NR4A1‐Driven Granulosa Cell Ferroptosis Contributes to Obesity‐Related Ovarian Dysfunction and Metformin‐Mediated Ovarian Protection

ABSTRACT

Obesity triggers ovarian dysfunction and female infertility accompanied by granulosa cell damage. Ferroptosis, iron‐dependent lipid peroxidation‐mediated cell death, participates in multiple reproductive diseases, yet the molecular cascade linking metabolic obesity to granulosa ferroptosis remains undefined. Using high‐fat diet obese mice, primary human granulosa cells from obese infertile patients, and KGN cell lines, we identified the nuclear receptor NR4A1 as a key mediator of obesity‐induced ferroptosis downstream of PI3K/AKT signaling. Elevated NR4A1 was associated with upregulated ACSL4 and suppressed GPX4, accompanied by iron overload, lipid peroxidation, and mitochondrial collapse, impairing folliculogenesis and oocyte quality. Metformin reduced NR4A1 activation in association with PI3K/AKT signaling modulation, thereby attenuating ferroptotic injury and contributing to the restoration of ovarian function. In vivo NR4A1 agonist administration attenuated the protective effects of metformin, while ovarian‐specific NR4A1 knockdown alleviated HFD‐triggered ovarian ferroptosis. Single‐cell RNA‐seq further suggested that metformin remodels oocyte redox and folliculogenic transcriptional networks. These findings support a role for the PI3K/AKT–NR4A1 signaling pathway in driving ferroptosis during obesity‐related ovarian dysfunction and suggest a potential anti‐ferroptotic mechanism of metformin in this context.

Schematic illustrating that NR4A1‐driven granulosa cell ferroptosis mediates obesity‐related ovarian dysfunction, while metformin‐mediated ovarian protection is achieved via suppression of the PI3K/AKT–NR4A1 ferroptotic signaling cascade.

1. Introduction

Obesity‐induced lipid overload has been associated with chronic redox imbalance and ferroptotic damage across multiple peripheral organs, including liver, adipose tissue, and gonads, and has become a major global public health challenge closely linked to metabolic disorders and female reproductive dysfunction [1, 2]. Clinical evidence consistently demonstrates that obesity significantly increases the risk of ovarian dysfunction, anovulation, and infertility in women, and adversely affects the outcomes of assisted reproductive technology [1, 2]. The ovary constitutes the core organ for maintaining female reproductive and endocrine homeostasis, in which granulosa cells play indispensable roles in follicular development, oocyte maturation, and steroid hormone secretion [3, 4]. Previous studies have shown that obesity contributes to abnormal follicular development, follicular atresia, and compromised oocyte quality by inducing insulin resistance, chronic low‐grade inflammation, and oxidative stress [5, 6, 7]. However, the precise mechanism by which obesity causes ovarian damage, particularly the identity and regulatory program of granulosa cell death, remains largely unclear.

Ferroptosis is a newly identified iron‐dependent form of regulated cell death characterized by iron overload, lipid peroxidation, reactive oxygen species accumulation, and mitochondrial structural damage [8]. Emerging evidence indicates that ferroptosis participates in the pathogenesis of multiple reproductive disorders, including polycystic ovary syndrome, endometriosis, and premature ovarian insufficiency [9, 10, 11, 12]. Nevertheless, whether ferroptosis is involved in obesity‐induced ovarian dysfunction and its key regulatory targets remain to be elucidated.

The PI3K/AKT pathway is essential for ovarian follicle development and granulosa cell homeostasis, yet its downstream effectors linking metabolic stress to ferroptosis remain undefined [13, 14, 15]. NR4A1, an orphan nuclear receptor, regulates granulosa cell function, but its role in obesity‐induced ferroptosis is unknown [16, 17, 18].

Metformin is a first‐line clinical agent for Type 2 diabetes. Beyond its hypoglycemic effects, metformin has been reported to improve ovarian function in obese women, although the underlying mechanism remains incompletely understood [19, 20, 21]. Whether metformin protects ovarian function by inhibiting ferroptosis via the PI3K/AKT–NR4A1 signaling pathway has not been reported.

Prior studies separately described obesity‐induced ovarian impairment, NR4A1‐mediated granulosa cell function, and GPX4/ACSL4‐dependent ferroptosis, but no work has integrated these modules to characterize a potential signaling relationship linking metabolic lipid overload to granulosa cell ferroptotic responses. In this study, we report three novel findings: (1) NR4A1 acts as a critical regulator of ferroptosis in obese granulosa cells; (2) the PI3K/AKT pathway is associated with modulation of NR4A1 expression and activity; (3) metformin exerts ovarian protective effects, at least in part, through modulation of NR4A1‐associated ferroptotic responses. We combined obese mouse models, human primary granulosa cells, pharmacological intervention, ovarian‐specific AAV knockdown, and multi‐omics sequencing to dissect this redox signaling pathway, identify a potential molecular target involved in obesity‐associated ovarian oxidative injury, and further characterize the ovarian actions of metformin.

2. Materials and Methods

2.1. Ethics Statement

This study was approved by the Ethics Committee for Scientific Research and Clinical Trials of The First Affiliated Hospital of Zhengzhou University (approval no. 2024‐KY‐0284‐002). All participants provided written informed consent prior to enrollment. Animal experiments were approved by the Institutional Animal Care and Use Committee (IACUC) of the First Affiliated Hospital of Zhengzhou University and conducted in accordance with national guidelines for laboratory animal care.

