}
According to Lakoff and Johnson \[[@CR70]\], metaphor is understood widely as a general function of the mind where a concept is described by its similarity to another, and it includes not only linguistic representations but also dreams, memories, and feelings.
**The inner safe place**
Imaginative techniques were applied frequently in the sessions. In one of the imaginative techniques borrowed from psychodynamic imaginative trauma therapy called the *"inner safe place"* \[[@CR57]\], the participants were guided to think of a place, which could be a real place or a fantasy place that conveyed them a sense of safety and security. Most participants wrote about their inner safe place in their journals, often including vivid and detailed descriptions, and they also mentioned the process of looking for their inner safe places and their relationship to it.
Extracts 10a "*"This time my safe place was the beach but without many people and even though it was hot, it wasn't overwhelming (...). It's warm and I only feel cold as I'm leaving because the sun has set, but I am happy, feeling the touch of that warmth on my skin." - P2S07*"
"*"My place of protection is my room, but with some changes. The window opened to a terrace with a grass floor (it's night and you can hear the noise of the leaves moved by the wind). Terrace, bed, table, window, grass, night, wind, leaves.(...) Music is heard in the room (...) and cats accompany me. They don't make a mess and they don't kill each other." -P4S08*"
"*"I can't see the ocean but it's inside me. Confusion. A coffeehouse in Paris, empty, and made from dark wood. There are many tables with many white cups, and emptiness in all of them. Paris looks nice...The Eiffel Tower, trees, bicycles, what a cliche...But it is in the exterior. I can observe the landscape through the big windows of the cafe. Calmness." -P4S09*"
While these two participants have different styles describing their safe place, some common elements can be detected in their writings: natural elements (sea, ocean, wind, etc.), emptiness, people or the absence of them. Both descriptions are rich in sensory elements, making the reader imagine these places vividly.
Extracts 10b "*"We imagined a safe place. It wasn't difficult for me to imagine it at all: the village. It was summer and we were in the village seated on a boat at sunset. It was all silent and he hugged me while we were watching the village. I felt calm, safe and loved." - P5S07*"
"*"Today we returned to the safe place but I hate that the village is not my safe place if P. isn't there, so I imagined another moment, when I was with T. but without focusing on her company." - P5S08*"
"*"My safe place turns out to be a person, I realized this today, since the last time we did this exercise, this person was present in my mind again so that I came to realize that it's not about the place, it's about the person. E. is my safe place. This scares me and frightens me. Each time we finished and A. told us to slowly say goodbye and that we could always return to it, I felt a tremendous pain because I don't have this certainty. So my safe place scares me, mostly because I'm afraid to lose it and since I feel that it is so safe and comfortable, that would be very painful." - P6S07*"
In these extracts, the focus of the inner safe place is shifted to another person. P5 enjoys the feeling of being loved by this person, and it seems that she finds it difficult to feel safe when this person is absent. She copes with this by imagining the company of a friend, and thus shifting her focus. Similarly, P6 has difficulties with the exercise of the safe place, when she comes to realize that she is projecting her sense of security onto a person, and she faces her fear of losing this person. In these cases, the safe place technique provided a source of self-knowledge for these participants.
Extracts 10c "*"What I did realize is that as we were given the instructions, I was trying to change or question "my safe place", like it was difficult to rationally trust my decision" - P1S08*"
"*"Today I felt confused, impotent and frustrated. My head was running a thousand per hour and I found it hard to focus on my safe place." - P6S08*"
"*"In today's session, I imagined the safe place again. \[...\] I didn't feel so safe. This made me nervous. So, dancing lightly, I imagined myself dancing on the terrain, making it mine. I felt relieved because it became \[safe\] again." - P7S09*"
These participants refer to the process of finding their safe place which was not always easy. P1 finds it difficult to maintain her focus on one safe place. P6's inability to focus her wandering mind gets in the way of accessing her safe place. When P7 does not feel safe, she uses symbolic movements and dance to regain her feeling of safety and she achieves it.
**Metaphorical expressions**
Some participants described their experience via metaphorical expressions.
Extracts 12 "*"(...) I remained like the flower in the middle, I haven't taken everything out, I haven't completely opened myself. I feel somewhat relieved but not liberated." - P6S06*"
"*"What is this mental capsule that I am in? It is so strong, no one can break it. It has a little window at eye level through which I can see. I am protected but it seems like a bad place, unhealthy, an exaggerated cover, or protection." - P4S11*"
In these extracts, the participants do not name the relevant emotions in order to describe their feelings, instead they use metaphors that help the reader get closer to that particular experience. With *"the flower in the middle"*, P6 is referring to a drawing that she made during this session, where she drew three flowers in the process of opening: the first one being closed, the second one half-opened, and the third one being completely opened. The inside vs. outside theme from the safe place can be recognized in P4's writing again. The strong capsule functions as a defense mechanism, and she recognizes its dysfunctionality (*unhealthy, bad, exaggerated cover*) as well as being protected by it.
### Domain 5: Reflections {#Sec18}
**Reflections on one's personal process** An evolution in the reflections of the participants was visible when they attended the sessions for the whole period. Below, a collection of extracts from the writings of two patients is presented.
Extracts 13a "*"We started by becoming aware of our bodies through a ball. It made me realize the long path that I have in front of me to recover my weight." - P5S05*"
"*"I am more open-minded now because I've learned that: If I don't want to do something (in this case \[coming in\] neither the clinic nor the sessions), before I wouldn't do it, but having continued, it's been good for me to change my mind." - P5S07*"
"*"As I am so well outside the center, at home and with P., ...I feel that my life starts to flow again. Coming to the clinic makes me feel as if I'm wasting time, since life is passing and I'm still shut-in, without being able to move." - P5S11*"
In these extracts, P5 changes attitude towards the treatment. The movement activity in an earlier session confronts her with her low body weight and makes her think of the "long path" in front of her. It is unclear whether this is a source of despair or motivation. Later, she reveals her skepticism about the treatment at first, both the Clinic in general and the DMT sessions, and that she has changed her mind. In the last quote, she expresses her frustration in the Clinic as she believes that her life is back on track and coming to the Clinic is what holds her back. She uses an observational and cognitively focused style in these extracts.
Extracts 13b "*"Fear, fear, fear. I want to do things and there it is. The present fucked me up. It is okay, maybe fucked up is too much because there is a way out and it is in me." - P4S10*"
"*"(...) I don't feel invisible as I am writing these lines, I don't feel like I'm hiding, but rather, I'm taking some time for myself, that I am connected with myself and at the same time I am seen, and this causes me some kind of insecurity." - P4S11*"
"*"Loneliness. I think that's what it's mostly about, next to other things. In the end, the core of all my worries is fear and the base of it loneliness, not feeling cared for, loved, as if I am not worthy of it or don't deserve it." P4S12*"
These extracts show P4's process throughout the last sessions as she gets in touch with difficult feelings such as fear and loneliness, and expresses them in her writings. Her expression suggests awareness and authenticity. In the last session, she develops an insight as she links her problems to feelings of worthlessness.
**On the DMT sessions**
As the therapeutic process was coming to an end, some participants included an overall view of the therapy in their reflective diaries, offering a qualitative evaluation of it, and revealing what it meant for them.
Extracts 14 "*"(...) these sessions have been very special for me. I was able to connect with myself more and see things or discover thoughts or internal worries that were unknown to me. I had the opportunity to feel closer to my companions and create a new type of connection where words haven't been necessary. That's why I'm very thankful to have had the opportunity to do dance therapy with A. - P6S12*"
"*"Even though I wish we had more sessions, this is the reality and I decide to stay with what I've learned, the skills, the moments and emotions. Thank you for accompanying me during such a hard time." - P4S12*"
"*"At first I thought this therapy wouldn't help me, that it was nonsense. It showed me much more than that. It helped me connect with myself, to get to know myself better, where are my limits, how to calm myself down when there are strong feelings. I am very thankful to you A. for all that you've achieved in me." - P5S12*"
"*"These sessions helped me a lot even though I did it for little time. When I entered demotivated, I left feeling very content and active. When I entered with a lot of anxiety, it helped me decrease it." - P7S12*"
Most participants expressed their thankfulness to the therapist. They reported that the DMT sessions promoted self-awareness, helping them connect with themselves, identify their feelings, and recognize their limits. P4 feels sorry that the therapy did not last longer. Also participants perceived that the sessions reduced their levels of anxiety and gave them energy and motivation. It is noteworthy that P5 attributes the source of change entirely to the therapist: *"all that you've achieved in me"*. Finally, the exploration of non-verbal interactions with the peers, the creation of new bonds and the accompaniment of the therapist were appreciated.
Discussion {#Sec19}
==========
The primary aim of this pilot study was to analyze the impact of a DMT intervention on body image and alexithymia in individuals with EDs. Secondly, a qualitative study arm was integrated in order to gain a deeper understanding of the therapeutic process. The following discussion of the quantitative outcomes should be regarded cautiously, as the small sample size does not allow making any generalisations of the obtained results. Instead, they should be taken as a departure point for future studies.
The participants who took part in the DMT group improved on the Appearance Evaluation and Body Areas Satisfaction scales, meaning that they felt more positively about their bodies after the DMT intervention. They scored lower on the Appearance Orientation scale after the therapy, which suggests that they were less preoccupied with their looks. These changes were both statistically and clinically significant. The decrease in the Overweight Preoccupation scale was not statistically significant but a medium to large effect size was observed, suggesting that there have been clinically significant improvements which were not statistically significant due to the small sample size. The control group that continued the treatment as usual did not change significantly in any of the body image aspects measured by the MBSRQ. A mixed-effects ANOVA model showed significant results in the interaction effect for Appearance Orientation and Body Areas Satisfaction. With the necessary caution because of the small sample size, these results suggest that it might be beneficial to include a treatment component to work with individuals on a bodily level in order to address body image issues. Considering that body image is an important maintenance and prognosis factor in EDs \[[@CR71]\], helping individuals construct a healthy body image is a crucial element that needs to be covered in any treatment modality.
There is a growing body of evidence suggesting that DMT has positive effects on body image in various populations \[[@CR49], [@CR50]\]. Regarding the field of EDs, one study has found DMT to improve body image and self-esteem, and decrease psychological distress, in obese women with emotional eating \[[@CR51]\]. However, despite the existence of rich theoretical contributions that dance movement therapists made to the field of EDs \[[@CR72]\] and several DMT case studies \[[@CR73]--[@CR75]\], to the authors' knowledge, no empirical studies evaluated the effects of a DMT intervention in individuals with AN, BN, and EDNOS. Therefore, the present pilot study has been an initial attempt to fill in an important research gap, especially bearing in mind the central role that body image plays in EDs. Hopefully, these promising preliminary results will encourage other researchers to conduct similar studies with larger samples.
The alexithymia scores of the DMT group tended to reduce or remain stable, while there was a tendency to increase in the control group, however these changes were not significant for either group. The stability in alexithymia in the DMT group is not surprising, concerning that the paucity of studies testing effects of diverse psychological treatments on alexithymia in patients with EDs raised diverging results \[[@CR15], [@CR76]\]. Even in studies that achieved to reduce alexithymia scores through treatment, the alexithymia levels often remained elevated, so that the clinical significance of these results is unclear \[[@CR76]\]. As alexithymia seems to be a trait that is resistant to change, longer treatment periods might be necessary to achieve significant results. On the other hand, the tendency to increase in alexithymia that was observed in the control group was unexpected. While this result should be interpreted with caution due to the small sample size (n = 5 for the control group), it can be an interesting research question whether cognitive behavioral techniques, when these do not specifically target alexithymia, increase alexithymia levels in patients with EDs, as it encourages the patients to take a distant stance from their personal experiences \[[@CR77]\].
The TAS-20 was used to measure alexithymia in the present study. While the TAS-20 is a widely used and accepted instrument, one study criticized its use with patients with EDs, as they found that patients with EDs with high alexithymia levels according to TAS-20 performed equally well as the control group on tasks of emotion identifying and reporting. They also found that depression and anxiety states can alter alexithymia scores \[[@CR78]\]. While there is no way of knowing if such an effect occurred in the current study, the qualitative analysis has shed more light on alexithymia and related issues in the participants of the DMT group.
The results of the qualitative analysis demonstrated difficulties in identifying feelings and a difficulty to be present in the moment. Some participants reacted defensively when they encountered strong feelings during the sessions. However, in some instances, they were able to face and accept these strong feelings. The DMT sessions seem to have provided a secure base for the participants to explore such feelings. In their overall evaluations of the intervention, participants reported that the sessions helped them be more present in the moment, identify their feelings, and get to know themselves better. Even though no significant results were obtained in the TAS-20 scores, these patient accounts suggest that the DMT sessions had an impact on some aspects of alexithymia.
Specific attention was given to the use of metaphors as patients' with EDs impaired symbolic thinking and their tendency to somatize their emotional experience have been identified as a challenge to traditional psychotherapy \[[@CR79]\]. The use of rich metaphors by some patients and their vivid descriptions of their "safe place" contradict with the impoverished inner fantasy life that is associated with alexithymia \[[@CR19]\]. The work with active imagination in movement is a core aspect of DMT. This process allows the patient to work with metaphors and symbolism, thus may help the patient access contents that were previously not conscious. These results suggest that the use of imagery techniques and creative movement in DMT, while putting an emphasis on body awareness, can help combat these limitations.
The participants reported that the intervention had a positive effect on their mood, helping them feel more cheerful or motivated, and less anxious, confused or angry. These qualitative results are in line with a previous study using self-report questionnaires that revealed an increase in energy and decrease in anger and confusion in adolescents in a day hospital after the DMT sessions \[[@CR80]\]. In the present study, the extracts from the patients' diaries also shed light on how the DMT sessions facilitated such change in mood. Most participants described a similar pattern of entering in the session in a depressed or anxious mood and initially had difficulties engaging in the activities. The therapist accepted and attuned to their initial mood while gently encouraging them to explore different movements. Furthermore, the use of music and materials such as balls and cloths facilitated playful interaction between the group members and raised their energy level. These effects can be related to neurohormonal effects of movement, as a previous study observed changes in dopamine and serotonin levels as well as a decrease in psychological distress after a 12-week DMT intervention in middle school students \[[@CR81]\].
However, it is important to note that such effects can only partially be explained by the benefits of physical activity, as one study comparing a single circle dance intervention to exercise on a home trainer bike found that depression levels decreased in the dance condition and not in the exercise condition \[[@CR82]\]. Dancing in a group seems to have more components that promote mental health than physical exercise only, like the physical and affective interaction between group members. What is more, physical exercise is a complicated issue in the treatment of EDs \[[@CR83]\] as patients often use excessive exercise for complex psychological reasons \[[@CR84]\]. DMT could provide an alternative to traditional physical exercise, since it seems to have the potential to promote body awareness, as well as benefit the body and mental health on a neurohormonal level. This would be valuable for ED treatment, as patients often suffer from comorbidities with depression and anxiety \[[@CR85]\]. Larger scale studies including biomarkers are needed in order to test these assumptions for patients with EDs.
Some interpersonal difficulties that are associated with EDs according to a recent systematic review \[[@CR86]\], have emerged in the participants' reflective diaries, such as interpersonal distrust, fear of intimacy, and difficulty expressing one's needs. At the same time, the empathetic interactions in the group, both with each other and with the therapist, were valued by the group members, similar to findings in a cognitive behavioral therapy group \[[@CR87]\]. A novelty in this study is the focus on the nonverbal aspect of these interactions. The reflective diaries showed clearly that the participants enjoyed getting to know one another on verbal and nonverbal levels. At the same time, some of the DMT techniques prompted them to reflect on their real-life relationships. Bearing in mind the importance of positive social relationships in the recovery of EDs \[[@CR71]\], and the various interpersonal problems that they often suffer from \[[@CR86]\], attending the interpersonal aspects has a crucial role in ED treatment. DMT can be an innovative way to address these issues, as it combines implicit and explicit learning, verbal and nonverbal interaction.
While the group format seems to have helped the participants enhance in relationship dimensions, it should be noted that individual DMT sessions would have its own advantages, such as the patients having more space to explore their movements freely. The privacy and confidentiality that individual therapy offers could allow patients to feel more secure and talk about personal issues in depth.
Interestingly, the participants made few references to body image issues in their reflective diaries. This might be because they were more focused on their lived bodies during the DMT sessions, and not so much on their appearances. It also suggests that the improvements in body image found in the MBSRQ scores happened through implicit ways. For instance, the writings indicate a pure enjoyment of the playful social interactions that the participants experienced in the sessions. Considering that patients with EDs tend to experience their bodies through external means such as weighing scales, mirrors, etc. \[[@CR79]\], the enjoyment through the lived body that DMT offers can be very valuable and this might be accountable for the improvement in body image found by the MBSRQ. This implicit aspect marks a significant difference between DMT and approaches that put an explicit focus on body image issues, such as cognitive-behavioral techniques. Lastly, most participants expressed their gratitude to the therapist in their reflective diaries. Overall, they were satisfied with the intervention proposed. There were no dropouts in the DMT group and the no-show rates were low, suggesting that the patients were committed to the therapy. This seems important considering the high dropout rates in EDs \[[@CR88]\].
Limitations {#Sec20}
-----------
The findings of this pilot study have to be considered in light of some important limitations. As authors such as Schweizer and Furley emphasize, prior to any research the appropriate sample size should be calculated in order to obtain more accurate, reliable and reproducible results. However, these authors also recognize that in certain circumstances this could make research difficult in certain areas of knowledge in which access to a sufficiently large sample is limited by numerous factors \[[@CR89]\].
The authors of the present study are extremely cautious about the quantitative results obtained, as Button et al. suggest. They consider the choice of a methodological design such as the mixed one to be an ethical procedure. Given the difficulty of having a sufficiently large sample, this makes it possible to incorporate other data collection and analysis tools that offer information of interest in relation to the subject matter being dealt with \[[@CR90]\].
Another issue that compromises the reliability of the quantitative results is the heterogeneity between the groups. The intervention and the control groups were homogeneous concerning their age and BMI, however the control group had been in the Clinic for a considerably longer period than the DMT group. A larger sample would be necessary to determine the influence that the time spent in the Clinic had on the results. Furthermore, the sample was heterogeneous concerning the ED subtype, with the majority in both conditions being EDNOS or AN, and only one BN patient in the control group. Therefore, it remains unknown if the diagnosis of a particular ED subtype had an effect on how the DMT sessions were received. The sessions were conducted by a DMT trainee, who was doing her clinical internship in the Clinic. She was at the moment in the beginning of her learning process and had no previous experience with this population. A therapist with more experience might have worked more efficiently and adapt better to the particularities of each participant or the ones of the group.
Another shortcoming was that the DMT group was a semi-open group. For the therapeutic process this implies working with individuals who can be in different moments of their personal process, which might present an additional challenge to the group therapist. A closed group would be beneficial for the patients in terms of developing trust in the group and getting used to working with their bodies. In terms of methodological rigor the variation in the number of sessions that each participant attended to, poses another variable that can not be controlled. The setting of the DMT group was adjusted to the needs of the private clinic where the study was conducted. In line with the Clinic's policies, two interns attended as observers the majority of the sessions, which presented a confounding variable that might have affected the results. Finally, this was an all-Spanish sample, and some interventions might not work in the same way with patients from other cultures, e.g. interventions that include touch and eye contact, as these aspects of nonverbal communication vary significantly across cultures. Practitioners should take the cultural context into account.
To the best of the authors' knowledge, there are no quantitative studies evaluating the impact of DMT interventions on the body image and alexithymia rates of patients with EDs. The authors had no reference point when designing this pilot study. The findings are promising despite the mentioned limitations, as significant improvements were observed in body image, and a slight tendency for improvement in alexithymia, compared to the control group. It would be interesting to be able to follow this line of research with larger samples which would minimise statistical analysis limitations. Last but not least, it should be taken into account that in bodywork approaches like DMT, a large sample does not make it possible to do in-depth work. In any case, mixed methodological research design is crucial, since the qualitative information collected from the patients themselves offers a better understanding of the process by giving space to each of their individual experiences. The integration of both approaches, quantitative and qualitative, should be, in the authors' opinion, the optimal way to proceed in this field.
Conclusions {#Sec21}
===========
To the authors' knowledge, the present pilot study is the first controlled trial to evaluate the impact of a DMT intervention on body image and alexithymia in a mixed group of patients with EDs, i.e., AN and EDNOS. Furthermore, the study has provided valuable insights into the experience of patients with EDs, using qualitative analysis of the reflective diaries written by the participants at the end of each session. The DMT group showed significant improvement in various body image scales, while there were no significant changes in the alexithymia scores. Moreover, the qualitative analysis revealed other factors of the DMT sessions that were appreciated by the participants, as they reported an increase in self-awareness, improvement in emotional states, as well as they valued verbal and non-verbal interaction with the group members and with the therapist. More research is necessary to confirm these promising findings. Combined, these results reveal the potentially significant impact that a DMT intervention might have in the treatment of patients with EDs. The combination of DMT with the traditionally proposed ED treatments might provide a more holistic and effective approach.
(ED)
: Eating disorder
(DMT)
: Dance and movement therapy
(TAS-20)
: Toronto alexithymia scale
(DSM-5)
: Diagnostic and statistical manual of mental disorders, fifth edition
(MBSRQ)
: Multidimensional body-self relations questionnaire
(SPSS)
: Statistical package for the social sciences
A. pseudo-initial for the therapist.
The references in movement qualities, from the therapist's diary, derived from Laban movement analysis that is used in DMT. They are related to the indulgence or the resistance of the participant and may indicate the way of relating with the other.
**Publisher's Note**
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Maria Savidaki and Sezin Demirtoka contributed equally to this work.
We are grateful to the patients of the Clinic for their participation in this study. We also thank the professionals of Clínica SETCA Barcelona, in particular Pilar López Ropero for her cooperation and support. We are thankful to the Autonomus University of Barcelona for giving us the chance to contact this study. We aslo express our gratitude to Ümit Çetin-Demirtoka, Manuel Carmona Delgado for their critical proof-reading of the manuscript and Israel John Thuissard for his statistical advices. We want to thank Violetta Matzorou for her active support in the manuscript preparation, Matthias Behrends for his theoretical contributions regarding the guided imagery techniques, and Alexandros Zacharakis for his generous technical support.
MS and SD collaborated in this study as part of their master's thesis under the supervision of RR. MS and RR contributed to the study design. MS organized the collaboration with the Clinic, and delivered the interventions. Quantitative data collection was facilitated by professionals of the Clinic. The handwritten reflective diaries were transcribed to digital text by MS and translated from Spanish to English by SD. SD conducted the data analysis and interpretation. RR contributed to the data analysis. SD and MS prepared the manuscript. RR supervised the whole study process. All authors critically reviewed the manuscript and approved the final version for submission.
Not applicable.
The quantitative data set and the full qualitative analysis including quotes may be requested from the corresponding author upon reasonable request.
Written approval for this study was provided to the authors by the ethical committee of the Clinic (00/01/2019). All participants signed a consent agreeing to participate in the study.
The intervention group signed a consent form agreeing to the anonymized publication of extracts from the reflective diaries that were used in this study.
The authors declare that they have no competing interests.
| {
"pile_set_name": "PubMed Central"
} |
Introduction
============
Despite the wide range of currently available strategies for chemotherapeutic treatment, prognosis of many types of cancer remains poor for many patients; therefore, there is a clear need for new therapies that improve outcomes.
Some of the most common drawbacks of cancer drugs are related to their low solubility, poor bioavailability, and undesirable toxic side effects, the last being a concern for doxorubicin (Dox).
Dox is an anthracycline used primarily in the treatment of childhood solid tumors, soft tissue sarcomas, lymphomas, and breast cancer.[@b1-ijn-10-3377]--[@b4-ijn-10-3377] The use of Dox is limited by significant side effects like cardiotoxicity, nephrotoxicity, hepatotoxicity, and also by its short circulation lifetime. Dox is a tetracyclic quinoid aglycone linked to an amino sugar that has a pKa of 8.3; thus, it is a hydrophilic weak amphipathic base usually incorporated in the aqueous interior of preformed vesicles such as liposomes, using ionic or pH gradients.[@b5-ijn-10-3377] However, these gradients are unstable and the incorporated drugs tend to leak out at a rate which strongly depends on the composition of the bilayer. Physical entrapment of Dox, mostly by hydrophobic effect, in self-assembled systems based on block copolymers, has been also reported,[@b6-ijn-10-3377]--[@b8-ijn-10-3377] as well as its incorporation in the cross-linked ionic core of some polymeric micelles through electrostatic interactions.[@b8-ijn-10-3377]
More recently, nanomaterials have arisen as a promising alternative to cancer therapeutics; they can enhance the delivery and treatment efficiency of anticancer drugs. In this context, to be used for drug delivery, nanomaterials must be biocompatible, have suitable biodegradation kinetics, be easy to process, possess adequate mechanical properties, and be compatible with the drug to be transported. In the last 20 years, many strategies based on nanoscale (10--200 nm) vehicles, such as liposomes,[@b9-ijn-10-3377] polymeric micelles,[@b10-ijn-10-3377],[@b11-ijn-10-3377] and polymer particles,[@b12-ijn-10-3377] have been increasingly used in a wide variety of approaches for therapeutic drug delivery devices.
In previous studies, we demonstrated that lipidic micelles composed of monosialoglycosphingolipid (GM1) spontaneously load high amounts of a hydrophobic cancer drug such as paclitaxel (Ptx).[@b13-ijn-10-3377] This micellar structure has already been proposed for loading small hydrophobic molecules that localize in the more internal region of the micelle, with the fatty acid chains of gangliosides.[@b14-ijn-10-3377] Moreover, GM1 micelles bind human serum albumin (HSA) via hydrophobic interactions.[@b15-ijn-10-3377] This offers additional advantages because many cancer cells overexpress gp60 receptors specific for albumin. This mechanism could facilitate the transport of the complexes with albumin to the tumor tissue where its interaction with the SPARC protein may result in intratumoral accumulation of albumin-bound paclitaxel.[@b16-ijn-10-3377],[@b17-ijn-10-3377]
In order to gain greater insights into the particular behavior of this micellar structure, we evaluated the interactions between a water-soluble cytotoxic drug such as Dox and GM1 ganglioside micelles and characterized the structures resulting from this interaction.
Furthermore, considering the special features of micelle polarity described earlier, we also evaluated whether this structure could simultaneously load hydrophobic and hydrophilic drugs into a single GM1 micelle and how the incorporation of these drugs regulates the subsequent interaction of the micelle with HSA. This would provide a transportation system which, in addition to being able to individually load drugs of very different polarities, it might be able to load drugs simultaneously, becoming a multidrug delivery system.
Materials and methods
=====================
Materials
---------
Purified monosialoganglioside GM1 from pig brain was obtained from TRB Pharma S.A. (Buenos Aires, Argentina), Dox--HCl (Dox) was from ELEA (Buenos Aires, Argentina), and Ptx was from Yunnan Smandbet Co. Ltd (Kunming, People's Republic of China). Flutax-1 (7-O-\[*N*-(4′-fluoresceincarbonyl)-l-alanyl\]taxol) was purchased from Calbiochem (San Diego, CA, USA). Purified HSA 20% (w/v) was obtained from Laboratorio de Hemoderivados, Universidad Nacional de Córdoba (Córdoba, Argentina). All other chemicals used were of analytical grade.
Methods
-------
### Standard procedure for studying the interactions between Dox and Ptx with GM1 micelles
Stock solutions of GM1 with a concentration of 250 mg mL^−1^ were prepared in bidistilled water 24 hours prior to use as described in Leonhard et al.[@b13-ijn-10-3377] Briefly, the solutions were maintained at 4°C for 24 hours. They were then centrifuged at 50,000× *g* for 15 minutes and the supernatant was filtered through 0.22 μm.
Stock solutions of Dox (24 mg mL^−1^) were prepared in 145 mM of NaCl. Solutions of Dox were slowly added to the solution of GM1 micelles in order to attain different Dox--GM1 molar ratios, ie, 1/15, 1/10, 1/5, 1/2.5, 1/1, 1/0.7. These mixtures were incubated at 4°C for 24 hours. Unbound Dox was removed by dialysis against 100 volumes of bidistilled water for 24 hours at 4°C.
Stock solutions of Ptx (50 mg mL^−1^) were prepared in dimethylsulfoxide (DMSO). The solutions were added to GM1 or Dox--GM1 micellar solutions and then stirred for 2 minutes before incubation at 4°C for 24 hours and dialyzation for 24 hours at 4°C against 100 volumes of bidistilled water to remove all DMSO. The removal of DMSO in the solution was determined spectrophotometrically by measuring absorbance at 227 nm at different times until this value remained constant.
When studying the interactions of Dox with Ptx--GM1, stock solutions of Dox described earlier were slowly added to Ptx--GM1 micelles in order to attain the desired Dox--GM1 molar ratios. These mixtures were incubated at 4°C for 24 hours. Unbound Dox was removed by dialysis against 100 volumes of bidistilled water for 24 hours at 4°C.
### Determination of GM1 concentration
Ganglioside concentrations were measured by the modified colorimetric resorcinol assay, described by Miettinen and Takki-Luukkainen.[@b18-ijn-10-3377] Briefly, 1 mL of resorcinol reagent was added to 1 mL of the samples and heated to 100°C for 15 minutes. (Resorcinol reagent: 2 mg of resorcinol powder dissolved in 0.1 mL of bidistilled water +0.8 mL of 37.9% \[w/v\] HCl +2.5 μL of 0.1 M CuSO~4~ + amount of bidistilled water so as to reach 1 mL). The samples were then allowed to cool and the chromophore developed was extracted with 2.5 mL of n-butyl acetate: n-butanol (85/15 by vol). After centrifugation at 2,500× *g* for 5 min, the supernatants were removed and measured spectrophotometrically at 580 nm.
### Determination of Dox concentration
Dox concentration was determined by absorbance at 490 nm using a calibration curve performed with a standard solution of Dox in 145 mM NaCl, as described by Abraham et al.[@b19-ijn-10-3377]
### Determination of Ptx concentration
Ptx was determined as described in Leonhard et al.[@b13-ijn-10-3377] Briefly, Ptx was extracted from micelles with 10 volumes of ethyl acetate. Samples were then centrifuged at 2,500× *g* for 5 minutes; the organic layer was transferred to a clean tube and evaporated to dryness at 40°C. The dried residue was solubilized in 1 volume of ethanol.
Ptx concentration was measured using a Curosil B C18 column (250×3.20 mm inner diameter, particle size 5 μm) and a Curosil B C18 guard column (30×4.60 mm inner diameter, particle size 5 μm) supplied by Phenomenex. The mobile phase was 60% (v/v) acetonitrile and 40% (v/v) bidistilled water. Flow rate was 0.7 mL min^−1^ and the eluent was monitored at 227 nm. Chromatography was performed at ambient temperature (20°C) and the calibration curve of Ptx was linear in a range from 5 to 150 μg.
### Structural characterization of Dox--GM1 and Dox--Ptx--GM1 mixed micelles
#### Chromatographic analysis
Samples were run on an Åkta Explorer 100 system (GE Healthcare Life Science, Buenos Aires, Argentina) as described in Leonhard et al.[@b13-ijn-10-3377] Briefly, samples containing GM1 (10 mg mL^−1^) were loaded into a Superdex 200^®^ column, equlibrated with 50 mM phosphate buffer (pH 7.0) and 150 mM NaCl at a rate of 0.4 mL min^−1^. The elution pattern of GM1 and Dox were monitored using an UV detector at 227 and 490 nm, respectively. GM1 and Dox levels were quantified as described. The molecular weights (MWs), corresponding to the various elution peaks, were extrapolated from a calibration curve made up of globular proteins of known MW as albumin (67,000 Da), lactoglobulin (35,000 Da), and human immunoglobulin G (160,000 Da) using the formula: log MW = a. eslution volume + b.
#### Determination of particle size by dynamic light scattering
Average particle size of the aqueous solutions of micelles with GM1 (10 mg mL^−1^) were measured by dynamic light scattering (DLS), which was performed on a Delsa™ Nano Submicron Particle Size and Zeta Potential Particle Analyzer at a fixed scattering angle of 165°. Data were analyzed by Delsa Nano Beckman Coulter software (version 2.2) with CONTIN analysis method.
### Physical stability of Dox--GM1 and Dox--Ptx--GM1 mixed micelles
#### Stability of micelles to high-speed centrifugation
Dox--GM1 and Dox--Ptx--GM1 micelles were centrifuged at 25,000, 50,000, or 100,000× *g* for 1 hour at 20°C in an XL-90 ultracentrifuge (Beckman Coulter Argentina S.A., Buenos Aires, Argentina). Afterward, the supernatants were dialyzed for 24 hours to remove unbound molecules before determining GM1, Ptx, and Dox concentrations as previously described.
#### Stability of micelles in solution
Dox--GM1 and Dox--Ptx--GM1 micelles were stored for 40 days at 4°C. Aliquots of samples were taken at various time periods and dialyzed for 24 hours prior to quantification of soluble Dox and Ptx as described. Dialysis conditions were as follows: membrane tubing was from SpectraPor with a MW cut-off of 10,000 Da, 1 mL of sample was dialyzed against 1,000 mL of bidistilled water. Dox and Ptx were quantified in the sample and dialyzed.
