Professor Hiromi Sanada, Department of Gerontological Nursing/Wound care Management, Graduate School of Medicine, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan. Tel.:/fax: +81-3-5841-3419;
Aging is accelerated, at least in part, by pathological condition such as metabolic syndrome (MetS), and various molecular pathways such as oxidative stress are common mediators of aging and MetS. We previously developed the aging-like skin model by single ultraviolet (UV) irradiation on the MetS model mice. Recent studies revealed that mineralocorticoid receptor (MR) signaling plays a pivotal role for various tissue inflammation and damages in MetS. Although previous studies reported that MR is expressed in the skin and that overexpression of MR in the skin resulted in the skin atrophy, the physiological or pathological functions of MR in the skin are not fully elucidated. Here, we show the involvement of MR signaling in the aging-like skin changes in our own model. Elevations of oxidative stress and inflammation markers were observed in the MetS mice, and the UV-evoked aging-like skin damages were attenuated by topical antioxidant. MR expression was higher in the MetS mouse skin, and notably, expression of its effecter gene Sgk1 was significantly upregulated in the aging-like skin in the UV-irradiated MetS mice. Furthermore, topical application of MR antagonist spironolactone suppressed Sgk1 expression, oxidative stress, inflammation, and the aging-like changes in the skin. The 2-week UV onto the non-MetS mice, the more usual photoaging model, resulted in the skin damages mostly equivalent to the MetS mice with single UV, but they were not associated with upregulation of MR signaling. Our studies suggested an unexpected role of MR signaling in the skin aging in MetS status.
In general, aging is a combination of physiological (or chronological) and pathological (or disease-based) processes (Goto, 2008). Two examples of the latter are photoaging of the skin irradiated by ultraviolet (UV) (Rabe et al., 2006) and metabolic syndrome (MetS). MetS causes cardiovascular, renal, and other tissue damages contributing to pathological precocious aging (Grundy et al., 2005; Goto, 2008). It is known that disruption of various lifespan determinant pathways such as silent information regulator (SIR)T1 and p99Shc leads to features of MetS in mice, and the same molecular mechanisms may be at least partially active in human with MetS (Fadini et al., 2011), suggesting close relationship between aging and MetS. We previously found that the skin of MetS model mice was more susceptible to UV than the normal mouse skin (Akase et al., 2012). Based on this finding, we established our own aging-like skin model using the single 90 min UV irradiation onto the MetS mouse skin and confirmed increased expression of tumor necrosis factor (TNF)-α, a dual marker for inflammation and aging, in this model (Akase et al., 2012).
Metabolic and hormonal disorders seen in MetS status involve dysregulation of adipocytokines, insulin resistance, and oxidative stress (Grundy et al., 2005; Roberts & Sindhu, 2009; Hopps et al., 2010). Among all, recent studies indicated that mineralocorticoid receptor (MR) signaling has a pivotal role in MetS-related tissue damages (Zennaro et al., 2009; Fujita, 2010; Nagase, 2010). MR and its classical ligand aldosterone mainly mediate sodium resorption in the distal parts of the nephron in the kidney, the intestine, the salivary, and sweat glands. In addition to this classical role, MR also functions in many other organs such as the heart, the vasculature, glomerular podocytes, and adipocytes, where MR mediates MetS-related inflammation and tissue damages via elevated oxidative stress (Nagase & Fujita, 2008; Odermatt & Atanasov, 2009; Zennaro et al., 2009; Nagase, 2010). Indeed, RALES (Randomized Aldactone Evaluation Study; Pitt et al., 1999) and EPHESUS (Eplerenone Post-AMI Heart Failure Efficacy and Survival Study; Pitt et al., 2003) showed that MR antagonists reduced cardiovascular events in human. We also previously reported that MR signaling mediated proteinuria and renal podocyte injury in the MetS model rats (Nagase et al., 2006, 2007). It is of note that MR has a high affinity for glucocorticoids as well as aldosterone (Arriza et al., 1987). Serum concentration of glucocorticoids (corticosterone in rodents) is much higher than that of aldosterone. MR activation by glucocorticoids is avoided by 11 β hydroxysteroid dehydrogenase type 2 (11βHSD2), which metabolizes glucocorticoids into inactive derivatives, in the classical target organs of aldosterone/MR. Activity of 11βHSD2 is much lower in the extrarenal organs, suggesting that glucocorticoids are main ligands in MR signaling in various tissue damaged in MetS (Odermatt & Atanasov, 2009; Zennaro et al., 2009; Farman et al., 2010). Other mechanisms of aldosterone-independent MR signaling include small G protein Rac1 activation, which transfers MR in the nuclei (Shibata et al., 2008).