2.2. Human Subjects and Granulosa Cell Collection

Women undergoing in vitro fertilization (IVF) at the Reproductive Medicine Center were recruited and stratified by body mass index (BMI) into a normal‐weight control group (BMI 18.5–25 kg/m², n = 30) and an obese group (BMI ≥ 28 kg/m², n = 30). Inclusion criteria for the control group: Age < 35 years; regular menstrual cycle (28–35 days); normal basal endocrine profiles; basal antral follicle count (AFC) > 5; normal karyotype; primary diagnosis of tubal obstruction or male‐factor infertility. Inclusion criteria for the obese group: Age < 35 years; obesity‐associated menstrual irregularity (e.g., LH/FSH ratio > 1, impaired follicle development); no diagnosis of polycystic ovary syndrome (PCOS) or other primary ovarian disorders; AFC > 5; normal karyotype; non‐ovarian factor infertility. Exclusion criteria included ovarian pathology, prior ovarian surgery, polycystic ovary syndrome (PCOS), uncontrolled endocrine disorders, and chromosomal abnormalities. Detailed clinical baseline characteristics of all participants (including age, BMI, basal FSH, LH, E₂, AMH, and total retrieved oocytes) are summarized in Table S2.

Follicular fluid was collected during oocyte retrieval. Granulosa cells were isolated, washed, and cultured in DMEM/F12 medium supplemented with 10% fetal bovine serum and 1% penicillin–streptomycin. For experiments using human granulosa cells, each patient‐derived cell preparation was considered an independent biological sample. Technical replicate measurements were averaged before statistical analysis.

2.3. Animal Model and Treatments

Eight‐week‐old female C57BL/6J mice (initial weight 18–23 g; the exact number of animals used in each experiment is indicated in the corresponding figure legends) were purchased from Vital River Laboratory Animal Technology Co. Ltd. (Beijing, China). Mice were housed under specific pathogen‐free (SPF) conditions at 22°C ± 2°C, 50% ± 10% humidity, and a 12 h light/12 h dark cycle with free access to food and water. After 1 week of acclimatization, mice were allocated to receive a standard control diet (4% fat) or a high‐fat diet (HFD, 60% fat; Medicience, Taizhou, China) based on a random allocation sequence generated by GraphPad Prism (version 9.0, GraphPad Software, San Diego, CA, USA). Body weight was monitored weekly, and food intake was recorded daily.

Successful HFD‐induced obesity model establishment was defined according to the following criteria: (1) body weight > 20% higher than control mice; (2) abdominal fat accumulation with waist circumference increased ≥ 15%; (3) elevated Lee's index [body weight (g)^(1/3) × 10 / body length (cm)]; (4) increased body fat percentage [(fat pad wet weight/body weight) × 100%].

Three intervention strategies were applied: (1) Metformin treatment: Metformin (30 mg/kg/day) was administered via dietary supplementation; (2) Metformin + CSN‐B: HFD mice receiving metformin were injected intraperitoneally with the NR4A1 agonist CSN‐B (13 mg/kg) every other day for 30 days; control mice received corn oil vehicle; and (3) Genetic intervention: Bilateral ovarian intrabursal injection of AAV‐shNC (control) or AAV‐shNR4A1 (10 μL/ovary, 2 × 10¹² vg/mL; Hanbio Tech, Shanghai, China) was performed to knock down NR4A1 in vivo. All functional evaluations, including ovarian histology assessment and follicle counting, were conducted by two independent investigators who were unaware of the group allocations.

2.4. Cell Culture and Treatments

Primary human granulosa cells and KGN cells were cultured in DMEM/F12 medium (Gibco, Waltham, MA, USA) supplemented with 10% fetal bovine serum (FBS; HyClone, Logan, UT, USA) and 1% penicillin/streptomycin (Leagene, Beijing, China) in a humidified incubator at 37°C with 5% CO₂.

Cells were treated with metformin (Sigma‐Aldrich, St. Louis, MO, USA), PI3K agonist 740Y‐P, PI3K inhibitor LY294002, NR4A1 antagonist DIM‐C‐pPhOH (DCp), or NR4A1 agonist CSN‐B (all from MedChemExpress, Shanghai, China) at the indicated concentrations and time points specified in the corresponding figure legends. Corresponding vehicle treatments were included as controls in all experiments. All in vitro experiments were performed with a minimum of three independent biological replicates (n = 3), with each sample assayed in technical triplicate.

2.5. Metabolic and Biochemical Assays

For serum lipid profiling, blood was collected from the retro‐orbital plexus under anesthesia. Serum triglycerides (TG) and total cholesterol (TC) were measured using commercial kits (Jiancheng Bioengineering Institute, Nanjing, China) on a Chemray 800 automated analyzer.

Cell viability was measured using Cell Counting Kit‐8 (CCK‐8; Boster Biological Technology, Wuhan, China) at 450 nm.

For steroid hormone measurement, primary granulosa cells were treated with 1 nM testosterone and 10 μM metformin for 48 h. Estradiol (E₂) and progesterone levels in culture supernatants were quantified using ELISA kits (HUABIO INC., Zhengzhou, China).

Oxidative stress markers, including glutathione (GSH) and malondialdehyde (MDA), were measured using commercial kits (Jiancheng Bioengineering Institute) at 405 nm and 532 nm, respectively.

2.6. Quantitative Real‐Time PCR (qRT‐PCR)

Total RNA was extracted using TRIzol reagent (Invitrogen, Carlsbad, CA, USA). cDNA was synthesized using the HiScript III All‐in‐one RT SuperMix (Vazyme, Nanjing, China). qRT‐PCR was performed using iTaq Universal SYBR Green Supermix (Bio‐Rad, Hercules, CA, USA) on a Bio‐Rad CFX96 system. Relative gene expression was calculated using the 2^(‐ΔΔCt) method with GAPDH as the internal control. Primer sequences are listed in Table S1.

2.7. Western Blotting

Total protein was extracted using RIPA lysis buffer containing protease and phosphatase inhibitors (Beyotime, Shanghai; Servicebio, Wuhan, China). Protein samples were separated by 10% SDS‐PAGE and transferred to PVDF membranes (Millipore, Bedford, MA, USA). Membranes were blocked with rapid blocking buffer (Bio‐Rad) or 5% BSA (Solarbio, Beijing, China) and incubated overnight at 4°C with primary antibodies against β‐actin, GPX4, ACSL4, NR4A1, p‐NR4A1, AKT, p‐AKT, PI3K, and p‐PI3K. After incubation with HRP‐conjugated secondary antibodies, signals were visualized using ECL substrate (Bio‐Rad) and quantified using Image Lab software.