#### Effect of freeze--thawing cycles and lyophilization
Dox--GM1 and Dox--Ptx--GM1 micelles were frozen at −80°C, and after 24 hours samples were kept at room temperature for about 2 hours until complete thawing, centrifuged at 15,000× *g* for 10 minutes and dialyzed for 24 hours before quantifying the concentrations of Dox, Ptx, and GM1 that remained soluble.
Moreover, lyophilized micelles were dissolved in their initial volume, filtered through a 0.22-μm pore, and dialyzed for 24 hours before measuring soluble Dox, Ptx, and GM1 as described.
### Chemical stability of Dox in Dox--GM1 and Dox--Ptx--GM1 micelles
We evaluated the chemical stability of Dox loaded into the micelles when subjected to a denaturing condition such as the alkaline environment where ester groups are hydrolyzed. For this, we incubated Dox--GM1 and Dox--Ptx--GM1 micelles at pH 10 and at room temperature (25°C±1°C). At various time periods, aliquots of the samples were taken and the amount of soluble Dox and Ptx was quantified as described. Stock solution of free Dox and Ptx was used as a control for the kinetics of chemical hydrolysis at alkaline pH.
### Characterization of the interaction of Dox--GM1 and Dox--PtxGM1 micelles with HSA
Dox--GM1 micelles of different molar ratios (1/20, 1/5, and 1/1), with or without preloaded Ptx (Ptx--GM1: 1/20 molar ratio) and with a GM1 concentration of 10 mg mL^−1^, were incubated for 24 hours with purified HSA (1/1 HSA--GM1 w/w ratio) at pH 3 and 37°C. Samples were then returned to pH 7 and the interaction was studied by chromatographic analysis and DLS as described.
### Determination of HSA concentration
The amount of albumin associated with the micelles was determined using a Coomassie Brilliant blue assay or by direct absorbance at 280 nm.[@b20-ijn-10-3377]
### In vitro biological effect of Dox in Dox--GM1 and Dox--GM1--HSA micelles on cell cultures
Hep2 (human epithelial carcinoma of larynx) cell lines were grown as described in Leonhard et al.[@b13-ijn-10-3377] Briefly, cells were grown in minimum essential medium supplemented with 10% of fetal bovine serum (NATOCOR, Córdoba, Argentina) at 37°C with 5% CO~2~. Then, the cell monolayers were incubated for 15, 30, 60, 120, and 240 minutes and for 24 hours at 4°C and 37°C with increasing concentration of Dox, Dox--GM1, and Dox--GM1--HSA micelles. After this incubation, the medium was removed and fresh medium was added.
After 24 hours at 4°C or 37°C, viable cells were measured by an 3-4,5 dimethylthiazol-2,5 diphenyl tetrazolium bromide staining assay. The half maximal inhibitory concentration (IC~50~) values were graphically estimated.
### Characterization of the cellular uptake of Dox and Ptx from Dox--GM1 and Dox--Ptx--GM1 micelles with and without HSA
Hep2 cell lines were grown as described. Dox, Ptx, Dox--GM1, and Dox--Ptx--GM1 micelles with and without HSA, with 1% of Flutax-1 (a fluorescently labeled derivative of Ptx) were diluted in minimum essential medium to a final concentration of 30 μg mL^−1^ of Dox and 10 μg mL^−1^ of Ptx.
Each sample was incubated with Hep2 cell monolayers at 70% confluence for 15, 30, and 60 minutes at 37°C. Cells were then washed out with phosphate-buffered saline, and the incorporated drugs were observed by fluorescent microscopy.
Results
=======
Interactions between Dox and GM1 micelles under different physicochemical conditions
------------------------------------------------------------------------------------
Gangliosides are double-tailed anionic glycosphingolipids with a complex polar head group of several sugar units. In particular, GM1 is a monosialoganglioside with four sugar units in its polar head group and due to thermodynamic and geometric constrains, it forms spherical micelles in aqueous media at a very low concentration (critical micellar concentration of 10^−8^--10^−10^ M).[@b13-ijn-10-3377],[@b21-ijn-10-3377]--[@b23-ijn-10-3377]
In order to start studying how Dox, a highly water-soluble drug, interacts with GM1 micelles, solutions of Dox were slowly added to a solution of GM1 micelles in order to attain different Dox--GM1 molar ratios, ie, 1/15, 1/10, 1/5, 1/2.5, 1/1, 1/0.7. These samples were incubated at 4°C for 24 hours and unbound Dox was removed by dialysis against bidistilled water as detailed in the Materials and Methods section. [Figure 1](#f1-ijn-10-3377){ref-type="fig"} shows that there is a spontaneous incorporation of Dox into GM1 micelles, reaching saturation around 10 mg mL^−1^ per 50 mg mL^−1^ of GM1.
The results demonstrate that these micelles can incorporate up to five times more Dox compared to current commercial products based on liposomal formulations containing 2 mg mL^−1^.[@b24-ijn-10-3377]
The encapsulation efficiency calculated from these varies according to the different Dox--GM1 molar ratios. Up to 1/2.5 molar ratio, the incorporation efficiency is greater than 95%, whereas at 1/1 and 1/0.7 ratios, it drops to 55% and 35%, respectively.
### Effect of medium pH on Dox--GM1 interaction
It was observed that the reduction of pH from 7 to 2 did not affect the ability of GM1 to incorporate Dox, demonstrating that Dox--GM1 interaction does not involve the sialic acid molecules of gangliosides (pKa \~2.6). In addition, it was observed that the presence of up to 0.5 M NaCl did not prevent Dox incorporation into GM1 micelles, again suggesting that the association involves mostly hydrophobic rather than electrostatic interactions (data not shown).
### Effect of temperature on Dox--GM1 interaction
Previous studies showed that the incorporation of Ptx in GM1 micelles is favored with increasing temperature to 55°C.[@b13-ijn-10-3377] Such condition produces a release of the water bound to the oligosaccharide chain resulting in a significant reduction of the hydrophilic portion which is associated with a small expansion of the hydrophobic portion of the micelle.[@b25-ijn-10-3377]--[@b28-ijn-10-3377]
We found that, unlike what was observed with Ptx, changes in temperature do not significantly affect the incorporation of Dox in GM1 micelles (data not shown). This result is in disagreement with what was obtained with Ptx, a highly hydrophobic drug, probably due to the fact that Dox is a water-soluble molecule whose entry into the micelle is insensitive to its changes in hydrophobicity.
Ability of GM1 micelles to load hydrophobic and hydrophilic drugs into the same structure
-----------------------------------------------------------------------------------------
Having demonstrated the ability of GM1 micelles to incorporate drugs with such different physicochemical characteristics, a highly hydrophobic drug like Ptx and a hydrophilic one like Dox, we wondered if it would be possible to incorporate both molecules simultaneously into the same micellar nanostructure. [Figure 1](#f1-ijn-10-3377){ref-type="fig"} shows that Ptx--GM1 micelles are able to incorporate large amounts of Dox, similarly to GM1 micelles, reaching a saturation level around 8 mg mL^−1^. In this case, the encapsulation efficiency is also higher than 95% up to the molar ratio Dox--GM1 1/2.5, while at 1/1 and 1/0.7 molar ratios it drops to 45% and 25%, respectively.
However, it was also noted that Ptx incorporation into GM1 is impaired by the presence of Dox in the micelles. Thus, when a low amount of Ptx is added (up to 0.5 mg mL^−1^), its incorporation into Dox--GM1 micelles is 63% lower than the incorporation into pure GM1 micelles. Curiously, the addition of larger amounts of Ptx (\<0.5 mg mL^−1^) caused precipitation of most of the Ptx, rendering mixed micelles with only around 10% of the Ptx than the amount that could be solubilized when mixed with Dox-free GM1 micelles ([Figure 2](#f2-ijn-10-3377){ref-type="fig"}).
These results clearly suggest that Dox induced a change in the micellar structure that hinders the subsequent entry of Ptx into the hydrophobic region of the micelles.
Structural characterization of Dox--GM1 and Dox--Ptx--GM1 micelles
------------------------------------------------------------------
In order to gain further insight into the structure and thermodynamic equilibrium of the micelles, we evaluated the elution profile of Dox--GM1 and Dox--Ptx--GM1 micelles at different molar ratios, using a size exclusion chromatography column, and their average size by DLS.
The chromatographic patterns obtained show that GM1 presents two elution peaks, one equivalent to a globular protein of 365 kDa and the other corresponding to the population of monomers that are in equilibrium with the micelles ([Figure 3A, B](#f3-ijn-10-3377){ref-type="fig"}).
The loading of Dox into GM1 micelles led to changes in the hydrodynamic radius that correlated with the amount of drug incorporated into the micelles. At molar ratios from 1/10 up to 1/5 (Dox--GM1), the incorporation of the drug produced a slight increase in the hydrodynamic radius of the micelles from 365 kDa to 390 kDa. However, when the amount of Dox incorporated into the micelles was increased up to 1/2.5 Dox--GM1 molar ratios, two populations of micelles with hydrodynamic radius of 370 and 210 kDa were found ([Figure 3A](#f3-ijn-10-3377){ref-type="fig"}).
These changes in the hydrodynamic radius as a function of the amount of Dox loaded into the micelles were also observed with its incorporation into Ptx--GM1 (1/20 molar ratio) micelles. Again, at 1/2.5 Dox--GM1 molar ratio, micelles seemed to be unstabilized, leading to two principal peaks with hydrodynamic radius of 370 and 210 kDa and to a gradient of new micellar populations of widely differing sizes ([Figure 3B](#f3-ijn-10-3377){ref-type="fig"}).
Interestingly, it was noticed that the elution peak corresponding to the monomers of GM1 disappeared completely after incorporating Ptx, Dox, or both drugs simultaneously, suggesting that the presence of either drug produces a displacement of the equilibrium toward the micellar state indicating that mixed micelles are thermodynamically more stable than pure GM1 micelles.
At this point, knowing the high loading capacity and the homogeneity of the Dox--GM1 1/5 molar ratio micellar population, we selected this proportion to continue our studies.
In agreement with the results obtained by chromatographic analysis of these mixed micelles, their average size as measured by DLS was not modified by the incorporation of Dox into GM1 or Ptx--GM1 micelles at 1/51 molar ratio ([Table 1](#t1-ijn-10-3377){ref-type="table"}).
Physical stability of Dox--GM1 and Dox--Ptx--GM1 micelles
---------------------------------------------------------
As shown in [Table 2](#t2-ijn-10-3377){ref-type="table"}, the amounts of Dox and Ptx that remained soluble after centrifugation of Dox--GM1 and Dox--Ptx--GM1 micelles at high speeds, was above 90% in all cases.
Furthermore, both micelles remained stable in solution for at least 2 months at 4°C and 25°C and neither Dox nor Ptx was released from the micelles after freeze--thawing cycles or lyophilization procedures (data not shown).
We also evaluated the stability of micelles loaded with Dox and Dox--Ptx dialyzed for 72 hours. It was observed that the entire initial drug loaded remained soluble within the dialysis bag. This result agrees with those observed when passing the mixed micelles through a size exclusion column indicating that the incorporation of either Dox or Ptx shifts the equilibrium toward the micellar form and that the mixed drug--GM1 complex structure formed is thermodynamically and kinetically stable in solution (data not shown).
Chemical stability of Dox in Dox--GM1 and Dox--Ptx--GM1 micelles
----------------------------------------------------------------
The incubation of Dox--GM1 and Dox--Ptx--GM1 micelles under pH conditions where the drugs are unstable (pH 10) resulted in about 17% and 25% of Dox hydrolyzed, respectively. This result should be compared with the percentage of free drug hydrolyzed that exceeds 90% after 6 hours ([Figure 4](#f4-ijn-10-3377){ref-type="fig"}).
These results differ from what happens with Ptx, in which case no hydrolysis of the drug was observed when incorporated into GM1 micelles.[@b13-ijn-10-3377] This different susceptibility to alkaline hydrolysis suggests that Dox could be located in a more external domain of the micelle, thus rendering the drug more exposed to the external aqueous environment.
Interaction of Dox--GM1 and Dox--Ptx--GM1 micelles with HSA
-----------------------------------------------------------
In previous reports we demonstrated that Ptx--GM1 micelles spontaneously associate with HSA.[@b15-ijn-10-3377] Further studies revealed that the amount of albumin that ends up being associated with micelles reached saturation at 1/1 HSA--GM1 w/w ratio and also that at 55°C and at pH 3, where HSA is reversibly denatured and exposes part of its hydrophobic residues, the interaction is favored.[@b29-ijn-10-3377]--[@b31-ijn-10-3377] Considering these results together and the fact that Dox might end up located in more external domains of GM1 micelles, we wondered whether the incorporation of Dox into GM1 or Ptx--GM1 micelles could have any effect on their ability to interact with HSA.
Although Dox--GM1 complexes also show the ability to spontaneously interact with HSA, the amount of protein that binds to the micelle in this case depends largely on the amount of incorporated drug. At 1/20 Dox--GM1 molar ratio, the amount of albumin that binds to the micelles reaches a 1/1 (w/w) HSA--GM1 saturation level. However, as the amount of Dox loaded in the micelle increases to 1/5 and 1/2.5 Dox--GM1 molar ratios, the percentage of bound albumin falls to 70% and 40%, respectively ([Figure 3C](#f3-ijn-10-3377){ref-type="fig"}). Similar results were observed for Dox--Ptx--GM1 micelles, as the amount of Dox increases, the quantity of albumin that binds to the micelle decreased ([Figure 3D](#f3-ijn-10-3377){ref-type="fig"}). In this case, the effect is even more pronounced and the amount of albumin that binds to these micelles is much lower than that which binds to the micelles loaded only with Dox falling to 40% for 1/20 and 1/5 Dox--GM1 molar ratios and to 20% for 1/1 molar ratio.
### Effect of albumin on the size of Dox--GM1 or Dox--Ptx--GM1 complexes
Another significant change observed in drug--GM1 mixed micelles after HSA binding was that the interaction induced a significant increase in the size of the structures, reaching an average size of 19.1±1.0 and 17.0±0.9 nm for Dox--GM1--HSA and Dox--Ptx--GM1--HSA complexes, respectively ([Table 1](#t1-ijn-10-3377){ref-type="table"}). This result was expected because HSA is a molecule that has a considerable size and therefore its association with the micelle produces an increase in the size of the micellar structure. In both cases, the binding of albumin to Dox--GM1 or Dox--Ptx--GM1 micelles seems to be driven by hydrophobic interactions, because it was not prevented or removed in the presence of 0.5 M NaCl (data not shown).
Characterization of the in vitro biological effect of Dox--GM1 and Dox--GM1--HSA micelles
-----------------------------------------------------------------------------------------
Hep2 cells were used to assess the biological effect of Dox--GM1 and Dox--GM1--HSA micelles compared to the effect of the free drug in 150 mM NaCl as a control. The results of [Figure 5](#f5-ijn-10-3377){ref-type="fig"}, performed at 37°C for 24 hours, show similar IC~50~ values of Dox whether the drug is as free drug in saline, in Dox--GM1, or in Dox--GM1--HSA micelles. It should be noted that studies conducted at shorter incubation periods (15, 30, 45, 60, and 240 minutes) gave similar results, without significant differences observed between the free drug and the micellar formulations. In addition, in studies performed at 4°C versus 37°C and after 60 minutes of cell incubation with 10 μg mL^−1^ of Dox or Dox--GM1, the cytotoxic effect of Dox was only observed in those cultures incubated at 37°C, whereas no effect was observed in those incubated at 4°C (data not shown), suggesting the involvement of active transport mechanisms in the uptake of both free drug and Dox--GM1 micelles.
Characterization of the cellular uptake of Dox and Ptx loaded into Dox--GM1 and Dox--Ptx--GM1 micelles with and without HSA
---------------------------------------------------------------------------------------------------------------------------
The natural fluorescence of Dox was used to evaluate the kinetics of the cellular uptake of Dox from GM1 micelles compared with the uptake of the free form of the drug. We also assessed whether the interaction of Dox--GM1 micelle with HSA affects the cellular uptake of Dox. The kinetic profile of naturally fluorescent Dox delivery from Dox--GM1 and Dox--GM1--HSA micelles to Hep2 cell line is shown in [Figure 6](#f6-ijn-10-3377){ref-type="fig"}. The images show that there are subtle differences in the cellular uptake of Dox from the different formulations, being slightly higher when the drug is free or associated to GM1 micelles without bound albumin. Regarding Dox location inside the cells, it is observed that in the case of the free drug and that loaded in GM1 micelles, the drug takes a little longer to reach its site of action, the cell nucleus, than in the case of the formulation containing HSA.
We also used a fluorescently labeled derivative of Ptx to assess its cellular uptake. As shown in [Figure 7](#f7-ijn-10-3377){ref-type="fig"}, the kinetic profile of Dox and Ptx delivery from Dox--Ptx--GM1 and Dox--Ptx--GM1--HSA micelles was similar to that from control solutions.
Discussion
==========
As previously mentioned, the use of Dox is limited by its serious side effects. Liposomal formulations of Dox have shown a strong reduction of the toxicity and an increase in blood circulation time. However, these formulations also generate a high cardiotoxic effect which strongly limits their use.[@b32-ijn-10-3377]--[@b34-ijn-10-3377] More recently, Dox was loaded into synthetic multicomponent structures such as pH-sensitive micelles or amphiphilic block copolymers that self-assemble into nanoscopic micelles and accommodate Dox in the hydrophobic core,[@b35-ijn-10-3377]--[@b41-ijn-10-3377] or under enzymatically activated pro-drug of Dox covalently bound to copolymers micelles.[@b42-ijn-10-3377]
In this work, we evaluated the interactions between GM1 nano-micelles and Dox and characterize the behavior and physicochemical properties of the structures resulting from this interaction. Our results show that GM1 micelles load spontaneously higher amount of Dox (10 mg mL^−1^) per unit volume than, for example, classical liposomal formulations of Dox (2 mg mL^−1^).[@b24-ijn-10-3377] This loading capacity of GM1 is not altered by changes in pH or in the ionic strength of the medium, implying that the hydrophobic effect is the main driving force involved in the interaction. Moreover, the ability of GM1 micelles to load Dox was not affected by temperature, which may be related to the fact that, as the Dox is a hydrophilic drug, the temperature rise in the aqueous medium does not produce a change in the repulsion between molecules, as seen in the case of Ptx.[@b13-ijn-10-3377]
We demonstrated that this Dox--GM1 complex is stable at different physical treatments such as freeze--thawing, high-speed centrifugation, and lyophilization--resuspension. Moreover, as in the case of Ptx,[@b13-ijn-10-3377] drug loading produces a shift in the monomer--micelle equilibrium to the aggregated form preventing disassembly of micelles due to dilution, providing greater stability.
DLS analysis of Dox--GM1 complexes in a molar ratio of 1/5 showed that these have an average size of about 12 nm, which is similar to that of polymeric micelles,[@b39-ijn-10-3377] but very different from the size of commercial liposomes, which have an average size around 100 nm. This confers a benefit over the latter, allowing micelles to evade elimination by the reticuloendothelial system, a common problem with liposomes. In this respect, many reports describe that the addition of a minimum amount of GM1 in the lipid mixture is sufficient to cause an increase in the average life of the liposome and also a differential tissue distribution,[@b43-ijn-10-3377],[@b44-ijn-10-3377] which could represent a potential advantage in the use of GM1 micelles.
In addition, the chromatographic profile of Dox--GM1 of different molar ratios shows that the size of micelles changes with increasing proportions of Dox in the structure. This result suggests differences in the stability of Dox--GM1 complexes at different molar ratios. Initially, the micelle enlarges with Dox loading; however, when a critical point is reached, the additional quantities of drug destabilize the structure, showing a blasting of structures of different molecular size. This phenomenon is observed with Dox but not with Ptx,[@b13-ijn-10-3377] suggesting that these modifications arise from changes in the areas occupied by Dox in the micelle. This is probably due to the fact that Dox is located in areas of the micelle described as critical for the stability of the aggregated structure formed by gangliosides.[@b45-ijn-10-3377]
Moreover, many reports[@b16-ijn-10-3377],[@b46-ijn-10-3377] describe that albumin increases the antitumor capacity through its binding to gp160, promoting endothelial transcytosis of the formulations. In this work we confirm that Dox--GM1 micelles retain the ability to spontaneously interact with albumin forming a ternary complex of Dox--GM1--HSA with a size of around 19 nm. However, in this case there is a competitive relationship between the amount of Dox incorporated into the micelle and the amount of albumin that binds, which suggests that Dox and albumin partially share common sites in the micelle.
On the other hand, considering published results regarding the ability of GM1 micelles to load Ptx added to those described here in relation to Dox--GM1 interaction, we confirm that GM1 micelle can load hydrophobic (Ptx) and hydrophilic (Dox) drugs together into the same micellar unit with a differential localization of Ptx and Dox in view of each polarity. The fact that supports the presence of this gradient of polarity is the existence of a specific loading sequence, first Ptx in the most inner part, and then Dox in a more superficial area of the structure. The initial loading of Dox determines a barrier precluding the subsequent entry of Ptx to the innermost of the micelle.
The in vitro studies to evaluate the cytotoxic effect of Dox showed that Dox--GM1 and Dox--GM1--HSA micelles have IC~50~ values similar to those of the free drug and, in turn, these IC~50~ values are similar to those obtained with other delivery systems of Dox.[@b47-ijn-10-3377],[@b48-ijn-10-3377]
Finally, the cellular uptake of Ptx and Dox from the micellar complexes was similar to that of the free form of these drugs, only slightly affected when the micelles are associated to albumin, proving that the superficial presence of the protein does not impair the interaction of the micelles with the cells and the subsequent drug release.
Conclusion
==========
The present study described a self-assembled GM1 micellar system able to load and release hydrophobic and hydrophilic drugs like Ptx, Dox, or both that, in addition, interact spontaneously with HSA. These results suggest that GM1 micelles could be used as a system that would allow multiple drug combinations with actives of very different physical and chemical characteristics in a single formulation.
In addition, GM1 micelles have several advantages as a drug delivery system such as homogeneous composition, self-assembly with very low critical micellar concentration (10^−8^--10^−10^ M), small size, spontaneous loading of hydrophobic and hydrophilic drugs as well as interaction with albumin.
The present work was supported by grants from Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET) to DMB and IDB (PIP 11220100100502 and PIP 11220090100732).
**Author contributions**
All authors contributed toward data analysis, drafting and revising the paper and agree to be accountable for all aspects of the work.
**Disclosure**
RVA, IDB, and DMB are researchers and VL is a research fellow of CONICET. All authors confirm that there are no other conflicts of interest associated with this work.
{#f1-ijn-10-3377}
{#f2-ijn-10-3377}
{#f3-ijn-10-3377}
{#f4-ijn-10-3377}
{#f5-ijn-10-3377}
{#f6-ijn-10-3377}
![Cellular uptake of Dox and Ptx from control solutions and from Dox--Ptx--GM1 and Dox--Ptx--GM1--HSA mixed micelles at 15, 30, and 60 min.\
**Notes:** The scale bars indicate 20 μm. Red color corresponds to fluorescent Dox and green color to fluorescent Ptx (7-O-\[*N*-(4′-fluoresceincarbonyl)-l-alanyl\]taxol, Flutax), while orange color refers to the overlapping of red Dox and green Ptx.\
**Abbreviations:** Dox, doxorubicin; GM1, monosialoglycosphingolipid; HSA, human serum albumin; Ptx, paclitaxel; min, minutes.](ijn-10-3377Fig7){#f7-ijn-10-3377}
######
Average size of micellar complexes by dynamic light scattering
Micelles Mean particle size (nm)
-------------------- -------------------------
GM1 12.9±1.0
Ptx--GM1 12.5±0.3
Dox--GM1 12.5±0.6
Dox--Ptx--GM1 12.6±0.5
Dox--GM1--HSA 19.1±1.0
Dox--Ptx--GM1--HSA 17.0±0.9
**Note:** Data are presented as mean ± standard deviation of the mean (n=5).
**Abbreviations:** Dox, doxorubicin; GM1, monosialoglycosphingolipid; HSA, human serum albumin; Ptx, paclitaxel.
######
Resistance to high centrifugation forces
Centrifugation Dox--GM1 micelles Dox--Ptx--GM1 micelles
---------------- ------------------- ------------------------ ----
25,000× *g* 98 98 98
50,000× *g* 94 95 94
100,000× *g* 90 91 93
**Notes:** Dox and Ptx were determined by absorbance at 490 nm and by high-performance liquid chromatography, respectively, on the supernatant, next to the centrifugation of mixed micelles at 25,000, 50,000, and 100,000× *g* for 1 hour at 20°C.
**Abbreviations:** Dox, doxorubicin; GM1, monosialoglycosphingolipid; Ptx, paclitaxel.
| {
"pile_set_name": "PubMed Central"
} |
1. Introduction {#sec1}
===============
Skin cancer is the most common cancer in the United States (US). It is estimated that close to 4 million skin cancer diagnoses (including basal cell and squamous cell carcinomas) are made every year \[[@B1]\]. Melanoma (an aggressive form of skin cancer) is diagnosed in more than 70,000 persons every year, creating a high health and economic burden with an estimated annual cost of \$3.5 billion \[[@B2]\]. Risk factors for skin cancer include sun sensitivity (sunburning easily, difficulty tanning), a history of excessive sun exposure, sunburns, use of artificial tanning, and a past history of skin cancer \[[@B1]\]. Most of skin cancer cases could be prevented by protecting the skin from excessive sun exposure and avoiding indoor tanning. Results from an analysis of national data showed that the majority of the US population reported infrequent incidence of sun protection behaviors \[[@B3]\]. Characteristics of groups reporting lower incidence of sun protection include being young (under the age of 40), having a lower education level, being a smoker or a risky drinker, and being less sensitive to the sun \[[@B3]\]. Health research should focus on the identification of psychosocial and modifiable variables to promote sun protection among groups at higher risk for skin cancer and in the general population.
Even when it has been documented that the Hispanic/Latino (referred to as Hispanic) population suffers from a disparity regarding certain cancers compared to non-Hispanic whites (referred to as whites), the lifetime risk of developing skin cancer is higher among whites than other racial groups. For melanoma, it is higher among whites (2.9% in men, 1.9% in women) than in Hispanics (0.52% in men, 0.51% in women) \[[@B1], [@B4]\]. A study conducted in Miami showed that, among 3000 cases of nonmelanoma skin cancer reviewed, 60.1% were diagnosed in whites and 38.4% were diagnosed in Hispanics \[[@B5]\]. Findings using the Southeastern Arizona Skin Cancer Registry showed that the rates for nonmelanoma skin cancer in whites were approximately 11 times greater than rates for Latinos \[[@B6]\]. A case control study of nonmelanoma cancer diagnoses in Hispanics (with whites as control) showed that 15.3% of Latino patients reported recurrence of their malignancy as compared to 31.3% of controls \[[@B7]\]. Also, a lower proportion of Latinos (34.0% versus 61.3% controls) had a current diagnosis or prior history of actinic keratosis. On the other hand, skin cancer has been associated with considerable morbidity and mortality in the Hispanic population. Compared with whites, Hispanics have lower 5-year melanoma survival rates, 76.6% versus 87.0% for men and 88.3% versus 92.3 for women \[[@B4]\]. Hispanics are more likely to have advanced and thicker melanomas at diagnosis when compared with whites \[[@B8]--[@B16]\]. A greater percentage of melanomas occurred among Hispanics in younger age groups (24.4% less than 40 years old) compared with blacks and whites, 15.8% and 14.3%, respectively \[[@B16]\]. Also, Hispanics tend to report lower frequency of skin-related visits to dermatologists than their white counterparts \[[@B17]\]. Data obtained from cancer registries of Puerto Rico, New York, New Jersey, and Connecticut show that Puerto Ricans living in the US report higher melanoma rates than those residing in Puerto Rico \[[@B18]\]. At the same time, there are variations in the behaviors reported by Hispanics and non-Hispanics. A systematic review examined the incidence of sun protection behaviors among Hispanics in the US \[[@B19]\]. Overall, the prevalence of these behaviors is both low and mixed. While a slightly lower share of Hispanics (9.5--29.9%) report usage of sunscreen either most of the time or always compared to 16.5%--35.9% of whites, Hispanics reported slightly higher rates of wearing hats either most of the time or always (23.9--25.0% versus 20--20.7%). Recent studies of sun protection behaviors show that around 53% of Hispanics stay in shade, and around 20% use protective clothing when outside on a warm sunny day either most of the time or always \[[@B20], [@B21]\]. Hispanics who are less acculturated report lower rates of sunscreen use than those who are more acculturated \[[@B21]\]. Still, little is known about skin cancer risk factors in the Latino population. It is critical to identify psychosocial and modifiable factors influencing skin cancer morbidity and mortality in Hispanics in the US.
The Community Preventive Service Task Force reviewed skin cancer prevention evidence from a Community Guide systematic review published in 2004 combined with more recent evidence \[[@B22], [@B23]\]. The review found that education interventions in primary and middle schools (Kindergarten--8th grade), which include strategies to integrate parents, caregivers, and teachers, decrease sun exposure, sun protection, and formation of new moles. Multicomponent, communitywide interventions including a combination of individual-directed strategies (e.g., activities to change the knowledge, attitudes, beliefs, or behaviors), mass media campaigns, and policy changes are recommended based on evidence of effectiveness in increasing sunscreen use, but results for effects on other protective behaviors are mixed. Results also suggest benefits in reducing sunburns among children. In addition, findings illustrate that other approaches, such as mass media alone, provider education and media-based education sessions in health care settings, and educational activities in high school and colleges, did not provide sufficient evidence to determine their applicability for skin cancer prevention. Many of these studies were conducted outside of the US (i.e., Australia and the United Kingdom), but the Task Force suggests that findings are likely to be applicable to the US because results were similar across countries. Various interventions and education initiatives in the recent past have targeted minorities with the intention of improving skin cancer, but these were not multicomponent initiatives \[[@B24]--[@B26]\]. A group of Hispanic women evaluated two educational videos to increase positive sun protection beliefs and behaviors \[[@B24]\]. There was an effect in skin cancer risk awareness postintervention, and participants reported they preferred the video emphasizing the benefits of sun protection for skin cancer prevention more than the video emphasizing its effect on photoaging. Little research has examined the association between sun protection behavioral outcomes and the health outcome of interest, that is, skin cancer incidence \[[@B23]\]. More research is needed to verify the efficacy of multicomponent, communitywide interventions addressing the effect of sun protection attitudes, perceptions, beliefs, and behaviors on increasing sun protection. In addition, research should evaluate its effect on decreasing sunburns (short-term effect) and skin cancer incidence (long-term effect) in the general public and in subgroups at particular risk for skin cancer.
This paper examines published studies that include health beliefs concerning skin cancer prevention and sun protection in Hispanics.
2. Materials and Methods {#sec2}
========================
2.1. Search Strategy {#sec2.1}
--------------------
We performed a search of the databases PubMed, PsycINFO, and CINAHL. All publication years and all search fields were included. The search was limited to articles in English and employed specific search keywords. One example of a search strategy used with the PubMED database is ((skin cancer) AND Hispanic) AND beliefs; ((skin cancer) AND Latino) AND beliefs; ((sun protection) AND Hispanic) AND beliefs; ((sun protection) AND Latino) AND beliefs. We decided to use broad search terms to make sure we would identify as many pertinent studies as possible. For our search, we decided to use the word "Hispanic" and "Latino" to indicate our population of interest, that is, US residents of Mexican, Cuban, Puerto Rican, Central American, South American, and other Spanish-speaking country origins. A search in PubMED demonstrated how research, with some exceptions (including US Census data and self-report), interchangeably uses these terms and lacks stratification of the members of this group \[[@B27]\]. A study by the Pew Hispanic Center found that more than half Hispanics (51%) have no preference for any of the two terms to describe their ethnicity \[[@B28]\]. At the same time, the term "Hispanic" was chosen to be used in this paper given that sun protection research applies this term more frequently compared with the term "Latino" (see [Table 1](#tab1){ref-type="table"} for list of the term(s) for ethnicity and raced used by each study included in this review). A manual secondary search of all bibliographies from relevant articles was performed to yield further relevant publications. We excluded studies conducted outside the US, as well as studies without data for Hispanic participants on the report of sun protection beliefs. Studies that compared the differences in sun protection beliefs between Hispanics and non-Hispanics were included as well.
2.2. Eligibility Criteria {#sec2.2}
-------------------------
Articles were reviewed for relevance with the criteria for inclusion being as follows. (1) The reports had to quantitatively examine and report (frequency, means, percentages, effect sizes, and/or odds ratio) sun protection beliefs in Hispanic samples, including constructs such as skin cancer risk/susceptibility and severity/seriousness, and sun protection beliefs (barriers and benefits of sun protection and skin cancer risk). (2) The number of Hispanic participants in the sample had to be stated. (3) Studies that reported the differences in sun protection beliefs between Hispanics and other racial and ethnic groups were used. Books, book chapters, meta-analyses, comments, and reviews were excluded.
