Cyclophilin D participates in the inhibitory effect of high‐fat diet on the expression of steroidogenic acute regulatory protein

Abstract Objective The high‐fat diet (HFD)–induced obesity is responsible for the testosterone deficiency (TD). However, the mechanism remains unknown. Mitochondrial homeostasis is proved to be important for maintaining the function of steroidogenic acute regulatory protein (StAR), the first rate‐limiting enzyme in testosterone synthesis. As the key regulator of mitochondrial membrane permeability, cyclophilin D (CypD) plays a crucial role in maintaining mitochondrial function. In this study, we sought to elucidate the role of CypD in the expression of StAR affected by HFD. Methods To analyse the influence of CypD on StAR in vivo and in vitro, mouse models of HFD, CypD overexpression and CypD knockout (Ppif −/−) as well as Leydig cells treated with palmitic acid (PA) and CypD overexpression plasmids were examined with an array of metabolic, mitochondrial function and molecular assays. Results Compared with the normal diet mice, consistent with reduced testosterone in testes, the expressions of StAR in both mRNA and protein levels in HFD mice were down‐regulated, while expressions of CypD were up‐regulated. High‐fat intake impaired mitochondrial function with the decrease in StAR in Leydig cells. Overexpression of CypD inhibited StAR expressions in vivo and in vitro. Compared with C57BL/6 mice with HFD, expressions of StAR were improved in Ppif −/− mice with HFD. Conclusions Mitochondrial CypD involved in the inhibitory effect of HFD on StAR expression in testes.


| INTRODUC TI ON
Testosterone is the most important androgen for male, exerting important physiological effects. 1 In adult men, testosterone deficiency (TD, total testosterone <300 ng/dL is most commonly used as the cut-off values for low testosterone) 2 not only has adverse effects on sexual and reproductive function 3 but also negatively affects glucose and lipid metabolism, cardiometabolic functions, body composition, bone mineral density and quality of life. [4][5][6] The global prevalence of TD in adult men ranges from 10% to 40%; among these, the incidence in the United States varies from 24% to 39%, and it ranges from 17% to 33% in China. 2 With the economic development and life improvement, people are more and more exposed to longtime and excessive high-fat diet (HFD) and sedentary lifestyle, which could increase bodyweight, induce glucose intolerance and insulin resistance, and cause dyslipidaemia. 7,8 In addition, long-time HFD has been reported to decrease testosterone levels 9,10 ; however, the mechanisms underlying have not yet been fully elucidated.
Testosterone, a steroid hormone, is mainly (95%) synthesized and secreted by the testicular Leydig cells. 11 In response to luteinizing hormone (LH), testicular testosterone biosynthesis is a multi-step process in which there are a series of steroidogenic proteins and enzymes, such as the steroidogenic acute regulatory protein (StAR), a cholesterol side-chain cleavage enzyme (P450scc or CYP11A) and 3β-hydroxysteroid dehydrogenase (3β-HSD). 12,13 StAR has been shown to accelerate the transport of cholesterol, raw materials for testosterone synthesis, from the outer mitochondrial membrane (OMM) to the inner mitochondrial membrane (IMM), which is the first and rate-limiting step of the testosterone biosynthesis. 14,15 Both the people suffer from congenital lipoid adrenal hyperplasia (lipoid CAH) caused by mutations in the StAR gene and StAR null mice almost completely fail to synthesize steroids, which indicates that the StAR protein is an indispensable ingredient in the process of steroid synthesis. 16 A previous study reported that both pharmacological inhibition of ATP synthesis and dissipation of mitochondrial membrane potential (Δψm) could inhibit steroidogenesis owing to the reduced expression of StAR, 17 proving that the StAR activity requires mitochondrial homeostasis. It is reported that HFD could down-regulate StAR expression in mice. 18 However, whether mitochondrial dyshomeostasis is involved in HFD-induced StAR expression down-regulation is still unclear.
Transient opening of the mitochondrial permeability transition pore (mPTP), a tunnel-like multiprotein complex spanning across the OMM and IMM, 19 in response to normal stimuli is important to maintain mitochondrial homeostasis. [20][21][22] However, under conditions of extreme detrimental stress, irreversible opening of mPTP leads to mitochondrial dysfunction characterized by the dissipation of Δψm, the uncoupling of oxidative phosphorylation followed with the inhibition of ATP generation, increased generation of reactive oxygen species (ROS), mitochondrial swelling, rupture and death. [21][22][23] Although molecular components of the mPTP have been still vague to date, 22 cyclophilin D (CypD), a peptidyl prolyl cis-trans isomerase F (PPIF) in mitochondrial matrix, has been demonstrated from gene deletion studies to be an indispensable regulator for mPTP opening. [24][25][26] At present, CypD has been shown to translocate from matrix to the IMM and bind to major component candidates of mPTP, including adenine nucleotide translocase (ANT), phosphate carrier (PiC) and oligomycin sensitivity conferring protein of the FoF1 ATP synthase (OSCP) 24,27,28 across the IMM, which is the foremost step in the process of facilitating PTP opening in response to various stress stimuli. 23,29 Xiaolei Wang et al have shown that HFD could elevate CypD protein expression levels and could induce the mitochondrial dysfunction in the mouse liver; meanwhile, CypD ablation could ameliorate HFD-induced hepatic mitochondrial stress, demonstrating that HFD could cause mitochondrial dysfunction by up-regulating the CypD expression in liver. 30 Therefore, we hypothesize that CypD-dependent mitochondrial dysfunction in Leydig cells of testes is involved in HFD-induced StAR expression deregulation.
Here, to testify and dissect the hypotheses, the expression of CypD and StAR was determined in vivo and in vitro with different administrations including HFD, CypD overexpression and CypD knockout. Meanwhile, through a series of measurements of mitochondrial function, the modulation of HFD on mitochondria is analysed. In the present study, we report for the first time that upregulation of CypD expression levels in Leydig cells of testes induced the mitochondrial dysfunction and participates in the HFD-induced StAR expression deregulation. CypD might be a new therapeutic target in the high-fat diet-induced TD. Ppif −/− mice (mice null for Ppif, the gene encoding CypD) on the same background with C57BL/6 mice were provided by Dr Heng Du.