2.8. siRNA Transfection

KGN cells were transfected with NR4A1‐specific siRNA or negative control siRNA (MedChemExpress) using Lipofectamine RNAiMAX (Thermo Fisher Scientific) in Opti‐MEM medium. Knockdown efficiency was verified by qRT‐PCR and Western blotting 48 h after transfection.

2.9. Histological and Morphological Analyses

Ovaries were fixed in 4% paraformaldehyde (Solarbio), embedded in paraffin, sectioned at 5 μm, and stained with hematoxylin and eosin (H&E). Follicles were classified and counted as primordial, primary, secondary, and antral follicles according to established morphological standards.

For transmission electron microscopy (TEM), ovarian tissues were fixed in glutaraldehyde, post‐fixed in 1% OsO4, dehydrated, and embedded in epoxy resin. Ultrathin sections (50–80 nm) were stained with uranyl acetate and lead citrate. Mitochondrial ultrastructure was observed using a Hitachi HT7800 transmission electron microscope (Hitachi High‐Tech Corp., Tokyo, Japan).

2.10. Oocyte Collection and Functional Assays

Mice were superovulated with 10 IU PMSG (Solarbio) followed by 10 IU hCG (China National Pharmaceutical Group) 44–46 h later. Cumulus‐oocyte complexes (COCs) were retrieved from the oviductal ampulla 14–16 h post‐hCG, denuded using hyaluronidase (Vitrolife, Gothenburg, Sweden), and cultured in KSOM medium.

Live oocytes were stained with ROS probe, FerroOrange (Fe²⁺), or JC‐1 (mitochondrial membrane potential; all from Beyotime or Dojindo). Fluorescence images were captured using a Zeiss LSM 880 laser scanning confocal microscope (Carl Zeiss, Germany) at defined excitation/emission wavelengths.

2.11. RNA Sequencing and Bioinformatics

For bulk RNA‐seq, total RNA (RIN > 7) from granulosa cells was sequenced on an Illumina NovaSeq 6000 platform (150 bp paired‐end). Raw reads were filtered, aligned to the GRCm38 genome, and quantified. Differentially expressed genes (DEGs) were identified using DESeq2 (|log₂FC| > 1, FDR < 0.05). KEGG pathway enrichment analysis was performed.

For single‐cell RNA‐seq (scRNA‐seq), single oocytes were lysed, and cDNA was synthesized using template‐switching reverse transcription. Libraries were constructed using the KAPA HyperPrep kit and sequenced on an Illumina NovaSeq 6000. DEG analysis (|log₂FC| > 1.5, FDR < 0.05) and KEGG enrichment were performed in R.

2.12. Ovarian in Situ Injection and in Vivo Imaging

Mice were anesthetized with 2% pentobarbital sodium (30 mg/kg, i.p.). Bilateral intra‐bursal injections of AAV‐shNR4A1 or AAV‐shNC (2 × 10¹² vg/mL) were performed. For in vivo fluorescence imaging, mice were anesthetized with isoflurane, and signals were captured using an IVIS Spectrum imaging system (PerkinElmer, USA) and analyzed using Living Image software.

2.13. Data and Code Availability

The raw sequence data have been deposited in the Genome Sequence Archive (Genomics, Proteomics & Bioinformatics 2025) in the National Genomics Data Center (Nucleic Acids Res 2025), China National Center for Bioinformation, Chinese Academy of Sciences (Human granulosa cells: https://ngdc.cncb.ac.cn/gsa‐human/browse/HRA018351; Mouse oocyte single‐cell RNA‐seq: https://ngdc.cncb.ac.cn/gsa/browse/CRA042650). All other data supporting the findings of this study are available within the article and its supplementary files. No custom code or algorithms were used.

2.14. Statistical Analysis

All data are presented as mean ± standard error of the mean (SEM). Data distributions were tested for normality using the Shapiro–Wilk test, and equality of variance was confirmed using Levene's test. Statistical analyses were performed using SPSS 18.0 (IBM) or GraphPad Prism 9.0. Comparisons between two groups were performed using unpaired Student's t‐test (for normally distributed data) or Mann–Whitney U test (non‐parametric data). One‐way ANOVA followed by Tukey's post hoc test was used for multiple group comparisons. Longitudinal body weight changes over time were analyzed using two‐way repeated‐measures ANOVA with Sidak's multiple comparisons test. A p‐value < 0.05 was considered statistically significant. The experimental unit and exact sample size for each experiment are specified in the corresponding figure legends. For cell‐based experiments, n refers to independent biological replicates, whereas technical replicates were averaged before statistical analysis. For animal experiments, the individual mouse was considered the experimental unit.

3. Results

3.1. HFD Feeding Induces Ovarian Structural Damage and Granulosa Cell Ferroptosis

Two‐month‐old female C57BL/6J mice were randomized to receive standard chow or a 12‐week high‐fat diet (HFD). Divergence in body weight between groups appeared at week 3 and increased throughout the feeding period. Final body weight measured 31.22 ± 1.03 g in HFD mice versus 23.13 ± 4.24 g in chow‐fed controls (p < 0.0001; Figure 1A,B). Serum triglycerides, total cholesterol, Lee's index, adipose tissue weight, and body fat percentage were higher in the HFD group (all p < 0.05; Figure S1A–D).

FIGURE 1.