Although it has been known that MR is also expressed in the skin (Bookout et al., 2006), functions of aldosterone/glucocorticoids/MR system in the skin have not been fully investigated (Farman et al., 2010). A recent report revealed that MR is present in the human dermal fibroblasts, and that aldosterone mediates deposition of dermal collagen fibers and elastic fibers in MR-dependent and MR-independent manners, respectively (Mitts et al., 2010). As for the epidermis, MR is expressed in the human epidermis as well as the hair follicles and sweat glands (Kenouch et al., 1994). Furthermore, another recent paper reported transgenic mice overexpressing MR gene under Keratin 5 promotor (K5-MR mice) (Sainte Marie et al., 2007). The K5-MR mice showed the atrophic epidermis, premature skin barrier function and alopecia, potentially resembling the senile skin. Although their report suggested involvement of MR signaling in the skin aging, this model is highly artificial with excess MR expression more than the physiological level. Roles of endogenous MR signaling in the aged skin are yet to be elucidated.
Based on the knowledge of MR involvement in the MetS pathophysiology, we speculated that endogenous MR signaling may mediate MetS-related tissue damage also in the skin, resulting in the aging-like skin changes in our UV-irradiated MetS mice model. The purpose of this study was to examine whether oxidative stress and inflammation were elevated in the MetS mice, and MR signaling are involved in our aging-like skin model using MetS mice. We determined upregulation of MR signaling in our model, and the aging-like skin changes were attenuated by application of antioxidant and MR antagonist spironolactone. These data suggested an unexpected, pivotal role of endogenous MR signaling in the skin aging.
Oxidative stress/inflammatory markers are elevated in metabolic syndrome (TSOD) mice skin, and 8-hydroxydeoxyguanosine (8OHdG) staining is enhanced in the aging-like skin changes evoked by single Uv irradiation on TSOD mice
In our aging-like skin model, we used 12-week-old male Tsumura-Suzuki obese diabetes (TSOD) mice (Hirayama et al., 1999; Suzuki et al., 1999; Akase et al., 2011, 2012). Twelve-week-old male Tsumura-Suzuki non-obesity (TSNO) mice were used as the non-MetS control. TSOD mice provide an animal model of MetS, with symptoms such as visceral fat accumulation, insulin resistance, glucose and lipid metabolism disorders, a lower metabolic rate, and hypertension. TSNO mice are derived from the same ancestors as TSOD mice but do not develop such symptoms (Fig. S1).
We irradiated single UV for 90 min (20 J m−2) onto the dorsal skin of the 12-week-old TSNO and TSOD mice. As reported previously (Akase et al., 2012), histological observation revealed that the skin of TSOD mice 24 h after UV irradiation showed greater susceptibility to UV (Fig. 1A–D) with remarkable infiltration of inflammatory cells (arrowheads in Fig. 1D). Taken together with TNF-α upregulation in the TSOD+UV group as reported previously (Akase et al., 2012), we designated this group as the aging-like skin model.
To examine whether elevated oxidative stress and inflammation were associated with the MetS status in TSOD mice, we executed real-time RT-PCR of an oxidative stress marker heme oxygenase 1 (Hmox1) (Maines, 1997) and an inflammation marker cyclooxygenase 2 (Cox2). The mRNA expression of Hmox1 was significantly upregulated in the TSOD and TSOD+UV groups than in the TSNO and TSNO+UV groups, respectively (Fig. 1E). The expression of Cox2 was also significantly higher in the TSOD and TSOD+UV groups than in the TSNO and TSNO+UV groups, respectively (Fig. 1F). These two markers were unchanged by UV irradiation.
In search of oxidative stress markers that are upregulated in our aging-like skin model, we executed immunostaining of 8-hydroxydeoxyguanosine (8-OHdG), another marker for oxidative stress. There was a faint 8-OHdG staining in the epidermis of TSOD and TSNO+UV groups (arrowheads in Fig. 1H,I). Notably, the TSOD+UV group showed remarkably stronger 8-OHdG staining in both the epidermis and dermis (Fig. 1J) compared with the other three groups (Fig. 1G–I). These data suggested that oxidative stress and inflammation were stronger in the TSOD skin than in the TSNO skin, and oxidative stress was further enhanced in the aging-like TSOD+UV group skin.