| Animals and diets
Ppif −/− mice were backcrossed to the C57BL/6 strain for ten generations to obtain stable ppif heterozygosis (HE) male and female mice.
Then, C57BL/6 mice (Ppif +/+ mice) and Ppif −/− mice in each brood were gained through sexual hybridization of stable HE male and female mice. At the 8th week of age, both these two kinds of mice were fed with either normal diet or high-fat diet, respectively. After 16 weeks, the mice were killed for the subsequent experiments.
In addition, Ad-PPIF (CypD overexpression adenovirus) and Ad-EGFP (empty vector adenovirus) were dissolved in sterile PBS and were injected through the caudal vein to the C57BL/6 mice at the eighth week of age. The adenovirus was given three times within a 6-day interval. After 3 weeks, these mice were killed for the subsequent experiments.
In the appropriate weeks of age mentioned above, the mice were anaesthetized with pentobarbital sodium and the blood samples were collected from eyes. Simultaneously, the testes were isolated and were preserved in liquid nitrogen rapidly or modified Davidson's fluid (mDF, containing 30% of a 37%-40% solution of formaldehyde, 15% ethanol, 5% glacial acetic acid and 50% distilled H 2 O) for paraffin sections. Some testes were rapidly cut into pieces of 1 mm × 1 mm × 1 mm sections and then were fixed with 2.5% glutaraldehyde in 0.1 M phosphate buffer for following transmission electron microscope analysis. For paraffin sections, testes preserved in mDF were embedded in paraffin and cut to 5-μm sections for immunofluorescence. For frozen sections, testes were cut at 5 μm thickness with the microtome portion of the cryostat, which were prepared for the subsequent Oil Red staining.

| The culture and treatment of mouse TM3 Leydig cells
The mouse TM3 Leydig cell line was derived from normal immature mouse testis and was purchased from Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences. TM3 cells were cultured in the petri dishes with a 1:1 (v:v) mixture of DMEM:F12 media supplemented with 0.5% antibiotics, 2.5 mmol/L l-glutamine, 0.5 mmol/L sodium pyruvate, 5% horse serum and 2.5% foetal bovine serum. The cells were cultured in a 5% CO 2 incubator at 37°C. For detection of the effect of lipid on the process of steroidogenesis in vitro, the cells were co-cultured with different dosages (0, 0.2, 0.4 mmol/L) of PA, respectively, for 24 hours. And then, the 1 U/ml hCG was added into the petri dishes 4 hours early before the cell collection.
Furthermore, to assess the effect of CypD overexpression on the testosterone synthesis, cells were transfected with pLV-PPIF (human CypD overexpression plasmid, Cyagen Biosciences Inc) using FuGENE ® HD Transfection Reagent as described by the manufacturer's instruction. The empty plasmids (pLV-EGFP) were used as control. The media were changed after 8-12 hours, and after 24-hour culture, cells were collected for further analysis. Moreover, cells were treated with 0 and 0.4 mmol/L of PA for 24 hours, with or without a 1 hours pretreatment of cyclosporin A (CSA, an inhibitor of CypD, 1 μmol/L). Similarly, cells were collected for following analysis.