HFD induces ovarian dysfunction and granulosa cell ferroptosis in female mice. Female C57BL/6J mice (n = 10 mice per group) were fed a standard chow diet (Ctrl) or high‐fat diet (HFD) for 12 consecutive weeks. (A) Longitudinal body weight growth curve over time (two‐way repeated‐measures ANOVA followed by Sidak's post hoc test). (B) Morphological comparison of mice after dietary intervention. (C) Ovarian H&E staining (scale bar = 200 μm). (D) Quantification of follicles at different developmental stages: Primordial (PFI), primary (PFII), secondary (SF), and antral (AF) follicles. (E) Glutathione (GSH) levels in granulosa cells. (F) Malondialdehyde (MDA) levels in granulosa cells. (G) Protein expression of ACSL4 and GPX4 in granulosa cells. (H) Mitochondrial ultrastructure in granulosa cells (upper scale bar = 5 μm; lower scale bar = 2 μm). Red arrows indicate swollen mitochondria with disorganized cristae. (I) Percentage of granulosa cells with abnormal mitochondrial morphology. (J) Representative fluorescence images of oocyte Fe²⁺ (scale bar = 20 μm). (K) Quantitative analysis of relative Fe²⁺ fluorescence intensity. Data are presented as mean ± SEM; statistical significance was determined by unpaired Student's t‐test unless otherwise specified, and p < 0.05 indicates significant differences. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

Estrous cycles spanned 5–7 days in HFD mice compared with 3–4 days in controls (p < 0.001; Figure S1E,F). The counts of primordial, primary, and antral follicles were lower in HFD ovarian tissue (p < 0.05; Figure 1C,D). Fewer oocytes were recovered after superovulation from HFD mice (p < 0.01; Figure S1G,H). Oocytes collected from HFD animals showed elevated ROS signal intensity, reduced mitochondrial membrane potential, and a larger proportion of mitochondria with disrupted cristae (all p < 0.01; Figure S1I–P).

Glutathione (GSH) concentrations were lower and malondialdehyde (MDA) concentrations higher in granulosa cells isolated from HFD ovaries (p < 0.01 and p < 0.001, respectively; Figure 1E,F). ACSL4 protein abundance increased, and GPX4 protein abundance decreased in HFD granulosa cells (Figure 1G). Transmission electron microscopy revealed swollen mitochondria with disorganized cristae in HFD granulosa cells, alongside a higher percentage of morphologically abnormal mitochondria (p < 0.01; Figure 1H,I). Oocyte intracellular Fe²⁺ fluorescence intensity was elevated in the HFD group (p < 0.05; Figure 1J,K).

3.2. Metformin Relieves HFD‐Triggered Ovarian Ferroptotic Injury in Vivo

Weekly body weight measurements showed reduced weight gain in HFD mice receiving metformin relative to vehicle‐treated HFD animals (p < 0.0001; Figure 2A). Lower final body weight was recorded in metformin‐treated HFD mice (Figure 2B). Serum triglycerides, total cholesterol, Lee's index, adipose mass, and body fat percentage decreased after metformin intervention (all p < 0.0001; Figure S2A–E). Disrupted estrous cyclicity observed in HFD mice was restored following metformin treatment (p < 0.01; Figure S2F,G).

FIGURE 2.

Metformin alleviates HFD‐induced ovarian dysfunction by suppressing granulosa cell ferroptosis. HFD‐fed mice (n = 8 mice per group) were treated with metformin (Met) or vehicle control. (A) Longitudinal weekly body weight monitoring over time (two‐way repeated‐measures ANOVA followed by Sidak's post hoc test). (B) Body morphology of mice after intervention. (C) Ovarian histology by H&E staining (scale bar = 200 μm). (D) Quantification of follicles at various developmental stages. (E) Representative fluorescence images of oocyte Fe²⁺ (scale bar = 20 μm). (F) Quantitative analysis of relative Fe²⁺ fluorescence intensity. (G) Transmission electron microscopy of mitochondrial morphology in granulosa cells (upper scale bar = 5 μm; lower scale bar = 2 μm). (H) Percentage of granulosa cells with abnormal mitochondria. (I–J) Protein expression of ACSL4 and GPX4 in granulosa cells. (K–L) GSH and MDA levels in granulosa cells. Data are presented as mean ± SEM; statistical significance was determined by one‐way ANOVA followed by Tukey's post hoc test unless otherwise specified, and p < 0.05 indicates significant differences. **p < 0.01 and ****p < 0.0001.

Follicle counts at all developmental stages were higher in metformin‐treated HFD ovaries (p < 0.05; Figure 2C,D). More oocytes were harvested after superovulation from metformin‐administered mice (p < 0.001; Figure S3A). Oocyte ROS fluorescence decreased, and mitochondrial membrane potential increased in the metformin group (p < 0.0001; Figure S3B–F). Fewer damaged mitochondria were observed in oocytes from metformin‐treated HFD mice (p < 0.01; Figure S3G,H).

Metformin reduced oocyte Fe²⁺ signal intensity (p < 0.0001; Figure 2E,F) and decreased the proportion of granulosa cells with abnormal mitochondrial ultrastructure (p < 0.001; Figure 2G,H). ACSL4 protein levels declined, and GPX4 protein levels rose in granulosa cells after metformin exposure (all p < 0.0001; Figure 2I,J). Granulosa cell GSH content increased, and MDA content decreased with metformin treatment (p < 0.01 and p < 0.0001; Figure 2K,L). Higher estradiol and progesterone concentrations were detected in culture supernatants of metformin‐treated primary human granulosa cells (p < 0.05 and p < 0.001; Figure S3I,J).

3.3. Metformin Reverses Ferroptotic Phenotypes in Primary Human Granulosa Cells From Obese Patients

Western blotting detected higher ACSL4 and lower GPX4 protein levels in granulosa cells from obese patients compared with normal‐weight subjects (both p < 0.05; Figure 3A,B). Metformin treatment reduced ACSL4 and restored GPX4 abundance in granulosa cells collected from obese individuals (p < 0.001 and p < 0.05; Figure 3C,D). GSH concentrations were lower and MDA concentrations higher in granulosa cells from obese patients (both p < 0.0001; Figure 3E,F).

FIGURE 3.