3. Results {#sec3}
==========
The first database searched was PubMED, followed by a search of PsycINFO and CINAHL. A total of 86 articles were identified from these databases, and 6 articles were identified from bibliographies, for a total of 92 records (see [Figure 1](#fig1){ref-type="fig"}). A search of duplicates was conducted, leaving 61 records that were title- and abstract-screened and 18 records that were screened in full. The title and abstract screening step evaluated the title and the abstract of each of the 61 articles to determine whether the abstracts met the following criteria: (1) informed about sun protection beliefs, (2) informed about the sample used (humans), (3) suggested that the study included Hispanics in the sample, (4) used English as publication language, and (5) indicated that the publication was peer-reviewed. As part of the full screening step, the manuscripts from the 18 abstracts were obtained and read in full to determine whether they met the eligibility criteria: (1) the records had to quantitatively examine and report sun protection beliefs in Hispanics, (2) the number of Hispanic participants in the sample had to be clearly specified, and (3) studies reporting differences in sun protection beliefs between Hispanics and other racial and ethnic groups were included in the review. These manuscripts were read in full, and eleven were included in the final analysis (see [Figure 1](#fig1){ref-type="fig"}). Data on the author, year of publication, sample characteristics, methodology used, measures selected, and quantitative results from each of the articles were abstracted and evaluated.
Findings are illustrated in [Table 1](#tab1){ref-type="table"}. Six studies included adult participants, and three studies included children and adolescents (students in middle school and high school). One study included both adolescents and adults, and one study did not report information regarding the inclusion of participants that were under 18 years old (participants were patients at a dermatology clinic). Three studies excluded participants with history of skin cancer, and five studies reported data on sun protection behaviors. One study was population-based. Two studies had survey materials available in Spanish, and one study was conducted entirely in Spanish. Two studies reported data on country origin/heritage. Five studies had relatively small Hispanic samples (less than 100 participants). Most studies (*n* = 10) were published during the last decade (2004--2014).
Skin cancer seriousness (severity, worry) was a belief considered in three studies \[[@B29]--[@B31]\]. Results from a population study showed that Hispanics and whites share similar levels of worry about skin cancer (*P* \> 0.05) \[[@B29]\]. One study reported a midrange score in terms of skin cancer worry in Hispanics. Another study found that perceived skin cancer severity was associated with incidence of total body examination (i.e., a head-to-toe examination of the skin performed by a physician used to identify suspicious growths that may be cancer or growths that may develop into skin), but not with incidence of skin self-examination (i.e., a head-to-toe examination of the skin performed by the individual, not a physician) \[[@B30], [@B31]\].
Most studies included in the review (*n* = 8) considered skin cancer susceptibility (or risk) beliefs. One study found that Hispanics believe they have lower than average risk of developing skin cancer, and that their level of risk is lower when compared with whites \[[@B32]\]. A second study reported that most Hispanics described their skin cancer susceptibility as average \[[@B33]\]. A third study also found lower perception of skin cancer risk when Hispanics were compared with whites but the difference was not significant when participants were asked to compare their own likelihood of getting skin cancer compared with the risk of an average person of the same age \[[@B29]\]. Two studies reported similar scores on their skin cancer risk and photoaging (changes in skin appearance induced by sun exposure) concern measures but used different scales for the scores \[[@B30], [@B34]\]. Participants were inclined to not agree or disagree (midrange score) with the following statement: "The natural color of my skin protects me from the sun" \[[@B30]\]. After quantifying qualitative information from an all-female sample, it was found that less than half of the participants believe they can develop skin cancer \[[@B24]\]. On the other hand, almost all participants were concerned about the effect of sun exposure on their appearance. Using a different sample as part of an experiment within the same study, participants reported an increase in skin cancer susceptibility "of Hispanics with fair skin and Hispanics with dark skin" (not their own susceptibility) after watching an educational video about sun protection behaviors. Another study asked Hispanics about the skin cancer susceptibility of people "in darker skin types," and results showed that a slightly lower proportion of Hispanics (78%) endorsed this statement compared with white (91%) and black (86%) participants \[[@B35]\]. Another study reported that perceived skin cancer risk is associated with skin self-examination \[[@B31]\].
A strong effect size was reported for association between perceived peer norms for sun exposure and barriers to sun safety in Hispanic middle school students \[[@B36]\]. Hispanics are more likely to believe that there is not much they can do to lower their risk of getting skin cancer and that there are too many recommendations to prevent this illness \[[@B26]\]. It was also reported that more than half of Hispanics believe that tanning makes people look more attractive and do not endorse the belief that tanning makes people older \[[@B37]\]. One study showed that Hispanics tend to marginally agree more with statements regarding sun protection benefits than barriers \[[@B30]\]. Participants also indicated what were the most important benefits and barriers to engage in sun protection behaviors, with "avoid getting sunburn" and "not part of my daily routine" as frequently endorsed statements. Another study showed a similar pattern in terms of sun protection beliefs \[[@B34]\]. An additional reason Hispanics endorse for failing to use sunscreen is because they consider themselves "dark skinned" \[[@B38]\].
4. Discussion {#sec4}
=============
This study examined published reports of sun protection beliefs in Hispanics, and we found eleven manuscripts that followed the established criteria. Results suggest that low skin cancer susceptibility is commonly found in this population and that Hispanics moderately perceive skin cancer as a serious health threat. Results also suggest that assessments of sun protection barriers and benefits vary significantly by study. Overall, findings illustrated that there are limited studies on psychosocial and modifiable factors that influence sun protection. Many of the studies included in this review have limited sample size or used samples that do not represent the heterogeneous Hispanic population (e.g., 70% of participants in one study were of Mexican origin) \[[@B30], [@B31]\]. Findings must be validated in larger, more comprehensive studies. Results also emphasize the need for comparable and consistent assessment regarding sun protection. This finding is consistent with previous skin cancer prevention results. An evaluation of interventions designed to educate primary care physicians about skin cancer showed a lack of uniformity across interventions and outcome assessments, preventing the direct comparison of intervention efficacy and the dissemination of effective components \[[@B39]\].
Skin cancer can be prevented by practicing sun protection, but skin cancer disparities might be associated with the perceptions, knowledge, attitudes, and beliefs Hispanics hold regarding skin cancer and sun protection. It has been found that individuals who express the benefits of sun protection are likely to report sun protection behaviors consistently more than those who communicate the barriers of sun protection \[[@B40]--[@B42]\]. Hispanics are more likely to believe there is little they can do to lower their chances of getting skin cancer, that there are so many recommendations about skin cancer prevention that they do not know which one to believe and believe that they are below average risk for skin cancer compared with whites. It is critical to understand the sets of beliefs that underlie sun protection among Hispanics and improve health promotion initiatives to decrease sun protection disparities.
Most common types of skin cancer are squamous cell carcinoma and basal cell carcinoma (SCC and BCC, resp.; nonmelanoma), and melanoma. These cancers have different causes and presentations. Basal cell carcinoma diagnoses are more common in Hispanics than squamous cell carcinoma and melanoma diagnoses \[[@B43], [@B44]\]. While person characteristics (e.g., light skin, sun sensitivity, and blistering sunburns early in life) and intermittent sun exposure are strong risk factors for the diagnosis of melanoma, both cumulative and intermittent sun exposure are the most common cause of basal cell carcinoma. In terms of presentation, melanoma usually involves sites not exposed to the sun, including palmar, plantar, and mucosal surfaces, and the lower extremities \[[@B45]\]. Areas such as the head and the neck regions seem to be more prone for basal cell carcinoma. Literature in sun exposure at the workplace indicates an elevated risk for SCC but is less conclusive for BCC \[[@B46]\]. A population-based control-study among individuals diagnosed with invasive melanoma found that frequent sunscreen use when not planning to be in the sun during the last 20 years was strongly associated with lower likelihood of melanoma \[[@B47]\]. Also, those who reported use of sun protection (not sunscreen) were at lower risk of developing melanoma, even if its use was inconsistent. Consistent with the compensation hypothesis of sunscreen use and increased sun exposure, optimal use of sunscreen SPF+15 was associated with highest amount of sun exposure. Research directly associating sun protection behaviors to decreased skin cancer risk is limited and inconsistent. The present study shows that research still struggles to investigate and understand the specific factors that might be associated with melanoma and nonmelanoma skin cancer incidence, and disparities in skin cancer. Research should clarify the association between the disease, the target population, and the particular mechanisms to prevent the disease.
Previous research shows a moderate level of awareness about skin cancer risk factors and prevention behaviors among Hispanics. Using a qualitative approach, forty Hispanics were asked about their understanding of skin cancer risk terminology \[[@B26]\]. Results illustrated that participants did not recognize possible indicators of skin cancer risk (e.g., painful sunburns). One study showed that more Hispanics do not use sunscreen because they perceive themselves as dark skinned when compared with whites and Asian/Pacific Islanders (29% versus 3% versus 11.4%, *P* \< 0.05) \[[@B38]\]. Results included in the present review emphasize the need for improved assessments of sun protection beliefs and to incorporate the evaluation of the meaning and significance Hispanics give to sun protection: do they know what protective clothing is?; what do they think about using sunscreen "all the time" when out in the sun, including cloudy days?; would they wear a hat at all?; do they know how their skin reacts to sun exposure? These are questions that future research should address if we want to report a more accurate analysis of sun protection. This review underscores the importance of developing culturally relevant, validated, and reliable measures of sun protection and skin cancer risk perception for Hispanic adults, adolescents, and children.
Our search was limited to peer-reviewed journals, which generally publish studies with significant results. It was also limited to results on sun protection beliefs in Hispanics living in the US. Findings cannot be generalized to studies conducted in other countries, and there is a possibility that important elements of sun protection were not captured in our review. A strength of this review is that it helped us realize that the full potentials of the assessment and applications of sun protection beliefs for the prevention of skin cancer in Hispanics remain largely unverified and untested. This finding should open the doors to many research initiatives to promote health in a growing minority population and to identify and understand factors that contribute to disparities in the incidence and mortality of cancer.
Hispanics are a diverse group that exhibit differences in terms of sun protection behaviors, sun sensitivity, level of acculturation, country of origin, access to health services, and socioeconomic status. Future research should develop comprehensive, culturally sensitive measures of sun protection beliefs, facilitators, and barriers. Measures should be grounded in theory, research evidence, and ethnographic study. Also, researchers must ensure that their recruitment strategy attains a more diverse sample than previous research. The National Cancer Institute states that overcoming cancer health disparities is one of the best opportunities we have for lessening the burden of cancer \[[@B48]\]. It is goal of the institute to improve the understanding of the causes of cancer health disparities as a way to eliminate them. It is critical to identify modifiable factors that can reduce skin cancer morbidity and mortality disparities in the US. There is a need for informed, culturally sensitive measures to assess sun protection in the Hispanic population in order to (1) directly inform the development of a study to investigate the ability of sun protection beliefs to predict the likelihood to engage in positive health outcomes (i.e., sun protection behaviors), (2) make informed assessments of the effect of sun protection on skin cancer morbidity and mortality, (3) contribute to the limited literature on sun protection in Hispanics, (4) inform the development of targeted public health recommendations and initiatives to increase sun protection, and (5) make a contribution to the identification and understanding of experiences Hispanics have regarding sun protection.
The authors acknowledge the support received by the Department of Oncological Sciences, Icahn School of Medicine at Mount Sinai (Grant R25-CA081137).
Conflict of Interests
=====================
The authors declare that there is no conflict of interests regarding the publication of this paper.
{#fig1}
######
Findings related to sun protection and skin cancer risk beliefs in Hispanics.
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Study Sample Design Term used for ethnicity and race Sun protection/skin cancer risk beliefs and findings (Hispanic) Comments about quality of study
----------------------------------------- ----------------------------------------------------------------------------------------- --------------------------------------------------------------------------------- ---------------------------------- ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Andreeva et al. (2008) \[[@B36]\] Total (*N* = 1,782)\ Cross-sectional\ Hispanic\ Results affect size in structural model (standardized solution): perceived peer norms for sun exposure and barriers to sun safety (0.459); barriers to sun safety and protanning attitudes (0.015), and barriers to sun safety and sun-safe behavior (0.142).\ Data were collected as part of sunny days, healthy ways program in Colorado, New Mexico, and Arizona. Information about validation of measurements in Hispanic sample was not reported. Original study reported low alpha reliability (less than 0.70) for measures of barriers to sun protection and peer norms for sun exposure.
Hispanic (*N* = 437)\ self-survey (no information about survey in Spanish) non-Hispanic Hispanics having lower scores on the tan-related measures and slightly higher scores on barriers.
*Adolescents*
Buster et al. (2012) \[[@B29]\] Total (*N* = 1,246)\ Population-based interviewer survey (no information about interview in Spanish) Hispanic\ Results comparison with whites: likelihood of future skin cancer (OR 1.41); likelihood of skin cancer compared with average person of the same age (OR 0.83); worry about skin cancer (OR 0.83); there is not much you can do to lower chance of getting skin cancer (OR 3.87); there are so many recommendations about preventing skin cancer; it is hard to know which ones to follow (OR 3.35). Data collected by Health Information National Trends Survey (HINTS). Statistical analyses were conducted with samples of different sizes (Hispanic *n* = 161 versus white *n* = 966). The HINTS is available in both Spanish and English (not mentioned in the paper), but there was no mention of possible influence of language preference.
Hispanic (*N* = 161)\ black, white
*Adults*
Cheng et al. (2010) \[[@B37]\] Total (*N* = 1,214)\ Cross-sectional\ Hispanic\ Percentages: tanning makes people look more attractive (true 61%); tanning makes people look older (True 27%). The survey was part of an educational intervention which included a pretest, a 30-minute lesson on sun protection, and a posttest. There was no mention of this intervention being available in Spanish and English, or assessment of cultural competence. Limitations include a population surveyed from only New Jersey public schools and small Hispanic sample.
Hispanic (*N* = 266)\ self-survey (no information\ black, white
*Adolescents* about survey in Spanish)
Coups et al. (2014) \[[@B30]\] Total (*N* = 787 Hispanic)\ Cross-sectional\ Hispanic Weighted means (range 1--5) and standard deviations: suntan benefits = 2.58 (1.08); sunscreen benefits = 3.71 (0.94); shade seeking benefits = 3.62 (0.96); sun protective clothing benefits = 3.72 (0.94); sunscreen barriers = 2.51 (0.78); shade seeking barriers = 2.77 (0.71); sun protective clothing barriers = 2.70 (0.83); skin cancer worry = 2.51 (1.13); perceived skin cancer risk = 3.71 (1.06); photoaging concerns = 3.72 (1.02); perceived natural skin protection = 2.61 (1.26). Participants completed an English- or Spanish-language online survey. Information about validity and reliability of measures by language preference was not provided. Authors mentioned survey items that were not already available in Spanish were translated, affecting the cultural appropriateness of the study for the Spanish-speaking Hispanic population. A large proportion of the sample was of Mexican origin (71%).
*Adults* self-survey/online (survey available in Spanish)
Coups et al. (2013) \[[@B21], [@B31]\] Total (*N* = 787 Hispanic)\ Cross-sectional\ Hispanic Correlates of sun protection beliefs and skin self-examination (and total body examination): perceived skin cancer risk AOR 1.34 (AOR 0.89); perceived skin cancer severity AOR 1.05 (1.91). Participants completed an English- or Spanish-language online survey. Information about validity and reliability of measures by language preference was not provided. Participants\' state of residence was as follows: California, *n* = 379; Texas, *n* = 231; Florida, *n* = 110; Arizona, *n* = 41; and New Mexico, *n* = 26.
*Adults* self-survey/online (survey available in Spanish)
Heckman and Cohen-Filipic\ Total (*N* = 74 Hispanic)\ Cross-sectional\ Hispanic Means (range 0--10 and 4--20) and standard deviations: how likely is it that you will develop skin cancer? = 3.70 (2.43); how likely is it that your skin will age too soon? = 3.69 (2.62); benefits of UV exposure (four items) = 11.10 (3.83); benefits of sun protection (four items) = 13.11 (3.87). This pilot study was part of an educational collaboration between a high school science department and a cancer center in the suburbs of Philadelphia. A small Hispanic sample (*n* = 74) was studied. Items from original study were developed from a sample of parents and children (*N* = white = 55%, African American = 26%, and Hispanic = 15%). The sun protection benefits scales were originally developed as mixed sun protection knowledge and attitude scale, and it had low alpha reliability (less than 0.70). The reviewed study showed better reliability. Participants could choose to receive information and surveys in either English or Spanish, but no information about appropriateness of the measures for culture or language preference was provided.
(2012) \[[@B34]\] *Adolescents and young adults* self-survey (survey available in Spanish)
Hernandez et al. (2014) \[[@B24]\] Total study 1 (*N* = 52 Hispanic); total study 2 (*N* = 80 Hispanic; 67 women, 13 men)\ Qualitative (quantitative results reported); experimental phase (video)\ Hispanic Study 1 frequencies: believes she can develop a skin cancer (Yes = 25/52); concerned about lentigines (Yes = 48/52) and wrinkles (Yes = 35/52).\ Effect of two short Spanish-language films on sun protection beliefs was tested (in Chicago, Illinois). One emphasized photoaging benefits of sun protection, while the second focused on its benefits for skin cancer prevention. Nine patients at a dermatology clinic, whose primary language was Spanish, were asked to view the videos and review the questionnaires before it being administered. The samples studied were small, and it is not clear how results would apply to English-speaking Hispanics. Authors developed the videos using primarily the opinions of women rather than men, and the male sample size in the intervention group was limited.
*Adults* (study conducted in Spanish) Study 2 frequencies: fair-skinned Hispanics are at risk for skin cancer (prevideo, agreement = 54/80; postvideo agreement = 72/80); dark-skinned Hispanics are at risk for skin cancer (preagreement = 44/80; postagreement = 69/80).
Imahiyerobo-Ip et al. (2011) \[[@B35]\] Total (*N* = 165)\ Cross-sectional\ Hispanic\ Frequency and percentage: believes that skin cancer can happen in darker skin types (29/37; 78%). A survey was administered to 165 patients seeking care from a dermatology practice in New York City. Limitations include the small sample size and the inclusion of patients who may have had a history of actinic keratoses.
Hispanic (*N* = 38)\ self-survey (no information\ white, African American,\
*Patients* about survey in Spanish) Asian, and others
Ma et al. (2007) \[[@B32]\] Total (*N* = 369)\ Cross-sectional\ White Hispanic\ Frequencies, percentages, and results comparison with whites: chances of developing skin cancer in the future "higher than average" (4.1%), "average" (19%), "lower than average" (51.6%), and "don\'t know" (25.3%). Logistic regression "average or above" (OR 0.6) after controlling for age, sex, skin type, and family history of skin cancer. A pilot survey study using 1 of the 33 public high schools located in the Miami-Dade County area of Florida. A self-administered, anonymous survey, which was derived from a tool used in a derivate of the national Nurses\' Health Study (94% white). Information about validation of measurements in Hispanic sample, or clarification of the term "average risk of skin cancer" was not reported.
Hispanic (*N* = 221)\ self-survey (no information\ white non-Hispanic
*Adolescents* about survey in Spanish)
Mahler (2014) \[[@B38]\] Total (*N* = 1183)\ Baseline self-survey (no information\ Hispanic\ Significant differences in percentages of responses when compared with whites after controlling for skin sensitivity. Sunscreen benefits: avoid getting too dark (Hispanic 15.1%, white 5.8%). Sunscreen barriers: it is too much trouble (Hispanic 16.4%, white 30.6%); I am dark skinned (Hispanic 29.1%, white 3%). The data were drawn from baseline questionnaires completed during 9 different sun protection experiments conducted in San Diego, California. Participants who indicated ever using sunscreen checked any of the listed sunscreen benefits/barriers. The authors mentioned that the list was developed through piloting, but no additional information is provided. No information about appropriateness of the measures for culture or language preference was provided. A small Hispanic sample was used for the statistical comparisons.
Hispanic (*N* = 65)\ about survey in Spanish) white, Asian/Pacific Islander
*Adults*
Pipitone et al. (2002) \[[@B33]\] Total (*N* = 153)\ Cross-sectional\ White Hispanic\ Frequencies perceived risk of melanoma or skin cancer "higher than average" (4%), "average" (59%), "lower than average" (22%), and "don\'t know" (15%). Prospective survey of a group of suburban employees to evaluate perceptions of skin cancer risk. Low participation in the study and small Hispanic sample. Information about validation of measurements in Hispanic sample, or clarification of the term "average risk of skin cancer" for comparisons was not reported.
Hispanic (*N* = 27)\ self-survey (no information\ white non-Hispanic
*Adults* about survey in Spanish)
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
OR: odds ratio; AOR: adjusted odds ratio.
[^1]: Academic Editor: Marianne Berwick
| {
"pile_set_name": "PubMed Central"
} |
Introduction
============
The quinolone derivative mefloquine is widely used in the treatment and prophylaxis of malaria in travelers to areas with chloroquine-resistant falciparum malaria (Palmer et al., [@B23]). In the Netherlands mefloquine is generally prescribed when traveling for longer than 6 weeks to a malaria endemic region. Mefloquine has gained popularity due to its weekly dosing regime and is used worldwide by all kinds of civil and military organizations.
A negative effect of mefloquine is that it occasionally causes neuropsychiatric adverse events. A number of studies on mefloquine tolerability in young and healthy populations have shown that mild neuropsychiatric adverse events, like headache, strange or vivid dreams, dizziness, anxiety, and sleeplessness, may occur in up to 32% of mefloquine users (Schlagenhauf et al., [@B27], [@B28]; van Riemsdijk et al., [@B37]), while more severe neuropsychiatric adverse events, like psychosis and bipolar disorder, have also been described (Croft and World, [@B7]; Piening and Young, [@B25]; Even et al., [@B11]; Toovey, [@B35]).
Although, it is known that mefloquine interacts with an array of protein targets and cellular signaling pathways no studies have been performed regarding the nature of the origin of these adverse events (Lim and Go, [@B18]; Gribble et al., [@B14]; Maertens et al., [@B20]; Kang et al., [@B15]; Dow et al., [@B10]; Weiss et al., [@B38]; Cruikshank et al., [@B8]). Mefloquine has been shown to block connexin 36 (Cx36) gap junction very potently (Cruikshank et al., [@B8]; Margineanu and Klitgaard, [@B21]). Cx36 is a plasma membrane protein responsible for the formation of gap junctions between neurons (Condorelli et al., [@B5]; Sohl et al., [@B33]) and has been found in many brain areas such as thalamus, hippocampus, basal ganglia, cerebral cortex, olfactory bulb, brainstem, and retina (Condorelli et al., [@B4]). Cx36 is highly expressed in the gap junctions of the inferior olive (IO) and these electrical synapses are the main source of intra-olivary electrical transmissions (De Zeeuw et al., [@B9]; Condorelli et al., [@B4]). The IO gives rise to the climbing fibers projecting to the Purkinje cells in the cerebellar cortex and functionally integrates sensory information and motor output resulting in well coordinated motor performance, motor learning, and cognitive processes (Simpson et al., [@B29]; Strick et al., [@B34]).
In mice that were injected with mefloquine in their IO, timing deficits of newly learned motor responses were found in an associative motor learning task (Van Der Giessen et al., [@B36]), indicating the importance of Cx36 gap junctions in the process of learning-dependent timing. Other studies revealed that the IO is involved in the cognitive property of perception of temporal complex rhythmic stimuli (Xu et al., [@B41]; Liu et al., [@B19]; Wu et al., [@B40]). The impact of mefloquine on the IO and on the olivocerebellar dependent behaviors has never been investigated in humans. Results from animal studies and the high prevalence of mefloquine associated neuropsychiatric adverse events led us to study the hypothesis that mefloquine alters olivocerebellar related behaviors in humans.
Therefore, we have investigated the effect of mefloquine on different aspects of motor performance, perceptual timing, and motor learning in humans. We examined the influence of mefloquine by studying voluntary motor timing (dart throwing task: Smeets et al., [@B30]), perceptual timing (rhythm perception task: Xu et al., [@B41]), reflex timing (reflex blinking: Koekkoek et al., [@B17]; Smit et al., [@B31]), and an associative motor learning task (eye-blink conditioning: Koekkoek et al., [@B17]; Smit et al., [@B31]).
Materials and Methods
=====================
Study population and experimental set up
----------------------------------------
We recruited volunteers, who were prescribed mefloquine (Roche, USA; the commercial name marketed as Lariam^®^) for malaria prophylaxis at the Travel Clinic Havenziekenhuis in Rotterdam, the Netherlands. Written informed consent was obtained from the participants or the parents of the participants when they were younger than 18 years. The experimental protocol was approved by the Erasmus MC Medical Ethics Committee (MEC-2004-205).
Participants who had used mefloquine in the preceding 2 months, who had risk factors for concentration impairment (e.g., use of opioids, hypnotics or tranquilizers during the study period or use of alcohol 4 h before testing) or who had illnesses resulting in altered motor behavior were excluded from this study. From the participants information was collected on gender, weight, height, age, weekly physical exercise, and previous mefloquine intake by use of a questionnaire (Table [1](#T1){ref-type="table"}). The participants were requested to take the standard dose of 250 mg mefloquine at the end of dinner, as food facilitates mefloquine absorption (Crevoisier et al., [@B6]). One day after the first mefloquine intake, users were carefully reviewed for adverse effects (by Thomas A. van Essen, see also Table [1](#T1){ref-type="table"}).
######
**Characteristics participants**.
Characteristics Performance task (n = 9) Learning task (no mefloquine; n = 11) Mefloquine (n = 8)
------------------------------- ---------------------------------- --------------------------------------- ----------------------------------
Men/women 6/3 7/4 4/4
Age (years) 40.7 24.3 25.0
Age range (years) 17--67^a^ 20--35^a^ 17--40^a^
Weight (kg) 74.2 65.5 68.8
Physical exercise (≥2 ×/week) 6 (67%) 7 (64%) 5 (63%)
Adverse events 3 (33%)[b](#tfn1){ref-type="fn"} -- 3 (38%)[b](#tfn2){ref-type="fn"}
Insomnia 2 1
Head ache 1 1
Abnormal dreams/nightmares 1 2
*Gender, mean age, age range, mean weight, amount of exercise level, and adverse events for each group*.
*^a^Age range is different between the motor performance and the motor learning participants*.
*^b^Numbers do not add up to total adverse events since some participants had more than one adverse event*.
The experimental set up consisted of performance tasks: (a) dart throwing task, (b) perceptual timing task, and (c) reflex blinking task. To test motor learning, the eye-blink conditioning task was used. The performance experiments were executed 2--4 days prior to the intake of mefloquine (baseline) and 24 h after the intake of the first mefloquine tablet. The learning experiments were only performed once per (naïve) subject, in order to prevent contamination of the results with a memory trace. The eye-blink conditioning task was performed in a group of participants 24 h after the intake of their first mefloquine tablet and in an age- and sex-matched control group who did not use mefloquine.
Performance experiments
-----------------------
### Dart throwing task
In this task participants threw 25 darts at a dartboard and were told to aim for the bull\'s eye while holding their underarm vertically during the course of each throw. They were placed on a standard range from the dart board (2 m) and each throw was followed by a short pause of 5 s.
Throwing imprecision can be divided in a horizontal and vertical component. The horizontal (E~x~) and vertical deviations (E~y~) of each dart from the center of the bull\'s eye were determined manually (Figure [1](#F1){ref-type="fig"}B). Under normal conditions, imprecision in release timing or throwing speed is not the main source of throwing imprecision, but the imprecision in direction of motion at the time of release is considered the main cause of throwing imprecision (Smeets et al., [@B30]). However, when either speed or timing precision is affected by mefloquine, an additional error source is introduced that will affect the precision of the dart throw. This error source will only affect the vertical component, but not the horizontal component, of the throw (Figure [1](#F1){ref-type="fig"}A). Therefore, we would expect an enhanced variability in the vertical deviation compared to the horizontal deviation after the use of mefloquine. The speed and timing sensitivity of the dart throw was determined by dividing vertical deviation from the target (E~y~) by the total deviation from the target (E~y~ + E~x~). The precision of the dart throw was determined by dividing the standard deviation of the vertical displacement from the target (sdE~y~) by the total standard deviation of the displacement from the target (sdE~y~ + sdE~x~).
{#F1}
### Rhythm perception task
Functional MRI studies revealed that the IO is activated during perception of complex rhythms and perception of a single stimulus with unexpected timing, and that the IO is sensitive to the temporal structure of sensory input (Xu et al., [@B41]; Liu et al., [@B19]; Wu et al., [@B40]). In this task participants have to compare successive auditory rhythms (Xu et al., [@B41]). Two successive rhythms were generated by two pairs of four tones. Each pair consisted of four tones of 500 Hz with a duration time of 5 ms and these tones were separated from each other by intervals of 333 ms. The pairs were separated by an interval of 2 s (Figure [2](#F2){ref-type="fig"}A). A rhythm asynchrony was introduced in either the first or second four-tone rhythm, consequently the other four-tone rhythm became the designated standard rhythm. Rhythm asynchrony was achieved by shifting the second tone of the rhythm closer to the first (i.e., shortening the inter-tone interval). This shortening of the inter-tone interval (perturbation) varied randomly throughout each test block between 0, 10, 25, 50, 75, and 100 ms. The task consisted of eight blocks and in each block the participants were exposed to all six perturbations. For each perturbation we determined the number of errors that were made per subject. The 0 ms perturbation served as a control of understanding the task and to asses to what extent the task was subject to guessing. After each series of two rhythms participants were asked whether the rhythms were identical or not. The participants were explicitly instructed not to guess. Due to the involvement of the IO in this task and the temporal structure of the task, we would expect more errors in discriminating the rhythms after the use of mefloquine.
{#F2}
### Reflex blinking
Participants were subjected to reflex blinking and eye-blink conditioning procedures. Eye blinks were recorded with the use of magnetic distance measurement technique as described by (Koekkoek et al., [@B17]; Smit et al., [@B31]). Briefly, a magnetoresistive sensor was attached at the rim of the orbit below the right lower eyelid, while a small gold plated neodymium magnet was attached to the tarsal portion of the right upper eyelid (Figure [3](#F3){ref-type="fig"}A). The signal was amplified by a pre-amplifier close to the sensor and further amplification could be adjusted per subject. The signal was digitized (1000 Hz) using National Instruments hardware. A custom made LabView (National Instruments, Austin, TX, USA) script controlled the monitoring of eyelid movement, presentation of the stimuli and capturing of data on disk. Via the puff nozzle, an air puff (unconditioned stimulus: US) was directed to the cornea close to the outer canthus of the eye at a distance of about 15 mm. The apparatus served to deliver a 20-ms air-puff with an intensity of 10--50 PSI at the source. The intensity and duration were manually adjusted to elicit a single blink reflex (unconditioned response: UR). In the reflex blinking sessions 15 blinks were randomly evoked in 1 s trials each. For each reflex blink three temporal aspects, the onset, peak time velocity, and peak time amplitude were determined. The onset of the blink is the time between the start of the US and the start of the eyelid movement, the peak time velocity is the time between the start of the US and the moment that the eyelid reaches the maximal velocity (which is determined after differentiating the eyelid trace), and the peak time amplitude is the time between the start of the US and the moment that the eyelid reaches the maximal amplitude (Figure [3](#F3){ref-type="fig"}A).
![**Reflex blinking is not affected by the intake of mefloquine**. **(A)** Left panel: Schematic drawing of eye-blink recording system. A gold plated neodymium magnet was attached to the edge of the upper eyelid (red circle), while a magnetoresistive sensor was attached at the edge of the orbit below the right lower eyelid (green circle). A blink was induced by a small air-puff (unconditioned stimulus: US). An increased magnetic force is detected by the sensor during closure of the eyelid. Right panel: eyelid movement recording of an unconditioned response (UR). Three temporal aspects of the reflex blink were determined: the onset, peak time velocity and peak time amplitude (red arrows). Figure is adopted from Smit et al. ([@B31]). **(B)** Histogram of eye-blink timing parameters before (blue bars) and 24 h after the intake of mefloquine (red bars). **(C)** Histogram of standard deviations of the measured eye-blink timing parameters before (blue bars) and 24 h after the intake of mefloquine (red bars). Histograms show mean and SEM (*n* = 9).](fnins-04-00191-g003){#F3}
Motor learning: eye-blink conditioning
--------------------------------------
Eye-blink conditioning is an associative learning task (Gormezano et al., [@B13]), in which a reflex blink (induced by US) is associated to a tone (conditioned stimulus: CS). Repetitive paired presentation of the CS with the US results in the generation of a conditioned response (CR), a closure of the eyelid upon presentation of CS (Figure [4](#F4){ref-type="fig"}A). The olivocerebellar system is involved in this conditioning of the eyelid response (Medina et al., [@B22]).