| Detection of body and perineal fat distribution
To compare body fat distribution of whole body and perineum between the ND and HFD groups, X-ray analysis was performed

| Measurement of lipid profile and sex hormone
The mouse serum was separated after centrifugation.

| Transmission electron microscope analysis
For transmission electron microscope analysis, fixed testicular sections were washed three times with phosphate buffer and were fixed in 1% osmium tetraoxide for 1 hour. The sections were dehydrated by a graded series of acetone and embedded in araldite. Ultrathin sections were cut with a diamond knife, stained with uranyl acetate and lead citrate, and viewed with an transmission electron microscope (JEM-1200EX, JEOL Co., Japan).

| Oil Red staining
To detect the lipid accumulation, the 5-μm-thickness frozen slices of testes and the slides of TM3 cells were stained with Oil Red and observed with the light microimaging system. Meanwhile, the area of Oil Red staining was quantified by ImageJ software.

| RNA isolation and quantitative reverse transcription-polymerase chain reaction (qRT-PCR)
Total RNA of the testicular tissues or TM3 cells was extracted by using RNAiso Reagent, which subsequently was subjected to reverse transcription kit (PrimeScript ® RT Reagent Kit (Perfect Real Time), TaKaRa, Japan) according to the manufacturer's protocol and was transformed to cDNA with common PCR instrument (Eppendorf Scientific. Inc) at 37°C for 15 minutes. The cDNA was used as a template for qRT-PCR conducted with the LightCycler480 (Roche). The cycling parameters were 95°C for 10 minutes, 40 cycles at 95°C for 15 seconds; 60°C for 60 seconds. The oligonucleotide primers used are shown in Table S1. Each specimen was analysed in triplicate. Transcript levels were normalized to those of β-actin, and relative expression levels were calculated using the 2 -ΔΔCt method.

| Immunoblotting analysis
Total protein was extracted from testicular tissues or Leydig cells

| Immunofluorescence
For immunofluorescence, fixed tissues were heat treated in Tris-EDTA buffer for antigen retrieval. After washed by PBS, sections were blocked with 5% goat serum albumin prior to the addition of the primary antibody: rabbit anti-StAR (1:100), and then incubated overnight at 4°C. After that, paraffin sections were incubated with the FITC-conjugated goat anti-rabbit IgG antibody (1:50, Sigma-Aldrich) for 1 hour at room temperature. The nuclei were stained with DAPI (Cat. C1006, Beyotime). After treatment at least five consecutive fields, each slice was observed with a fluorescence microscope (Axio Imager A2, Zeiss, Jena, Germany).

| Cell Mito stress test
The mitochondria energy metabolism of TM3 cells was determined with Agilent Seahorse XF96 Extracellular Flux Analyzer

| Mitochondrial ATP assay
The ATP levels of Leydig cells were measured by using a firefly luciferase based on ATP assay kit (S0026, Beyotime) according to the manufacturer's instructions. The fluorescence luminance (RLU) was measured with the luminometer (Multimode Reader LB942 TriStar2, BERTHOLD, Germany). The unit of ATP was corrected to nmol/mg with concentration of mitochondrial protein.

| Assay of reactive oxygen species generation in mitochondria
After treatment with PA in 96-well plates, TM3 cells were washed with PBS three times. As per the manufacturer's direction, 100 μL MitoSOX working solution was added into each well and then the plate was placed in the incubator at 37°C away from light for 10 min.
And then, mitotracker for the mitochondria detection presenting green fluorescence was added into the medium in proportion as 1:10 000. After 10-minute incubation the same as the above, the working solution was removed and the plate was washed with PBS three times again. Next, every 100 μL of Hochest mixed solution with DMEM basic medium (1:10) for the cell nucleus detection revealed as blue fluorescence was added and the plate was placed in the incubator at 37°C away from light for 3 minutes. The fluorescence situation was observed with the fluorescence microscope.