Metformin inhibits ferroptosis in human granulosa cells from obese patients. Primary granulosa cells were collected from normal‐weight (NOR) and obese patients (n = 10 patients per group), with or without metformin treatment (10 μM, 48 h). KGN cells were treated with H₂O₂ (100 μM, 48 h) or RSL3 (10 nM, 48 h) to establish ferroptosis models. (A, B) Protein levels and quantification of ACSL4 and GPX4 in granulosa cells from NOR and obese patients. (C, D) ACSL4 and GPX4 expression in obese patient granulosa cells after metformin treatment. (E, F) GSH and MDA levels in NOR and obese patient granulosa cells. (G, H) Protein levels of ACSL4 and GPX4 in H₂O₂treated KGN cells with metformin. (I) Viability of KGN cells treated with H₂O₂ or RSL3 plus metformin. Data represent mean ± SEM from n = 10 individual patient samples for clinical data or n = 3 independent biological replicates. Statistical significance was determined by unpaired Student's t‐test for two‐group comparisons or one‐way ANOVA followed by Tukey's post hoc test for multi‐group comparisons, and p < 0.05 indicates significant differences.

H₂O₂ exposure increased ACSL4 and suppressed GPX4 expression in KGN cells; co‐incubation with metformin reversed these protein changes (p < 0.0001; Figure 3G,H). Reduced cell viability was recorded in KGN cells treated with H₂O₂ or RSL3, and viability improved with metformin co‐treatment (p < 0.01; Figure 3I).

3.4. Transcriptomic Profiling Identifies NR4A1 as a Metformin‐Responsive Gene Associated With PI3K/AKT Signaling

Bulk RNA‐seq was performed on granulosa cells from four groups: Normal‐weight, normal‐weight plus metformin, obese, obese plus metformin. Distinct gene expression patterns were visible across groups (Figure 4A). KEGG enrichment analysis of differentially expressed genes highlighted enrichment within PI3K/AKT‐related pathways (Figure 4B). Heatmap visualization revealed altered expression of multiple PI3K/AKT‐associated transcripts in granulosa cells from obese patients (p < 0.05; Figure 4C). NR4A1 transcript abundance increased in the obese group and decreased after metformin administration (p < 0.05 and p < 0.01; Figure 4D).

FIGURE 4.

Transcriptomic screening identifies NR4A1 as a key target of metformin. Bulk RNA‐seq was performed in granulosa cells from NOR, NOR + metformin, Obese, and Obese + metformin groups. (A) Heatmap of gene expression profiles in four groups. (B) KEGG pathway enrichment of differentially expressed genes. (C) Heatmap of PI3K/AKT pathway‐related genes in four groups. (D) Heatmap of differentially expressed PI3K/AKT pathway‐related genes in Obese and Obese + metformin groups. (E–H) Protein levels of AKT, p‐AKT, and NR4A1 and their quantitative analysis. (I) NR4A1 mRNA level verified by qPCR. Data represent the mean ± SEM from n = 5 independent biological replicates. Statistical significance was determined by one‐way ANOVA followed by Tukey's post hoc test, and p < 0.05 indicates significant differences.

Total AKT protein levels did not differ between experimental groups (p > 0.05). The p‐AKT/AKT ratio and NR4A1 protein expression were elevated in granulosa cells from obese patients (p < 0.05 and p < 0.001; Figure 4E–H). qPCR quantification demonstrated higher NR4A1 mRNA levels in cells from obese subjects (p < 0.001; Figure 4I).

3.5. PI3K/AKT Signaling Modulates NR4A1 to Control Granulosa Cell Ferroptosis

NR4A1 siRNA transfection reduced NR4A1 mRNA and protein abundance in KGN cells (all p < 0.0001; Figure 5A–C). The p‐PI3K/PI3K ratio decreased after NR4A1 knockdown (p < 0.01), while the p‐AKT/AKT ratio remained unchanged (p > 0.05; Figure S4A,B). Treatment with the PI3K agonist 740Y‐P elevated p‐PI3K/PI3K and p‐AKT/AKT ratios, whereas the PI3K inhibitor LY294002 reduced both metrics (all p < 0.0001; Figure 5D; Figure S4C,D). The p‐NR4A1/NR4A1 ratio rose with 740Y‐P exposure and fell with LY294002 treatment (p < 0.0001; Figure 5D; Figure S4E). NR4A1 transcript levels shifted following modulation of PI3K/AKT activity (p < 0.05; Figure S4F).

FIGURE 5.

PI3K/AKT signaling regulates NR4A1 to mediate granulosa cell ferroptosis. KGN cells were transfected with NR4A1 siRNA (20 nM) or treated with PI3K agonist 740Y‐P (20 μM), PI3K inhibitor LY294002 (25 μM), NR4A1 agonist CSN‐B (10 μM), NR4A1 antagonist DCp (15 μM), AKT agonist SC79 (10 μM), and metformin (10 μM) as indicated. (A–C) NR4A1 mRNA and protein levels after siRNA knockdown. (D) Protein levels of p‐NR4A1, NR4A1, p‐PI3K, PI3K, p‐AKT, and AKT upon pathway modulation. (E) ACSL4 and GPX4 expression after CSN‐B treatment in granulosa cells from normal patients. (F, G) GSH and MDA levels in CSN‐B‐treated cells. (H–J) Ferroptotic markers in obese patient cells treated with DCp. (K) Ferroptosis markers in granulosa cells treated with the AKT agonist SC79 or NR4A1 antagonist DCp. (L, M) GSH and MDA levels in granulosa cells treated with the AKT agonist SC79 or NR4A1 antagonist DCp. (N–P) NR4A1, GPX4, GSH, and MDA levels in granulosa cells from obese patients treated with metformin, 740Y‐P, DCp, or CSN‐B. Data represent mean ± SEM from n = 3 independent biological cell preparations or n = 10 individual patient samples. Statistical significance was determined by one‐way ANOVA followed by Tukey's post hoc test, and p < 0.05 indicates significant differences.

CSN‐B (NR4A1 agonist) upregulated ACSL4 and downregulated GPX4 in normal granulosa cells (all p < 0.0001). GSH levels declined, and MDA levels increased after CSN‐B incubation (p < 0.0001 and p < 0.01; Figure 5E–G). The NR4A1 antagonist DCp reversed ACSL4 and GPX4 shifts in granulosa cells from obese patients (Figure 5H–J; Figure S4H). The AKT agonist SC79 increased ACSL4 and decreased GPX4; co‐treatment with DCp partially restored baseline protein levels (all p < 0.001; Figure 5K, Figure S4I). Corresponding decreases in GSH and increases in MDA were observed after SC79 treatment, and these changes were attenuated by DCp co‐incubation (Figure 5L,M).