{#F4}
For this task the participants were equipped with a headphone, video goggles (Logitech, USA), and a puff nozzle. The headphone and video goggles provided additional entertainment in order to ensure that attention deficits and anxiety would not interfere with the eye-blink conditioning session (Smit et al., [@B31]). In the eye-blink conditioning sessions, participants received eight blocks of eight 2 s trials. Each block consisted of one air-puff only trial (US only) as described above, one trial in which the tone (CS) was presented alone (CS only) and six trials in which tone and air-puff were paired. In these paired trials the 20 ms air-puff ended simultaneously with the 520 ms tone by presenting it with a 500 ms delay after onset of the tone (Figure [4](#F4){ref-type="fig"}A). For each CS only and paired trial the presence or absence of a CR was determined. For each subject and experiment we calculated the percentage of CRs within the blocks and as a whole. The eight blocks were preceded by the first block, called block 0, during which participants were presented tone-only trials. In this block, we confirmed that the participants did not yet learned the ability to make a conditioned eye-blink response.
Statistical analyses
--------------------
With regard to the performance experiments, all comparisons were done with paired, two-tailed *t* test. For the learning experiments (i.e., eye-blink conditioning), we used the unpaired, two-tailed *t* test. Furthermore, we fitted this learning data set with the following learning curve: the exponential function *y* = *a* − *a* exp^−*x*/*b*^, where *y* is the percentage of CRs, *x* the block number, and *a* and *b* are the learning coefficients (*a* represents learning capacity and *b* learning speed) of the model. The goodness of fit was indicated by *R*^2^ values. All mean values were accompanied by standard error of the mean (SEM) unless stated otherwise.
Results
=======
In the study period of July 2009 to April 2010, 28 individuals were included. Nine participants volunteered for the performance experiments and 19 for the motor learning experiments (8 used mefloquine; 11 individuals served as control participants \[no mefloquine use\]). Table [1](#T1){ref-type="table"} summarizes the general characteristics of the individuals participating in the performance and motor learning experiments. Six participants reported adverse events after the intake of mefloquine (see Table [1](#T1){ref-type="table"}).
Performance of voluntary motor timing, perceptual timing, and reflex timing is not affected in mefloquine users
---------------------------------------------------------------------------------------------------------------
To find out whether mefloquine leads in humans to deficits in olivocerebellar related performances, we tested our participants before and after the intake of mefloquine (i.e., off and on medication), using specific tasks in which the IO is thought to be involved.
Voluntary motor timing was assessed by a dart throwing task. Throwing imprecision can be divided in a vertical and horizontal component. Imprecision in release timing or throwing speed is more reflected in the vertical deviation than in the horizontal deviation to the target (Figure [1](#F1){ref-type="fig"}A). Our participants revealed a significant larger vertical deviation than horizontal deviation under both conditions (Figure [1](#F1){ref-type="fig"}C, meflo: *p* = 0.01, no meflo: *p* = 0.01, *t* test), indicating the more complex control of the vertical component over the horizontal component of the throw. The vertical deviation as well as the horizontal deviation was not affected by the use of mefloquine (Figure [1](#F1){ref-type="fig"}C, *p* = 0. 47 and *p* = 0.73 respectively, *n* = 9, *t* test). In order to correct for non-specific inter-trial variability, we normalized the vertical deviation by the total deviation of each throw. These normalized vertical deviations did not reveal a significant difference (Figure [1](#F1){ref-type="fig"}D, *p* = 0.45, *n* = 9, *t* test). Neither did the normalized variation in the vertical deviation, which is an indicator for vertical throwing precision (Figure [1](#F1){ref-type="fig"}E, *p* = 0.34, *n* = 9, *t* test).
Since the IO has been suggested to play a role in the perception of complex rhythms (Xu et al., [@B41]), we investigated whether a rhythm perception task in mefloquine users would result in reduced ability to distinguish between rhythms (Figure [2](#F2){ref-type="fig"}A). No significant differences were found with respect to the number of errors made per perturbation between the two conditions (Figure [2](#F2){ref-type="fig"}B). The 25 ms perturbation yielded the largest difference, although not significant (*p* = 0.25, *n* = 9, *t* test).
The effect of mefloquine on an involuntary motor performance task was investigated by studying the reflex blink (Figure [3](#F3){ref-type="fig"}A). Reflex blinks obtained before and after the intake of mefloquine were analyzed kinetically. The onset, the peak time velocity and the peak time amplitude did not reveal a significant difference between participants "off" mefloquine compared to "on" mefloquine (Figure [3](#F3){ref-type="fig"}B, *p* = 0.78, *n* = 9, *t* test). The precision in timing of the eye-blink was investigated by testing the standard deviation of these parameters. Although, the standard deviations were all larger when the participants were under the influence of mefloquine, they were not significantly larger (Figure [3](#F3){ref-type="fig"}C, *p* = 0.59, *n* = 9, *t* test).
In all performance tasks the variation was more closely examined with the coefficient of variation (CV). The resulting CVs were not significantly different either (unpublished observations). Taken together, these results show that mefloquine has no influence on these three different IO-related performance tasks.
Eye-blink conditioning is affected in mefloquine users
------------------------------------------------------
To investigate whether the use of mefloquine leads to motor learning deficits in humans, we tested mefloquine users and control participants (i.e., no mefloquine use) using an eye-blink conditioning task (Figure [4](#F4){ref-type="fig"}A).
Compared to the control participants, mefloquine users showed a significant lower number CRs that were acquired during training and after training (Figures [4](#F4){ref-type="fig"}B,C, *p* = 0.02, control participants: *n* = 11 and mefloquine users: *n* = 8, *t* test). The percentage of CRs was on average 74.0 ± 4.7% in the control group and 46.8 ± 11.0% in the group of mefloquine users. The learning curves of the mefloquine group and the control group are plotted in Figure [4](#F4){ref-type="fig"}D. Before the task started, it was confirmed that volunteers did not respond to the tone and did respond to the US with an UR. Therefore, in block 0 no conditioned blinking has been observed. In blocks 1, 2, 4, 5, and 6 the number of CRs was significantly lower in the group of mefloquine users compared to the control group (respectively *p* = 0.03, 0.01, 0.01, 0.04, and 0.03, *t* test), whereas blocks 3, 7, and 8 did not reveal significantly difference (respectively *p* = 0.14, 0.14, and 0.08, *t* test). Curve fitting the data of both groups resulted in well-associated learning curves ($R_{\text{control}}^{\text{2}} = 0.98$, *p* \< 0.01 and $R_{\text{mefloquine}}^{\text{2}} = 0.87$, *p* \< 0.01). The fitted learning curve generated two learning coefficients: *a* and *b*. The learning coefficient *a* represents learning capacity, whereas the learning coefficient *b* represents learning speed. The coefficients *a* and *b* of the learning curves were both significantly different between the control group and the group of mefloquine users (control: *a* = 84.7 ± 0.7 and *b* = 1.53 ± 0.05 vs mefloquine: *a* = 70.8 ± 7.2 and *b* = 3.96 ± 0.71, both *p* \< 0.05, *t* test). Mefloquine users learned this task 2.6 times slower than the participants that did not use mefloquine and their learning capacity was also reduced by 16% compared to the non-users. The average onset latency of the CRs was similar in the mefloquine group and control group (control: 877 ± 14.8 ms vs mefloquine: 906 ± 16.6 ms, *p* = 0.23, *t* test). UR onset across groups did not differ either (control: 1062 ± 6.1 ms vs mefloquine: 1075 ± 4.5 ms, *p* = 0.13, *t* test), excluding that possible deficits in reflex pathways can contribute to the observed reduced motor learning. Conclusively, these data show that the use of mefloquine affects motor learning.
Discussion
==========
In this study we investigated the possibility that mefloquine induces deficits in olivocerebellar related behaviors in young and healthy volunteers. We show that mefloquine has no effect on dart throwing, rhythm perception and reflex blinking, but has an effect on the conditioning of eye-blink responses to a tone (associative learning). In this learning-dependent motor task, mefloquine users were impaired in their learning speed as well as in their learning capacity. Mefloquine reduced their learning speed by 2.6 times and their learning capacity by 16%.
Animal studies in which mefloquine was injected in the IO and studies on Cx36-deficient mice have shown comparable motor learning problems (Frisch et al., [@B12]; Van Der Giessen et al., [@B36]). The training process in these animals was slowed down substantially, but they revealed no impaired general motor performance (Kistler et al., [@B16]; Van Der Giessen et al., [@B36]) similar to our findings in humans. The lack of effect of mefloquine on general performance is most likely due to the specific function of Cx36 within the IO. Blocking Cx36 by mefloquine does not affect the performance, it makes the performance only unadjustable. In order to affect these performances other mechanisms within the IO need to be altered (Rondi-Reig et al., [@B26]; Welsh, [@B39]).
With regard to our experimental design, the question arises whether a single dose of mefloquine is capable of altering brain functions. Mefloquine has no problem crossing the blood--brain barrier (Baudry et al., [@B2]) and there is evidence that mefloquine can accumulate in the brain (Barraud de Lagerie et al., [@B1]). This is, for instance, supported by the observation of a higher risk of neuropsychiatric adverse events in the first days after first mefloquine intake (Schlagenhauf et al., [@B27], [@B28]; van Riemsdijk et al., [@B37]), suggesting a direct effect of mefloquine on the brain. Some of these neuropsychiatric adverse events were also observed in our mefloquine using participants (Table [1](#T1){ref-type="table"}). Furthermore, the reported IC~50~ of mefloquine for Cx36 is 0.3 μM (Cruikshank et al., [@B8]) and this is much smaller than the mefloquine blood level concentration achieved with one dosage of 250 mg mefloquine (Pennie et al., [@B24]), indicating that the level of mefloquine in the brain is high enough to alter brain functions.
Of course, one could argue that the altered brain function associated with mefloquine might be caused by mechanisms other than interference with Cx36 gap junction coupling. However, the IC~50~ values for mefloquine inhibiting ATP-sensitive potassium channels (IC~50~ = 3 μM), L-type calcium channels (IC~50~ ≈ 10 μM), delayed rectifier channels (IC~50~ = 1 μM), and volume- (IC~50~ = 1 μM) and calcium-activated chloride channels (IC~50~ = 3 μM) are all higher than those required for Cx36 blockade (Coker et al., [@B3]; Gribble et al., [@B14]; Maertens et al., [@B20]; Kang et al., [@B15]). So far, Cx36 gap junctions are the most mefloquine-sensitive molecules in the brain and the mefloquine blocking effect in the IO will lead to motor learning deficit (Van Der Giessen et al., [@B36]).
Finally, one could also suggest that the present findings might be explained by a confounding involvement of other Cx36 containing brain areas apart from the IO. Despite the fact that for all the tasks used in this study the involvement of the IO is necessary (Rondi-Reig et al., [@B26]; Xu et al., [@B41]; Van Der Giessen et al., [@B36]), we cannot directly exclude a possible role of other Cx36 containing brain areas. It is, though, noteworthy that an altered attention due to mefloquine might have affected the dart throwing and the rhythm perception task; it cannot explain the results on eye-blink and eye-blink conditioning tasks, because not being aware of the stimuli does not prevent eye blinking or eye-blink conditioning (Smith et al., [@B32]). Thus, the most likely explanation for our results is the selective blocking of Cx36 gap junctions in the IO.
The demonstration that mefloquine, a marketed anti-malarial drug, imposes a deficit in motor learning capabilities in man might come as a surprise in the field of adverse events of malaria prophylaxis and treatment. Although this adverse effect is not defined on direct clinical grounds (i.e., people do not report the adverse effect reduced motor learning as a result of mefloquine usage), it can have clinical relevance. For example, the use of mefloquine may limit the speed and capacity to rehabilitate motor functions after traumatic injuries. Moreover, knowledge of this effect is also relevant for civil and military organizations that detach people to malaria endemic regions where they have to rely heavily on fine-tuning their motor skills or learn new motor skills. Therefore, this adverse effect should be further investigated and recognized by clinicians.
Conflict of Interest Statement:
===============================
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
We are grateful for the contribution of the participants. We thank the nurses of the Travel Clinic in Rotterdam for their assistance during the recruitment of the volunteers. We also acknowledge the help of Dr. A. Smit and J. Willems with the eye-blink experiments. This work was supported by the Netherlands Organization for Scientific research (ZonMw: grant 971.96.347).
[^1]: Edited by: Ian M. Stanford, Aston University, UK
[^2]: Reviewed by: Carol Holland, Aston University, UK; Stephen D. Hall, Aston University, UK
[^3]: This article was submitted to Frontiers in Neuropharmacology, a specialty of Frontiers in Neuroscience.
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#s1}
============
Ethiopia is a rapidly developing country that is burdened and held back by a high prevalence of communicable disease. Of the so-called 'big three' killer diseases, HIV/AIDs, tuberculosis, and malaria, malaria is the most frequent cause of out-patient presentation and in-patient admission nationwide and is second only to respiratory tract infections as a cause of death in children [@pntd.0000197-Federal1]. Co-endemic with the 'big three' in Ethiopia are the 'neglected tropical diseases' of which trachoma is the most geographically widespread and cause of greatest morbidity.
The Ethiopian national blindness and low vision survey conducted in 2006 suggests that Ethiopia is the most trachoma affected country in the world. The entire rural population of approximately 65 million people are at risk of blindness from trachoma. At any time there are an estimated 9 million children with clinical signs of active disease, 1.2 million adults with trachomatous trichiasis, and 354,000 persons with blindness or low vision attributed to trachoma [@pntd.0000197-Berhane1]. In addition to the effects on vision and the high likelihood of developing blindness if unoperated [@pntd.0000197-Burton1], trichiasis is a terrible condition in which the eyelashes rub against the surface of the eye ball, leaving sufferers in constant and disabling pain. The irritation and pain caused by the lashes on the surface of the eye ball and cornea is exacerbated by smoke, dust and bright light which prevents people from conducting their normal routine activities such as cooking over solid fuel fires, farming in dusty environments, gathering firewood and collecting water [@pntd.0000197-Frick1]. Of the ten states in Ethiopia, Amhara Regional State is disproportionately affected by trachoma, bearing an estimated minimum of 45% of the national trichiasis burden and with approximately one in twenty of all adults suffering from trichiasis [@pntd.0000197-Berhane1]. It is recognized that the geographic distribution of trachoma within the state is not uniform with some health posts reporting trachoma as the primary cause of out-patient consultations and for others, trachoma ranks below malaria and respiratory tract infection as a cause of out-patient consultation. Trachoma is a barrier to development in Amhara and controlling this disease is a state priority.
Trachoma control in Amhara, and the whole of Ethiopia, is based on the World Health Organization (WHO) endorsed SAFE strategy, in which S is corrective lid surgery for patients with trichiasis, A is antibiotic treatment for individuals with signs of active disease and for mass drug administration to at-risk populations, F is facial cleanliness and hygiene promotion to prevent transmission, and E is environmental improvements such as provision of sanitation and water that address trachoma risk factors [@pntd.0000197-WHO1].
Nationally, it has been recognized that malaria is a major barrier to the development. Ethiopia including Amhara , is prone to unstable and epidemic malaria [@pntd.0000197-Abeku1],[@pntd.0000197-Cox1],[@pntd.0000197-Negash1]. The Federal Ministry of Health has launched a control program of unprecedented scale and scope to relieve this burden. The program is based on personal protection/vector control, and effective case detection and treatment. It is being targeted to the entire Ethiopian population at risk of malaria (estimated at 50 million people) [@pntd.0000197-Federal2]. The program has four pillars: distribution of free long-lasting insecticidal nets (LLINs) at the target rate of two per household in malaria-endemic areas; targeted indoor residual spraying (IRS) in high transmission areas; serum-based rapid diagnostic tests (RDTs) available at all health facilities; and treatment with artemesinin combination therapy for *Plasmodium falciparum* malaria and chloroquine for uncomplicated malaria caused by *Plasmodium vivax*. In Amhara, the malaria is seasonal and follows the rain; presentation of malaria cases is most common from September to January, with a peak in November and December. The survey was conducted in December at the end of the normal peak transmission period.
Recently, there have been a series of publications proposing that large single disease programs be integrated to improve efficiency and reduce costs [@pntd.0000197-Molyneux1], that programs specifically targeting the neglected tropical diseases be integrated and expanded [@pntd.0000197-Hotez1], that neglected tropical disease programs be integrated with the 'big three' [@pntd.0000197-Molyneux2], and that deworming programs be integrated with malaria control [@pntd.0000197-Druilhe1]. It was in this climate that the Amhara Regional Health Bureau partnered with The Carter Center to plan an integrated malaria and trachoma control program that would simultaneously target two of the most devastating communicable diseases in the state: malaria and trachoma.
Amhara (population approximately 19.6 million [@pntd.0000197-BFEOD1]) is divided administratively into 10 zones, which are themselves divided into 140 woredas (district equivalents), and 3,231 kebeles (groups of villages with approximately 5,000 population) ([Figure 1](#pntd-0000197-g001){ref-type="fig"}, Map). Within kebeles the lowest administrative unit is the state team (now known as development teams) which are groups of about 50 families, usually around 250 people, who have an elected representative. Although national direction comes from the Federal Ministry of Health, overall coordination of programs has been devolved to the states, and planning for implementation is conducted by the woreda representatives at the level of the zone. In order to facilitate planning of the interventions and to enable evaluation at both state and zonal level, we conducted an integrated malaria and trachoma survey that was powered to provide prevalence estimates at the zonal level. The integrated survey had the following objectives.
{#pntd-0000197-g001}
*Malaria:* 1) Estimate malaria parasite prevalence by species of parasite and age and gender of host at the zonal level; and 2) Estimate the coverage of indoor residual spraying, use of mosquito nets and use of long-lasting insecticidal nets.
*Trachoma:* 3) Estimate the prevalence of trachomatous inflammation follicular (WHO grade TF) in children aged 1--9 years at the zonal level; 4) Estimate prevalence of trachomatous trichiasis (WHO grade TT) in adults aged over 15 years by zone and sex; and 5) Estimate other household risk factors for trachoma including latrine ownership, access to water supply, and facial cleanliness of children aged 1--9 years.
The data from the integrated survey would allow calculation of programmatic needs for both components of the integrated program: malaria and trachoma. For malaria the number of long lasting insecticidal nets required to meet the distribution targets and information on current net usage for targeting health education messages; for trachoma the estimated backlog of TT surgeries, the population requiring mass drug administration with antibiotic and health education, and the need for household latrines and access to water.
Methods {#s2}
=======
Sample size calculation {#s2a}
-----------------------
The sample size was calculated based on the assumption that prevalence of TT would be the lowest of the indicators that we wished to measure and we would determine prevalence estimates at the zonal level. For each zone we assumed a prevalence of TT in adults of 5%, 2.0% precision, 95% confidence limit, and a design effect of 2. We estimated that we needed to examine at lest 1,000 adults per zone. We assumed adults were 50% of the population, so that the total population sampled would need to be 2,000, and that household size was, on average, 5. This gives an estimated sample of 400 households per zone.
Bahir Dar town was excluded from the sampling frame as were two special woredas 'Debre Markos Ketema' in East Gojjam and 'Debre Tabor' in South Gondar zone. The latter two were excluded because, in accordance with the Regional Health Bureau definition, less than 10% of their population lived in malarious areas; however, these two woredas comprised only 0.4% of the overall population.
To select the 400 households in each zone we used a multi-stage cluster random sampling design. In each zone, eight woredas were selected using probability proportional to size and, within the selected woredas, we selected two kebeles also using probability proportional to size. Within the kebele, five state teams (which are all approximately similar in size) were selected by lottery, literally drawing the names out of a hat at the woreda office. In the final stage, five households were selected from the 50 in the state team using the random walk method.
To determine malaria parasite prevalence, we assumed a population prevalence of 8%, 2% precision, 5% level of significance and 95% confidence limit, a design effect of 1.2 and 15% non-response to give an estimated sample size of 1,000 people or 200 households in each zone. Consequently, we arbitrarily determined that even-numbered households in the overall sample would be recruited for malaria blood films.
In summary, the integrated survey had ten domains (the zones). Within each zone there were 16 clusters, each of 25 households to give a sample of 400 households or 2,000 people per zone. Overall there were 160 clusters of 25 households. Each zone had a different population, so that persons sampled within a zone represented a different proportion of the total population. State-wide prevalence estimates could be calculated from the weighted domain estimates. The survey was conducted in December 2006 at the end of the malaria transmission season.
Household questionnaire {#s2b}
-----------------------
The survey questionnaire was based on the Malaria Indicator Survey Household Questionnaire , modified for the local conditions and to include risk factors for trachoma [@pntd.0000197-WHO1]. The questionnaire was translated and printed in Amharic language and field-tested in a non-survey kebele to determine the validity of the pre-coded answers. Interviews were conducted with the head of household, or another adult if the head of household was absent or unable to respond for any reason. If interviews were conducted with someone other than the head of household then the respondent was requested to answer as though he or she were the head of household. The data collection form had three parts: household questionnaire; malaria parasite prevalence; and trachoma survey.
In the household questionnaire respondents were asked about: their source of drinking water; time to collect water; toilet facilities (latrine presence, if reported, was verified by observation); proxy indicators of wealth; room construction materials; indoor residual spraying; presence and type of mosquito net (verified by observation); demographic information on residents; and where people slept. We defined the source of drinking water as being 'safe' if it was a capped spring, protected hand-dug well, tube well, borehole, cart with small tank, or piped water. Other water sources were described as 'unsafe' and were unprotected springs, unprotected hand-dug wells, and surface water. Our proxy indicators of wealth were electrification of the household, possession of a functioning radio set, and possession of a functioning television set.
Malaria parasite prevalence {#s2c}
---------------------------
Consenting residents of even-numbered households were recruited for the malaria parasite prevalence survey. Participants had both a rapid diagnostic test which gave an on-the-spot diagnosis and provided thick and thin blood films for microscopy. The rapid diagnostic test used was ParaScreen (Zephyr Biomedical Systems, [www.tulipgroup.com](http://www.tulipgroup.com)), this test is able to detect both *P. falciparum* and other plasmodia species (in Amhara most likely *P. vivax*). The test uses approximately 100 µl of blood and is readable after ten minutes. Participants with positive rapid tests were offered treatment according to national guidelines, CoArtem® for *P. falciparum* infection, chloroquine for other *Plasmodium* infection, and clinic-based quinine therapy for self-reported pregnant women [@pntd.0000197-Federal3].
Two blood slides, each composed of thick and thin films, were taken for each participant by a clinical technician according to standard WHO-approved protocol [@pntd.0000197-WHO2]. Slides were labelled and air-dried horizontally in a carrying case in the field, and stained with Giemsa at the nearest health facility when the team returned from the field. Usually, field teams returned to the clinic each evening but when working in inaccessible areas, which required walking up to eight hours each way, they were obliged to sleep in the field and stain the slides the following day. To ensure maximum participation, households with absentees were revisited a second time on the same day to recruit those missing at the first visit.
Blood slides were read at a reference laboratory in Addis Ababa and classified qualitatively as either negative, *P. falciparum* positive, *P.vivax* positive, or mixed infection. One hundred high power fields of the thick film were examined before calling a slide negative. If positive, the thin film was read to determine the species. Parasite density was not quantified. To ensure accuracy, all positive slides and a random sample of 5% of the negative slides were re-examined by a separate microscopist, who was blinded to the diagnosis of the first slide-reader. The second slide from each participant was used if the first was broken or unreadable. The identity of survey participants who had positive blood slides was sent back to the field teams for follow-up and appropriate treatment, where necessary.
Trachoma survey {#s2d}
---------------
Trachoma grading was carried out by Integrated Eye Care Workers (IECW) who were experienced in using the WHO simplified grading [@pntd.0000197-Thylefors1]. This scheme comprises of 5 stages: trachomatous inflammation-follicular (TF), trachomatous inflammation-intense (TI), trachomatous scarring (TS), trachomatous trichiasis (TT) and corneal opacity (CO) (Box 1). Minimum accepted inter-observer agreement was set at 80% and reliability assessed in two stages. In the first stage, potential examiners identified trachoma grades using the WHO set of trachoma slides [@pntd.0000197-WHO3]. Those examiners who achieved at least 80% agreement then proceeded to the second stage of field evaluation. During field evaluation a reliability study comprising 50 persons of varying age and sex were selected by the senior examiner to represent all trachoma grades. Each potential examiner evaluated all 50 subjects independently and recorded their findings on a pre-printed form. Inter-observer agreement was then calculated for each trainee using the senior examiner\'s observation as the 'gold standard'. Examiners achieving at least 80% inter-observer agreement after the field evaluation were included as graders.
All persons living within each selected household who gave verbal consent were examined using a torch and a x2.5 magnifying binocular loupe. Each eye was examined first for in-turned lashes (TT), and the cornea was then inspected for corneal opacities (CO). The upper conjunctiva was subsequently everted and examined for inflammation (TF and TI) and scarring (TS). Both eyes were examined. Signs had to be clearly visible in accordance with the simplified grading system in order to be considered present. Alcohol-soaked cotton-swabs were used to clean the examiner\'s fingers between examinations. Individuals with signs of active trachoma (TF and/or TI) were offered treatment with 1% tetracycline eye ointment. TT patients were referred to health centres where free eyelid surgery was available.
Data to determine whether children aged 1--9 years had a 'clean face' were collected during the eye examinations. Facing the child, the observer looked for the presence of ocular and nasal discharge, recording each separately as a dichotomous variable. Ocular discharge was defined as any material of any colour or consistency in the corner of the eyes, or matting of the eyelashes caused by such a discharge (tears, medication and make-up were excluded). Nasal discharge was defined as the presence of wet exudate of any colour below one or both nostrils.
Quality control, data entry and analysis {#s2e}
----------------------------------------
Forms were checked by the supervisor in the field and inconsistencies verified with the respondent. Data were double entered by different entry clerks and compared for consistency using Census and Survey Processing System (U.S. Census Bureau Washington DC, USA). Statistical analysis was conducted using Stata™ 9.2 (Stata Corporation, College Station, Texas, USA). Descriptive statistics were used to describe the characteristics of the sample. Sampling probabilities were calculated for woredas, kebeles and state teams. Sampling weights were then derived as the inverse of the product of sampling probabilities at the woreda, kebele and state team levels. Point estimates and confidence intervals were derived using the SURVEY (SVY) routine in Stata which controls for clustering and allowed for adjustments for the sampling design as well as weighting for sampling probability [@pntd.0000197-STATA1].
To give greater precision in the estimates of trichiasis burden, TT prevalence was modelled for sex-specific ten-year age groups using logistic regression. Prevalence of TT was calculated for ten-year age groups for males and females separately for each zone. Ten-year age population structures by sex were obtained for each zone from the Amhara Regional Health Bureau [@pntd.0000197-BFEOD1] and applied to the ten-year age group TT prevalence estimates for males and females. The 95% confidence intervals of the point prevalence estimates were multiplied by the respective population structure estimates to derive the lower and upper bounds of the TT burden. All zonal estimates and corresponding upper and lower bounds were summed to derive the state-wide estimate of those requiring TT surgery.
Ethical consideration {#s2f}
---------------------
The protocol received ethical approval from the Emory University Institutional Review Board (IRB 1816) and the Amhara Regional Health Bureau. Verbal informed consent to participate in interviews and trachoma screening was sought from the heads of the household, each individual and the parents of children aged 10 years and younger in accordance with the tenets of the declaration of Helsinki. Signed informed consent was sought from each individual and parents of children aged 10 years and younger in accordance with the tenets of the declaration of Helsinki for blood films. Personal identifiers were removed from the data set before analyses were undertaken.
Results {#s3}
=======
Characteristics of study households and participants {#s3a}
----------------------------------------------------
A total of 4,122 households were selected for the survey of which 21 were excluded since no one was present: 4,101 (99.5%) were included in the analysis ([Figure 2](#pntd-0000197-g002){ref-type="fig"}). The characteristics of the study households are shown in [Table 1](#pntd-0000197-t001){ref-type="table"}, and the location of the sites surveyed are represented on [Figure 1](#pntd-0000197-g001){ref-type="fig"}. The overall mean household size was 4.6 persons (95% confidence interval \[CI\] 4.5--4.7) with household size ranging from 1 to 17. The main source of drinking water came from 'safe' sources for 34.4% (95% CI 27.8--64.3) of households and the round-trip time to collect water was \<30 minutes for 74.1% (95% CI 69.1--78.5). Household pit latrines were present in 24.3% (95% CI 19.9--29.3) of households with 75.7% (95% CI 70.7--80.1) reporting open defecation. Our proxy indicators of wealth (electricity, radio, TV) were reported in 5.7% (95% CI 3.0--10.4), 24.8% (95% CI 22.2--27.6), 1.2% (95% CI 0.8--2.0) of households, respectively. The majority of households, 57.1% (95% CI 52.3--61.8), had a thatch roof; walls made from sticks, 90.5% (95% CI 87.2--93.0); and floors of compacted earth, 73.3% (95% CI 68.8--77.4). Indoor residual spraying in the last 12 months was reported in 14.8% (95% CI 10.1--21.1) of households. Any mosquito net was reported and verified by observation in 34.7% (95% CI 28.3--41.6) of households with a range of 0 to 5 nets observed. Of all the nets seen in all the households, 54.6% were LLINs, and 16.1% (95% CI 12.1--21.2) of households had at least one LLIN. The mean number of LLINs per household was 0.30 (95% CI 0.2--0.3).[](#pntd-0000197-g003){ref-type="fig"}
![The WHO simplified grading scheme for assessment of trachoma.\
Reproduced from Thylefors, 1987 [@pntd.0000197-Thylefors1].](pntd.0000197.g002){#pntd-0000197-g002}
{#pntd-0000197-g003}
10.1371/journal.pntd.0000197.t001
###### Characteristics of study households
{#pntd-0000197-t001-1}
Household characteristics
--------------- ------------ ----- --------------------------- ----- ------ ------ ------ ------ ------ ------ ------
Amhara Region 19,391,698 160 4,101 4.6 57.1 75.7 25.9 34.7 16.1 0.3 14.8
Zones
North Gondor 3,241,161 16 392 4.7 55.1 85.4 60.3 50.5 33.4 0.5 9.8
Waghemira 375,440 16 410 5.0 91.7 94.3 53.1 52.8 52.0 0.8 8.4
South Gondor 2,243,477 16 415 5.0 74.0 76.6 3.6 55.6 4.1 0.1 13.6
North Wollo 1,636,699 16 404 4.5 79.1 49.8 12.8 47.2 30.1 0.4 16.6
West Gojjam 2,674,974 16 400 4.5 15.8 82.8 39.0 33.8 13.7 0.2 24.5
Awi 1,090,879 16 395 5.0 46.7 54.1 25.7 21.3 9.4 0.1 13.7
East Gojjam 2,470,060 16 459 4.2 21.7 81.6 26.7 5.6 0 0 8.8
South Wollo 2,878,970 16 403 4.1 81.5 74.2 10.8 6.1 4.0 0.04 2.5
Oromiya 588,943 16 400 5.8 76.9 59.4 36.1 99.9 53.6 1.2 57.4
North Shewa 2,191,096 16 423 5.2 62.4 91.1 21.4 19.2 11.2 0.2 9.3
**\*:** Bureau of Finance and Economic Development for the year 2006/2007
HH, household; LLIN, Long lasting insecticidal net
[Table 2](#pntd-0000197-t002){ref-type="table"} shows the individual characteristics of study participants and mosquito net usage. A total of 19,710 people were enumerated of whom 41 were excluded from analysis due to missing data on age or sex. Of the 19,669 people included in the sample, the overall mean age was 21.9 years (95% CI 21.5--22.3) and 48.7% were male. The overall proportion of people reporting sleeping under any mosquito net the previous night was 25.8% (95% CI 21.2--30.9). Among the population with particular vulnerability to malaria, the under five year-olds and pregnant women, sleeping under a mosquito net was reported for 29.2% (95% CI 24.1--34.9) and 33.6% (95% CI 25.4--42.8), respectively. Persons reporting sleeping under an LLIN the previous night were: 12.5% (95% CI 9.4--16.5) overall. Of the vulnerable groups, 425/2,929 children aged less than 5 years, 14.5% (95% CI 10.8--19.2); and 46/315 pregnant women, 14.6% (95% CI 10.0--21.0) reported sleeping under an LLIN last night.
10.1371/journal.pntd.0000197.t002
###### Characteristics of study participants and net usage
{#pntd-0000197-t002-2}
Participants characteristics Proportion of people reporting sleeping under a net last night (%)
--------------- ------------------------------ -------------------------------------------------------------------- ------ ------ ------ ------ ------ ------ ------ ------
Amhara Region 19,669 21.9 48.7 25.9 25.8 29.2 33.6 12.5 14.5 14.6
Zones
North Gondor 1,892 20.5 47.8 18.6 34.3 38.7 62.4 23.0 27.4 38.1
Waghemira 2,018 21.7 51.2 53.6 38.7 45.5 78.8 37.6 43.9 77.3
South Gondor 2,094 21.0 52.3 24.1 37.3 39.0 59.5 1.7 2.4 2.0
North Wollo 1,812 23.3 48.5 32.9 31.0 31.8 0.0 21.6 20.6 0.0
West Gojjam 1,812 21.2 47.7 25.3 21.3 24.3 27.3 8.5 10.4 18.5
Awi 1,965 20.2 48.6 74.9 15.5 18.1 10.6 7.4 7.2 4.1
East Gojjam 1,949 22.2 45.6 42.8 2.1 4.0 6.4 0.0 0.0 0.0
South Wollo 1,653 23.6 48.1 3.2 4.5 4.1 14.7 2.9 2.1 4.8
Oromiya 2,289 20.7 48.9 22.2 91.7 97.1 96.5 48.3 51.1 43.6
North Shewa 2,185 23.0 48.9 23.7 14.3 14.8 12.6 8.7 8.6 3.5
LLIN, Long lasting insecticidal net
[Figure 2](#pntd-0000197-g002){ref-type="fig"} shows the sample population and those recruited for trachoma examination and malaria parasite prevalence. Trachoma examination was conducted in 17,242 (87.7%). A total of 9,140 people in even numbered households were eligible for malaria testing of whom 7,745 were included in the analysis (84.7%).