| Statistical analyses
Statistical analyses were conducted using GraphPad Prism (v.5, GraphPad Software Inc, CA). All samples were determined by an independent-sample t test or one-way analysis of variance (ANOVA) followed by the least significant difference (LSD) multiple comparison test. All data were expressed as mean ± SEM, and the difference was considered to be statistically significant if P < 0.05.

| General conditions and lipid deposition in Leydig cells of mice with high-fat intake
To analyse the effects of HFD on the testosterone levels, 8-weekold mice were given HFD or ND for 16 weeks. Compared with the F I G U R E 1 General conditions and lipid deposition in Leydig cells of mice with high-fat intake. A, Comparison of bodyweight of mice fed with ND or HFD during 16 weeks. B, Ratio of fat distribution of whole body or perineum in the ND or HFD group at the 16th week of feeding. C, Representative photographs of somatotype, testes and epididymal fat of mice fed with ND or HFD at the end of the 16th week. D, Comparison of serum lipid levels between the ND and HFD groups. E, Comparison of serum sex hormone levels between the ND and HFD groups, including testosterone (T) and LH at the end of the 16th week. T: n = 10; LH: n = 5. F, Representative Oil Red staining sections and the area quantitative analysis showing lipid deposition of interstitial tissue of testes from mice in the ND or HFD group at the 16th week. G, Intratesticular testosterone levels of mice in the ND or HFD group. H, Representative Oil Red staining and the area quantitative analysis showing lipid accumulation in the Leydig cells with treatments of 0, 0.2 and 0.4 mmol/L of PA, respectively. I, Testosterone concentrations in media of TM3 cells exposed to 0, 0.2 and 0.4 mmol/L of PA for 24 hours. All data were represented as mean ± SEM, n = 4-11 for each group. ** P < 0.01 vs ND or 0 PA group, and ## P < 0.01 vs 0.2 PA group were considered highly significant difference from the control ND group, the weight of mice in the HFD group increased significantly from the 4th week after being fed with high fat and continued to increase with the prolonged feeding time. By the 16th week, the average weight tended to be stable in the HFD group and it was 1.77 times of the ND group (P < 0.01; Figure 1A. The body fat distribution analysis with X-ray further testified that the average percentage of body fat in perineum, around the testes, was 46.08% in the HFD group versus 18.70% in the ND group, showing the HFD group was 2.46 times of the ND group (P < 0.01, Figure 1B Figure 1D. As shown in Figure 1E, serum sex hormone detection revealed no significant change in both testosterone and LH in the HFD group compared with the ND group. In consideration of testicular Leydig cells as the main place to synthesize and secrete testosterone in male, the role of HFD on intratesticular testosterone levels was further assessed. Oil Red staining of the testes showed that the lipid accumulations were found mainly in the Leydig cells. Meanwhile, there was more visible lipid deposition in interstitial tissue of testes in the HFD group than that in the ND group and quantitative analysis supported the above conclusion (P < 0.01, Figure 1F. It was noticeable that there was an obvious decline in the intratesticular testosterone levels in the HFD group in contrast to the ND group (P < 0.01, Figure 1G).  Figure 1I).

F I G U R E 2 High-fat intake down-regulates the expression of StAR in Leydig cells. A, qRT-PCR analysis of the expression of StAR, P450scc
and 3β-HSD mRNA in testes from mice in the ND or HFD group. B, Immunoblot and quantitative analysis of StAR protein in testes of mice in the ND or HFD group. C, Representative immunofluorescence assay showing the localization and quantitative analysis of StAR in testicular interstitial tissue from mice in the ND or HFD group. Green fluorescence represented StAR and blue fluorescence represented cell nucleus. D, qRT-PCR analysis showing the relative expression of StAR mRNA and E, immunoblot and quantitative analysis of StAR protein abundance in TM3 cells treated with 0, 0.2 and 0.4 mmol/L of PA, respectively. Representative images for quantitative immunoblot are shown. ND, n = 3 and HFD, n = 3. Data were represented as mean ± SEM. * P < 0.05 and ** P < 0.01 vs ND or 0 PA group were considered highly significant difference from the control

| High-fat intake down-regulates the expression of StAR in Leydig cells
StAR assists the transportation of cholesterol from the OMM to the IMM, which is the first and rate-limiting step in steroidogenesis process. And then, P450scc across the IMM catalyses cholesterol conversion to pregnenolone which next diffuses to the smooth endoplasmic reticulum where 3β-HSD facilitates pregnenolone to form the more biologically active androstenedione, the immediate precursor of testosterone. All the above three proteins or enzymes play key roles during the steroidogenesis process. Herein, the effects of HFD on StAR, P450scc and 3β-HSD were detected in mRNA levels firstly.  Figure 2E).