Metformin lowered NR4A1, raised GPX4, increased GSH, and reduced MDA in granulosa cells from obese patients (all p < 0.0001). Co‐treatment with 740Y‐P or CSN‐B attenuated these metformin‐induced molecular shifts (all p < 0.0001; Figure 5N–P, Figure S4J).

3.6. NR4A1 Agonist Attenuates Metformin‐Mediated Ovarian Protection in Obese Mice

Intraperitoneal injection of CSN‐B elevated p‐NR4A1 protein abundance in mouse granulosa cells (p < 0.05; Figure 6A, Figure S5A). Follicle counts at all developmental stages were higher in metformin‐treated HFD mice; co‐administration of CSN‐B reduced follicle numbers (p < 0.01; Figure 6B,C). More oocytes were recovered from metformin‐treated HFD mice after superovulation, and oocyte yields decreased with combined metformin + CSN‐B treatment (p < 0.001; Figure 6D, Figure S5B). Mitochondrial structural abnormalities observed in HFD oocytes were alleviated by metformin, and this improvement was lost after CSN‐B co‐treatment (Figure 6E, Figure S5C).

FIGURE 6.

NR4A1 activation abolishes metformin‐mediated ovarian protection in vivo. HFD‐fed mice (n = 8 mice per group) were treated with metformin alone or combined with NR4A1 agonist CSN‐B (13 mg/kg, i.p. every other day for 30 days). (A) p‐NR4A1 protein level in mouse granulosa cells. (B, C) Ovarian H&E staining (scale bar = 200 μm) and follicle counting. (D) Number of oocytes retrieved by superovulation (scale bar = 120 μm). (E) Transmission electron microscopy of oocyte mitochondrial ultrastructure (scale bar = 2 μm). Red arrows indicate mitochondria exhibiting pathological features such as shrinkage and disordered or absent cristae. (F, G) GSH and MDA levels in granulosa cells. (H) ACSL4 and GPX4 protein expression. (I) Granulosa cell mitochondrial morphology (scale bar = 2 μm). Data represent the mean ± SEM from n = 8 mice per group. Statistical significance was determined by one‐way ANOVA followed by Tukey's post hoc test, and p < 0.05 indicates significant differences.

Metformin reduced oocyte ROS and Fe²⁺ fluorescence signals in HFD mice; CSN‐B reversed both effects (all p < 0.01; Figure S6A–D). JC‐1 red fluorescence intensity and the red/green ratio increased in oocytes from metformin‐treated animals, and these metrics declined with CSN‐B co‐administration (all p < 0.001; Figure S6E–H).

Metformin increased granulosa cell GSH and decreased MDA concentrations in HFD mice; CSN‐B co‐treatment reversed both biochemical changes (all p < 0.0001; Figure 6F,G). Metformin‐mediated downregulation of ACSL4 and upregulation of GPX4 were eliminated by CSN‐B co‐administration (all p < 0.0001; Figure 6H, Figure S7A). The percentage of granulosa cells with damaged mitochondria decreased with metformin and increased again after CSN‐B co‐treatment (Figure 6I, Figure S7B).

3.7. Ovarian NR4A1 Knockdown Mitigates HFD‐Induced Ovarian Dysfunction

Ovarian intrabursal AAV‐shNR4A1 or AAV‐shNC injection was performed in HFD‐fed mice after 11 weeks of dietary intervention (Figure 7A). In vivo fluorescence imaging demonstrated detectable ovarian‐associated fluorescence six weeks after injection (Figure 7B). Body weight gain slowed in HFD mice receiving AAV‐shNR4A1 (p < 0.0001; Figure 7C,D).

FIGURE 7.

NR4A1 knockdown alleviates obesity‐induced ovarian dysfunction by inhibiting ferroptosis. HFD‐fed mice (n = 8 mice per group) received ovarian intrabursal injection of AAV‐shNC or AAV‐shNR4A1 (10 μL/ovary, titer 2 × 10¹² vg/mL). (A) After 11 weeks of dietary intervention, successfully modeled mice received an ovarian in situ injection of either AAV‐shNR4A1 or AAV‐shNC (scrambled control). (B) In vivo imaging of AAV transduction. (C, D) Longitudinal body weight curve over time (two‐way repeated‐measures ANOVA followed by Sidak's post hoc test) and gross morphology. (E) Ovarian protein levels of NR4A1 and GPX4. (F, G) GSH and MDA levels in granulosa cells. (H, I) Ovarian histology (scale bar = 100 μm) and follicle quantification. (J, K) Transmission electron microscopy of granulosa cell mitochondrial ultrastructure (scale bar = 2 μm) and abnormal ratio. (L) KEGG analysis of scRNA‐seq DEGs in HFD vs. Ctrl oocytes. (M) Pathway enrichment in HFD + Met vs. HFD oocytes. Data represent the mean ± SEM from n = 8 mice per group for animal measurements or n = 3 independent preparations for sequencing. Statistical significance was determined by one‐way ANOVA followed by Tukey's post hoc test unless otherwise specified, and p < 0.05 indicates significant differences. *p < 0.05, ****p < 0.0001.

NR4A1 protein levels decreased, and GPX4 protein levels increased in ovarian tissue after AAV‐shNR4A1 injection (all p < 0.001; Figure S8A,B). Granulosa cell ACSL4 expression fell, GSH concentrations rose, and MDA concentrations dropped following ovarian NR4A1 knockdown (all p < 0.01; Figure 7E–G, Figure S8C). More oocytes were harvested from AAV‐shNR4A1‐treated HFD mice after superovulation (p < 0.05; Figure S8D,E). Secondary and antral follicle counts increased in ovaries with NR4A1 silencing (Figure 7H,I). The proportion of granulosa cells with abnormal mitochondrial ultrastructure declined after AAV‐shNR4A1 injection (Figure 7J,K).

