Women who attended and were treated at the Reproductive Medicine Centre of Hebei Reproductive Health Hospital (Shijiazhuang, China) between June 2021 and June 2024 were included in the present study. Patients with PCOS (n=36) were diagnosed according to the Rotterdam criteria (21), and at least two of the following features were detected: Oligo-ovulation/anovulation, biochemical or clinical hyperandrogenism, and polycystic ovarian morphology on ultrasound. The exclusion criteria for the control and PCOS groups included the following: i) Ovarian cysts; ii) diminished ovarian reserve; iii) endometriosis; iv) hydrosalpinx; v) chromosomal abnormalities; or vi) chronic systemic conditions that could influence gene expression in GCs. The control group (n=41) included age-matched women with tubal factor infertility or who were undergoing fertility treatment due to male factor infertility, and with no endocrine abnormalities.

Oocyte-cumulus complexes were retrieved by transvaginal ultrasound-guided follicular aspiration 36 h after human chorionic gonadotropin administration. The FF was collected from the same follicular aspirates after removal of the oocyte-cumulus complexes. Oocyte-cumulus complexes were subsequently washed in several flat dishes with G-Mopse PLUS (Vitrolife AB) to remove residual mural GCs, haemocytes and cellular debris. Cumulus complexes were identified under a microscope and mechanically flattened in Petri dishes. Cumulus GCs (CGCs) surrounding the oocytes were then carefully dissected.

Following centrifugation at 400 × g for 10 min at room temperature, the CGCs were processed using standard ultrastructural techniques. The samples were fixed in 2.5% glutaraldehyde at 4°C for 2 h and pre-embedded in 1% agarose (Ted Pella Inc.) to stabilize the pellet. Secondary fixation was performed with 1% osmium tetroxide for 2 h at room temperature, followed by sequential dehydration in a graded ethanol series. The specimens were then infiltrated and embedded in Araldite^® epoxy resin. For TEM analysis, CGC samples from three randomly selected patients per group were examined. Ultrathin sections (60 nm) were prepared and allowed to air-dry overnight at room temperature. For staining, the sections were double-stained with 2% uranyl acetate in saturated alcoholic solution and lead citrate, each for 15 min at room temperature. Multiple ultrathin sections were prepared from each sample, and imaging areas were randomly selected at low magnification by electron microscopy technicians who were blinded to group allocation. Representative images were subsequently captured at higher magnifications to observe mitochondrial morphology and autophagic structures. A total of 10–15 cells were randomly selected from each sample. Autophagosomes were counted at the same magnification, and the results are presented as the mean ± SD for each group.

The human GC line KGN (cat. no. CL-0603), derived from a GC tumour, was obtained from Procell Life Science & Technology Co., Ltd. Both CGCs and KGN cells were maintained in DMEM/F-12 (cat. no. D6421; Sigma-Aldrich; Merck KGaA) supplemented with 10% heat-inactivated foetal bovine serum (FBS; cat. no. A5256701; Gibco; Thermo Fisher Scientific, Inc.) and 1% penicillin-streptomycin under standard culture conditions (37°C, 5% CO₂). Cellular morphology and proliferation kinetics were monitored daily using phase contrast microscopy. At 80–90% confluence, adherent cells were dissociated with trypsin-EDTA and subcultured at a 1:3 ratio to maintain logarithmic growth.

KGN cells were stimulated with dihydrotestosterone (DHT; cat. no. 15874; Cayman Chemical Company) at various concentrations (0, 10, 50, 100, 500, 1,000 and 2,000 nM) at 37°C for 24 h, and the optimum concentration was selected for subsequent experiments. In addition, KGN cells were exposed to MT (cat. no. M5250; Sigma-Aldrich; Merck KGaA) at concentrations of 0, 0.1 and 10 pM, 1 and 100 nM at 37°C for 48 h, and the optimum concentration (10 pM) was selected for subsequent experiments. In addition, KGN cells were first exposed to 100 nM DHT for 24 h followed by treatment with or without MT (10 pM) for an additional 48 h at 37°C. To investigate the effect of MT on SIRT3 expression in human ovarian CGCs, CGCs obtained from control subjects and patients with PCOS were cultured in vitro. The cells were treated with 10 pM MT at 37°C for 48 h, after which, the cells were harvested, and SIRT3 protein expression levels were measured using western blot analysis.