Topical application of antioxidant ameliorates the aging-like skin changes in the TSOD+UV mice
To examine whether oxidative stress really mediates the aging-like skin changes seen in the TSOD+UV group more precisely, we next applied antioxidant on the back skin before and after UV irradiation. We selected Tempol as antioxidant, because it was used for reducing proteinuria and the renal podocyte injury in the MetS rat model in our previous studies (Nagase et al., 2006, 2007). Only the vehicle was applied to the control TSOD+UV group. The aging-like skin inflammation was again observed in the TSOD+UV+Vehicle group (Fig. 2A,B), and Tempol application ameliorated these changes to the nearly normal condition (Fig. 2C). 8OHdG immunostaining revealed that oxidative stress was stronger in the TSOD+UV+Vehicle group (Fig. 2D,E), and it was clearly reduced by Tempol application (Fig. 2F). The mRNA expressions of Hmox1 and Cox2 genes were also significantly reduced in the TSOD+UV+Tempol group than in the TSOD+UV+Vehicle group (Fig. 3G,H).
Taken together, these results suggested that oxidative stress may substantially contribute to the mechanisms of the aging-like skin changes in the TSOD mice with single UV.
MR expression is upregulated in the TSOD skin and MR signaling is activated in the aging-like skin in the TSOD+UV mice
We next focused on MR signaling as a mechanism of enhanced oxidative stress in the aging-like skin changes, because we previously reported that MR signaling mediates oxidative stress in the renal damages in the MetS rats (Nagase et al., 2006). We confirmed localization of MR immunofluorescence in the TSOD and TSOD+UV group skin (Fig. 3A–L), particularly in the epidermis (arrowheads in Fig. 3B) and around the hair follicles (arrows in Fig. 3B and F), as reported previously in human (Kenouch et al., 1994). MR is a nuclear receptor, and it is transferred into the nucleus when its signaling is activated. Double stained nuclei with MR and 4′,6-diamidino-2-phenylindole (DAPI) were observed in the TSOD/TSOD+UV groups in the epidermis (arrowheads in Fig. 3D) and in the area surrounding the hair follicles (arrows in Fig. 3D and H). Notably, MR was strongly stained in the damaged area of the superficial skin, where 8OHdG was present, in the TSOD+UV group (Fig. 3J). Dense cluster of nuclei was observed in this area (Fig. 3K), possibly suggesting that MR was positive in the infiltrated inflammatory cells (Fig. 3L).
To observe nuclear localization of MR around the hair follicle more precisely in the TSOD+UV skin, we executed observation using a confocal microscope. A double stained nucleus with MR and DAPI was focused in the X-Y plane (Fig. S2A). Three dimensional images with the Y-Z plane (Fig. S2B) and the X-Z plane (Fig. S2C) clearly demonstrated that the double staining was not due to overlap of the distinct two red and blue spots, but MR was really translocated into the nucleus.
Interestingly, the mRNA expression level of MR in the TSOD and TSOD+UV groups was significantly upregulated than in the TSNO group (Fig. 3P). MR upregulation was not significantly related to UV irradiation in either the TSNO or TSOD mice (Fig. 3P). However, the expression of a well-known MR downstream gene, serum- and glucocorticoid-inducible kinase 1 (Sgk1) (Nagase et al., 2006, 2007; Shibata et al., 2008), was significantly enhanced in the TSOD+UV group compared with the other three groups (Fig. 3Q), clearly indicating MR signaling activation in the TSOD+UV group.
Serum concentration of aldosterone, a classical ligand for MR, was significantly elevated in the TSOD mice than in the TSNO mice (Fig. S3A). Serum concentration of corticosterone, which has a similar affinity for MR, was also significantly elevated in the TSOD mice than in the TSNO mice (Fig. S3B). The mRNA expression of 11βHsd2 gene in the skin of TSNO and TSOD mice was significantly lower than in the kidney (Fig. S3C), suggesting that the main ligand of MR is corticosterone in the mouse skin.
Taken altogether, these results suggested that endogenous MR expression was upregulated in the TSOD groups, and MR signaling was definitely activated in the aging-like skin of the TSOD+UV group, possibly by the corticosterone as the main ligand.