Malaria prevalence {#s3b}
------------------
The malaria parasite prevalence by blood slide microscopy is shown on [Table 3](#pntd-0000197-t003){ref-type="table"}. A total of 7,745 blood slides were examined with good concordance between first and second reading. The overall malaria parasite prevalence in Amhara was 4.6% (95% CI 3.8--5.6) with prevalence by zone ranging from 2.4% (95% CI 1.5--4.0) in Oromiya to 6.1% (95% CI 4.5--8.5) in South Gondor. There were no differences in the proportion of people with positive blood slides by age group: age \<5 years, 5.4%; age 5--14 years, 3.8%; age 15--49 years, 4.5%; and age \>50 years, 4.2%. The malaria species seen most frequently was *P. falciparum*, 52.2% of positive slides had *P. falciparum* only and 8.7% were mixed *P. falciparum and P. vivax*. *Plasmodium vivax* only was seen on 41.3% of the positive slides. The overall ratio of *P. falciparum* to *P. vivax* was 1.2 with zonal estimates ranging from 0.9 to 2.1.
10.1371/journal.pntd.0000197.t003
###### Prevalence of malaria by blood slide microscopy
{#pntd-0000197-t003-3}
Domain Number examined Malaria parasite prevalence (%) Pf:Pv ratio
--------------- ----------------- --------------------------------- ------------- ----- ----------------- -----
Amhara Region 7,745 2.4 1.9 0.4 4.6 (3.8--5.6) 1.2
Zones
North Gondor 703 3.1 2.8 0.0 5.9 (4.2--8.2) 1.1
Waghemira 809 2.0 1.0 0.2 3.1 (1.5--6.4) 1.9
South Gondor 744 2.7 2.5 0.8 6.1 (4.3--8.5) 1.1
North Wollo 835 1.3 1.5 0.4 3.1 (1.5--6.2) 0.9
West Gojjam 713 1.5 1.7 0.3 3.4 (1.7--6.7) 0.9
Awi 799 2.4 2.8 0.4 5.6 (2.4--12.3) 0.9
East Gojjam 699 2.9 1.6 0.4 4.9 (2.4--9.6) 1.7
South Wollo 758 3.6 1.5 0.5 5.6 (3.2--9.4) 2.1
Oromiya 825 1.0 1.1 0.3 2.4 (1.5--4.0) 0.9
North Shewa 860 2.6 2.0 0.2 4.8 (2.7--8.4) 1.3
CI, confidence interval; Pf, *Plasmodium falciparum*; Pv, *Plasmodium vivax*
**\*:** Total parasite prevalence do not equal the sum of the species prevalence in some rows due to rounding
Trachoma prevalence {#s3c}
-------------------
The key trachoma prevalence indicators are shown in [Table 4](#pntd-0000197-t004){ref-type="table"}. The overall prevalence of TF in children aged 1--9 years was 32.7% (95% CI 29.2--36.5). Prevalence of TF in children by zone ranged from 12.6% (95% CI 7.8--19.7) in South Wollo to 60.1% (95% CI 50.4--69.0) in Waghemira. There was no sex difference in TF prevalence: Odds Ratio (OR) = 1.0 (95% CI 0.9--1.2). The overall prevalence of TT in persons aged 15 years and above was 6.2% (95% CI 5.3--7.4). Point estimates by zone of TT prevalence in adults ranged from 2.4% (95% CI 1.4--4.1) in Oromiya to 10.0% (95% CI 6.3--15.6)in West Gojjam. After adjusting for age, adult women were three times more likely to have TT than men: OR = 3.1; (95% CI 2.3--4.1). Trichiasis was also observed in children aged less than 15 years with an overall prevalence of 0.3% (95% CI 0.2--0.5) in the age group 0--14 years, ranging from 0% in North Gondor to 0.8% (95% CI 0.3--1.8) in North Wollo.
10.1371/journal.pntd.0000197.t004
###### Key trachoma prevalence indicators: trachomatous inflammation--follicular (TF) and trachomatous trichiasis (TT)
{#pntd-0000197-t004-4}
Domain TF in children aged 1--9 years TT in children aged 0--14 years TT in people aged 15 and above
--------------- -------------------------------- --------------------------------- -------------------------------- ------- ----- ----------- ------- ------ -----------
Amhara Region 5,485 32.7 29.2--36.5 8,121 0.3 0.2--0.5 9,121 6.2 5.3--7.4
Zones
North Gondor 466 34.7 24.4--46.8 700 0 730 4.3 2.8--6.6
Waghemira 581 60.1 50.4--69.0 918 0.5 0.2--1.5 1,030 6.3 3.9--9.9
South Gondor 589 28.9 20.1--39.6 887 0.1 0.01--0.4 904 3.8 2.5--5.7
North Wollo 539 51.9 35.4--68.0 739 0.8 0.3--1.8 971 9.4 7.2--12.1
West Gojjam 500 33.1 25.3--42.0 774 0.4 0.1--1.3 874 10.0 6.3--15.6
Awi 588 38.9 22.7--57.9 866 0.1 0.01--0.4 893 5.4 4.0--7.3
East Gojjam 548 48.3 44.4--52.2 798 0.3 0.1--0.8 881 7.1 5.4--9.4
South Wollo 484 12.6 7.8--19.7 701 0.3 0.1--1.4 931 3.2 2.2--4.6
Oromiya 663 28.7 19.6--39.8 958 0.1 0.02--0.8 964 2.4 1.4--4.1
North Shewa 527 23.2 14.1--35.9 780 0.3 0.1--1.1 943 9.0 6.7--11.9
CI, confidence interval
Trichiasis burden estimates {#s3d}
---------------------------
Estimates of trichiasis burden are summarized in [Table 5](#pntd-0000197-t005){ref-type="table"}. The number of people with TT in Amhara was estimated to be 643,904 (lower and upper bounds = 419,274--975,635). Consistent with the increased odds of TT in women, the TT burden in females (of all ages) was estimated to be 2.2 fold compared to males. For planning purposes, the number of persons requiring corrective trichiasis surgery in Amhara was estimated to be 645,000.
10.1371/journal.pntd.0000197.t005
###### Trichiasis burden estimates by gender
{#pntd-0000197-t005-5}
Domain Male Female Total
--------------- --------- --------- --------- --------- --------- --------- --------- --------- ---------
Amhara Region 199,929 122,889 321,230 443,975 296,386 654,405 643,904 419,274 975,635
Zones
North Gondor 21,996 13,378 35,843 49,669 31,605 76,663 71,665 44,984 112,506
Waghemira 4,716 2,807 7,773 10,127 6,391 15,581 14,844 9,198 23,354
South Gondor 15,579 9,285 25,908 35,344 22,278 55,051 50,923 31,563 80,960
North Wollo 29,631 19,314 44,762 63,163 45,666 86,186 92,795 64,980 130,948
West Gojjam 42,729 23,540 74,338 88,644 56,656 133,936 131,373 80,196 208,273
Awi 9,531 6,337 14,274 21,699 15,286 30,557 31,230 21,622 44,832
East Gojjam 24,751 16,195 37,634 56,644 39,415 80,682 81,395 55,610 118,316
South Wollo 17,353 10,214 29,286 44,541 28,447 68,778 61,893 38,662 98,064
Oromiya 2,193 1,214 3,930 5,694 3,298 9,640 7,886 4,512 13,570
North Shewa 31,449 20,605 47,483 68,451 47,342 97,331 99,900 67,947 144,813
Lower and upper bounds represents the 95% confidence interval of the point estimates
Discussion {#s4}
==========
This integrated survey of malaria and trachoma demonstrates that malaria is endemic in all zones of Amhara, an area of unstable malaria transmission, and that malaria parasites can be demonstrated at the end of the transmission season in a non-epidemic year. The key malaria indicators of mosquito net coverage and proportion of different population groups sleeping under long-lasting insecticidal nets (LLINs) demonstrated in the survey will be used as a baseline for planning and evaluation of the regional malaria control program. The survey has underscored the public health significance of blinding trachoma and allowed the calculation of intervention targets per zone with unprecedented precision. Conducting an integrated survey has maximized the benefits of organizing a complex survey without significant increases in the time required in the field or the effort of the field teams.
Selecting the sample and conducting field work posed considerable logistical challenges. The mountainous nature of Amhara makes transportation particularly difficult and field teams were obliged to walk for up to eight hours in order to reach some of the selected clusters. In accordance with the state definition of "malarious", three urban centers (comprising 0.4% of the total population) were excluded from the sampling frame. By integrating malaria and trachoma, prevalence estimates, indicators, and risk factors for both diseases could be obtained for the cost of conducting one disease survey with an incremental addition of one person per field team. If the base survey were for malaria, the field team would require a trained trachoma examiner, if the base survey were for trachoma, the field team would require a clinical technician to take the blood slides. All interviewing, record-keeping, and form-checking would be constant.
In order to achieve the required sample size for the estimation of trichiasis, we required twice the number of people needed to estimate malaria parasite prevalence. We arbitrarily selected even-numbered households for malaria parasite prevalence and all households for trachoma. Since there were twenty-five households sampled and numbered sequentially per cluster, this resulted in systematically sampling 12 households for malaria parasite prevalence per cluster, or slightly less than half, and not half of all households. Such inadvertent compromises are likely to occur in integrated surveys and extremely careful planning is required to ensure that the overall objectives required to measure two or more outcomes are achieved. This integrated survey required considerably more pre-planning and coordination than a single disease survey, yet a relatively minor increase in implementation effort. The overall benefits of integration far outweighed the total effort required to plan, conduct and finance two separate surveys.
Amhara is prone to unstable epidemic malaria and the survey was conducted at the end of the transmission season in a non-epidemic year. The data presented here provide the first comprehensive representative baseline data of malaria parasite prevalence, species ratios and malaria indicators for the region. The parasite prevalence is consistent with that reported for unstable transmission areas in Ethiopia by Newman *et al* [@pntd.0000197-Newman1]. The Ethiopian Demographic and Household Survey conducted in 2005 estimated the proportion of households in Amhara that had any mosquito net to be 3.8% and the average number of LLINs per household to be 0.0 [@pntd.0000197-DHS1]. Our data suggest that by the end of 2006 considerable progress has been made in the Amhara Regional State malaria control program, as these indicators have risen to 34.7% of households with any net and a mean of 0.3 LLINs per household. In addition, these data show what needs to be accomplished if the national target of having everybody at risk of malaria sleeping under an insecticidal net, with a mean of two LLINs per household by the end of 2007 are to be achieved.
We included the use of rapid diagnostic tests (RDT) in the field component of the survey in order to maximize the likelihood that participants who were parasite positive received prompt and effective treatment. After the blood slides were read, it was possible to link back to the database and determine the proportion of participants positive by microscopy who were also positive by RDT. There was not absolute concordance between microscopy, which is considered the gold standard, and the RDTs. Participants with negative RDTs and positive blood slides were followed up back to the field and provided treatment according to national guidelines. The discordance will be presented in a separate paper.
Data collected in the survey are all indicative of the prevalence of the diseases and extent of indicators and risk factors. The prevalence of malaria is time-dependent and will vary from month to month and year to year. There is no known seasonality in presentation of clinical signs of trachoma in Ethiopia; however, the application of the simplified grading system is subjective. Clinical examiners went through a two-stage training and were required to have greater than 80% concordance with the gold standard before joining the field teams, which minimizes but does not exclude the risk of observer bias. It has been argued that nucleic acid amplification techniques of conjunctival swabs are the best way of determining the prevalence of ocular chlamydial infection, although this is currently advocated by the WHO for programmatic use [@pntd.0000197-Dawson1].
The 2005/2006 National Survey on Blindness, Low Vision and Trachoma in Ethiopia had a total of 174 clusters of which 33 were in Amhara [@pntd.0000197-Berhane1]. The sampling framework was based on probability proportional to size. A total of 4,609 individuals were surveyed but trichiasis was only reported for the adult population (persons aged 15 years and above) who comprise around 50% of the total population. The national survey estimated the TT prevalence in adults in Amhara to be 5.2%. Our survey included 160 clusters and a total of 17,242 persons surveyed for trichiasis. By having 10 domains and 16 clusters in each domain, rather than one domain and 33 clusters used in the national survey, we are able to estimate the trichiasis burden with greater precision, and have the possibility of focusing interventions to the zones of greatest need. Planning for interventions to operate the TT surgical backlog is conducted at the level of the zone. The estimates derived in this survey enable realistic preparation and planning. The threshold for district-wide intervention with mass antibiotic administration, hygiene promotion and environmental change is 10% TF in children aged 1--9 years [@pntd.0000197-WHO3]. The national survey estimated TF in children aged 1--9 years to be 39.1% in Amhara. Our survey found an overall weighted prevalence of 32.7%, ranging from 12.6% to 60.1% by zone, but more importantly for program planning, the threshold of 10% was exceeded in all ten zones.
The next steps in terms of integrating the delivery of the control program is to determine where integration makes sense in terms of logistics and to continue with separate activities where there are no potential synergistic gains from integration. For the distribution of LLINs, the FMOH policy of providing two free nets per household is not suitable for integration with the strategy for delivering azithromycin for trachoma control, since only one person per household needs to present themselves for nets whereas the entire household need to present themselves and be measured for mass drug administration. Specific training for health extension workers and village volunteers makes sense logistically and can be integrated. The program will use the findings from the malaria indicator questionnaire and trachoma risk factor questionnaire to design combined health education materials that include promotion of positive existing behaviours that will control malaria and trachoma.
The purpose of this paper is to present selected malaria indicators, malaria parasite prevalence, and key trachoma indicators to facilitate planning and evaluation of the Amhara Regional Health Bureau\'s integrated malaria and trachoma control program. Further analysis is underway to characterise the risk factors and identify the best opportunities for promotion of the integrated control program.
We thank Dr. Fiona Matthews from the Medical Research Council Biostatistics Unit at the Institute of Public Health, Cambridge, for statistical advice during the data analysis; Mr. Frew Demeke and Ms. Kelly Callahan from The Carter Center for considerable logistical support; Dr. Eshetu Gurmu from Addis Ababa University Population Studies and Research Center for the coordination of data entry and cleaning; and Ms. Elizabeth Cromwell from The Carter Center for assistance in preparing the manuscript. We are indebted to all the survey participants who gave freely of their time in the survey.
The authors have declared that no competing interests exist.
The work was funded by The Carter Center Malaria Control Program and staff were allocated to the survey by the Amhara Regional Health Bureau. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
[^1]: Conceived and designed the experiments: PE JN PG AWM FR. Performed the experiments: JN EB PG YE TG TE AG AWM MZ AM FR. Analyzed the data: PE JN PG YE TE FR. Wrote the paper: PE JN. All authors reviewed the manuscript.
| {
"pile_set_name": "PubMed Central"
} |
The spectrum of pathogenic mutations found in genes related to cancer development can vary depending on the ethnic groups that are being studied. Specific pathogenic mutations associated to particular ethnic groups show a high frequency due to founder effects, or population bottleneck and consequent inbreeding. As a result, rare pathogenic mutations become more common within the population over time ([@B6]). Inbreeding contributes to linkage disequilibrium in genomic regions that are segregated together for many generations with specific alleles in *loci* placed closest to the mutation site in that specific population ([@B5]).
A well-known example of founder effects is that of *BRCA* pathogenic mutations in the Ashkenazi Jewish population ([@B19]). At least 2.6% (1/40) of this population carry one of the three founder pathogenic mutations described for *BRCA1* (OMIM \#113705) and *BRCA2* (OMIM \#600185): *BRCA1c.66_67delAG* (p.Glu23fs), *BRCA1 c.5266dupC* (p.Gln1756fs, former named 5382insC), and *BRCA2c.5946delT* (p.Ser1982fs) ([@B19]; [@B1]; [@B6]; [@B9]; [@B20]). Although these pathogenic mutations were first identified in Ashkenazi Jews, *BRCA1 c.5266dupC*, was later described in other populations. It has already been identified in many countries of Central and Eastern Europe ([@B2]; [@B8]) and also recurrently described in the Brazilian population ([@B11]; [@B3]; [@B7]), representing 20% of the BRCA1 pathogenic variants reported in a recent survey ([@B13]).
The comparison of haplotypes between families sharing the same mutation allows to distinguish whether high-frequency alleles derive from a single mutational event (a founder mutation), or if they have arisen independently more than once in a population ([@B9]; [@B12]). In a previous report, [@B3] showed that seven unrelated carriers of this mutation share the same haplotype of genetic marker alleles flanking *BRCA1*. However, data available in Brazil regarding the frequency of different pathogenic variants in individuals at risk of hereditary breast/ovary cancer and the availability of samples from the c.5266dupC mutation carriers limited the conclusions at that time with respect to a possible founder effect. In addition, [@B3] did not evaluate the ancestry of the carriers.
In this work, we present a haplotype analysis in an expanded set of Brazilian carriers of *BRCA1 c.5266dupC*. In addition, considering that the Brazilian population is highly admixed, with genetic backgrounds derived from Europeans, Africans and Amerindians, ancestry analysis was carried out to assess the contribution of these different backgrounds to the genetic diversity present in the carriers
Fourteen unrelated heterozygous probands harboring *BRCA1 c.5266dupC* and 26 relatives (carriers or non-carriers of eight families) were recruited for this study from collaborating research centers located in the cities of Rio de Janeiro (5 probands), Barretos (5 probands), and Porto Alegre (4 probands) between 2004 to 2016. All research procedures followed ethical guidelines and were approved by the local Ethics Committee (009/07). The five probands/families from Rio de Janeiro were also previously analyzed in [@B3].
Haplotypes were characterized based on three SNPs and four Short Tandem Repeat (STR) markers along \~581 kb of chromosome 13 encompassing the *BRCA1* locus ([Figure S1](#suppl02){ref-type="supplementary-material"}). SNPs were analyzed by PCR amplification and DNA sequencing, as described by [@B3].
Four microsatellite *loci* were used for genotyping: D17S855 (intragenic marker in intron 20), D17S1325 and D17S1326 (3' markers) and D171321 (5' marker); primers are listed in [Table S1](#suppl01){ref-type="supplementary-material"}. For all microsatellite loci, PCR amplifications were performed in final volumes of 25 μL, with 2.0 mmol/L of MgCl2, 125 μmol/L of each dNTP, 20 pmol of each primer, 1x PCR buffer, 1 U of Platinum *Taq* DNA polymerase (Invitrogen) and 50 ng of genomic DNA. Reactions were submitted to 30 cycles of 94 °C for 15 s, 60 °C for 15 s, and 72 °C for 20 s. Forward primers were labeled with carboxyfluorescein, and PCR products were analyzed in an ABI-PRISM 3730 automatic sequencer (Applied Biosystems, Foster City, CA). Scoring of allele size was achieved using the internal size standard GeneScan --500 LIZ^®^ (Applied Biosystems). Allele size was estimated using Peak Scanner^TM^ software v1.0 (Applied Biosystems).
To estimate genetic ancestry, 46 ancestry-informative markers (AIMs) were selected. These markers were used to investigate the contribution of African, European, East Asian, and Native American populations to the genetic background of the probands. AIMs were genotyped in one multiplex PCR assay followed by capillary electrophoresis, as previously described ([@B15]). *Structure* software ([@B16]) was used to estimate the ancestral components of the samples, and the results were validated by *Admixture* software. Genetic ancestry analysis was carried out for the 14 index cases.
Of all *BRCA1 c.5266dupC* carriers, 21 (63.6%) had been diagnosed with breast and/or ovarian cancer at ages ranging from 22 and 63 years old (median = 43 years) ([Table 1](#t1){ref-type="table"}). The age of cancer-unaffected mutation carriers (n=12) ranged from 22 to 66 years old (median = 47). For five of the 14 probands only the personal cancer history was available, for the remainder at least one relative developed breast or ovarian cancer ([Table 1](#t1){ref-type="table"}).
###### Clinical features of *BRCA1 c.5266dupC* carriers.
Carrier Relationship Gender Cancer type Tumor location Age at diagnostic
--------- ----------------------- -------- ------------- ---------------- -------------------
RJ-01 Proband F Ovarian Bilateral 47/47
RJ-02 Proband F Breast Bilateral 36/41
RJ-02C 3^rd^ degree relative F Breast Bilateral 47
RJ-02F 3^rd^ degree relative F Breast Unilateral 49
RJ-02G 4^th^ degree relative F Healthy \- \-
RJ-03 Proband F Breast Bilateral 47/47
RJ-04 Proband F Breast Bilateral 33/38
RJ-05 Proband F Breast Unilateral 33
RJ-05A 1^st^ degree relative F Healthy \- \-
SP-01 Proband F Breast Bilateral 36
SP-01A 2^nd^ degree relative F Breast Bilateral 46
SP-01B 3^rd^ degree relative F Healthy \- \-
SP-01C 5^th^ degree relative F Healthy \- 35
SP-02 Proband F Breast Unilateral 63
SP02-A 1^st^ degree relative F Healthy \- \-
SP-03 Proband F Breast Bilateral 37
SP-03A 1^st^ degree relative F Breast Unilateral 41
SP-03B 1^st^ degree relative M Healthy \- \-
SP-04 Proband F Breast Bilateral 31
SP-04A 1^st^ degree relative F Healthy \- \-
SP-05 Proband F Ovarian \- 49
SP-05A 1^st^ degree relative F Healthy \- \-
SP-05B 1^st^ degree relative F Breast Unilateral 47
SP-05C 1^st^ degree relative F Healthy \- \-
SP-05D 3^rd^ degree relative M Healthy \- \-
SP-05E 3^rd^ degree relative F Healthy \- \-
RS-01 Proband F Breast Bilateral 36/-
RS-01A 3^rd^ degree relative F Ovarian Bilateral 52/54
RS-02 Proband F Breast Bilateral 23/44
RS-03 Proband F Breast Bilateral \-
RS-04 Proband F Breast Bilateral 35/45
RS-04A 1^st^ degree relative F Healthy \- \-
RS-05B 3^rd^ degree relative F Breast Bilateral 49/64
Fourteen probands carrying *BRCA1 c.5266dupC* and 26 relatives (carriers and non-carriers) were haplotyped. The same haplotype associated with *c.5266dupC* was segregating within all the families analyzed, revealing no recombinants in a region of 0,68 Mb ([Figure S2](#suppl03){ref-type="supplementary-material"}). On the other hand, this haplotype was not found in non-carrier relatives analyzed (n=7). Ancestry analyses showed that the European component was predominant among the probands, with an average of 81.15% ([Figure 1](#f1){ref-type="fig"}).
{#f1}
The high frequency of breast cancer observed in our sample, especially bilateral cancer (n=13/21), was also reported in other studies and recently associated to its location (within the breast cancer cluster region of the gene) ([@B19]; [@B9]; [@B17]; [@B18]; [@B20]). The segregation of the same haplotype within *BRCA1 c.5266dupC* in all carrier relatives analyzed, reinforced the founder effect of this mutation in the Brazilian population.
Our data is in accordance with previous results showing a European component that exceeds 70% in the South and Southeast of Brazil ([@B10]), and support the European origin of *BRCA1 c.5266dupC* in Brazil. The predominant European ancestry observed in the carriers studied here is in line with the proposition of [@B9] for the Scandinavian origin of this mutation, followed by its dispersion through Central Europe 400-500 years ago. As was stated before ([@B3]), a better explanation for the presence of this mutation in the Brazilian population is the immigration from Central Europe diring the 19^th^ century encouraged by Brazilian officials ([@B14]), particularly to the Southeast and South regions of Brazil ([@B10]), where the mutation is nowadays more frequently found.
There are some limitations to the present study. Unfortunately, the sample size was small, considering that *c.5266dupC* corresponds to 20.2% of all Brazilian *BRCA1* pathogenic variants ([@B13]), although, our samples account for patients of three Brazilian states. Nonetheless, it is the largest published study revealing a single haplotype between carriers in Brazil. This unique haplotype validates the founder effect for the *c.5266dupC* insertion in Brazil, and the ancestry data reveal the contribution of Central Europe for the Brazilian genetic background. The frequency of this mutation is shown to be relevant especially among patients of the southern and utheastern Brazilian regions, where the European ancestry contribution is large. Our study also shows that *c.5266dupC* is associated with the appearance of bilateral breast tumors, which confirms what was previously observed by other authors ([@B4]). Considering a scenario of limited resources, low cost screening focused on this recurrent pathogenic variant could be offered for patients and their families of European ancestry. However, this strategy is not adequate in view of the diversity of pathogenic *BRCA1* and *BRCA2* variants found in Brazil and the admixed ethnic origin of its individuals.
This work was supported by National Institute for Cancer Control (INCT para Controle do Câncer; http://www.inct-cancer-control.com.br), Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq, Brazil; grant numbers: 305873/2014-8 and 573806/2008-0), and Fundação Carlos Chagas Filho de Amparo à Pesquisa do Estado do Rio de Janeiro (FAPERJ, Brazil; grant number: E26/170.026/2008).
The following online material is available for this article:
Table S1- Primers sequences used for genotyping.Click here for additional data file.
Figure S1- Seven molecular markers used in haplotype analysis.Click here for additional data file.
Figure S2- Pedigree of probands' families and haplotype analyses.Click here for additional data file.
[^1]: **Author Contributions:** RG and BLS conceived the study, carried out micro-satellite analysis, and participated in writing the manuscript. RM, CBON, BA and PAP contributed with patient samples, family cancer history, and participated in writing the manuscript. PSF and EIP carried out the ancestry analysis and participated in writing the manuscript. MAMM conceived the study, contributed with patient samples, family cancer history, and participated in writing the final version of the manuscript. All authors read and approved the final version.
[^2]: **Conflict of Interest:** The authors declare no conflict of interest.
[^3]: **Associate Editor:**Houtan Noushmehr
| {
"pile_set_name": "PubMed Central"
} |
An incorrect value for Mosquitoes Natural Mortality rate was given in [Table 1](#tab01){ref-type="table"} of the paper by E. Massad, J. Rocklov and A. Wilder-Smith. \[[@ref1]\] [Table 1](#tab01){ref-type="table"} is republished here with the correct value. Table 1.Model\'s parameters, biological meaning, values and sourcesParameterMeaningValueSource*a*Average Daily Biting rate0·164\[15\]*b*Fraction of actually infective bites0·088Fitted to dataμ~H~Humans Natural Mortality rate3·5 × 10^−5^ days ^−1^\[16\]*r* ~H~Birth rate of humans8 days ^−1^\[16\]κ~H~Humans Carrying Capacity16 × 10^6^\[16\]α~H~Dengue Mortality in Humans10^−3^ days ^−1^\[17\]γ~H~Humans recovery rate0·143 days ^−1^\[17\]*p* ~S~Susceptible eggs hatching rate0·15 days ^−1^\[18\]*d* ~1~Winter modulation parameter0·07assumed*d* ~2~Winter modulation parameter0·06assumedγ~M~Mosquitoes Latency rate0·143 days ^−1^fFrequency of seasonal cycles2·8 × 10^−3^ days ^−1^assumedμ~M~Mosquitoes Natural Mortality rate0·09 days ^−1^\[19\]α~M~Dengue Mortality in MosquitoesNegligible---*r* ~M~Oviposition rate50 days ^−1^\[19\]*p* ~I~Infected eggs hatching rate0·15 days ^−1^\[19\]*g*Proportion of infected eggs0·5assumedκ~E~Eggs Carrying Capacity9·8 × 10^7^assumedμ~E~Eggs Natural Mortality rate0·1 days ^−1^\[19\]*c*A.aegypti susceptibility to dengue0·087Fitted to data
| {
"pile_set_name": "PubMed Central"
} |
Objectives
==========
To compare the accuracy of an integrated fiberoptic monitoring system in measuring blood volume (BV) with standard method using chromium-51-tagged erythrocytes in septic shock.
Design
======
Prospective animal laboratory study.
Measurement and main results
============================
Twenty anaesthetised, and mechanically ventilated pigs (20.9 ± 1.9 kg) were investigated over a period of 8 h. Septic shock was induced with faecal peritonitis (1 g/kg body weight autologous faeces). A central venous catheter was used for the injection of the indicator dyes. BV was measured by detecting indocyanin green by reflection densitrometry using a fiberoptic thermistor tipped catheter inserted into right carotid artery (4F PV 2024, Pulsion Medical Systems). Haemodynamic treatment scheme was aimed at maintenance of a central venous pressure of 12 mmHg. Data were analysed using Bland-Altman analyses, linear regression and correlation. Forty data pairs of simultaneous BV-measurements were yielded during haemodynamic consistency with a mean BV measured by integrated fiberoptic monitoring system of 66.6 ± 20.3 ml/kg (range: 24.5--122.6 ml/kg). Mean BV measured by chromium-51-tagged erythrocytes was 76.1 ± 17.9 ml/kg (range: 49.7--121.6 ml/kg). Linear regression equation was: BV by integrated fiberoptic monitoring system = 0.65 × BV: chromium-51-tagged erythrocytes + 17.6; *r* = 0.57, *P* \< 0.01. The mean bias was 9.6 ml/kg (95% confidence interval: 3.7--15.4 ml/kg), with limits of agreement of --26.5 to 45.6 ml/kg and a precision of 16.8 ml/kg.
Conclusion
==========
In this model of porcine septic shock we could show a significant correlation in blood volume measurement between fiberoptic monitoring system and chromium-51-tagged erythrocytes. The relatively wide limits of agreement might be due to pronounced circulatory alterations including slow mixing compartments, prolonged equilibration and sequestration in septic shock.
| {
"pile_set_name": "PubMed Central"
} |
To the Editor:
{#cpt1196-sec-0001}
A recent drug--drug interaction (DDI) study reported a 3.89--4.67‐fold increase in exposures of dasabuvir by clopidogrel,[1](#cpt1196-bib-0001){ref-type="ref"} and suggested a discrepancy between observed data and previously published physiologically based pharmacokinetic (PBPK) predictions.[2](#cpt1196-bib-0002){ref-type="ref"} We have published prospective predictions of dasabuvir and clopidogrel DDI using PBPK, which suggested a 2.8--4.5‐fold increase at steady state. The PBPK simulation design represented a clinically relevant scenario in which dasabuvir and ritonavir are coadministered with clopidogrel maintenance dosage (75 mg q.d.) at steady state (**Figure** [1](#cpt1196-fig-0001){ref-type="fig"} **a**). The model sensitivity analysis suggested a maximum increase in dasabuvir exposures to 2.6‐fold for peak plasma concentration (*C* ~max~) and 4.5‐fold for area under the curve (Figure [1](#cpt1196-fig-0001){ref-type="fig"}b).[2](#cpt1196-bib-0002){ref-type="ref"}
{ref-type="ref"} vs. clinical study by Itkonen *et al*.[1](#cpt1196-bib-0001){ref-type="ref"} (**a**) PBPK simulation study design of a clinically relevant scenario; all drugs at steady state, interaction assessment with CLOP maintenance dosage of 75 mg q.d., DAS 250 mg b.i.d., and RTV boosting dose of 100 mg q.d., on simulated study D5 of coadministration. (**b**) Results from the previously published PBPK predictions of a base case DDI scenario and sensitivity analysis, including worst case. (**c**) Itkonen *et al*.[1](#cpt1196-bib-0001){ref-type="ref"} clinical study design for maximizing DDI; interaction assessment with CLOP 300‐mg loading dose, higher RTV dosage of 100 mg b.i.d., and DAS 250‐mg single dose on study D1 of coadministration. (**d**) PBPK simulation results replicating Itkonen *et al*.[1](#cpt1196-bib-0001){ref-type="ref"} clinical study design using a previously published PBPK model by Shebley *et al*.[2](#cpt1196-bib-0002){ref-type="ref"} AUCR, ratio of area under the curve with inhibitor relative to without inhibitor; CLOP, clopidogrel; CmaxR, ratio of peak plasma concentration with inhibitor relative to without inhibitor; D, day; DAS, dasabuvir; DDI, drug‐drug interaction; PBPK, physiologically based pharmacokinetic; RTV, ritonavir.](CPT-105-320-g001){#cpt1196-fig-0001}
Itkonen *et al*.[1](#cpt1196-bib-0001){ref-type="ref"} suggested an apparent underprediction of DDI by the PBPK model. However, when differences between the PBPK simulation and clinical study designs were taken into consideration (**Figure** [1](#cpt1196-fig-0001){ref-type="fig"} **a,c**), the clinical data appear to validate our PBPK predictions. PBPK simulations predicted effects of clopidogrel 75‐mg q.d. maintenance dosage on dasabuvir 250 mg b.i.d. with ritonavir 100 mg q.d. at steady state. Itkonen *et al*.[1](#cpt1196-bib-0001){ref-type="ref"} evaluated clopidogrel 300‐mg loading dose on dasabuvir 250‐mg single dose, with and without ritonavir 100 mg b.i.d. With clopidogrel 300 mg, plasma concentrations of the acyl‐β‐glucuronide metabolite are threefold to fourfold higher compared with 75 mg. The impact of clopidogrel higher dose on dasabuvir exposures was replicated by the PBPK model, suggesting that the observed higher interaction relative to that predicted previously could be explained by higher exposures of the acyl‐β‐glucuronide (**Figure** [1](#cpt1196-fig-0001){ref-type="fig"} **d**). The ritonavir regimen was shorter in the clinical study, leading to relatively lower induction of Cytochrome P450 2C8 (CYP2C8), alternatively explaining differences between the results.