Cells were transfected with SIRT3 overexpression plasmids (pCMV6-Entry backbone, Myc-DDK-tagged) purchased from OriGene Technologies, Inc. (cat. no. RC200190), or with an empty pCMV6-Entry plasmid in the negative control (NC) group, using Lipofectamine^® 3000 (cat. no. L3000015; Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's instructions. Briefly, when the KGN cell density reached 70–80%, transient transfection was performed; 4 µg plasmid solution was diluted with serum-free medium and gently mixed, and P3000 (supplied with Lipofectamine 3000) was added. Lipofectamine 3000 was equilibrated to room temperature and gently mixed with serum-free Opti-MEM^® (cat. no. 31985; Gibco; Thermo Fisher Scientific, Inc.), followed by incubation at room temperature for 5 min to allow complex formation. Subsequently, the plasmid DNA solution was diluted in an equal volume of Opti-MEM and combined with the lipid mixture (Lipofectamine 3000 and Opti-MEM). The DNA-lipid complexes were allowed to form through a 10-min incubation at room temperature, after which, they were added dropwise to the cell culture dish using a sterile pipette. The dish was gently agitated to ensure an even distribution of the transfection mixture and was then transferred to a humidified 37°C incubator containing 5% CO₂ for 4 h. Following this incubation period, the transfection medium was replaced with complete growth medium supplemented with 10% FBS, and the cells were maintained under standard culture conditions for an additional 24 h prior to an efficiency assessment through western blot analysis. For the inhibitor experiments, the cells were treated with the SIRT3 inhibitor 3-TYP (50 µM cat. no. 29660; Cayman Chemical Company) for 4 h at 37°C.

KGN cells were treated with 100 nM DHT at 37°C for 24 h to establish the DHT-induced model, followed by transfection with the SIRT3 overexpression plasmid for at 37°C for 4 h. After transfection, the medium was replaced, and the cells were cultured for an additional 48 h. Then, chloroquine (CQ; 50 µM; cat. no. C6628; Sigma-Aldrich; Merck KGaA) was added for 4 h at 37°C before cell collection. The protein expression levels of LC3 and P62 were subsequently detected by western blotting.

A CCK-8 assay (cat. no. CA1211; Beijing Solarbio Science & Technology Co., Ltd.) was used to measure the viability of CGCs, according to the manufacturer's instructions. Briefly, the cells were seeded on 96-well plates at a density of 2×10⁴ cells/well, after which, 10 µl CCK8 reagent was added to the culture medium and the plate was incubated for 1.5 h. Finally, the absorbance was detected at 450 nm using a microplate reader (SpectraMax^® iD3; Molecular Devices, LLC). The inhibition rate was calculated as follows: Inhibition rate (%)=[(OD control-OD treatment)/(OD control-OD blank)] ×100. All CCK-8 assays were performed in triplicate and repeated three times.

The collected FF was centrifuged at 400 × g for 10 min at room temperature, after which the cell-free supernatant was collected as the FF sample and stored at −80°C for subsequent assays. The MT content of FF was measured using a Human MT ELISA Kit (cat. no. EH3344; Wuhan Fine Biotech Co., Ltd.), according to the manufacturer's instructions.

KGN cells (1×10⁴ cells/well) were loaded with 10 µmol/l dichlorodihydrofluorescein diacetate (DCFH-DA; Beyotime Biotechnology) by incubation at 37°C for 20 min. Subsequently, the fluorescence intensity was quantified using a SpectraMax iD3 multifunctional microplate reader with excitation/emission wavelengths set at 488/525 nm.

The MMP of KGN cells was measured using the JC-1 fluorescent probe (Cayman Chemical Company). KGN cells (1×10⁵ cells/well) were seeded in 24-well plates. After 24 h of culture, the cells were subjected to cell transfection or exposure to 10 pM MT as aforementioned, and then labelled with 10 µmol/l JC-1 for 20 min in an atmosphere containing 5% CO₂ at 37°C. Fluorescence images were captured using an Image Xpress confocal microscope (Molecular Devices, LLC) to visualize both green (monomeric) and red (aggregated) fluorescence. For each group, 5–6 fields were randomly selected at low magnification by researchers who were blinded to the experimental groups to avoid selection bias. The total red and green fluorescence intensities in each field were measured using ImageJ software (version 1.51; National Institutes of Health), and MMP is expressed as a red/green fluorescence ratio. The data were obtained from at least three independent experiments.