MR blockade by topical spironolactone reduces oxidative stress and ameliorates the aging-like skin changes in the TSOD+UV mice
We examined further whether endogenous MR signaling has a substantial role in the aging-like skin change in the UV-irradiated TSOD mice using MR antagonist spironolactone. Spironolactone was topically applied to the back skin before and after UV irradiation, and only the vehicle was again applied on the control TSOD+UV group. Histologically, application of spironolactone remarkably reduced the aging-like skin changes to the nearly normal condition (Fig. 4A–C). 8OHdG immunohistochemistry also revealed reduced intensity of staining by spironolactone (Fig. 4D–F).
The mRNA expressions of Sgk1 gene was significantly diminished by spironolactone (Fig. 4G), suggesting that endogenous MR signaling was well blocked by this intervention. Hmox1 and Cox2 expressions were also significantly decreased by spironolactone (Fig. 4H,I). Interestingly, Sgk1 expression was also reduced by Tempol (Fig. 4G), suggesting that oxidative stress itself upregulates MR signaling. These results suggested that MR signaling has an integral role for mechanism of the UV-induced aging-like skin changes in the TSOD mice via oxidative stress.
Long-term UV on the TSNO mice also induces aging-like skin changes but It is not associated with upregulation of MR signaling
Finally, we performed long-term UV irradiation on the TSNO mice to observe to what extent our aging-like skin model using the TSOD+single UV is equivalent to the more usual photoaging protocol with long-term UV onto the non-MetS animals. We firstly confirmed that the single UV on the TSNO mice with or without Tempol and spironolactone caused only negligible effects (Fig. S4). Then, we irradiated total of 20 J m−2 UV for 15 min x 6 times on the TSNO mice for the period of 2 weeks, which is reasonably long as UV duration in the photoaging models using Hairless mice (Oberyszyn et al., 1998; Inomata et al., 2003).
The histological changes were observed after 2 week of UV with inflammatory cell infiltration (Fig. 5A,B) and clear 8HdG staining (Fig. 5C,D) on the epidermal surface of the TSNO mice. Hmox1 (Fig. 5E) and Cox2 (Fig. 5F) expressions were upregulated, suggesting that this protocol of 2-week UV on the TSNO mice produced the skin damages like our TSOD+single UV model. However, interestingly, expressions of MR and Sgk1 were unchanged after long-term UV on the TSNO (Fig. 5G,H), suggesting the possibility that MR signaling is involved especially in the MetS-related skin aging.
This is the first study to demonstrate an unpredicted, essential role of endogenous MR signaling in the aging-like skin changes in the MetS model animals.
Oxidative stress and inflammation are common mediators of MetS and aging. (Franceschi et al., 2000; Golden et al., 2002; Frisard & Ravussin, 2006; Dröge & Schipper, 2007; Franceschi et al., 2007; Goto, 2008; Roberts & Sindhu, 2009; Hopps et al., 2010). Upregulation of Homx1/Cox2 in the TSOD/TSOD+UV groups compared with the TSNO/TSNO+UV groups was irrelevant of UV irradiation. We interpret these results in the following ways: (i) TSOD mice are in proinflammatory status due to MetS resulting in Hmox1/Cox2 upregulation even without UV. (ii) Hmox1 upregulation after UV might be transient and recovered to the basal level after 24 h, as reported in the human fibroblasts in vitro (Keyse & Tyrrell, 1989). (iii) Induction of oxidative stress and inflammation occurred only in the superficial skin by UV, but the sampling for RT-PCR included the total layer of the skin, resulting in the negligible Homx1/Cox2 upregulation in the UV groups. Further studies are needed to clarify which possibility is correct. Nevertheless, we consider that the TSOD+UV group can be regarded as the aging-like model (Akase et al., 2012), because the infiltration of inflammatory cells and intense 8OHdG staining were observed. 8OHdG is a degradate of damaged DNA by oxidative stress and is chemically stable over 24 h (Matsumoto et al., 2008), more precisely reflecting elevated oxidative stress than Homx1 in the TSOD+UV group.
Involvement of oxidative stress in our model was further supported by administration of antioxidant Tempol. Lower Hmox1 expression in the TSOD+UV+Tempol group than in the TSOD group might be due to the reduction of basal oxidative stress level. Although toxicity of Tempol cannot be fully ruled out, it is not likely to be toxic in the TSNO experiments (Fig. S4). Cox2 upregulation in the TSOD+UV+Vehicle group compared with the TSOD group (Figs 2H and 4I) suggests possible effect of vehicle, but further analyses should be necessary to elucidate this point.