When Itkonen *et al*.[1](#cpt1196-bib-0001){ref-type="ref"} compared clinical results with PBPK predictions, the authors failed to highlight this important factor of higher clopidogrel dose (300‐mg loading dose) on observed DDI with dasabuvir compared with that predicted by PBPK at lower dosage (75 mg q.d.). The authors highlighted potential concern of corrected QT interval (QTc) prolongation attributable to DDI; however, they did not acknowledge that the 1.82‐fold increase in dasabuvir *C* ~max~ was with clopidogrel 300 mg, which is used clinically as a loading dose only on day 1 of treatment.[3](#cpt1196-bib-0003){ref-type="ref"} This suggests that clopidogrel maintenance dosage of 75 mg q.d. may cause lower increases in dasabuvir exposures, as previously demonstrated by the PBPK model.[2](#cpt1196-bib-0002){ref-type="ref"}
Maximum effects on dasabuvir exposures at the higher 300‐mg loading dose of clopidogrel were 1.40--1.82‐fold on *C* ~max~ and 3.89--4.67‐fold on area under the curve, a range well captured in the sensitivity analysis of the PBPK model by Shebley *et al*.[2](#cpt1196-bib-0002){ref-type="ref"} Thus, dasabuvir contraindication with strong CYP2C8 inhibitors remains appropriate.[4](#cpt1196-bib-0004){ref-type="ref"} These data demonstrate the powerful impact of PBPK modeling in predicting DDIs, deciphering study design differences, and evaluating complex interactions.
Funding {#cpt1196-sec-0002}
=======
AbbVie, Inc., provided financial support for the studies and participated in the design, study conduct, analysis, and interpretation of data as well as the writing, review, and approval of the publication.
Conflict of Interest {#cpt1196-sec-0003}
====================
M. Shebley is an employee of AbbVie, Inc., and may hold stock or stock options.
| {
"pile_set_name": "PubMed Central"
} |
1. Introduction {#sec1-foods-09-00341}
===============
Food adulteration is an illegal activity of food production, which may threaten food quality and safety. On one hand, the nutritional value of food is limited because of the reduction of effective components in food. On the other hand, the adulterants may have a detrimental effect on human health. Several scandals concerning food adulteration have been reported around the world \[[@B1-foods-09-00341],[@B2-foods-09-00341],[@B3-foods-09-00341]\]. Honey is one of the most commonly adulterated foods because of its economical purpose and wide use. There are two main approaches for honey adulteration. One is to mix pure honey with sugar-based adulterants, and the other is to adulterate high-quality honey with inferior honey. These two cases will be explored in this study.
The adulterant usually has a similar constituent or characteristic with the pure honey, and it is hard to distinguish from the appearance. Several studies concerning honey adulteration detection have been reported. Amiry et al. \[[@B4-foods-09-00341]\] discriminated adulterated honey (mix pure honey with date syrup and invert sugar syrup) with linear discriminant analysis. Different parameters including color indices, rheological, physical, and chemical parameters were used as variables for discrimination. Physical and chemical parameters achieved the best results, with accuracy above 95%. The results highlighted the use of physical and chemical parameters to detect honey adulteration. In addition, Arroyo-Manzanares et al. \[[@B5-foods-09-00341]\] used gas chromatography-ion mobility spectrometry to detect sugar cane or corn syrup adulterated honey; seven out of nine commercial honeys were classified as adulterated samples. Traditionally, the chemical features of honey are detected with wet chemical analysis, which is time and labor consuming. Hence, several rapid analytical methods based on electronic and optical techniques were proposed by other researchers, e.g., electronic nose \[[@B6-foods-09-00341]\], electronic tongue \[[@B6-foods-09-00341]\], fluorescence spectroscopy \[[@B7-foods-09-00341]\], visible-near infrared spectroscopy \[[@B8-foods-09-00341],[@B9-foods-09-00341]\]. The 'fingerprint information' of honey could be rapidly obtained by these sensors, and the adulterated honey could be distinguished with the help of chemometric methods.
For its part, laser-induced breakdown spectroscopy (LIBS), which allows elemental analysis, may be useful for honey authenticity. The elemental information of honey can be obtained through analyzing the atomic emission spectroscopy from plasma which is induced by a laser. It has the advantages of fast detection, multi-elemental analysis, and environmentally friendly feature \[[@B10-foods-09-00341]\]. As a novel approach in food, it has been used for regional discrimination \[[@B11-foods-09-00341]\] and elemental detection \[[@B12-foods-09-00341],[@B13-foods-09-00341],[@B14-foods-09-00341]\]. Because LIBS spectrum often contains numerous variables, chemometric methods are usually used to figure out the useful information and establish models for food adulteration detection. Recently, LIBS was used to classify the botanical origins of honey, and detect rice syrup adulterated samples \[[@B15-foods-09-00341]\]. However, the adulterant content in honey should be further quantified. Herein, LIBS combined with partial least squares regression was used as an analytical tool for fast quantification of honey adulterant content.
In this study, acacia honey mixed with high fructose corn syrup (HFCS) and rape honey were analyzed by LIBS. The specific objectives were to: (1) analyze the LIBS spectral features of pure honey and adulterants; (2) determine the feature variables that are related to adulteration; (3) quantify the adulterant content with univariate and multivariate analysis.
2. Materials and Methods {#sec2-foods-09-00341}
========================
2.1. Sample Preparation {#sec2dot1-foods-09-00341}
-----------------------
Honey including acacia honey (Guanshengyuan Co., Ltd, Shanghai, China) and rape honey (Yaoquan Food Co., Ltd, Yunnan, China) were collected from main producers in China, and two kinds of HFCS with different fructose concentrations (F55 and F90) were purchased from markets. HFCS F55 contains 55% fructose, and HFCS F90 contains 90% fructose. In this case, acacia honey was considered as pure honey, and HFCS (F55 and F90) and rape honey were used as adulterants.
Honey adulteration was prepared by mixing the acacia honey with HFCS F55, HFCS F90, and rape honey. To establish models for quantifying adulterant content, acacia honey was adulterated with HFCS and rape honey at 21 different percentages (0%, 5%, 10%, 15%, 20%, 25%, 30%, 35,% 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 95%, and 100%). In addition, adulterated samples for external prediction were prepared at 13 different adulteration rates, i.e., 0%, 8%, 16%, 24%, 32%, 40%, 48%, 56%, 64%, 72%, 80%, 88%, and 96%. The adulteration rates of 0% and 100% indicated pure acacia honey and pure adulterant, respectively. All sample adulteration was performed in three replications, so there were 63 samples for calibration, and 39 samples for prediction. After mixing, all samples were kept in a water bath at 37 °C for 12 h to ensure homogeneity.
2.2. LIBS Measurement {#sec2dot2-foods-09-00341}
---------------------
A laboratory-assembled LIBS device was used for honey adulteration detection. The detailed description of the device was introduced in our previous published article \[[@B16-foods-09-00341]\]. First, 8 g of sample was added in 12-well plates and placed in a X-Y-Z moving stage. A pulse laser (Vlite 200, Beamtech, Beijing, China) operated at 532 nm was used to ablate the sample with energy of 80 mJ. Then, emission light from induced plasma was transferred into an Echelle spectrograph (ME 5000, Andor, Belfast, UK), and detected by an intensified charge coupled device (ICCD, DH334T-18F-03, Andor, Belfast, UK). To improve the signal-to-background ratio, the delay time, integral time, and relative gain of ICCD camera were set at 2 µs, 10 µs, and 26. Single shot scanning was performed in an ablation region of 10 mm × 10 mm with resolution of 1 mm. Hence, 100 successive spectra were collected for each sample, the spectra were averaged to minimize the sample inhomogeneity. Because of the advantages of LIBS, no sample preparation was needed, and the total detection time for one sample was less than two minutes.
2.3. Data Analysis {#sec2dot3-foods-09-00341}
------------------
Because the peak in LIBS spectrum corresponds to the emission from a certain element or molecule band, the observed peak intensity was used as the variable for analysis. To establish a model for quantifying adulterant content, PLSR was used. In addition, several feature selection methods based on PLSR were used to determine the key LIBS emissions that related to the adulterant content.
PLSR is a commonly and widely used multivariate method for quantitative analysis. It projects the raw variables into new dimensions with the maximal variation, and regresses the first few new variables (latent variable, LV) with respond value \[[@B17-foods-09-00341]\]. In this case, the raw variables were peak intensities of main emissions, and the respond value was the adulterant content in honey. Before modeling, the auto scale preprocessing method, which used mean-centering followed by dividing each variable by the standard variation of the variable, was used to correct the scaling of each variable. Ten-folds random cross-validation was used to determine the number of LV, and prevent the overfitting. In addition, the straightforward implementation of a statistically inspired modification of the PLS (SIMPLS) algorithm was used to calculate the PLS model parameters \[[@B18-foods-09-00341]\].
Three feature selection methods including genetic algorithm (GA), variable importance in projection (VIP), and the selectivity ratio (SR) were used in this case. GA is a subset search algorithm that was inspired by biological evolution theory and natural selection \[[@B19-foods-09-00341]\]. The subset of relevant variables selected by GA is then fitted with PLSR to evaluate the performance, and determine the feature variables. Different from GA, the variable selection based on VIP and SR is carried out by using a threshold of some parameters from the PLSR model. VIP calculates the accumulation of PLS weights, and SR defines the ratio between explained variance and the unexplained variance in the PLS model. The larger values of VIP and SR, the greater contribution of the variable. For the criteria of variable selection, VIP follows the rule of 'greater than one rule', and SR follows the F-test (95%) criterion \[[@B20-foods-09-00341]\]. In this case, the variables with VIP value greater than 1 and SR value greater than 1.532 were selected as important variables.
After modeling, some measures should be used to evaluate the performance. In this case, model performance was evaluated with correlation coefficient (*r*) and root-mean-square error (RMSE). The *r* value measures the relationship between predicted adulterant content and actual value, and the RMSE value measures the predictive error. The larger the *r* value and the smaller the RMSE value, the better the model performance. All data analyses were carried out in the MATLAB (v2019b, The MathWorks Inc., Natick, MA, USA).
3. Results and Discussion {#sec3-foods-09-00341}
=========================
3.1. LIBS Spectral Characteristics {#sec3dot1-foods-09-00341}
----------------------------------
Before quantification, LIBS spectral characteristics of acacia honey, rape honey, HFCS F55, and HFCS F90 were first analyzed ([Figure 1](#foods-09-00341-f001){ref-type="fig"}). All the LIBS spectra ranged from 240 to 860 nm. In general, the average LIBS spectra for different samples were similar except some emissions in certain spectral range. It was credited to the similar constituent of honey and HFCS. In general, honey contains 75% saccharides (mainly glucose and fructose), 15% water, amino acids, and minerals, etc. HFCS mainly contains glucose and fructose. According to the concentration of fructose, the HFCS can be divided into three categories: F42 (42% fructose), F55 (55% fructose), and F90 (90% fructose). Hence, the main components ablated by laser in both honey and HFCS were glucose and fructose. As shown in [Figure 1](#foods-09-00341-f001){ref-type="fig"}, the emissions from C, H, O, and N were observed in all samples. The molecular band CN that usually appears in an organic sample when analyzed in air atmosphere was also found in this case.
Some differences in elemental emissions could be observed between honey and HFCS. It was obvious that emissions from Mg, Ca, and K appeared in the spectra of honey, while it cannot be found in the spectra of HFCS. It indicated that the concentrations of Mg, Ca, and K in honey were significantly higher than those in HFCS. In addition, there was no obvious difference between acacia honey and rape honey, except relatively stronger emission of Na in acacia honey. These elemental differences might be used to differentiate the adulterants. However, it was hard to quantify the adulterant content simply by analyzing spectrum. Hence, some modeling methods were further used to quantify the adulterant content.
3.2. Univariate Analysis {#sec3dot2-foods-09-00341}
------------------------
Univariate analysis was used to explore the relationship between adulterant content and single variable and quantify the adulteration. In this case, the peak intensities of main emissions from samples were used for analysis. Univariate analysis was performed by regressing the peak intensity of each emission with the adulterant content, and *r* and RMSE were used to evaluate the results. The corresponding element for each emission could be identified with the National Institute of Standard and Technology (NIST, Gaithersburg, Maryland, USA) database \[[@B21-foods-09-00341]\]. [Table 1](#foods-09-00341-t001){ref-type="table"} shows the results of univariate analysis between main emission lines and adulterant content. Forty-three univariate models were established. The variables contained emissions from C, Si, Mg, Ca, Na, K, N, H, O, and CN. Four variables with emissions of 748.47, 794.83, 795.17, and 822.43 nm were marked with unknown, because they could not be identified with the NIST database or references.
In general, the models for quantifying adulterant content of HFCS F90 had the best results with higher *r* and lower RMSE. It indicated that high concentration of fructose in HFCS led to greater spectral difference and contributed to the univariate analysis. In addition, for HFCS F90 and HFCS F55, the emissions from Mg II 279.58, 280.30 nm, Mg I 285.25 nm, Ca II 393.37, 396.89 nm, Ca I 422.70 nm, Na I 589.03, 589.63 nm, and K I 766.57, 769.97 nm had compact relationship with the adulterant content, with *r* \> 0.9 and RMSE \< 11.0%. For rape honey, models based on emissions from Na I 589.03 and 589.63 nm had good results, with *r* of 0.919 and 0.903, and RMSE of 12.0% and 13.0%. It indicated that emissions from mineral elements played an important role in adulteration quantification. It also verified the LIBS spectral difference between acacia honey and adulterants.
3.3. Quantification of Adulterant Content Based on Multivariate Analysis {#sec3dot3-foods-09-00341}
------------------------------------------------------------------------
Multivariate analysis was further used to quantify the adulterant content. First, all variables in univariate analysis were used as the inputs of PLS models. As seen in [Table 2](#foods-09-00341-t002){ref-type="table"}, PLS models based on all variables achieved good results for all three types of adulteration. The *r* values for HFCS F55, HFCS F90, rape honey in the prediction set were 0.962, 0.980, 0.988, and the RSME values were 15.6%, 16.6%, 4.7%, respectively. The latent variables for these three models were 4, 4, 5, which were determined by cross validation. The results of PLS models were better than those of univariate analysis. It also verified the advantages of multivariate analysis. The combination of information from multiple emissions contributed to the adulterant content quantification.
In addition, results of PLS models based on feature variables (selected by GA, VIP, and SR) are also shown in [Table 2](#foods-09-00341-t002){ref-type="table"}. In general, prediction results after feature selection were similar or better than those based all variables. The irrelevant variables in models might worsen the modeling performance \[[@B22-foods-09-00341],[@B23-foods-09-00341]\], which also verified the necessity of feature selection. Only one exception happened for the GA-PLS model in HFCS F55 quantification. The RMSE value in prediction set was 0.320, which is greatly worse than that without feature selection (0.156). It might be credited to the selected variables by the GA method. As shown in [Figure 2](#foods-09-00341-f002){ref-type="fig"}, lots of irrelevant variables were selected. The GA method might not be suitable for feature selection in the honey adulteration with HFCS F55. With the consideration of variable number and prediction performance, the models marked with bold achieved the best results. The RMSE value for HFCS 55, HFCS F90, and rape honey in the prediction set were 8.9%, 8.2%, and 4.8%, respectively. In addition, similar results were achieved in 10-folds cross-validation, and RMSE value for HFCS 55, HFCS F90, and rape honey were 8.5%, 6.5%, and 4.6%, respectively.
We also compared the variables selected with GA, VIP, and SR methods ([Figure 2](#foods-09-00341-f002){ref-type="fig"}). Row 1, 5, 9 showed the correlation coefficient between each variable and adulterant content of HFCS F55, HFCS F90, and rape honey, respectively. The values of correlation coefficient were in the range of 0 to 1. Other rows represented the variables selected by GA, VIP, and SR methods. Selected variables were represented in blue, and non-selected variables were in white. As shown in [Figure 2](#foods-09-00341-f002){ref-type="fig"}, VIP and SR methods chose the variables with a high correlation coefficient, while some variables with a low correlation coefficient were selected by the GA method. It was related to the principal of feature selection methods. For the GA method, the variables were randomly combined and verified by PLSR. The variables were selected based on the results of PLSR modeling. For VIP and SR methods, the contribution of each variable was considered in the selection \[[@B20-foods-09-00341]\]. The variables selected by the GA method might be easily affected when testing with external samples. In addition, VIP and SR methods had some common variables, while the number of selected variables was different. It might be credited to the different threshold measure of each method. Hence, VIP and SR methods might be recommended for feature selection in quantification of honey adulterant content.
The scatter plot of the best model for quantifying adulteration ratio of HFCS 55, HFCS 90, and rape honey is shown in [Figure 3](#foods-09-00341-f003){ref-type="fig"}. Among these three models, the quantification for rape honey achieved the best result, with *r* and RMSE of 0.988 and 4.8% in the prediction set. The samples in calibration and prediction sets distributed closely around the regression lines, and the regression lines almost went through original point. The emissions from Mg II 279.58, 280.30 nm, Mg I 285.25 nm, Ca II 393.37, 396.89 nm, Ca I 422.70 nm, Na I 589.03, 589.64 nm, and K I 766.57, 769.97 nm, which were the feature variables in the rape honey quantification, were also included in the other two models. It indicated that these variables might play an important role in honey adulteration analysis.
4. Conclusions {#sec4-foods-09-00341}
==============
In this study, LIBS combined with chemometric methods was used to detect honey adulteration. The adulterant content of acacia honey (adulterated with HFCS 55, HFCS 90, and rape honey) was successfully quantified. SR and VIP methods detected effectively the most relevant variables for adulteration determination. The emissions from Mg II 279.58, 280.30 nm, Mg I 285.25 nm, Ca II 393.37, 396.89 nm, Ca I 422.70 nm, Na I 589.03, 589.64 nm, and K I 766.57, 769.97 nm were considered as feature variables and played an important role in modeling. The importance of these variables was also verified in univariate analysis. The SR-PLSR, VIP-PLSR, and VIP-PLSR achieved the best results for detecting an adulteration ratio of HFCS F55, HFCS 90, and rape honey, with RMSE of 8.9%, 8.2%, and 4.8%, respectively. The results indicated the promising possibility of using LIBS and chemometric methods for quantification in honey adulteration. In addition, some research concerning model transfer could be explored, and more types of acacia honey as well as adulterants could be included in modeling in further study, which might be helpful for practical application.
Conceptualization F.Z. and F.L.; Data curation, J.P.; Formal analysis, J.P.; Funding acquisition, J.P. and Z.Z.; Investigation, W.X.; Methodology, J.P.; Project administration, F.Z. and F.L.; Resources, J.J. and Z.Z.; Software, W.X.; Validation, W.X., F.Z. and F.L.; Writing -- original draft, J.P.; Writing -- review & editing, J.P., W.X., J.J., Z.Z., F.Z. and F.L. All authors have read and agreed to the published version of the manuscript.
This research was funded by National Key R&D Program of China, grant number 2018YFD0700502, Zhejiang Provincial Key Research and Development Program, grant number 2017C02027, And China Postdoctoral Science Foundation, grant number 2019M652143.
The authors declare no conflict of interest.
{#foods-09-00341-f001}
{#foods-09-00341-f002}
{#foods-09-00341-f003}
foods-09-00341-t001_Table 1
######
Results of univariate analysis based on peak intensities of main emissions.
----------------------------------------------------------------------------------------------
No. Observed\ Element HFCS F55 HFCS F90 Rape Honey
Wavelength (nm)
----- ----------------- ----------- ---------- ---------- ------------ ------- ------- -------
1 247.88 C I 0.493 26.4% 0.823 17.2% 0.066 30.2%
2 250.72 Si I 0.176 29.8% 0.204 29.6% 0.154 29.9%
3 251.45 Si I 0.180 29.8% 0.214 29.6% 0.153 29.9%
4 251.64 Si I 0.204 29.7% 0.222 29.5% 0.166 29.8%
5 251.94 Si I 0.193 29.7% 0.205 29.6% 0.159 29.9%
6 252.44 Si I 0.195 29.7% 0.210 29.6% 0.149 29.9%
7 252.88 Si I 0.200 29.7% 0.210 29.6% 0.158 29.9%
8 279.58 Mg II 0.932 10.9% 0.936 10.6% 0.516 25.9%
9 280.30 Mg II 0.922 11.7% 0.934 10.8% 0.441 27.2%
10 285.25 Mg I 0.959 8.6% 0.959 8.6% 0.517 25.9%
11 288.20 Si I 0.194 29.7% 0.227 29.5% 0.161 29.9%
12 385.07 CN 4-4 0.550 25.3% 0.828 17.0% 0.487 26.4%
13 385.49 CN 3-3 0.576 24.8% 0.820 17.3% 0.454 27.0%
14 386.17 CN 2-2 0.596 24.3% 0.821 17.3% 0.442 27.2%
15 387.13 CN 1-1 0.473 26.7% 0.828 17.0% 0.460 26.9%
16 388.33 CN 0-0 0.514 26.0% 0.824 17.2% 0.466 26.8%
17 393.37 Ca II 0.957 8.8% 0.948 9.6% 0.694 21.8%
18 396.89 Ca II 0.959 8.6% 0.951 9.3% 0.652 22.9%
19 422.70 Ca I 0.953 9.2% 0.942 10.1% 0.707 21.4%
20 589.03 Na I 0.937 10.6% 0.973 6.9% 0.919 12.0%
21 589.64 Na I 0.936 10.6% 0.975 6.8% 0.903 13.0%
22 656.33 H$\alpha$ 0.617 23.8% 0.538 25.5% 0.243 29.4%
23 715.77 O I 0.316 28.7% 0.766 19.5% 0.227 29.5%
24 742.45 N I 0.220 29.5% 0.739 20.4% 0.268 29.2%
25 744.30 N I 0.197 29.7% 0.738 20.4% 0.278 29.1%
26 746.92 N I 0.162 29.9% 0.742 20.3% 0.248 29.3%
27 748.47 Unknown 0.507 26.1% 0.632 23.4% 0.220 29.5%
28 766.57 K I 0.943 10.1% 0.960 8.4% 0.756 19.8%
29 769.97 K I 0.931 11.1% 0.959 8.6% 0.750 20.0%
30 777.47 O I 0.183 29.8% 0.760 19.7% 0.215 29.6%
31 794.83 Unknown 0.316 28.7% 0.758 19.7% 0.215 29.6%
32 795.17 Unknown 0.299 28.9% 0.773 19.2% 0.206 29.6%
33 818.57 N I 0.170 29.9% 0.740 20.4% 0.256 29.3%
34 818.86 N I 0.217 29.6% 0.736 20.5% 0.265 29.2%
35 820.10 N I 0.232 29.5% 0.746 20.2% 0.237 29.4%
36 821.14 N I 0.221 29.5% 0.744 20.2% 0.247 29.3%
37 821.68 N I 0.244 29.4% 0.725 20.8% 0.250 29.3%
38 822.28 N I 0.067 30.2% 0.782 18.9% 0.305 28.8%
39 822.43 Unknown 0.290 29.0% 0.706 21.4% 0.303 28.8%
40 824.32 N I 0.291 29.0% 0.719 21.0% 0.285 29.0%
41 844.73 O I 0.252 29.3% 0.743 20.3% 0.260 29.2%
42 856.86 N I 0.325 28.7% 0.729 20.7% 0.275 29.1%
43 859.49 N I 0.357 28.3% 0.706 21.5% 0.316 28.7%
----------------------------------------------------------------------------------------------
Note: The shade color of the table represents the performance of univariate analysis. The shade color of being green indicates the best compact relationship (*r* = ±1) and the lowest predictive error (RMSE = 0).
foods-09-00341-t002_Table 2
######
Multivariate analysis results based on partial least square regression (PLSR) and feature selection methods.
-----------------------------------------------------------------------------------------------------------------------
Adulterant Method No. of LV No. of Var. Calibration C.V. Prediction
-------------- -------- ----------- ------------- ------------- ----------- ------------ ----------- ---------- -------
HFCS\ PLSR 4 43 0.977 6.5% 0.965 8.0% 0.962 15.6%
F55
GA-PLSR 4 12 0.983 5.6% 0.978 6.4% 0.794 32.0%
VIP-PLSR 5 16 0.982 5.7% 0.966 8.1% 0.938 18.6%
**SR-PLSR** **1** **11** **0.965** **7.9%** **0.960** **8.5%** **0.966** **8.9%**
HFCS\ PLSR 4 43 0.973 7.0% 0.964 8.2% 0.980 16.6%
F90
GA-PLSR 5 19 0.979 6.1% 0.972 7.3% 0.985 11.3%
**VIP-PLSR** **5** **15** **0.982** **5.7%** **0.977** **6.5%** **0.980** **8.2%**
SR-PLSR 5 20 0.981 5.9% 0.973 7.0% 0.982 9.4%
Rape honey PLSR 5 43 0.993 3.6% 0.990 4.3% 0.988 4.7%
GA-PLSR 4 21 0.994 3.3% 0.990 4.4% 0.988 4.7%
**VIP-PLSR** **3** **10** **0.991** **4.1%** **0.989** **4.6%** **0.988** **4.8%**
SR-PLSR 1 2 0.912 12.4% 0.874 15.0% 0.943 11.3%
-----------------------------------------------------------------------------------------------------------------------
Note: No. of LV: number of latent variables; No. of var.: number of variables; C.V.: cross-validation; *r*: correlation coefficient; RMSE: root-mean-square error; GA: genetic algorithm; VIP: variable importance in projection; SR: selectivity ratio.
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#S0001}
============
Non-small-cell lung cancer (NSCLC) including adenocarcinomas and squamous cell carcinomas, ranks among the top 10 list with high prevalence and mortality worldwide. NSCLC was a highly malignant tumor with severe poor prognosis, distant invasion and migration.[@CIT0001] Although the effect of invasion and migration has been increasingly concerned, it was still the overwhelming cause of mortality in patients with NSCLC.[@CIT0002] A comprehensive understanding of the mechanisms and relative signal pathways activated in NSCLC cells was thus essential for the development of novel anti-cancer therapeutic strategy on tumor-associated treatment.
Non-coding RNA (ncRNA), which does not encode any proteins, is a group of RNAs including rRNA, tRNA, snRNA, microRNA and many other RNAs with extensively known functions and long non-coding RNAs (lncRNAs) with unexplored functions.[@CIT0003] It was once thought that lncRNAs were merely a matter of transcriptional noise, and did not involve in the cellular life circles.[@CIT0004] However, accumulating reports showed that lncRNAs participated in various cellular biological processes and might contribute in tumor development through promoting or suppressing cell proliferation, invasion, metastasis, apoptosis, and differentiation.[@CIT0005]--[@CIT0009] Owing to their impact on certain cancers, lncRNAs can be divided into two categories: oncogenes and tumor suppressor. Overexpression of those oncogenic lncRNAs in normal epithelial cells might result in carcinogenesis,[@CIT0010],[@CIT0011] in contrast, down-regulation of tumor suppressor lncRNAs led to tumor formation.[@CIT0012],[@CIT0013] It is apparent that the tissue-specific expression phenomenon of *BRAF-activated non-coding* RNA (BANCR) existed in various human cancers. Down-regulation of BANCR was associated with the progression of cancers and capable of modifying the tumor volume and weight, clinical stage and TNM stage of cancer patients. Notably, BANCR was responsible for viability, proliferation, migration, invasion, and apoptosis of cancer cells.
The aim of the study was to probe the effect of *BANCR* on NSCLC development in vitro and in vivo. Furthermore, we displayed that BANCR expression change in different NSCLC cells presented an influence on their viability, metastasis, and apoptosis. An alteration of E-cadherin, N-cadherin, Vimentin, Bcl-2 and BAX at protein and RNA level, confirmed the function of BANCR on metastasis and apoptosis of NSCLC cells. Our study expanded the understanding of the role of BANCR as a NSCLC suppressor and might facilitate the development of lncRNA-targeted cancer diagnostics and therapeutics.
Materials and method {#S0002}
====================
Patients {#S0002-S2001}
--------
Twenty-seven cases of NSCLC patients in this study ranged from 25- to 65-year-old, with a mean of 53-year-old. These patients were diagnosed with NSCLC (stages I, 8 cases; II, 16 cases; and III, 3 cases.) based on histopathological evaluation. Clinicopathological characteristics, including TNM staging, were recorded. No local or systemic treatment was conducted in these patients before surgery. All collected tissue samples were immediately snap-frozen in liquid nitrogen and stored at --80°C until required. The corresponding NSCLC paraneoplastic tissues were acquired at least 1.00 cm^3^ from the neoplastic tissue. All cases were initial pneumonectomies and randomly chosen from the pneumonectomies performed over a 1--2 years duration in the Hongqi Hospital, Mudanjiang Medical college, P.R. China between April 2012 and April 2016. The experimental protocol was approved by the local Ethical Committee of the Hongqi Hospital, Mudanjiang Medical college, P.R. China. Patients tissue samples were conducted in accordance with Declaration of Helsinki. Written informed consent for sample usage in this research was obtained from both the patients and clinicians. All samples were reviewed and diagnosed by two pathologists independently.
Cells {#S0002-S2002}
-----
Six NSCLC adenocarcinoma cell lines (A549, SPC-A1, H1299, H1650, H1975, and PC-9) normal human pneumonocytes 16HBE were purchased from the Institute of Biochemistry and Cell Biology, Chinese Academy of Sciences. The cells were cultured in RPMI-1640 medium or DMEM supplemented with 10% FBS, 100 U/mL penicillin and 100 mg/mL streptomycin at 37 ºC and 5% CO~2~.
Construction of expression vector for BANCR {#S0002-S2003}
-------------------------------------------
The *BANCR* sequence was constructed into the pcDNA3.0 vector. Ectopic expression of BANCR was achieved through pcDNA3.0-BANCR transfection using lipofectamine-2000, with an empty pCDNA3.0 vector was served as a negative control (NC). The expression levels of BANCR were measured by real-time quantitative PCR.
*In vivo* tumorigenesis in NSCLC mouse model {#S0002-S2004}
--------------------------------------------
Forty male BALB/c nude mice (20--22 g) were purchased from Animal Center of the Chinese Academy of Science. For mouse model establishment and tumor growth assay, a total number of 2 × 10^6^ SPC-A1 cells transfected with empty vector pcDNA3.0 or plasmid pcDNA3.0-BANCR were firstly resuspended in PBS, and then subcutaneously injected into the right flank of nude mice (n=6 per group). Tumor length and width were measured every 3 days after injection. Four weeks later, tumor volume was calculated as length × (width^2^/2). Mice were sacrificed at 28 days after the injection, and the tumors were weighed.
The animal experiment was performed in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health and approved by the Scientific Ethics Committee of Mudanjiang Medical college (permit number: MDJMC20160708005).
Immunohistochemistry staining (IHC) {#S0002-S2005}
-----------------------------------
Tumor tissues were embedded in paraffin and subjected to IHC staining. Tissues were deparaffinized and hydrated and then permeabilized in 0.5% Triton X-100 in PBS for 10 mins. IHC staining for BRAF-antibody (1:250) was performed using the indirect avidin biotin-enhanced horseradish peroxidase method according to the manufacturer's instructions (Vector Laboratories, Burlingame, CA, USA). After developing, all sections were observed by microscope (20x) and analyzed using the Image-Pro Premier software offline (v.9.0) program (Media Cybernetics).
CCK-8 assay {#S0002-S2006}
-----------
Cell viability of NSCLC cell lines was determined by Cell Counting Kit CCK-8/WST-8 assay, which was performed following standard procedure in a 96-well plate as manufacturer's instructions. In brief, cells with a quantity of 5 × 10^3^ cells/well were seeded in a 96-well plate and grown to 80% confluence. Then, either BANCR or its empty controls were transfected into NSCLC cell lines. At 0, 24, 48 and 72 hrs post-transfection, CCK-8 reagent was added into each well. After 1 hr of incubation, cellular viability was determined by measuring the absorbance of the converted dye at 450 nm.