JC‐1 red fluorescence intensity and the red/green ratio were higher in oocytes from AAV‐shNR4A1‐treated HFD mice (p < 0.05; Figure S9A–D). Oocyte Fe²⁺ and ROS signal intensity decreased after ovarian NR4A1 knockdown (all p < 0.001; Figure S9E–H).

Single‐cell RNA‐seq identified 131 differentially expressed genes (DEGs) when comparing oocytes from control and HFD mice (Figure S10A). DEGs associated with oocyte development and antioxidant defense were downregulated, while ferroptosis‐ and apoptosis‐related transcripts were upregulated in HFD oocytes (Figure S10B). KEGG enrichment of these DEGs revealed prominent MAPK and cAMP pathway shifts (Figure 7L). A total of 193 DEGs were detected in oocytes from HFD mice with versus without metformin treatment (Figure S10C). Metformin upregulated transcripts linked to folliculogenesis and antioxidant activity and downregulated ferroptosis/apoptosis‐related genes (Figure S10D). KEGG pathway enrichment showed transcriptional remodeling linked to nuclear signaling after metformin exposure (Figure 7M).

4. Discussion

Obesity‐related ovarian dysfunction is a major contributor to female infertility, yet the precise mode of granulosa cell death triggered by chronic metabolic stress remains poorly characterized [1, 2, 22, 23]. In the present study, we systematically characterized the pathological cascade linking high‐fat diet‐induced obesity to ovarian injury and characterized potential molecular mechanisms underlying the ovarian effects of metformin. Our data indicate that NR4A1 acts as a key mediator of granulosa cell ferroptosis under lipid overload, participating in a PI3K/AKT–NR4A1 signaling pathway to drive ferroptosis and mediate metabolic stress‐induced gonadal damage.

Multiple clinical and preclinical studies have documented that obesity disrupts estrous cyclicity, impairs follicular maturation, and degrades oocyte quality [5, 6, 7]. Consistent with those observations, our HFD mouse model exhibited comprehensive ovarian dysfunction accompanied by features consistent with ferroptotic stress: Reduced GSH, elevated MDA, increased ACSL4, suppressed GPX4, and damaged mitochondrial ultrastructure. Ferroptosis is an iron‐dependent form of regulated cell death widely studied in metabolic and degenerative pathologies [8]. This work provides novel in vivo and human primary cell evidence that granulosa ferroptosis is closely involved in obesity‐induced ovarian impairment, bridging lipid metabolic stress and gonadal redox failure.

NR4A1 is an orphan nuclear receptor with established roles in cellular metabolism, survival, and programmed cell death [16, 17, 18]. Prior reproductive research has described NR4A1 modulation of granulosa cell physiology, but no prior work has linked NR4A1 activity to ferroptotic signaling. In our study, NR4A1 protein and mRNA levels were robustly elevated in granulosa cells from both obese mice and obese infertile patients, and metformin reversed this upregulation. Functional pharmacological and genetic manipulation suggested that NR4A1 is involved in promoting granulosa cell ferroptosis, whereas NR4A1 inhibition alleviates lipid‐overload‐triggered ferroptotic phenotypes. These results expand the understanding of NR4A1 from a metabolic regulator to an active participant governing iron and lipid redox homeostasis, enriching the biological repertoire of NR4A1 in metabolic biochemistry.

Dysregulated PI3K/AKT signaling in obese ovarian tissue disrupts granulosa cell homeostasis, yet the downstream effector connecting this pathway to ferroptosis has not been fully elucidated [13, 14, 15]. Our findings support the PI3K/AKT–NR4A1 signaling pathway as a mediator of granulosa cell ferroptosis under obesity‐related metabolic stress. Whether this pathway functions similarly in other tissues remains to be determined. This dissection clarifies the uncharacterized downstream branch of PI3K/AKT signaling mediating metabolic reproductive injury and supports NR4A1 as a druggable molecular target for obesity‐related ovarian dysfunction.

Metformin is a first‐line antihyperglycemic drug with documented ovarian benefits in obese infertile populations, though its tissue‐protective molecular mechanisms remain incompletely resolved [19, 20, 21]. The present study indicated that metformin helps restore ovarian function by targeting the PI3K/AKT–NR4A1 signaling pathway to attenuate granulosa cell ferroptosis in both mouse ovarian tissue and human primary granulosa cells. In vivo NR4A1 agonist administration markedly attenuated metformin's ovarian protective phenotypes, supporting the role of NR4A1 as a key downstream mediator of metformin's gonadal effects. Single‐cell transcriptomic profiling further illustrated that metformin remodels oocyte transcriptional networks to upregulate folliculogenic and antioxidant genes while repressing ferroptotic and apoptotic transcripts. These findings suggest that modulation of ferroptotic responses may contribute to the ovarian effects of metformin in the setting of obesity. Whether these effects are independent of metformin's systemic metabolic actions requires further investigation.

This study has several limitations. First, while our data establish a functional connection between NR4A1 and the ACSL4/GPX4 axis, the underlying genomic binding or transcriptional co‐factors remain to be fully elucidated in future research. Second, the precise downstream phosphorylation sites or structural modifications of NR4A1 modulated by the PI3K/AKT pathway warrant further investigation. Third, the clinical sample size is relatively modest; future multi‐center and large‐cohort studies are warranted to strengthen translational relevance.

In summary, our research supports a signaling relationship underlying obesity‐induced ovarian dysfunction: NR4A1 promotes granulosa cell ferroptosis as a downstream effector of PI3K/AKT signaling upon lipid overload. Metformin exerts ovarian protective effects partially by suppressing this NR4A1‐dependent ferroptotic cascade. These findings highlight NR4A1 as a metformin‐sensitive key regulator of granulosa ferroptosis and provide a candidate therapeutic target for clinical intervention of obesity‐associated female infertility.