Active mitochondria were labelled using Mito-Tracker Red CMXRos (Invitrogen; Thermo Fisher Scientific, Inc.). KGN cells (1×10⁵ cells/well) were seeded in 24-well plates. After 24 h of culture, the cells were subjected to cell transfection or exposure to 10 pM MT as aforementioned, and were then incubated with 10 µM MitoTracker Red CMXRos staining solution at 37°C for 20 min. After incubation, the cells were washed three times with PBS (Wuhan Servicebio Technology Co., Ltd.) and maintained in fresh culture medium. The fluorescence intensity was detected with an Xpress confocal microscope. For each group, 5–6 fields were randomly selected at low magnification by researchers who were blinded to group allocation. The mean red fluorescence intensity per field was measured using ImageJ software (version 1.51) to determine the mitochondrial mass. The data were obtained from at least three independent experiments.

CGCs and KGN cells were homogenized in RIPA buffer (cat. no. RW0001; Hebei Ruipate Bio & Technology Co., Ltd.) and the protein concentration was determined using a BCA kit. Protein samples (5–8 µg) were then separated by SDS-PAGE on 10% gels and transferred to a polyvinylidene fluoride membrane (Sigma-Adrich; Merck KGaA). Subsequently, a blocking buffer containing 5% non-fat dried milk (Inner Mongolia Yili Industrial Group Company Ltd.) was used for 1.5 h at room temperature. The membrane was then incubated with primary antibodies against SIRT3 (1:1,000; cat. no. ab217319; Abcam), optic atrophy 1 (OPA1; 1:500; cat. no. ET1705-9; Huabio), DRP1 (1:1,000; cat. no. IPB0168; Baijia), phosphorylated (P)-DRP1 (Ser⁶¹⁶) (1:1,000; cat. no. AF8470; Affinity Biosciences), mitofusin 2 (MFN2; 1:1,000; cat. no. 9482S; Cell Signaling Technology, Inc.), FIS1 (1:500; cat. no. ET7109-17; Huabio), P62 (1:2,000; cat. no. PM045; MBL International Co.), LC3 (1:500; cat. no. PM036; MBL International Co.), and β-actin (1:5,000; cat. no. AC026; ABclonal Biotech Co., Ltd.) overnight at 4°C. Following the primary antibody incubation, the membrane was incubated with HRP-goat anti-rabbit IgG (1:3,000; cat. no. RS0002; ImmunoWay Biotechnology Company) or HRP-goat anti-mouse IgG (1:3,000; cat. no. S1001; Hebei Ruipate Bio & Technology Co., Ltd.). The protein bands were visualized using an ECL detection reagent (Shanghai Yeasen Biotechnology Co., Ltd.) on the Minichemi 320 chemiluminescence imaging system (Beijing Sage Creation Science Co., Ltd.) and protein levels were semi-quantified with ImageJ (version 1.51) software.

Data are presented as the mean ± SD for normally distributed data, or as median (interquartile range) for non-normally distributed data. Statistical analyses were performed using SPSS 26.0 software (IBM Corp.). Normality of the data distribution was assessed using the Shapiro-Wilk test. For data following a normal distribution with homogeneous variances, parametric tests were applied; comparisons between two groups were performed using an independent sample t-test, whereas comparisons among multiple groups were performed using one-way ANOVA followed by Tukey's post hoc test. For data that did not follow a normal distribution or with unequal variances, nonparametric tests were applied; comparisons between two groups were performed using Mann-Whitney U test, whereas comparisons among multiple groups were performed using Kruskal-Wallis test followed by the Dunn-Bonferroni test for post hoc pairwise comparisons. Each experiment was repeated using at least three independent primary CGC samples (biological replicates, n ≥3 per group), and each biological replicate was measured in two technical replicates. Due to inherent limitations in obtaining primary CGCS, sample sizes varied across different detection markers; detailed sample sizes are provided in the Results section and corresponding figure legends. P<0.05 was considered to indicate a statistically significant difference.

The clinical characteristics of the study population are summarized in Table I. A total of 36 women with PCOS and 42 age-matched controls were included in the analysis. No significant differences were observed between the two groups regarding age, body mass index, baseline oestradiol and progesterone levels, metaphase II oocyte rate, fertilization rate or high-quality embryo rate. However, distinct endocrine profiles were observed: Compared with in the control group, the PCOS group had significantly elevated baseline testosterone levels, increased luteinizing hormone (LH) concentrations and a significantly higher LH/follicle-stimulating hormone (FSH) ratio. Conversely, basal serum FSH and prolactin levels were notably lower in patients with PCOS.