BrdU immunofluorescence assay {#S0002-S2007}
----------------------------
A549, SPC-A1,H1299, H1650, H1975, and PC-9 cells were seeded on cover glasses placed in a 6-well plate. After transfection with BANCR vector or empty control for 48 hrs, BrdU stock solution at 10 mg/mL in PBS was diluted 1000× in the culture medium and incubated for 60 mins. After washing with PBS, cells were fixed in 4% paraformaldehyde for 20 mins and permeabilized with 0.3% Triton X-100 for 10 mins. After blocking with 10% FBS for 1 hr, cells were incubated with a primary rabbit antibody against BrdU (1:200) overnight at 4°C, and then incubated with the corresponding secondary antibody coupled to a fluorescent marker, Cy3, at room temperature for 2 hr. After DAPI stain and PBS washing, the coverslips were mounted on to glass slides and visualized using a fluorescence microscope (Olympus 600 auto-biochemical analyzer; Olympus Corporation, Tokyo, Japan) with Image-Pro Plus software for image analysis, and 10 microscopic fields were taken for BrdU calculating.
Transwell migration assays {#S0002-S2008}
--------------------------
After 24 hrs of transfection, NSCLC cells were harvested by trypsinization and washed once with D-Hanks solution. To measure cell invasion, 8-μm pore size culture or Matrigel inserts were placed into the wells of 24-well plates. In the lower chamber, 400 μL F-12 containing 10% FBS and 20 ng/mL of HGF was added. Then, 1 × 10^5^ cells were added to the upper chamber. After 20 hrs of incubation, the cells that migrated through the pores were stained with crystal violet and observed under the microscope.
Hoechst 33342 staining {#S0002-S2009}
----------------------
The various NSCLC cells with a density of 5 × 10^5^ cells/mL were plated in 12-well plates prior to transfection. After 48 hrs of transfection, the plates were washed three times with PBS, then 500 µL Hoechst 33342 solutions was added to each well followed by incubation for 30 mins at 37°C in the dark. Nuclear DNA staining was observed using a fluorescence microscope (Olympus). A total number of 200 cells were randomly counted from 10 fields and the fluorescence staining percentage of positive cells was expressed as the ratio of apoptotic cells with respect to the total amount of counted cells.
Annexin V-FITC/PI analysis {#S0002-S2010}
--------------------------
The A549, SPC-A1, H1299, H1650, H1975, and PC-9 cells were harvested at 48 hrs post transfection. Annexin V-FITC/PI apoptosis detection kit was utilized to detect cell apoptosis according to the manufacturer's instructions, and the percentage of apoptotic cells was calculated using a Beckman Coulter FACS flow cytometer (Beckman Coulter).
Caspase-3/7 activity detection {#S0002-S2011}
------------------------------
Caspase-3/7 activity was measured using a synthetic rhodamine labeled caspase-3/7 substrate performed immediately after the detection of viability in the same wells.
Western blot {#S0002-S2012}
------------
Cells were lysed in RIPA buffer (50 mM Tris-HCl, 150 mM NaCl, 1% NP-40, 0.1% SDS, pH 8.0) supplemented with a protease inhibitor cocktail. The protein concentration was measured using a BCA Protein Quantitation Kit. Protein samples were separated by 10% SDS-PAGE and electroblotted onto 0.45 µm Immobilon polyvinylidene difluoride membranes. The membranes were blocked with 5% BSA in PBST for 1 hr at room temperature and incubated overnight at 4°C with the respective antibodies: rabbit anti-BRAF (1:2,500), rabbit anti-Bcl-2 (1:2,000), rabbit anti-Bax (1:1,000), or mouse anti-GAPDH antibody (1:5,000). Membranes were incubated for 1 hr at room temperature with Amersham ECL peroxidase-linked secondary antibodies: goat anti-mouse IgG (1:10,000) or goat anti-rabbit IgG (1:10,000). WB immune-reactivity was detected using a Super Signal West Femto Maximum Sensitivity Substrate Kit (Thermo) with a C-DiGit Blot Scanner.
RNA extraction and real-time PCR {#S0002-S2013}
--------------------------------
Total RNA was extracted from NSCLC cells after different treatments or lung tissues using Trizol reagent according to manufacturer's instruction. The levels of BANCR were determined using SYBR Green incorporation on Roche Light-Cycler 480 Real-Time PCR system (Roche, Germany), with GAPDH as an internal control for BANCR, Bcl-2, and Bax. The sequences of primers used were listed as follows:
BANCR F: 5ʹ-ACAGGACTCCATGGCAAACG-3ʹ,
BANCR R: 5ʹ-ATGAAGAAAGCCTGGTGCAGT-3ʹ;
Bax F: 5ʹ-CCCGAGAGGTCTTTTTCCGAG-3ʹ,
Bax R: 5ʹ-CCAGCCCATGATGGTTCTGAT-3ʹ;
Bcl-2F: 5ʹ-CTTTGAGTTCGGTGGGGTCA-3ʹ,
Bcl-2R: 5ʹ-GGGCCGTACAGTTCCACAAA-3ʹ;
GAPDHF: 5ʹ-GGAAAGCTGTGGCGTGAT-3ʹ,
GAPDHR: 5ʹ-AAGGTGGAAG AATGGGAGTT-3ʹ.
Quantitative real-time PCR was performed in 20 μL volumes with SYBR Green PCR Master Mix at 95°C for 10 mins and 40 cycles at 95°C for 15 s, 60°C for 30 s and 72°C for 30 s, using Light Cycler 480. The amount of target (2^−ΔΔ^CT) was obtained by normalizing to endogenous reference and relative to a calibrator (average of the control samples).
Statistical analysis {#S0002-S2014}
--------------------
Results are presented as mean ± SD. Comparisons between groups were made by one-way ANOVA or two-tailed Student's *t* test. Differences were considered statistically significant at *P*\<0.05.
Result {#S0003}
======
BANCR expression in human and mouse NSCLC cells {#S0003-S2001}
-----------------------------------------------
To identify whether *BANCR* are critical regulators for NSCLC development, *BANCR* expression was examined in six different NSCLC cell lines (A549, H1975, H1299, H1650, SPC-A1, and PC-9), NSCLC mouse model and human NSCLC lung tissues. Real-time PCR and western blot (WB) for *BANCR* expression determination revealed that *BANCR* was significantly reduced in these six NSCLC cells ([Figure 1A](#F0001){ref-type="fig"}). To confirm the findings of these NSCLC cells, expression of *BANCR* was measured in the pneumonocytes of normal mouse and lung cancer cells from NSCLC mouse. We found that BANCR was also nearly undetectable in NSCLC tumors of mouse model, compared to these normal pneumonocytes in NC group ([Figure 1B](#F0001){ref-type="fig"}). In addition, the expression levels of *BANCR* were also evaluated from 30 human NSCLC lung tumor tissues and 15 adjacent normal tissues, *BANCR* expression deficiency was observed in human lung tissue during the study ([Figure 1C](#F0001){ref-type="fig"}). Generally, our data suggested that NSCLC cells or tissues displayed a decreased expressing of *BANCR*, comparing with normal cells or lung tissues.Figure 1*BANCR* expression level in NSCLC cells, NSCLC mouse, and human NSCLC lung tissue. (**A**) Decreased expression of *BANCR* in human NSCLC cell lines was examined by real-time PCR (upper panel) and WB for detecting the amount of BRAF protein expression (lower panel), compared to normal human pneumonocytes. (**B**) Reduced expression level of *BANCR* in human NSCLC tumors (n=30) versus normal lung tissue (n=15). IHC was performed to identify the expression of BRAF protein in sections of both tumor and healthy tissue (right panel). (**C**) Level of *BANCR* in lungs of NSCLC mouse models was reduced, compared to normal mice via IHC method. IHC was performed to identify the expression of BRAF protein in sections of both tumor and healthy tissue (right panel). Data represent mean ± SD. \**P*\<0.05, \*\**P*\<0.01.**Abbreviations:** *BANCR*, *BRAF activated non-coding RNA*; IHC, immunohistochemistry; NSCLC, non-small-cell lung cancer; WB, western blot.
Effect of BANCR overexpression on NSCLC cell viability {#S0003-S2002}
------------------------------------------------------
BANCR and NC were transfected into A549, H1975, H1299, H1650, SPC-A1, and PC-9 cells, respectively, to assess the capacity of BANCR on the growth and viability of NSCLC cells. The expression of *BANCR* was confirmed up-regulated in the above six cells after transfection with BANCR in comparison to cells transfected with NC by real-time PCR ([Figure 2A](#F0002){ref-type="fig"}). CCK-8 assay was performed to assess the proliferation of six cell lines. Compared with the NC group, BANCR caused a decreased cell growth rate with a concentration of 100 nM in the cells at 0, 24, 48, and 72 hrs after transfection ([Figure 2B](#F0002){ref-type="fig"}). BrdU immunofluorescence assay was also performed to measure the cell viability at 48 hrs after transfection. In contrast, BANCR transfection inhibited the NSCLC cells DNA synthesis by approximately 30%, 50%, 40%, 50%, 40%, and 30%, respectively, in comparison to NC group ([Figure 2C](#F0002){ref-type="fig"}). Totally, the *BANCR* overexpression reflected an inhibitory effect in a time-dependent manner in these cells.Figure 2Effect of *BANCR* overexpression on NSCLC cell viability. (**A**) The expression of *BANCR* in each NSCLC cells transfected with *BANCR* expressing vector and empty plasmid was measured by real-time PCR. (**B**) The cell viability of six NSCLC cells was measured by CCK-8 assay at 0, 24, 48, and 72 hrs post transfection. (**C**) At 48 hrs post transfection, the cell proliferation of NSCLC cells expressing *BANCR* or not was determined by BrdU incorporation assays. Data in NC group was setting as 100%. Data represent mean ± SD.\**P*\<0.05, \*\**P*\<0.01. **Abbreviations:** *BANCR*, *BRAF activated non-coding RNA*; NC, negative control; NSCLC, non-small-cell lung cancer.
BANCR mediates invasion of NSCLC cells {#S0003-S2003}
--------------------------------------
Compelling studies reported invasion of NSCLC cells was major issue for mortality during the NSCLC development and progression. To determine whether *BANCR* influenced the invasion of NSCLC cells, Transwell migration was carried out after transfection with the *BANCR* and NC plasmids transfection in A549, H1975, H1299, H1650, SPC-A1, and PC-9 cells. The results showed that abnormal overexpression of *BANCR* resulted in an obvious suppression of cell invasion in NSCLC cells as shown in the Transwell migration assay. In Transwell migration assays, ectopic expressing of *BANCR* in cells presented an inhibitory ability on invasion of these six cell lines, especially for H1299, H1975, H1650, and SPC-A1 cells ([Figure 3A](#F0003){ref-type="fig"}, [B](#F0003){ref-type="fig"}). The results suggest that *BANCR* overexpression suppressed the invasive property of NSCLC cells in vitro.Figure 3*BANCR* overexpression suppressed NSCLC cells invasion. After *BANCR* plasmids transfection, invasive capacity of six NSCLC cell lines was measured by Transwell migration assay (**A**and**B**). Representative data from three independent experiments represent mean ± SD. \**P*\<0.05, \*\**P*\<0.01.**Abbreviations:** *BANCR*, *BRAF activated non-coding RNA*; NC, negative control; NSCLC, non-small-cell lung cancer.
Ectopic overexpression of BANCR up-regulates apoptosis level of NSCLCs {#S0003-S2004}
----------------------------------------------------------------------
Reduced apoptotic activity in NSCLC cell lines was a crucial property of NSCLC. In the study, an evaluation whether *BANCR* played a vital role in NSCLC cellular apoptosis was performed, Hoechst 33342 staining and Annexin V-FITC/PIflow cytometry were performed in A549, H1975, H1299, H1650, SPC-A1, and PC-9 cells transfected with *BANCR* and other plasmids. The phenomenon of *BANCR* expressing cells revealed that a higher amounts of positive Hoechst 33342-stained cells at 48 hrs post-transfection, compared with control groups ([Figure 4A](#F0004){ref-type="fig"}, [B](#F0004){ref-type="fig"}). Meanwhile, overexpression of BANCR caused an increased apoptotic activity in NSCLC cell lines has been detected via flow cytometry examination ([Figure 4C](#F0004){ref-type="fig"}, [D](#F0004){ref-type="fig"}).Figure 4Ectopic overexpression of *BANCR* enhanced apoptotic level of NSCLC cell lines. (**A** and**B**) Hoechst 33342 staining was carried out in each group of NSCLC cells expressing *BANCR* or control cells. Magnification, ×200. Apoptotic rate of positive Hoechst 33342 staining in each group of NSCLC cells was displayed in upper panel. (**C**and **D**) Annexin V-FITC/PI staining and flow cytometry was performed to evaluate the amount of apoptotic cells. The upper and lower right quadrant of each plot represented early apoptotic cells. Apoptotic rate analysis of NSCLC cells in each group was displayed in lower panel. Data represent mean ± SD.\**P*\<0.05, \*\**P*\<0.01, \*\*\**P*\<0.001 versus control group.**Abbreviations:** *BANCR,* *BRAF activated non-coding RNA*; NC, negative control; NSCLC, non-small-cell lung cancer.
Because *BANCR* suppressed cell viability and promoted NSCLC cells apoptosis, its role in regulating the expression of apoptosis-related proteins has been further studied. Caspase-3 and caspase-7 are typical apoptotic markers; therefore, we tested the caspase-3/7 activity in these NSCLC cell lines to evaluate the effect of *BANCR* on apoptosis. The result showed that *BANCR* is able to induce the apoptosis and caspase activity at 72 hrs post-transfection ([Figure 5A](#F0005){ref-type="fig"}). Bcl-2 and Bax, typical anti- and pro-apoptotic proteins, were also examined by WB and real-time PCR. As shown in [Figure 5B](#F0005){ref-type="fig"} and [C](#F0005){ref-type="fig"}, *BANCR* transfection led to an expression decline on Bcl-2, in contrast, it increased Bax expression when compared with NC group in both protein and mRNA level ([Figure 5D](#F0005){ref-type="fig"}, [E](#F0005){ref-type="fig"}). These findings also revealed a mechanism of how *BANCR* regulated the apoptosis and were coincided with previous findings.Figure 5Overexpression of *BANCR* regulated expression level of apoptosis-associated proteins. (**A**) BANCR overexpression increased Caspase-3/7 activity in six NSCLC cell lines. (**B** and**D**) Western blot and (**C**and**E**) real-time PCR were performed to assess the Bcl-2 and Bax expression in protein and mRNA level, which was regulated by *BANCR* overexpression. Data represent mean ± SD.\**P*\<0.05, \*\**P*\<0.01 versus NC group. **Abbreviations:** *BANCR, BRAF activated non-coding RNA*; NC, negative control; NSCLC, non-small-cell lung cancer.
BANCR displays the strong therapeutic effect on NSCLC in mice {#S0003-S2005}
-------------------------------------------------------------
To determine the effect of *BANCR* on xenograft tumor formation, BALB/c mice were subcutaneously injected with SPC-A1 cells constitutively expressing *BANCR* or normal SPC-A1 cells, and monitored daily for tumor growth. At 28 days after the injection, these mice were sacrificed and the lung tissues were excised and weighed ([Figure 6A](#F0006){ref-type="fig"}). The data unraveled that the average volumes of BANCR-expressing tumors grew at a comparatively slower rate than the tumors in the control group, and the mean tumor weight was significantly less than those in control groups ([Figure 6B](#F0006){ref-type="fig"} and [C](#F0006){ref-type="fig"}).Figure 6*BANCR* suppresses xenograft tumor formation. (**A**) SPC-A1cells with stable expression of *BANCR* or not were injected subcutaneously into nude mice (n=8 per group). Representative photographs of nude mice at day 30 post-inoculation were displayed. Mice were sacrificed and tumor block were weighted at day 28 post inoculation. (**B**) Tumor growth curve during 28 days post inoculation was displayed. The volume of the largest tumor detected in the study was 0.82 cm^3^ (**C**) Tumors were weighted after removal from each group. **Abbreviations:** *BANCR,* *BRAF activated non-coding RNA*; NC, negative control.
Discussion {#S0004}
==========
The development of tumor was a synergetic process of many oncogenes activation as well as tumor suppressor genes stimulation, for instance, epigenetic changes in the tumor pathogenesis. Although NSCLC has been extensively investigated in the previous studies, the underlying mechanisms of NSCLC were still poorly understood. During the study, we found a specific lncRNA, *BANCR*, which was down-regulated with its expression in various NSCLC cell lines. It showed the capacity of the inhibition in key properties of NSCLC cells. Overexpression of BANCR in six NSCLC cell lines (A549, H1975, H1299, H1650, SPC-A1, and PC-9 cells) inhibited cell viability, invasion and simultaneously promoted their apoptosis. In vivo experiment revealed that lung tissue of NSCLC mouse model injected with BANCR-expressing cells exhibited the robust therapeutic ability on NSCLC tumor growth. Our findings suggested that BANCR suppressed NSCLC development via regulation of cell viability, invasion, and apoptosis, in which provided evidence that *BANCR* can be served as therapeutic agent on NSCLC treatment.
MiRNAs and lncRNAs were regarded as two new categories of RNAs and formed an essential component among the repertoire of ncRNAs.[@CIT0014] Recently, accumulating studies have demonstrated the crucial roles of lncRNAs and miRNAs shed an insight into the carcinogenesis and cancer-associated molecular biology.[@CIT0015] Several lncRNAs enhanced the importance in lung cancer, such as HOTAIR, NEAT1, and PVT1. HOTAIR supported invasion and migration of lung cancer cell via interaction between chromatin remodeling factor LSH and HOTAIR to affect the ratio of FOXA1 to FOXA2.[@CIT0016] NSCLC development and progression was also proved to be promoted by NEAT1 via acting as a competing endogenous RNA for the hsa-miR-377-3p.[@CIT0017] lncRNA LINC00152 mediated NSCLC cell proliferation through inhibition of IL24 expression, which was regulated by binding with EZH2.[@CIT0018] Down-regulation of another lncRNA SNHG20 caused a significant inhibition of NSCLC cell proliferation and migration in vitro and suppression of NSCLC tumor growth in vivo.[@CIT0019] Among these oncogenic lncRNAs and tumor suppressor, lncRNA *BANCR* has been drawn sufficient attention. BANCR is frequently overexpressed or down-regulated in the tumor tissues, and played bidirectional roles in the progression of cancers, such as endometrial cancer,[@CIT0020] esophageal squamous cell carcinoma,[@CIT0021] hepatocellular carcinoma,[@CIT0022] gastric cancer,[@CIT0023] retinoblastoma,[@CIT0024] colorectal cancer,[@CIT0025] melanoma.[@CIT0026] Down-regulation of *BANCR* was detected in 30 tissues of NSCLC patients in comparison with 12 adjacent normal lung tissues. The data concealed that the cell viability, metastasis, and apoptosis of A549, H1975, H1299, SPC-A1, and PC-9 cells were significantly inhibited after transfection with *BANCR*. Sun et al also found that BANCR expression was significantly decreased in 113 NSCLC tumor tissues compared with normal tissues. Their work demonstrated that *BANCR* expression was highly associated with larger tumor size, advanced pathological stage, and metastasis. Ectopic expression of *BANCR* impaired cell viability and invasion, leading to the inhibition of metastasis in vitro and in vivo.[@CIT0027] In our study, it mainly focused on the relation between *BANCR* expression and NSCLC cell apoptosis. BANCR Overexpression was shown to play a pivotal role in reinforcing the apoptotic level in different NSCLC cell lines and NSCLC tissue in vivo via the regulation of Bcl-2 and BAX expression.
Conclusion {#S0005}
==========
Taken together, our study revealed that a multi-functional anti-tumor effect of lncRNA *BANCR* in major malignant properties of three NSCLC cell lines and NSCLC progressing in a mouse model. Furthermore, the endogenous factors which *BANCR* targeted to and how it contributed to the NSCLC pathogenesis still remain unknown. Hence that, to screen and identify the targets for biotin-labeled *BANCR* in NSCLC cell lines, a modified HITS-CLIP method combined by RNA-Trap might be applied in our further investigation. This method would provide us a precise and comprehensive interpretation for the role of *BANCR* in NSCLC development.
Disclosure {#S0006}
==========
The authors report no conflicts of interest in this work.
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#sec1}
============
For the earliest cellular organisms,^[@ref1],[@ref2]^ replication must have occurred through primitive mechanisms without elaborate enzymatic machinery.^[@ref3],[@ref4]^ Template-directed nonenzymatic primer extension^[@ref5]^ is an important experimental model system for the study of chemically driven genome replication. In this model, chemically activated mononucleotides or short oligonucleotides^[@ref6]^ bind to a template,^[@ref7]^ followed by monomer polymerization^[@ref8]^ or oligomer ligation^[@ref9]^ to assemble the template complement. Imidazoles, among other heteroarenes,^[@ref10],[@ref11]^ have been broadly utilized in the chemical activation of mono-^[@ref12],[@ref13]^ and oligonucleotides.^[@ref6],[@ref14]^ Much effort has been invested in enabling mixed sequences, especially those containing adenosine and uridine,^[@ref15]^ to be copied efficiently and with high fidelity. Among these efforts, a small, focused screen of substituted imidazoles as monomer leaving groups revealed 2-aminoimidazole-activated monoribonucleotides as superior primer extension substrates compared to imidazole- and 2-methylimidazole-activated monomers.^[@ref14]^ On the other hand, activated trinucleotides that bind downstream of the extending nucleotide were discovered to be much better catalysts than activated monomers, not only providing rate enhancements of at least 2 orders of magnitude, but also enabling copying of templates containing all four canonical nucleotides in one pot.^[@ref6]^ Additional work has been performed to determine and enhance the fidelity,^[@ref16],[@ref17]^ regioselectivity,^[@ref18]−[@ref20]^ and thermodynamics^[@ref21]−[@ref23]^ of these reactions. Plausible prebiotic conditions conducive to imidazole synthesis have also been probed.^[@ref24],[@ref25]^
For decades, the classical S~N~2-type mechanism was thought to constitute the sole mechanistic pathway for primer extension by nucleotide phosphoroimidazolides:^[@ref3]^ the nucleophilic 3′-hydroxyl group of the primer would attack the imidazole-activated phosphate of the incoming monomer via an in-line mechanism, and concomitantly displace the imidazole leaving group. Activated monomers^[@ref26]^ and oligomers^[@ref6],[@ref14]^ binding downstream of the polymerization site play important catalytic roles, as their presence strongly accelerates the rate of primer extension. This rate acceleration was initially hypothesized to result from noncovalent leaving group--leaving group interactions between consecutively bound activated substrates, which preorganize the upstream monomer and lower the activation barrier for subsequent phosphodiester bond formation. However, emerging evidence suggests that two activated monomers can react with each other to form a highly activated imidazolium-bridged dinucleotide intermediate (Np-MeIm-pN, [Figure [1](#fig1){ref-type="fig"}](#fig1){ref-type="fig"}), and the observed rate acceleration afforded by a downstream activated monomer (or oligomer) may be due to covalent nucleophilic catalysis.^[@ref27]^ Hence, we have sought to ascertain which of these two potential pathways is primarily responsible for catalysis by downstream activated nucleotides.
).](ja-2017-11623b_0001){#fig1}
Nonhydrolyzable nucleotide analogues have been indispensable tools in crystallographic and mechanistic studies of polymerases and ligases.^[@ref28]^ For example, nonhydrolyzable nucleotide triphosphates, such as α,β-methylene^[@ref29]^ and α,β-difluoromethylene^[@ref30]^ nucleoside triphosphates have been cocrystallized with T7 RNA polymerase^[@ref29]^ and DNA polymerase β,^[@ref30]^ respectively. Also, Raines and co-workers used α,β-methylene GTP to show that the scission of the GTP α--β-phosphoanhydride bond is critical for RNA ligation by the archaeal RNA ligase RtcB.^[@ref31]^ Drawing inspiration from these strategies, we have developed several nonhydrolyzable 2-MeImpN analogues as mechanistic probes of nonenzymatic RNA template copying. We previously reported the synthesis of guanosine 5′-(3-methylpyrazolyl) phosphonate (PZG, [Figure [1](#fig1){ref-type="fig"}](#fig1){ref-type="fig"})^[@ref32]^ and guanosine 5′-(2-methylpyrrolyl)phosphonate (PYG, [Figure [1](#fig1){ref-type="fig"}](#fig1){ref-type="fig"})^[@ref14]^ as nonhydrolyzable analogues of 2-MeImpG. Our crystal structures of RNA-PZG complexes revealed^[@ref32]^ that the noncovalent interactions between monomers and RNA templates do not always obey canonical Watson--Crick geometry under crystallization conditions. More importantly, we were not able to observe any noncovalent interactions between the leaving groups of consecutively bound PZG nucleotides. We suspected that the low p*K*~a~ of the pyrazolyl group of PZG might have led to chemical differences from 2-MeImpG that prevented the formation of catalytically relevant interactions in the crystal structure. In a parallel approach, we used *P*^1^,*P*^3^-diguanosine-5′-triphosphate (GpppG) as a stable analogue of the corresponding activated Gp-MeIm-pG dinucleotide for crystallographic studies, and showed that it could have the appropriate geometry to bind to cognate RNA primer--template complexes via Watson--Crick base pairing, with its 5′--5′-bridge preorganized for subsequent S~N~2 attack by the primer 3′-hydroxyl group.^[@ref33]^
Here we report the synthesis of guanosine 5′-(4-methylimidazolyl)phosphonate (ICG, [Figure [1](#fig1){ref-type="fig"}](#fig1){ref-type="fig"}), the third and closest nonhydrolyzable isosteric analogue of 2-MeImpG. For all three analogues (ICG, PYG, and PZG), we report the p*K*~a~ of the leaving group mimics in the context of a full nucleotide, along with the binding constants of 2-MeImpG and all three pseudo-activated nucleotides for RNA primer--template complexes. Finally we report measurements of the pH-dependent catalytic activity of all four nucleotides in nonenzymatic primer extension. None of the three analogues shows catalytic effects under mild alkaline primer extension conditions, yet ICG, which has a p*K*~a~ and RNA affinity similar to those of 2-MeImpG, is a modest catalyst under acidic pH. Using NMR and mass spectrometry, we demonstrate that ICG reacts with an activated C monomer to form the 5′--5′-imidazole-bridged guanosine-cytosine dinucleotide (CpICG) with very high regioselectivity. The purified GC-dinucleotide, with a labile nitrogen--phosphorus and a stable carbon--phosphorus linkage flanking the imidazole bridge, reacts with cognate RNA primer--template complexes to afford extended primers only when the labile nitrogen--phosphorus linkage, but not the stable carbon--phosphorus linkage, is proximal to the primer 3′-hydroxyl group. These results are consistent with our recent mechanistic proposal that 2-MeImpN-driven primer extension proceeds via the formation of 5′--5′-imidazolium-bridged dinucleotide intermediates.
Results {#sec2}
=======
Synthesis and p*K*~a~ Determination of Guanosine 5′-(4-Methylimidazolyl)phosphonate (ICG) {#sec2-1}
-----------------------------------------------------------------------------------------
Our synthetic strategy for the construction of ICG was parallel to our previously described synthesis of PZG.^[@ref32]^ Briefly, the commercially available 4-bromo-5-methylimidazole **1** was protected on N3 with a benzenesulfonyl group to give **2**,^[@ref34]^ followed by installation of the phosphonate moiety via a palladium-catalyzed carbon--phosphorus coupling^[@ref35]^ reaction to afford **3**. The protected C--P-linked imidazolylphosphonate was then coupled to the 5′-hydroxyl of guanosine to afford **5**, by first removing the phosphonate alkyl groups (**4**),^[@ref36]^ followed by Mitsunobu phosphonylation.^[@ref37]^ Subsequent deprotection of **5** then furnished the desired ICG product ([Scheme [1](#sch1){ref-type="scheme"}](#sch1){ref-type="scheme"}).
 for the ^1^H, ^13^C, and ^31^P NMR spectra of ICG.](ja-2017-11623b_0010){#sch1}
The p*K*~a~ of the leaving group of monomers in the downstream (i.e., primer+2) position has been shown to strongly influence the rate of reaction of the primer with an activated monomer in the adjacent (i.e., primer+1) position,^[@ref14]^ suggesting that the protonation state of the downstream monomer plays an important role in catalysis. Hence, we sought to determine the protonation p*K*~a~ of the leaving group of 2-MeImpG and the heteroaryl-phosphonate moieties of nonhydrolyzable 2-MeImpG analogues in the context of the full nucleotide ([Figure [2](#fig2){ref-type="fig"}](#fig2){ref-type="fig"}A), and to correlate the experimentally determined p*K*~a~ values with catalysis under primer extension conditions. We used proton NMR spectroscopy for p*K*~a~ determination^[@ref38],[@ref39]^ by monitoring the chemical shifts of the aromatic protons of 2-MeIm and its mimics as a function of pD ([Figure [2](#fig2){ref-type="fig"}](#fig2){ref-type="fig"}B). The resulting sigmoidal curves ([Figure [2](#fig2){ref-type="fig"}](#fig2){ref-type="fig"}C) were fit to an equation derived from the Henderson--Hasselbach formula ([Section S3.3, Supporting Information (SI)](http://pubs.acs.org/doi/suppl/10.1021/jacs.7b11623/suppl_file/ja7b11623_si_001.pdf)), followed by correction as reported by Krezel and Bal^[@ref40]^ ([Section S3.1, SI](http://pubs.acs.org/doi/suppl/10.1021/jacs.7b11623/suppl_file/ja7b11623_si_001.pdf)) to estimate the p*K*~a~ values of the four monomers ([Figure [1](#fig1){ref-type="fig"}](#fig1){ref-type="fig"}). Of the nonhydrolyzable analogues, the protonation p*K*~a~ of the imidazolyl moiety of ICG (p*K*~a~ = 6.3) is most similar to that of the 2-MeIm leaving group of 2-MeImpG (p*K*~a~ = 6.7, literature value^[@ref41]^ = 7.09; [Figure S1, SI](http://pubs.acs.org/doi/suppl/10.1021/jacs.7b11623/suppl_file/ja7b11623_si_001.pdf)). The p*K*~a~ values of the heteroaryl groups of PZG and PYG are at least 4 units lower than that of ICG ([Figures S2 and S3, SI](http://pubs.acs.org/doi/suppl/10.1021/jacs.7b11623/suppl_file/ja7b11623_si_001.pdf)). It is interesting to note that the p*K*~a~ values of ICG and 2-MeImpG are at least 1 unit lower than the reported p*K*~a~ (7.86) of free 2-methylimidazole^[@ref42]^ despite the expected stabilization of the protonated state by the adjacent negative charge of the phosphate or phosphonate. Apparently, the inductive electron withdrawing effect of phosphate or phosphonate groups has a greater influence and serves to lower the p*K*~a~ of the leaving groups relative to free 2-methylimidazole.