Mineralocorticoid receptor is known to be expressed in the skin and its appendages (Kenouch et al., 1994; Bookout et al., 2006; Mitts et al., 2010), and the K5-MR transgenic mice showed aging-like skin atrophy (Sainte Marie et al., 2007). MR upregulation in the skin in the MetS model mice was our novel finding. Although MR mRNA expression level was again unchanged by UV, MR immunostaining was strongly positive in the damaged skin in the TSOD+UV group, just like the 8OHdG staining patterns. This suggests that MR positive inflammatory cells were recruited to the damaged skin (Usher et al., 2010). It is also known that MR mRNA and protein expression levels are not necessarily parallel due to the posttranslational modification (Odermatt & Atanasov, 2009). Further studies are needed addressing this point. Considering the nuclear localization of MR around the hair follicle and Sgk1 upregulation together, it is reasonable to consider that MR signaling is activated in the TSOD+UV group. We speculate that corticosterone, rather than aldosterone, may be a dominant ligand in the skin, because 11βHsd2 expression was significantly lower in the skin than in the kidney.
Mineralocorticoid receptor antagonist spironolactone has affinities also for glucocorticoid receptor or androgen receptor, with lesser specificity for MR than the more selective MR antagonist eplerenone (Baxter et al., 2004; Mulatero et al., 2006; McManus et al., 2008). However, a previous report suggested that the inhibition of glucocorticoid receptor by spironolactone might be negligible (Mulatero et al., 2006). The dose of spironolactone in our study (20%) was extremely high compared to the 5% topical spironolactone for human acne (Afzali et al., 2012). However, dose for human may be determined for avoiding possible side effects (gynecomastia etc.). Considering poor permeability of the skin, we think that our dose of spironolactone is reasonable for animal experiments. MR activation, elevated oxidative stress, and tissue inflammation were also reported in the visceral adipose tissue in obese mice (Guo et al., 2008; Hirata et al., 2009), suggesting that there is a common molecular scenario in the skin and adipose tissue, as well as other target organs in MetS status.
The long-term UV damages of the TSNO skin without expression change of MR signaling genes suggest that UV-induced oxidative stress may be mediated by other pathways than MR signaling in the non-NS skin. In other words, it is suggested that MR signaling may have a special role in MetS-related pathological aging of the skin. Future studies of this long-term UV model, including application of Tempol and spironolactone, will provide us further insights regarding this point.
TSNO and TSOD mice
The 12-week-old male TSNO and TSOD mice were obtained from the Institute of Animal Reproduction (Ibaraki, Japan). The animal care conditions included room temperature 23 ± 1°C, humidity 55 ± 5%, and lights-on at 9 am and lights-off at 9 pm. The animals were allowed free access to sterilized water. Animals were given ad libitum MF (Oriental Yeast Co., Ltd., Chiba, Japan). All animals were handled in accordance with the guidelines of the National Research Council (1996). This study was approved by the experimental animal ethics committees of the University of Tokyo.
Single UV irradiation
Ultraviolet irradiation was performed as described previously (Akase et al., 2012). Briefly, we shaved hair from the backs of the 12-week-old TSNO and TSOD mice using a hair remover and exposed the naked skin to UV irradiation (290–380 nm, maximum 312 nm) for 90 min using a UV lamp at a distance of 20 cm (Kyokko Denki, Tokyo, Japan), for a total dose of 20 J m−2. We sacrificed the animals 24 h after UV irradiation. We obtained irradiated skin samples from the backs of the mice. We maintained mice in the TSNO and TSOD control groups for 24 h without UV irradiation and then collected skin samples as described previously.
Topical application of tempol and spironolactone
1M Tempol (Sigma-Aldrich, St Louis, MO, USA) and 20% spironolactone (Wako) were dissolved in the vehicle (propylene glycol/glycerin/water = 10:25:45). The vehicle, Tempol, and spironolactone were topically applied on the back of the TSNO and TSOD mice at 200 μL per 5 cm2 three times per each animal; one hour before the 90-min UV irradiation, just after the irradiation and four hours after the irradiation.