To investigate mitochondrial dynamics in CGCs, western blot analysis was performed on samples obtained from 9 patients with PCOS and 9 control patients. Owing to the limited protein yield from CGCs, the effective sample size for certain markers was smaller than that of the initial cohort (n<9), and specific sample sizes are indicated in the corresponding results. The protein levels of DRP1, MFN2, OPA1, FIS1 and P-DRP1 (Ser⁶¹⁶) were detected by western blotting. The results revealed that the expression levels of MFN2 and OPA1 were significantly lower in the PCOS group than in the control group (Fig. 1A). Conversely, the expression level of FIS1 and the ratio of P-DRP1 (Ser⁶¹⁶)/total DRP1 were significantly increased in the PCOS group. These findings indicated a pronounced imbalance in mitochondrial dynamics in CGCs derived from patients with PCOS, characterized by increased fission and decreased fusion. In addition, SIRT3 protein levels were significantly lower in the CGCs from the PCOS group compared with those from the control group, as determined by western blotting, suggesting that SIRT3 downregulation may contribute to mitochondrial dysfunction in PCOS GCs (Fig. 1C).

To evaluate autophagic activity in the CGCs of patients with PCOS, autophagic vesicles were observed under electron microscopy (Fig. 1B). The results revealed an increased number of autophagic vesicles within the ovarian CGCs of patients with PCOS, suggesting that autophagy may be increased in the CGCs of patients with PCOS.

Furthermore, the expression levels of autophagy-related markers, LC3 and P62, were analysed. Compared with those in the control group, the PCOS group exhibited a significantly increased LC3II/I ratio and significantly decreased P62 expression levels (Fig. 1C), indicating that autophagic activity was enhanced in the CGCs of patients with PCOS.

The granulosa-like tumour cell line KGN, which retains its steroidogenic capacity and serves as a robust model for human GC research (22), was used to establish an in vitro PCOS model through DHT induction, a widely adopted approach in PCOS studies (23–25). To optimize the treatment conditions, KGN cells were exposed to increasing concentrations of DHT (0–2,000 nM) for 24 h. CCK-8 analyses revealed that DHT inhibited cell viability in a concentration-dependent manner (Fig. 2A). Notably, treatment with 100 nM DHT reduced cell viability by ~50%. Western blot analysis was then performed to assess the protein expression levels of SIRT3, mitochondrial fission/fusion markers (OPA1, DRP1 and MFN2) and P62 in KGN cells treated with various concentrations of DHT (Fig. 2B). Consistent with the expression patterns observed in GCs from patients with PCOS, 100 nM DHT markedly downregulated SIRT3, DRP1, OPA1, MFN2 and P62 expression compared with in the untreated control group. Therefore, this concentration was selected for subsequent experiments.

To investigate the functional role of SIRT3 in DHT-treated KGN cells, SIRT3 overexpression was used to assess mitochondrial function. Firstly, successful SIRT3 overexpression in KGN cells was confirmed, with protein levels increasing ~2-fold relative to those in the NC cells (Fig. 2C), thereby validating the transfection efficiency.

SIRT3 modulates MMP and dynamics (26); therefore, to evaluate the regulatory role of SIRT3 in mitochondrial function, three experimental groups were analysed: Untreated control, DHT-treated and DHT + SIRT3 groups. The MMP was assessed using JC-1 staining. Cells with a low MMP contained JC-1 monomers and exhibited green fluorescence signals, whereas those with a high MMP showed red fluorescence signals due to JC-1 polymer formation. JC-1 staining revealed a significant reduction in the MMP in DHT-treated cells compared with that in the control group, whereas this effect was partially reversed by SIRT3 overexpression (Fig. 2D). By contrast, MitoTracker staining revealed no significant differences in mitochondrial fluorescence intensity across groups, suggesting that mitochondrial mass remained unchanged (Fig. 2E).