{#fig2}
Affinity of 2-MeImpG and Nonhydrolyzable Analogues to RNA Duplex {#sec2-2}
----------------------------------------------------------------
We assessed the affinity of 2-MeImpG and its nonhydrolyzable analogues for cognate binding sites on an RNA primer--template complex by isothermal titration calorimetry.^[@ref43],[@ref44]^ Specifically, 4.4 equiv of 2-MeImpG or analogue was titrated into 1.5 mM of a previously reported interrupted RNA duplex in which the binding site for guanosine mononucleotides is sandwiched between the primer 3′-terminus and a downstream helper oligonucleotide, in the presence of pH 7-buffered NaCl solution (500 mM Na^+^) at 20 °C ([Figure [3](#fig3){ref-type="fig"}](#fig3){ref-type="fig"}A). The helper oligonucleotide primarily serves to enhance the affinity of monomer or analogue to the RNA primer--template complex, thus enabling affinity be measured by ITC.^[@ref21]^ We now show that the affinities of the nonhydrolyzable 2-MeImpG analogues for the sandwich duplex are comparable, with estimated *K*~a~ values between 2.8 × 10^3^ and 4.5 × 10^3^ M^--1^; this is slightly higher than the *K*~a~ of 2-MeImpG ((2.1 ± 0.4) × 10^3^ M^--1^, [Figure [3](#fig3){ref-type="fig"}](#fig3){ref-type="fig"}B). The estimated stoichiometric constant *n* of RNA:analogue binding for all three monomer analogues was ∼0.8, consistent with 1:1 analogue-to-RNA binding. The free energy of RNA binding for the analogues ranged between −4.6 and −4.9 kcal mol^--1^, suggesting that the chemical characteristics of the heteroaryl groups of the analogues have only minimal effects on the free energy of binding with RNA. Notably, the presence of 50 mM Mg^2+^ did not alter the affinity of ICG for RNA. We suspect that the high ionic strength of both solutions negates any effect Mg^2+^ may have on the binding of the analogue to RNA duplexes.
 for the LC-HRMS analysis of the sandwich oligonucleotides; see [Figure S5](http://pubs.acs.org/doi/suppl/10.1021/jacs.7b11623/suppl_file/ja7b11623_si_001.pdf) and [Table S1 (SI)](http://pubs.acs.org/doi/suppl/10.1021/jacs.7b11623/suppl_file/ja7b11623_si_001.pdf) for data derived from individual ITC titrations. All ITC experiments were performed at 20 °C.](ja-2017-11623b_0003){#fig3}
Effects of 2-MeImpG and Nonhydrolyzable Analogues on RNA Primer Extension {#sec2-3}
-------------------------------------------------------------------------
With these insights into the chemical and thermodynamic properties of the 2-MeImpG analogues, we turned to the question of catalysis of the reaction between the primer and an adjacent monomer, by analogues bound to the template at the primer +2 position. We measured the pseudo-first-order rate of conversion of primer to extended products on a template with a GCCCAA-5′ overhang ([Figure [4](#fig4){ref-type="fig"}](#fig4){ref-type="fig"}A; [Figure S6, SI](http://pubs.acs.org/doi/suppl/10.1021/jacs.7b11623/suppl_file/ja7b11623_si_001.pdf)), using a reaction mixture that contained a large excess of the 2-MeImpC (30 mM). 2-MeImpG and its nonhydrolyzable analogues were used at 30 mM, because higher concentrations led to observable precipitation, especially for ICG and GMP; the catalytic metal ion Mg^2+^ was present at 100 mM. We varied the pH of the primer extension reactions from mildly acidic (pH 5.5) to slightly alkaline (pH 8) to probe the effect of pH on catalysis by the analogues. Interestingly, the rate of primer extension with 2-MeImpC in the absence of any downstream guanosine nucleotides (blue bars, [Figure [4](#fig4){ref-type="fig"}](#fig4){ref-type="fig"}B) was slightly higher than in the presence of unactivated GMP (black bars, [Figure [4](#fig4){ref-type="fig"}](#fig4){ref-type="fig"}B). We attribute this small effect to the potential for weak binding of 2-MeImpC at the primer +2 position in the absence of a competing G nucleotide. To avoid this confounding effect, we compare the catalytic effects of 2-MeImpG and the three analogues to the baseline rate observed in the presence of GMP. As expected 2-MeImpG in the primer +2 position exhibits a strong catalytic effect at all pH values from 5.5 to 8.0 (red bars, [Figure [4](#fig4){ref-type="fig"}](#fig4){ref-type="fig"}B). In general, all primer extension reactions exhibit higher rates at alkaline pH than in neutral or acidic pHs, presumably due to the need to deprotonate the primer 3′-hydroxyl. We then asked whether binding of ICG, PYG or PZG at the primer +2 position could enhance the rate of primer extension with 2-MeImpC (green, purple, and orange bars, respectively; [Figure [4](#fig4){ref-type="fig"}](#fig4){ref-type="fig"}B). Regardless of pH, the rate of primer extension was not improved in the presence of PYG or PZG, compared with GMP. Therefore, neither PZG nor PYG affords any catalysis to primer extension.
 for representative denaturing PAGE gel images and first-order rate plots for the primer extensions depicted in (B), as well as the saturation curves of 2-MeImpG- and ICG-catalyzed primer extensions at pH 6.3 and 8.0.](ja-2017-11623b_0004){#fig4}
On the other hand, ICG, which has an imidazolyl p*K*~a~ close to the p*K*~a~ of 2-MeImpG, did exhibit a pH-dependent catalytic effect on primer extension ([Figure [4](#fig4){ref-type="fig"}](#fig4){ref-type="fig"}B). At a primer extension pH of 8, the presence of ICG led to only a very small increase in the rate of primer extension of 20% (compared to the 60-fold effect of 2-MeImpG). However, when the extension reactions were carried out at neutral pH, the catalytic effect of ICG increased to ∼2.5-fold (vs 40-fold for 2-MeImpG), while at pH 6.3, ICG exhibited a 15-fold rate enhancement relative to GMP, rivaling the ∼20-fold effect of 2-MeImpG. The *k*~obs~ of ICG-assisted primer extension is similar from pH 8.0 to 6.3, while the *k*~obs~ of 2-MeImpG-assisted primer extension decreased ∼37 fold from pH 8.0 to pH 6.3. At pH 5.5, primer extension in the presence of GMP could barely be detected even after 20 h. Even 2-MeImpG in the primer +2 position yielded less than 5% extended primer after 20 h. However, the catalytic effect of ICG surpassed that of 2-MeImpG and resulted in readily observable primer extension, leading to 20% of the primer +1 product after 20 h ([Figure S7, SI](http://pubs.acs.org/doi/suppl/10.1021/jacs.7b11623/suppl_file/ja7b11623_si_001.pdf)). Thus, the rate enhancement afforded by ICG is much more pronounced at acidic pH, while the rate enhancement afforded by 2-MeImpG is most pronounced at alkaline pH.
To see if the increasing catalytic effect of ICG at lower pH (and conversely, the declining catalytic effect of 2-MeImpG at lower pH) was due to changes in template binding, we measured the catalytic effect as a function of concentration at pH 6.3 and 8.0. Primer extensions at pH 6.3 were repeated at four different concentrations of ICG (0, 3, 10, 30, 60 mM). The estimated *k*~max~ was 0.075 ± 0.008 h^--1^, and the *K*~M~ was 7 ± 3 mM. From the rate vs concentration plot ([Figure [4](#fig4){ref-type="fig"}](#fig4){ref-type="fig"}C), it is apparent that the rate-enhancement afforded by ICG approaches a maximum when the concentration of ICG is at, or above 20 mM. Since both ICG and 2-MeImpG achieve their maximal catalytic effect at 30 mM at pH 6.3 ([Figure S13, SI](http://pubs.acs.org/doi/suppl/10.1021/jacs.7b11623/suppl_file/ja7b11623_si_001.pdf)), their different catalytic abilities cannot be attributed to differential binding. In contrast, at pH 8.0, the rates of primer extension are unaffected by ICG concentration (0 mM ICG; rate ≈ 0.07 h^--1^; [Figure S12, SI](http://pubs.acs.org/doi/suppl/10.1021/jacs.7b11623/suppl_file/ja7b11623_si_001.pdf)). Since at pH 8 ICG and 2-MeImpG bind with similar affinities to primer/template complexes, the pH-dependence of catalysis by ICG does not reflect pH-dependent changes in binding, but likely reflects a direct pH-dependence of the catalytic mechanism.
How can we explain the observation that ICG becomes a better catalyst of primer extension at progressively lower pH, while 2-MeImpG gradually loses catalytic ability at low pH? Given the observation of a 5′--5′-imidazolium-bridged dinucleotide in primer extension reactions, we wondered if ICG might catalyze primer extension in a similar fashion, with the imidazolyl group of ICG (highlighted in blue, [Figure [5](#fig5){ref-type="fig"}](#fig5){ref-type="fig"}) reacting with 2-MeImpG to form a mixed N--P- and C--P-linked imidazole-bridged intermediate ([Figure [5](#fig5){ref-type="fig"}](#fig5){ref-type="fig"}, middle panel).^[@ref27]^ Due to the lower p*K*~a~ of ICG than 2-MeImpG, the ratio of protonated 2-MeImpG to unprotonated ICG is increased at lower pH, while a higher proportion of ICG remains in the neutral, nucleophilic form. Therefore, we sought to examine whether the proposed 5′--5′-imidazole-bridged dinucleotide was formed during the ICG-catalyzed primer extension.
{#fig5}
Detection of 5′--5′-Imidazole-Bridged Mixed-Linkage Dinucleotide via LCMS {#sec2-4}
-------------------------------------------------------------------------
Initial attempts to identify the 5′--5′-imidazole-bridged mixed-linkage dinucleotide were performed with analytical reversed-phase liquid chromatography-coupled mass spectrometry (LCMS), using ICG and dideoxycytidine 5′-phosphoro-2-methylimidazolide (2MeImpddC) as the substrate. 2MeImpddC was chosen in lieu of 2-MeImpC because its dideoxyribose moiety affords enhanced hydrophobicity, allowing 2MeImpddC, as well as its ICG-conjugated products to be chromatographically resolved from ICG. Additionally, removal of the cytidine ribose 2′- and 3′-hydroxyl groups eliminates the possibility of 2′-5′- and 3′-5′-linked CC dinucleotide formation, thus facilitating chromatographic separation of possible products. 2MeImpddC was conveniently prepared from Zalcitabine with a three-step protection--activation--deprotection sequence ([Figure [6](#fig6){ref-type="fig"}](#fig6){ref-type="fig"}A). First, 30 mM ICG and 30 mM 2MeImpddC were incubated at pH 6.3 in the absence of Mg^2+^. After 2 h, the incubated sample was chromatographically analyzed under reverse phase conditions (300 pmol each for ICG and 2MeImpddC). ICG and 2MeImpddC were well resolved, and only one new peak appeared, with a retention time of 19.11 min ([Figure [6](#fig6){ref-type="fig"}](#fig6){ref-type="fig"}B) and an *m*/*z* of 699.33 ([Figure [6](#fig6){ref-type="fig"}](#fig6){ref-type="fig"}C), which corresponds to the expected *m*/*z* of monoanionic ddCMP-ICG dinucleotide. To confirm that the ddCMP-ICG dinucleotide carries the desired 5′--5′-imidazole bridge, we performed tandem MS fragmentation on the precursor ion with the *m*/*z* = 699.33 (with a *m*/*z* window of ±1). The fragmentation voltage was chosen such that the precursor ion was only partially fragmented so that the major fragment *m*/*z* peak could indicate which bond within the precursor ion is the most labile ([Figure [6](#fig6){ref-type="fig"}](#fig6){ref-type="fig"}D). Notably, the major fragment peak, with an *m*/*z* of 416.17 (peak 1, [Figure [6](#fig6){ref-type="fig"}](#fig6){ref-type="fig"}D), is suggestive of fragmentation of the 5′--5′-imidazole-bridged ddCMP-ICG dinucleotide, with heterolytic cleavage occurring at the O--P bond between the guanosine and the imidazole bridge moieties ([Figure [6](#fig6){ref-type="fig"}](#fig6){ref-type="fig"}E). Other possible regioisomers of ddCMP-ICG dinucleotide, such as the 3′-5′-, 2′--5′-, and nucleobase-linked dinucleotide, could not afford fragments with *m*/*z* of 416.17. This observation thus provides the first clue that ICG could react with activated monomers to afford 5′--5′-imidazole-bridged dinucleotides.
 of SI for a more detailed assignment of the MS^2^ spectrum. (E) The proposed structure of the 5′--5′-imidazole-linked ddCMP-ICG dinucleotide, the observed *m*/*z* signals, and the corresponding CpICG fragments. 2-MeImpddC had been coincubated with ICG, PYG, and PZG, and the formation of 5′--5′-heteroarene-bridged dinucleotide had been monitored by LCMS and NMR. See [Figures S15--S20 (SI)](http://pubs.acs.org/doi/suppl/10.1021/jacs.7b11623/suppl_file/ja7b11623_si_001.pdf) for detailed analyses.](ja-2017-11623b_0006){#fig6}
Structure Elucidation of CpICG by NMR {#sec2-5}
-------------------------------------
Encouraged by the LC-MS results, we then asked whether 2-MeImpC and ICG could react to form the 5′--5′-imidazole-bridged, mixed linkage CG dinucleotide (CpICG, [Figure [5](#fig5){ref-type="fig"}](#fig5){ref-type="fig"}, middle panel). 50 mM ICG was incubated at pH 6.3 with 50 mM 2-MeImpC in the absence of Mg^2+^, and the progress of dinucleotide formation was monitored by ^31^P NMR spectroscopy overnight ([Figure [7](#fig7){ref-type="fig"}](#fig7){ref-type="fig"}A). As seen in the *t* = 0 h spectrum, acquired immediately after mixing, two major ^31^P singlet peaks were seen at −9.89 and 2.30 ppm, corresponding to 2-MeImpC and ICG, respectively. Three additional minor singlet peaks were observed at −8.95, 2.25, and 7.54 ppm, and no other significant ^31^P peaks are observed. These peaks increased in intensity as the incubation proceeded, with the peaks at −8.95 and 7.54 ppm maintaining similar intensity throughout the entire NMR time course, suggesting that they might arise from the same molecule. The peak at 2.25 ppm is assigned to CMP by internal referencing, which results from the hydrolysis of 2-MeImpC ([Figure S21, SI](http://pubs.acs.org/doi/suppl/10.1021/jacs.7b11623/suppl_file/ja7b11623_si_001.pdf)). The remaining two ^31^P peaks at −8.95 and 7.54 ppm were tentatively assigned to the two phosphorus atoms of the newly formed GC dinucleotide. To establish the regiochemistry of the newly formed GC dinucleotide, and rule out the possibility of 2′-5′-, 3′-5′-, or guanine-phosphate linkages, we examined the reaction of PYG with 2-MeImpC. The pyrrolylphosphonate group of PYG is known to be non-nucleophilic under near-neutral conditions. Throughout an overnight ^31^P NMR time course, the only new detectable product was CMP ([Figure [7](#fig7){ref-type="fig"}](#fig7){ref-type="fig"}B), suggesting that the nucleophilic ribose hydroxyls and nucleobase exocyclic amine of PYG do not react with 2-MeImpC over the time course of primer extension assays. Since the only reactive site on ICG is the imidazolylphosphonate group, the CpICG dinucleotide should indeed contain a 5′--5′-imidazole bridge.
 for the complete ^31^P stacked spectrum of the NMR time course analyses depicted in panels (A)--(D), respectively. See [Figure S25 (SI)](http://pubs.acs.org/doi/suppl/10.1021/jacs.7b11623/suppl_file/ja7b11623_si_001.pdf) for the complete ^31^P NMR time course analyses of 2-MeImpC-PZG coincubation. See [Figures S37----S40 (SI)](http://pubs.acs.org/doi/suppl/10.1021/jacs.7b11623/suppl_file/ja7b11623_si_001.pdf) for detailed LCMS time course analyses of the formation and hydrolysis of CpICG at pH 6.3 and 8.0. See [Figure S41 (SI)](http://pubs.acs.org/doi/suppl/10.1021/jacs.7b11623/suppl_file/ja7b11623_si_001.pdf) for the fitted rate of CpICG formation and degradation derived from the LCMS time course analyses.](ja-2017-11623b_0007){#fig7}
Because ICG catalyzed primer extension at pH 6.3 but not at pH 8.0, we asked whether the 5′--5′-imidazole-bridged CpICG dinucleotide could form under basic conditions. Using ^31^P NMR spectroscopy, we monitored the overnight reaction of coincubated ICG and 2-MeImpC at pH 8.0 in the absence of Mg^2+^. As with the pH 6.3 NMR time course, three new singlet peaks increased in intensity over time. The three new peaks were shifted downfield to varying degrees, presumably due to the change in pH ([Figure [7](#fig7){ref-type="fig"}](#fig7){ref-type="fig"}C). The peak at 3.85 ppm was assigned to CMP ([Figure S22, SI](http://pubs.acs.org/doi/suppl/10.1021/jacs.7b11623/suppl_file/ja7b11623_si_001.pdf)) suggesting that the remaining two peaks belong to CpICG. This was confirmed with the same PYG control experiment as used above ([Figure [7](#fig7){ref-type="fig"}](#fig7){ref-type="fig"}D). To confirm that the GC dinucleotide formed at pH 6.3 and 8.0 is indeed the same chemical entity, the dinucleotide prepared from overnight 2-MeImpC-ICG coincubation at pH 8.0 was chromatographically purified under reverse-phase conditions, followed by lyophilization and ^31^P NMR acquisition at pH 6.3. The two singlet ^31^P peaks thus obtained have chemical shift values close to those obtained in the pH 6.3 ^31^P NMR time course (within ∼±0.2 ppm; [Figure S26, SI](http://pubs.acs.org/doi/suppl/10.1021/jacs.7b11623/suppl_file/ja7b11623_si_001.pdf)), thus confirming that the change in dinucleotide phosphorus chemical shifts is due to pH change.
The proposed structure of CpICG, as well as the presence of the 5′--5′-imidazole-bridged connectivity, was validated by a combination of 1D homonuclear and 2D heteronuclear coupled NMR spectroscopy ([Figure [8](#fig8){ref-type="fig"}](#fig8){ref-type="fig"}A,B). We looked specifically for the additional ^31^P-induced splitting of the carbon resonances associated with the bridging imidazole of CpICG. To identify these carbon resonances, we used both ^1^H--^13^C coupled gc2HSQCse and gHMBCAD spectroscopy. The doublet ^1^H peak at 7.87 ppm was assigned to the imidazolyl methine proton, while the ^1^H peak at 2.44 ppm was assigned to the methyl protons exocyclic to the imidazole bridge ([Figure S27, SI](http://pubs.acs.org/doi/suppl/10.1021/jacs.7b11623/suppl_file/ja7b11623_si_001.pdf)). As shown in the partial ^1^H--^13^C coupled gc2HSQCse spectrum, the imidazolyl methine proton is correlated with the doublet of doublet resonance at 141.8--142.0 ppm ([Figure [8](#fig8){ref-type="fig"}](#fig8){ref-type="fig"}C,F). This is consistent with the expected splitting pattern induced by both the N--P and C--P phosphorus atoms. On the other hand, the partial ^1^H--^13^C coupled gHMBCAD spectrum clearly demonstrates that the exocyclic methyl protons of the bridging imidazole are coupled to only two carbon resonances, which must be the imidazole bridge carbons that are two and three bonds away ([Figure [8](#fig8){ref-type="fig"}](#fig8){ref-type="fig"}D,F). These resonances at 137.0--137.4 and 131.5--133.4 ppm were also doublets of doublets, consistent with the 5′--5′-phosphate-imidazole-phosphonate linkage regiochemistry. The corresponding carbon signals of ICG only have doublet splitting because the molecule only contains one phosphorus atom ([Figure S35, SI](http://pubs.acs.org/doi/suppl/10.1021/jacs.7b11623/suppl_file/ja7b11623_si_001.pdf)). Finally, ^1^H--^31^P coupled gHMBCAD spectroscopy of CpICG (estimated ^*n*\>1^*J*~P,H~ = 8 Hz), clearly showed that the three protons of the imidazole exocyclic methyl group are coupled to both phosphorus atoms of the imidazole-flanking phosphate and phosphonate ([Figure [8](#fig8){ref-type="fig"}](#fig8){ref-type="fig"}E). These data, together with the ^31^P NMR time courses shown in [Figure [7](#fig7){ref-type="fig"}](#fig7){ref-type="fig"}, strongly suggest that ICG reacts with 2-MeImpC to form the 5′--5′-imidazole-bridged CpICG.
 for the full ^1^H, ^13^C, and ^31^P spectra of CpICG, and [Figures S30--S33 (SI)](http://pubs.acs.org/doi/suppl/10.1021/jacs.7b11623/suppl_file/ja7b11623_si_001.pdf) for the full 2D spectrum of ^1^H--^1^H gCOSY, ^1^H--^13^C gc2HSQCse and gHMBCAD, and ^1^H--^31^P gHMBCAD spectra of CpICG.](ja-2017-11623b_0008){#fig8}
Primer Extension Reactions with CpICG {#sec2-6}
-------------------------------------
Having established that 5′--5′-linked CpICG can form by the reaction of ICG with 2-MeImpC, we asked whether this dinucleotide can bind to its cognate RNA template and react with an upstream primer. We carried out the primer extension experiment as in [Figure [4](#fig4){ref-type="fig"}](#fig4){ref-type="fig"}, using the same RNA primer--template complex and reaction conditions as used for the pseudo-first-order kinetic analysis, but with 20 mM purified CpICG in lieu of the 2-MeImpC + ICG combination ([Figure [9](#fig9){ref-type="fig"}](#fig9){ref-type="fig"}A; [Figure S42, SI](http://pubs.acs.org/doi/suppl/10.1021/jacs.7b11623/suppl_file/ja7b11623_si_001.pdf)). At both pH 6.3 and pH 8.0, the primer +1 product can be clearly observed by 15 min, with its abundance increasing at longer reaction times. This result strongly suggests that CpICG can bind to a cognate RNA template and react with the adjacent primer 3′-hydroxyl group. As a control experiment, we reversed the sequence of the templating nucleotides at the primer +1 and +2 positions (underlined nucleotides, [Figure [9](#fig9){ref-type="fig"}](#fig9){ref-type="fig"}A), from 3′-[GC]{.ul}CCAA-5′ to 3′-[CG]{.ul}AA-5′. If CpICG binds to this "reversed" template, then the unreactive C--P linkage of the imidazole bridge would be pointing toward the primer 3′-OH nucleophile, instead of the labile N--P linkage, and no primer extension should occur. Indeed, reactivity is completely abolished with this template, even at the highest tested CpICG concentration (20 mM, [Figure [9](#fig9){ref-type="fig"}](#fig9){ref-type="fig"}A).
 for representative denaturing PAGE gel images, first-order rate plots, and the saturation curves of CpICG-driven primer extensions at pH 6.3 and 8.0 at different CpICG concentrations.](ja-2017-11623b_0009){#fig9}
To assess the affinity of CpICG for its cognate RNA primer--template duplex, we determined the *K*~M~ of this dinucleotide under primer extension conditions ([Figure [9](#fig9){ref-type="fig"}](#fig9){ref-type="fig"}B). We measured the pseudo-first-order rate of primer extension as a function of concentration of CpICG at pH 6.3 and 8.0. At pH 6.3, the estimated *k*~max~ and *K*~M~ were 0.28 ± 0.01 h^--1^ and 10.6 ± 0.8 mM, respectively (*R*^2^ = 0.996). At pH 8.0, the estimated *k*~max~ and *K*~M~ were 0.35 ± 0.02 h^--1^ and 12 ± 1 mM, respectively (*R*^2^ = 0.994). The pH change does not significantly alter the affinity of CpICG for its cognate template, and the estimated *k*~max~ of CpICG-driven primer extension was comparable under both acidic and alkaline conditions, consistent with the similar rates of ICG-catalyzed primer extension with 2-MeImpC at pH 6.3 and 8.0 (*k*~obs~ = 0.069 ± 0.002 h^--1^ and *k*~obs~ = 0.066 ± 0.003 h^--1^, respectively; [Figure [4](#fig4){ref-type="fig"}](#fig4){ref-type="fig"}B and [Figures S11 and S12, SI](http://pubs.acs.org/doi/suppl/10.1021/jacs.7b11623/suppl_file/ja7b11623_si_001.pdf)).
Discussion {#sec3}
==========
The mechanism by which nonenzymatic RNA polymerization occurs has been of interest for decades. Since the early 90s, it has been known that the presence of an activated monomer downstream of the reacting nucleotide catalyzes primer extension.^[@ref26],[@ref45]^ Our laboratory recently confirmed and extended these early results, showing that the presence of an activated 5′-phosphate on the downstream nucleotide (or oligonucleotide) catalyzed the reaction between the primer and the primer-adjacent nucleotide. Noncovalent leaving group--leaving group interactions, or simply steric or Coulombic repulsions between the leaving groups, were thought to catalyze primer extension by bringing the activated phosphate of the reacting monomer closer to the primer 3′-hydroxyl, or helping to orient the leaving group for in-line nucleophilic substitution. Indeed, our previously reported RNA-PZG crystal structures show that the presence of a downstream PZG nucleotide decreases the distance from the primer 3′-OH to the adjacent PZG phosphate from 6.5 to 4.5 Å,^[@ref32]^ and we have obtained similar results in a more recent study with ICG.^[@ref46]^ However, no leaving group-leaving group interactions were observed in either study. If noncovalent leaving group interactions are a critical component of downstream monomer-induced catalysis, PYG, which has a neutral pyrrolyl leaving group mimic (p*K*~a~ \< 2), should catalyze upstream primer extension to some degree but no catalysis is observed ([Figure [4](#fig4){ref-type="fig"}](#fig4){ref-type="fig"}B). Since 4.5 Å is too far for phosphodiester bond formation, other factors must come into play.
Our laboratory has proposed a novel mechanism for primer extension in which the phosphoroimidazolide moieties of two 2-MeImpN first react with each other to form a 5′--5′-imidazolium-bridged dinucleotide intermediate, followed by template binding, primer extension by one nucleotide and regeneration of the downstream 2-MeImpN.^[@ref27]^ The Richert lab has also noted the high reactivity of imidazolium-bridged dinucleotides in primer extension.^[@ref23]^ Primer extension reactions driven by partially purified Cp-MeIm-pC display extremely high (\>30 h^--1^) initial rates of primer extension, at a concentration much lower than the \>50 mM concentrations of 2-MeImpC commonly employed,^[@ref27]^ suggesting that catalysis involving a 5′--5′-imidazolium-bridged dinucleotide may be the predominant mechanism of primer extension.
In the present study we synthesized ICG as a close nonhydrolyzable analogue of 2-MeImpG, which we initially hoped to use as a probe of catalytically relevant noncovalent interactions. However, no such interactions have been observed. In addition, we were puzzled by the fact that ICG does act as a modest catalyst of primer extension, but only at mildly acidic pH. We then found that ICG could react with 2-MeImpC to form the corresponding 5′--5′-imidazole-bridged GC dinucleotide (CpICG). Indeed, the rates of ICG-catalyzed primer extensions, carried out with 30 mM of the monomers, can be reproduced with only 3 mM of the purified 5′--5′-imidazole-bridged CpICG (*k*~obs~ = 0.068 ± 0.004 and 0.074 ± 0.001 h^--1^ for pH 6.3 and 8.0, respectively; [Figures S43 and S44, SI](http://pubs.acs.org/doi/suppl/10.1021/jacs.7b11623/suppl_file/ja7b11623_si_001.pdf)), suggesting not only that CpICG is more reactive than the corresponding 2-MeImpC monomer under primer extension conditions, but also that the *in situ* formation of activated 5′--5′-imidazole-bridged dinucleotides could account for the rate of ICG-catalyzed primer extensions shown in [Figure [4](#fig4){ref-type="fig"}](#fig4){ref-type="fig"}. These results complement our recent reports^[@ref27],[@ref33],[@ref47]^ supporting the hypothesis that the formation of 5′--5′-imidazolium-bridged dinucleotides in nonenzymatic RNA primer extensions is an integral step for 2-MeImpN-driven primer extension.
A few characteristics of CpICG are notable. First, the ^31^P NMR analysis of the products of 2-MeImpC-ICG coincubation suggest that only one regioisomer out of the four theoretically possible GC dinucleotides (imidazole-, 2′--5′- and 3′--5′-phosphodiester-, and nucleobase-linked) is formed. As shown in the 2-MeImpC-PYG coincubation experiment ([Figure [7](#fig7){ref-type="fig"}](#fig7){ref-type="fig"}B,D), the ribose hydroxyls and guanine exocyclic amines of the analogues do not detectably react with the electrophilic phosphoroimidazolide moiety of 2-MeImpC. Therefore, formation of the C--P/N--P mixed-linkage dinucleotide occurs with remarkable regioselectivity, with the nonhydrolyzable imidazolylphosphonate of ICG being the exclusive nucleophile during dinucleotide formation. This high regioselectivity simplifies the synthesis and purification of other C--P/N-P mixed-linkage dinucleotide products. Since the identity of the nucleobase is expected to have minimal impact on the yield or regioselectivity of the mixed linkage dinucleotide formation, the preparation of all 16 possible 5′--5′-imidazole-linked, mixed-linkage dinucleotides should be possible. Second, because of the zwitterionic character of the imidazolium bridge, the 2-MeImpN-derived 5′--5′-bridged dinucleotides are inherently reactive, as evidenced by their short half-lives under acidic conditions. Despite careful control of starting material stoichiometry and rigorous purification procedures, preparation and purification of the Cp-MeIm-pC dinucleotide proved to be challenging, affording enriched samples with a maximum of 35% Cp-MeIm-pC.^[@ref27]^ In contrast, the CpICG dinucleotide described herein is quite stable, especially under basic conditions. Even in the presence of a high concentration of Mg^2+^, CpICG degrades only minimally after 10 h under alkaline conditions ([Figures S39 and S40, SI](http://pubs.acs.org/doi/suppl/10.1021/jacs.7b11623/suppl_file/ja7b11623_si_001.pdf)), and lyophilization of an aqueous CpICG solution did not lead to substantial hydrolysis. The stability of CpICG, which enabled our extensive 1D and 2D NMR studies of the 5′--5′-imidazole bridge connectivity, will also enable future kinetic, thermodynamic, and crystallographic studies of primer extension. We are currently pursuing crystallographic studies of RNA--CpICG complexes, with the goal of revealing the overall conformation of template-bound CpICG, the noncovalent interactions modulating RNA--CpICG base pairing, the relative orientation and distance between the 5′--5′-imidazole-linkage and the primer 3′-terminus, and other mechanistically relevant structural features.
The most puzzling aspect of ICG catalysis of primer extension with 2-MeImpC was the observation of a moderate catalytic effect at acidic pH, together with the fact that the rate of primer extension was slow but almost pH independent from 6.3 to 8.0. Here we present a possible explanation for these observations. CpICG-driven primer extension requires at least two major mechanistic steps: (1) nucleophilic attack of the primer 3′-OH on the imidazole bridge from the labile N--P bond side, and (2) displacement of the imidazolylphosphonate group to liberate a neutral ICG analogue. At high pH, the 5′--5′-bridging 2-methylimidazole of CpICG is uncharged; direct nucleophilic displacement of the imidazole phosphonate bridge, without prior imidazole protonation or concomitant proton transfer, would result in an unfavorable anionic imidazolyl-phosphonate leaving group. Therefore, the imidazolyl group must first be protonated and become cationic, before nucleophilic attack by the primer 3′-OH takes place. Under acidic conditions, the 5′--5′-imidazole bridge of CpICG is more likely to be protonated (estimated p*K*~a~ ≈5.1, [Figure S45, SI](http://pubs.acs.org/doi/suppl/10.1021/jacs.7b11623/suppl_file/ja7b11623_si_001.pdf)), but deprotonation of the primer 3′-OH will be correspondingly more difficult. We suggest that the need to deprotonate the primer 3′-OH while simultaneously protonating the imidazolyl-phosphonate leads to a low primer extension rate regardless of overall pH, together with a catalytic effect of ICG at acidic pH when the intermediate can be protonated. On the other hand, preliminary data suggested that PZG does react with 2-MeImpC to give the 5′--5′-pyrazole-bridged CpPZG ([Figure S25, SI](http://pubs.acs.org/doi/suppl/10.1021/jacs.7b11623/suppl_file/ja7b11623_si_001.pdf)), yet PZG is unable to catalyze primer extensions to any significant extent ([Figure [4](#fig4){ref-type="fig"}](#fig4){ref-type="fig"}B). This lack of reactivity may be due to the low protonation p*K*~a~ of the 5′--5′-pyrazole bridge (likely to be 1.94 or less; [Figure [1](#fig1){ref-type="fig"}](#fig1){ref-type="fig"} and [Figure S2, SI](http://pubs.acs.org/doi/suppl/10.1021/jacs.7b11623/suppl_file/ja7b11623_si_001.pdf)), which renders cationic activation of the pyrazole bridge unlikely when primer extension is carried out at pH ≥5.5.
In conclusion, we have described the synthesis and properties of ICG, the closest nonhydrolyzable 2-MeImpG analogue reported to date. ICG acts as a modest catalyst of primer extension under acidic conditions, primarily via the formation of a 5′--5′-imidazole-bridged dinucleotide which carries both N--P and C--P linkages. This relatively stable mixed-linkage dinucleotide serves as a substrate in primer extension reactions, consistent with our recent mechanistic proposal that the formation of 5′--5′-imidazolium-bridged dinucleotides is an integral step for 2-MeImpN-driven primer extensions. The affinity of these dinucleotides for cognate and noncognate RNA primer--templates, as well as the crystal structures of RNA--dinucleotide complexes, are currently being determined.
The Supporting Information is available free of charge on the [ACS Publications website](http://pubs.acs.org) at DOI: [10.1021/jacs.7b11623](http://pubs.acs.org/doi/abs/10.1021/jacs.7b11623).Figures S1--S45, Table S1, methods, and references ([PDF](http://pubs.acs.org/doi/suppl/10.1021/jacs.7b11623/suppl_file/ja7b11623_si_001.pdf))
Supplementary Material
======================
######
ja7b11623_si_001.pdf
^∇^ C.P.T. and L.Z. contributed equally.
The authors declare no competing financial interest.
We gratefully acknowledge Mr. Gregory J. Heffron and Dr. Charles A. Sheehan, of Harvard Medical School East Quadrangle Nuclear Magnetic Resonance Core Facility, for their assistance with NMR spectroscopy. We thank Dr. Victor S. Lelyveld for assistance with high-resolution liquid chromatography-coupled mass spectrometry. We thank Prof. Seung Soo Oh, Dr. Daniel Duzdevich, Dr. Kyle Strom, Dr. Li Li, and Dr. Derek O'Flaherty for insightful discussions and critiques of the manuscript. J.W.S. is an Investigator of the Howard Hughes Medical Institute. This work was supported in part by a grant from the NSF (CHE-1607034) to J.W.S. and a grant (290363) from the Simons Foundation to J.W.S.
| {
"pile_set_name": "PubMed Central"
} |