Author Contributions

H.W., C.W., S.G., and G.Y. conceived the study, performed most experimental investigations, formal analysis and data curation, established the methodology, completed experimental validation, and drafted the original manuscript. J.F., J.H., R.J., J.Z., H.J., and W.S. took part in the experimental investigation. S.S., X.H., L.Q., Z.W., Y.L., N.S., and X.Z. assisted in methodology construction. G.Y. conceptualized the overall project, supervised the whole research, administered the project, acquired funding, revised and edited the manuscript, and approved the final version for submission. All authors reviewed and approved the final manuscript.

Funding

This study was supported by the National Natural Science Foundation of China (U1904138), the Natural Science Foundation of Henan Province (262300421646), the Key Scientific Research Project of Colleges and Universities in Henan Province (26A320024), the Provincial and Ministerial Co‐construction Project of Henan Medical Science and Technology Program (SBGJ202502063), and the Open Research Fund of the National Health Commission Key Laboratory of Birth Defects Prevention (NHCKLBDP202407).

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1: Metabolic and reproductive impairments in HFD‐induced obese mice. (A) Serum triglyceride (TG) and total cholesterol (TC) levels. (B) Lee's index. (C) Wet weight of adipose tissue. (D) Body fat percentage. (E) Distribution of estrous cycle stages (P: Proestrus; E: Estrus; M: Metestrus; D: Diestrus). (F) Total length of the estrous cycle. (G) Representative images of superovulated oocytes (scale bar = 120 μm). (H) Quantification of oocyte yield. (I) Representative fluorescence images of oocyte ROS (scale bar = 20 μm). (J) Quantitative analysis of relative ROS fluorescence intensity in oocytes. (K) JC‐1 staining for oocyte mitochondrial membrane potential (scale bar = 20 μm). (L–N) Quantitative analysis of JC‐1 green, red, and red/green fluorescence ratio. (O) Oocyte mitochondrial ultrastructure (upper scale bar = 5 μm; lower scale bar = 1 μm). (P) Percentage of oocytes with abnormal mitochondrial morphology. Data are presented as mean ± SEM, and p < 0.05 indicates significant differences.

Figure S2: Metformin improves metabolic and estrous cycle disorders in HFD‑fed mice. (A) Serum triglyceride (TG) and total cholesterol (TC) levels. (B) Lee's index. (C, D) Fat wet weight. (E) Body fat percentage. (F–G) Estrous cycle distribution and total length in Ctrl, Met, HFD, and HFD + Met mice. Data are presented as mean ± SEM, and p < 0.05 indicates significant differences.

Figure S3: Metformin restores oocyte quality and steroid hormone secretion. (A) Number of superovulated oocytes (scale bar = 120 μm). (B) Oocyte ROS fluorescence intensity (scale bar = 50 μm). (C–F) Oocyte mitochondrial membrane potential (JC‑1 staining; scale bar = 50 μm) and quantitative analysis. (G, H) Oocyte mitochondrial ultrastructure (scale bar = 2 μm) and abnormal ratio. (I, J) Estradiol and progesterone secretion in primary human granulosa cells. Data are presented as mean ± SEM, and p < 0.05 indicates significant differences.

Figure S4: PI3K/AKT signaling regulates NR4A1 phosphorylation and transcription. (A, B) p‑PI3K/PI3K and p‑AKT/AKT ratios after NR4A1 silencing. (C, D) p‑PI3K/PI3K and p‑AKT/AKT levels after 740Y‑P or LY294002 treatment. (E) p‑NR4A1/NR4A1 ratio upon PI3K/AKT modulation. (F) NR4A1 mRNA level after pathway intervention. (G–J) Quantitative analysis of ferroptosis‑related proteins in corresponding treatment groups. Data are presented as mean ± SEM, and p < 0.05 indicates significant differences.

Figure S5: CSN‑B activates NR4A1 and blocks metformin‑mediated ovarian protection. (A) Quantitative analysis of p‑NR4A1/NR4A1. (B) Number of superovulated oocytes per mouse. (C) Quantitative analysis of oocyte mitochondrial abnormality ratio. Data are presented as mean ± SEM, and p < 0.05 indicates significant differences.

Figure S6: NR4A1 activation reverses metformin‐induced oocyte oxidative stress and mitochondrial function. (A, B) Oocyte ROS staining and quantitative analysis. (C, D) Fe²⁺ fluorescence intensity and quantitative analysis. (E–H) JC‑1 staining and quantitative analysis. Scale bar = 50 μm. Data are presented as mean ± SEM, and p < 0.05 indicates significant differences.

Figure S7: NR4A1 activation abolishes metformin‑induced granulosa cell ferroptosis. (A) Quantitative analysis of ACSL4 and GPX4 protein levels. (B) Percentage of granulosa cells with abnormal mitochondria. Data are presented as mean ± SEM, and p < 0.05 indicates significant differences.

Figure S8: NR4A1 knockdown regulates ferroptosis‑related proteins and improves ovarian function. (A) Quantitative analysis of ovarian NR4A1 and GPX4. (B, C) Protein levels of NR4A1, ACSL4, and GPX4 in granulosa cells. (D, E) Morphology (scale bar = 60 μm) and number of superovulated oocytes. Data are presented as mean ± SEM, and p < 0.05 indicates significant differences.

Figure S9: NR4A1 knockdown restores oocyte mitochondrial function and reduces oxidative stress. (A) Oocyte JC‑1 staining. (B–D) JC‑1 green and red fluorescence and red/green ratio. (E–H) Oocyte Fe²⁺ and ROS fluorescence and quantitative analysis. Scale bar = 20 μm. Data are presented as mean ± SEM, and p < 0.05 indicates significant differences.

Figure S10: Single‑cell RNA‑seq analysis of oocytes following HFD and metformin treatment. (A) Volcano plot of DEGs between HFD and Ctrl oocytes. (B) Heatmap of DEGs between HFD and Ctrl oocytes. (C) Volcano plot of DEGs between HFD + Met and HFD oocytes. (D) Heatmap of DEGs between HFD + Met and HFD oocytes.

Table S1: Sequences of qRT‑PCR primers used in this study.

Table S2: Comparison of general characteristics between the two groups.

Data Availability Statement

All data generated or analyzed during this study are included in this published article and its Supporting Information.