| High-fat intake has detrimental effects on the mitochondria in Leydig cells
Mitochondria are the essential organelles of ATP synthesis as well as testosterone synthesis, which dysfunction would disturb StAR activity. We further examined the effects of HFD or PA on the mitochon- suggesting that the mitochondrial function of the PA group was destroyed (P < 0.01, Figure 3B. Simultaneously, consistent with the detection of ATP in the Mito Stress Test, the ATP production measurement with kit displayed that ATP production of mitochondria in 0.4 mmol/L of the PA group was significantly lower than that in the control group (P < 0.05, Figure 3C. The results of MitoSOX staining showed that ROS production of the PA group was significantly more than that of the control group Figure 3D. It has been reported that

| High-fat intake up-regulates the expression of CypD in Leydig cells
Under the condition of stress, the overly increase in CypD expression caused the exceedingly opening of mPTP followed with mitochondrial dysfunction, which had a detrimental effect on the StAR activity. In order to detect whether CypD is involved in the HFDmediated mitochondrial dysfunction, we further determined the expression of CypD of testes, respectively, in mRNA and protein levels. Compared with the ND group, CypD expression in mRNA levels increased with significant statistical difference (P < 0.05; Figure 4A). Similarly, in protein levels, CypD expression in the HFD group was up-regulated and there was a statistical difference in the ratio of CypD protein level to β-actin compared with the ND group (P < 0.05, Figure 4B). Although in vitro the increase in CypD mRNA expression of TM3 cells was not statistically significant (P > 0.05, Figure 4C), the protein expression of CypD in the PA-treated group (0.2 or 0.4 mmol/L) was enhanced in contrast to the control group (P < 0.01, Figure 4D).

| CypD overexpression down-regulates expression of StAR
To further testify the role of CypD in regulating StAR expression, StAR expression was analysed after overexpression of CypD in vivo and in vitro. The CypD overexpression mouse model was made with Ad-PPIF by caudal vein injecting. As shown in Figure 5A, with increasing expression of CypD mRNA in testes, StAR mRNA levels decreased (P < 0.01). The levels of protein expressions were consistent with its mRNA levels (P < 0.05, Figure 5B). Then, the use of pLV