Histology and immunohistochemistry
The 5-μm-thick skin sections were deparaffinized in xylene, rehydrated in ethanol, and washed in distilled water. Hematoxylin and eosin staining was performed on the skin samples.
8OHdG immunostaining was performed as previously described (Akase et al., 2012). Briefly, we incubated the slides with anti-8OHdG mouse monoclonal antibody (Japan Institute for the Control of Aging, Shizuoka, Japan; diluted 1:200) at room temperature for 60 min. After washing, we subsequently incubated sections with biotin-conjugated anti-mouse IgG antibody (Vector Laboratories, Burlingame, CA, USA; diluted 1:1000) for 30 min at room temperature. We detected immunoreactions using a VectaStain ABC Kit (Vector Laboratories) with 3,30-diaminobenzidine tetrahydrochloride substrate (Nacalai Tesque, Kyoto, Japan) and counterstained using hematoxylin.
Mineralocorticoid receptor fluorescent immunostaining was performed using anti-rat MR mouse monoclonal antibody (clone 1D5, a gift from Dr. Celso Gomez-Sanchez, University of Mississipi; diluted 1:50), cyanine 3 tyramide signal amplification (TSA) systems (Perkin Elmers Inc., Waltham, MA, USA), and DAPI-Fluoromount-G (SouthernBiotech, Birmingham, AL, USA) as nuclear staining. The samples were observed using a fluorescent microscope (Axioplan 2 imaging, Carl Zeiss Light Microscopy, Göttingen, Germany). MR immunostaining for confocal microscopy was performed using anti-MR mouse monoclonal antibody (Perseus Proteomics, Tokyo, Japan; diluted 1:50), the M.O.M Immunodetection Kit (Vector Laboratories) and TEXAS RED AVIDIN DCS (Vector Laboratories; diluted 1:100) as fluorescent label. The samples were observed using a confocal microscope (DMIRE2 and TCS SP2, Leica Microsystems, Wetzlar, Germany).
Quantitative real-time RT-PCR
Total RNA was extracted from the skin and mouse kidney samples using an RNeasy Lipid Tissue Mini Kit (QIAGEN, Hilden, Germany). Reverse transcription (RT) was performed using a MJ Mini thermal cycler (Bio-Rad, Richmond, CA, USA) and a High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA, USA). For quantitative PCR, the amplification of the target specific region of cDNA was performed using THUNDERBIRD CYBR qPCR Mix (TOYOBO, Osaka, Japan) by 40 cycles of 95°C for 30 s, 95°C for 5 s, and 60°C for 30 s, and monitored with a real-time PCR system (LightCycler, Roche, Indianapolis, IN, USA). The gene expression levels were quantified by the comparative Ct method and normalized for the expression of the internal control, glyceraldehyde-3-phosphate dehydrogenase (Gapdh). The primer sequences are shown in Supporting Data Table S1.
Long-term UV irradiation onto the TSNO mice
The 14-week-old male TSNO mice were used. UV irradiation was performed for 15 min per once, 3 times per week, and for 2 weeks. Thus, total dose of 20 J m−2 was the same as the single UV experiments. Hair was removed every time of UV irradiation. The animals were sacrificed 24 h after the last UV, and the tissues were processed as described previously. The control TSNO mice were maintained for 2 weeks without UV and then sacrificed.
Results were shown as mean ± SD. Student's t-test was used for analyses between the two groups. ANOVA and the Bonferroni corrections were used for analyses among the more than two groups. Values of P < 0.05 were considered to be statistically significant. All statistical analyses were performed using SPSS ver.16.0 (Japan IBM, Tokyo, Japan).
The authors express sincere thanks to Dr. Celso Gomez-Sanches for providing us a MR monoclonal antibody, Dr. Hitomi Eto for her technical assistance for the confocal endoscope and to Professor Kiyoshi Kita for instructing us to use the real-time RT-PCR system. This work was supported by a grant-in-aid for Scientific Research from MEXT (Ministry of Education Culture, Sports, Science and Technology) (No. 21659494). There is no conflict of interest.
TN involved in study design and drafting the manuscript. TA involved in data collection, data analysis, statistical analysis, and the drafting manuscript. HS, TM, KY, MN, GN, and JS were involved in study design. AI carried out quantitative PCR. LH carried out histological analysis. M Asada carried out RT-PCR. TS and M Aburada involved in analysis of metabolic parameters.