DHT treatment significantly reduced the protein levels of the mitochondrial fusion proteins OPA1 and MFN2 and increased the P-DRP1 (Ser⁶¹⁶)/DRP1 ratio (Fig. 2F). Notably, SIRT3 overexpression in DHT-treated cells upregulated OPA1 and MFN2 expression (P<0.05 vs. DHT) and decreased the P-DRP1 (Ser⁶¹⁶)/DRP1 ratio, suggesting that SIRT3 may enhance mitochondrial function. The expression levels of autophagy-associated proteins were assessed by western blotting. DHT treatment significantly reduced the protein levels of the autophagic substrate P62 (Fig. 2G), indicating increased autophagic flux. Notably, SIRT3 overexpression in DHT-treated cells significantly increased P62 expression compared with DHT alone, suggesting impaired autophagic flux. LC3-II levels showed an increasing trend in both DHT and DHT + SIRT3 groups, but the changes were not statistically significant. As shown in Fig. 2G, overexpression of SIRT3 in DHT-treated cells led to elevated P62 levels, which appears to exert an opposite effect to DHT. It remains unclear whether SIRT3 overexpression inhibits autophagic flux, selectively regulates mitophagy without affecting general autophagy, or induces compensatory changes in P62 synthesis. To definitively address this issue, autophagic flux was analysed in KGN cells using CQ. CQ blocks lysosomal degradation, so the extent of LC3-II and P62 accumulation in the presence of CQ is proportional to autophagic flux; a greater accumulation reflects higher autophagic flux. The cells were divided into the following six groups: Control, control + CQ, DHT, DHT + CQ, DHT + SIRT3 and DHT + SIRT3 + CQ, and the accumulation of LC3II and P62 in the presence of CQ was compared. The accumulation of LC3II and P62 in the DHT + CQ group was significantly greater than that in the control + CQ group, confirming that DHT increased autophagic flux (Fig. 2H). Furthermore, the accumulation of LC3II and P62 in the DHT + SIRT3 + CQ group was significantly lower than that in the DHT + CQ group, indicating that SIRT3 overexpression inhibited the DHT-mediated increase in autophagic flux.

To assess MT levels in patients with PCOS, the concentrations of MT in the FF were measured using ELISA. Compared with in the FF from control patients, the FF from patients with PCOS had significantly lower MT levels (172.7±51.63 vs. 205.02±59.87 pg/ml; Fig. 3A).

To investigate the regulatory effect of MT on SIRT3 expression, KGN cells were cotreated with 100 nM DHT and various concentrations of MT; the results revealed that treatment with 10 pM MT significantly increased SIRT3 expression compared with DHT (Fig. 3B).

In human CGCs, SIRT3 protein levels were significantly lower in PCOS-derived CGCs than in control CGCs. By contrast, MT treatment of CGCs form patients with PCOS (PCOS + MT group) restored SIRT3 expression to near-control levels (Fig. 3C).

To assess the effect of MT on mitochondrial homeostasis, the intracellular ROS levels in DHT-induced KGN cells were measured by DCFH-DA staining. Notably, ROS levels were increased in the DHT group, whereas this increase was normalized by cotreatment with MT (Fig. 3D). Mito Tracker Red staining revealed no significant differences in mitochondrial mass among the groups (control, control + MT, DHT and DHT + MT; Fig. 3E). However, JC-1 staining revealed a decrease in the MMP in DHT-treated cells, which was effectively restored by MT (Fig. 3F).

Western blot analysis of mitochondrial dynamics-related proteins revealed that DHT increased P-DRP1 (Ser⁶¹⁶) levels, and reduced MFN2 and OPA1 levels, all of which were normalized by MT cotreatment (Fig. 3G). These results suggested that MT may ameliorate the DHT-induced abnormalities in mitochondrial dynamics.

DHT exposure decreased P62 and SIRT3 levels, and increased the LC3-II/I ratio, whereas MT treatment reversed these effects (Fig. 3H). Together, these findings suggested that MT alleviates DHT-induced mitochondrial dysfunction by modulating SIRT3 expression and regulating mitophagy.

To further verify whether MT regulates mitochondrial dynamics by upregulating SIRT3 expression, KGN cells were treated with or without DHT for 24 h, then treated with the SIRT3-specific inhibitor 3-TYP for 4 h, followed by treatment with MT for 48 h. KGN cells were divided into four groups: Control, DHT, DHT + MT and DHT + MT + 3-TYP, and the expression levels of mitochondrial dynamics-associated proteins were assessed by western blotting. The results showed that the protein expression levels of OPA1 and MFN2 in the DHT + MT group were significantly higher than those in the DHT group, but were significantly reduced after the addition of the SIRT3 inhibitor 3-TPY (Fig. 4A). The ratio of P-DRP1 (Ser⁶¹⁶)/DRP1 in the DHT + MT group was significantly lower than that in the DHT group; however, the addition of the SIRT3 inhibitor did not alter this ratio. This revealed that MT failed to reverse the DHT-induced abnormalities in mitochondrial fusion when SIRT3 expression was inhibited, suggesting that SIRT3 is essential for MT-mediated regulation of mitochondrial fusion.

To verify whether MT alleviates DHT-induced autophagic dysfunction through upregulation of SIRT3 expression, the expression of autophagy-related markers was assessed by western blotting. Following SIRT3 inhibition, MT was unable to suppress excessive autophagy (Fig. 4B). These findings further support that MT may restore mitochondrial dynamics and autophagic activity by increasing SIRT3 expression.