| Inhibition of CypD improves the inhibiting role of lipotoxicity on the StAR
To clarify whether HFD-induced CypD overexpression is a key point in the process of down-regulation StAR, the CypD complete knockout (Ppif −/− ) mouse model was applied in this study. In Figure 6A, the immunoblot revealed that, in comparison with the Ppif +/+ HFD group, the CypD in the Ppif −/− HFD group was certainly inhibited totally while the StAR in protein level has no significant difference, suggesting that HFD could not down-regulate StAR expression in the context of CypD deficiency.
The above assumption was further examined in vitro. It has been reported that cyclosporin A (CSA) was one of the nonspecific inhibitors on the mitochondrial oxidation and disturbed the opening of F I G U R E 4 High-fat intake upregulates the expression of CypD in Leydig cells. A, qRT-PCR analysis of the relative expression of CypD mRNA and B, representative images for immunoblot and quantitative analysis of CypD protein abundance of testes of mice in the ND or HFD group. C, qRT-PCR analysis showing the relative expression of CypD mRNA and D, representative images for immunoblot and quantitative analysis of CypD protein abundance in TM3 cells treated with 0, 0.2, 0.4 mmol/L of PA, respectively. Data were presented as mean ± SEM, n = 3 for each group. * P < 0.05 vs ND or 0 PA group was considered highly significant difference from the control the mPTP. It was used to demonstrate whether CSA could alleviate PA-induced StAR down-regulation by inhibiting the CypD expression in Leydig cells. As shown in Figure 6B and 6D, the mRNA and protein expressions of CypD in CSA pretreatment group were less than those in the control group, indicating that CSA inhibited CypD expression, while both mRNA and protein levels of StAR did not show differences between the control group and pretreatment group Figure 6C and 6D, suggesting that CSA alone has no influence F I G U R E 5 CypD overexpression down-regulates expression of StAR. A, qRT-PCR analysis of the relative expression of CypD and StAR mRNA and B, representative images for immunoblot and quantitative analysis of CypD and StAR protein abundance of testes of mice in the Ad-EGFP or Ad-PPIF group. C, Representative images for immunoblot and quantitative analysis of CypD and StAR protein abundance of TM3 cells in the pLV and pLV-PPIF group. The ratio of CypD and StAR protein level to β-actin was calculated. Data were presented as mean ± SEM, n = 3 for each group. * P < 0.05 and ** P < 0.01 vs Ad-EGFP group were considered highly significant difference from the control F I G U R E 6 Inhibition of CypD improves the inhibiting role of lipotoxicity on the StAR. A, Representative images for immunoblot and quantitative analysis of CypD and StAR protein abundance of testes of Ppif +/+ or Ppif −/− mice with HFD (n = 3). B and C, qRT-PCR analysis of the relative expression of CypD and StAR mRNA of TM3 cells exposed to PA with or without the CSA pretreatment. D, Representative images for immunoblot and densitometry analysis of CypD and StAR protein of TM3 cells exposed to PA with or without the CSA pretreatment. Data were represented as mean ± SEM. * P < 0.05 and ** P < 0.01 vs ppif +/+ HFD or control group, # P < 0.05 and ## P < 0.01 vs PA group were considered highly significant difference from the control on StAR. Interestingly, in the CSA + PA group in which TM3 cells were pretreated with CSA and subsequently exposed to PA, the protein level of CypD was less than that in the control group and was similar to that in the CSA group, showing that PA could not elevate again CypD expression inhibited by CSA in the protein level. In the same time, although both mRNA and protein levels of StAR in the CSA + PA group were still lower than the control group, its expression levels increased with statistical difference in comparison with the PA group Figure 6C

| D ISCUSS I ON
Clinical studies suggested a decreased testosterone level in serum in reproductive-age men with diet-induced obesity. 31 The Ppif +/+ mice and Ppif −/− mice were bred by sexual hybridization of Ppif ± male and female mice. Ppif −/− mice were born at the expected Mendelian frequency, developed normally, and did not have any detectable phenotype and viability anomalies. 25,26 We observed that there was apparent increase of StAR expression levels in Ppif −/− mice in the HFD group compared with those in Ppif +/+ mice in the HFD group, suggesting that the CypD through deletion could help StAR to resist to long-standing HFD down-regulated effects. Simultaneously, the pretreatment with CSA before PA treatment was performed to TM3 cells and the results indicated that the inhibition of CypD via CSA could alleviate PA noxious implication on the StAR expression. We therefore identified a possible mechanism that ablation of CypD attenuated the HFD-induced StAR down-regulation. However, the concrete mechanism through which CypD overexpression inhibits StAR expression is unclear and needs further investigation.
In conclusion, the present study demonstrated that CypD was involved in inhibition of StAR expression and activity by HFD.
However, there were still some deficiencies in this study. In the present study, the CypD overexpression model was performed by tail vein injection of Ad-PPIF, and the knockout model was global gene knockout, in which we could not rule out the influences of holistic factors. It has been reported that the roles of CypD in metabolic homeostasis were tissue-specific. 21 In the future, the Leydig cell-specific deletion of CypD models will be used.
Moreover, how CypD regulates the mPTP switch in Leydig cells still needs to be further explored. CypD plays important roles in tissue ischaemia-reperfusion injury, apoptosis and metabolic homeostasis in some cells; therefore, the studies of roles of CypD in testosterone synthesis as well as in metabolic homeostasis in Leydig cells could provide therapeutic target for HFD-induced testosterone deficiency.

ACK N OWLED G EM ENTS
We acknowledge Dr Heng Du for providing CypD knockout mice.

CO N FLI C T O F I NTE R E S T
There are no conflicts of interest to declare.

AUTH O R CO NTR I B UTI O N S
Y. C. and GQ designed the experiments. SX, LD and ZY performed experiments, testing and data analyses. YC and WX provided technical expertise. YC and SX wrote the manuscript. SX and WX organized the mice. X.X measured levels of testosterone and LH. SM and YJ administrated mice fed with HFD or ND. LD and ZM provided cell treatment skills. All authors edited the manuscript and provided comments.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data used to support the findings of this study are available from the corresponding author upon request.