Gypenosides ameliorate polycystic ovary syndrome via gut microbiota modulation: An integrated gut microbiome and ovarian transcriptome study

Summary

Polycystic ovary syndrome (PCOS) is a prevalent endocrine disorder with complex metabolic and reproductive manifestations. Gut microbiota dysbiosis has emerged as a key factor in PCOS, yet the causal role of microbiota-targeted interventions remains to be fully defined. In a letrozole-induced rat model, we show that gypenosides (GPs)—bioactive saponins from Gynostemma pentaphyllum—alleviate weight gain, hormonal imbalance, and estrous cycle disruption, alongside reducing systemic oxidative stress and inflammation. Integrated ovarian transcriptomics and gut microbiome 16S rRNA sequencing reveal that GPs modulate ovarian gene expression related to inflammation and cellular function while reshaping gut microbiota by enriching beneficial genera including Lactobacillus and Romboutsia. Fecal microbiota transplantation (FMT) establishes causality: PCOS-derived microbiota transfers disease traits to normal recipients, whereas GPs-conditioned microbiota alleviates PCOS phenotypes in recipients. These findings demonstrate that GPs act through a “gut microbiota-oxidative stress-ovary” axis, supporting GPs as a promising microbiota-targeted intervention for PCOS management.

Graphical abstract

Highlights

• •Gypenosides alleviate key PCOS phenotypes in a rat model
• •GPs reduce ovarian inflammation and oxidative stress in PCOS
• •GPs reshape gut microbiota by enriching beneficial bacteria
• •FMT confirms gut microbiota mediates GPs therapeutic effects

Biological sciences; Microbiology; Omics

Introduction

As one of the most common endocrine disorders, polycystic ovary syndrome (PCOS) affects 10%–18% of women of childbearing age worldwide at present, with an annually increasing morbidity.^1,2 It presents a range of clinical features, including hyperandrogenism, menstrual irregularities, oligomenorrhea, and anovulation, in addition to disturbances in metabolism homeostasis, reproductive function, dermatological manifestations, and psychological symptoms.^3,4 PCOS imposes an oppressive impairment in life quality which is a direct consequence of its dual reproductive and metabolic dysfunctions. Importantly, the co-occurrence of these features is driven by several pathological processes, including metabolic imbalance, oxidative stress, and systemic chronic inflammation.^5,6 This insight reveals that effectively targeting these core processes is crucial for advancing treatment beyond fertility restoration, a therapeutic goal in contrast to current clinical management. As the etiology of PCOS remains incompletely understood, clinical management primarily relies on a combination of lifestyle modifications and symptomatic interventions. Current treatments mainly aim to regulate menstrual cycles, reduce hyperandrogenism, improve metabolic parameters, and induce ovulation, with the objective of relieving clinical symptoms.^4,7 However, most pharmacological therapies are unable to effectively alleviate all clinical manifestations of PCOS. In addition, long-term medication use is often associated with gastrointestinal discomfort, drug resistance, and symptom recurrence after discontinuation,^8,9 pointing to the need for advancing mechanism-based, precise interventions in PCOS.

Although a complete understanding of its pathogenic mechanisms is still evolving, PCOS is widely considered a multifactorial disorder involving genetic, epigenetic, and environmental interactions. In recent years, the gut microbiota has been recognized as a key regulator of metabolic and endocrine homeostasis and gained increasing attention in pathogenesis of multiple diseases, such as metabolic and endocrine disorders.¹⁰ Studies have shown that women with PCOS exhibit significant gut microbiota dysbiosis, characterized by alterations in microbial diversity and composition (with some studies reporting reduced diversity while others show no significant change) of specific bacterial genera, decreased levels of beneficial probiotics, and an increase in opportunistic pathogens. These changes are reflected in reduced α-diversity, altered β-diversity, and an altered Firmicutes-to-Bacteroidetes ratio.^11,12 Such microbial disturbances are significantly correlated with hormonal levels and inflammatory markers in PCOS patients.¹³ Importantly, gut microbiota dysbiosis is not merely a consequence of PCOS but also a key driver of disease progression. For instance, fecal microbiota transplantation (FMT) from healthy donors to PCOS recipients has been shown to alleviate PCOS-like phenotypes,¹⁴ providing direct evidence for a causal role of gut microbiota in PCOS pathogenesis. Mechanistically, gut microbiota dysbiosis may contribute to PCOS pathogenesis not only by influencing the hypothalamic-pituitary-gonadal axis via the gut-brain axis but also through microbial metabolite signaling.^15,16 Furthermore, alterations in the hepatic-intestinal circulation efficiency of sex hormones can affect their systemic levels, thereby exacerbating reproductive endocrine disorders.¹⁷ These findings provide a scientific basis for targeting the gut microbiota as a therapeutic strategy for PCOS.^18,19 In particular, dysregulation of key microbial metabolites—such as short-chain fatty acids, bile acids, and tryptophan derivatives—has been shown to interfere with insulin signaling pathways (e.g., PI3K/AKT^20,21 and NF-κB^22,23), alter the enterohepatic circulation of sex hormones, and consequently exacerbate reproductive endocrine disorders. These mechanistic insights provide a scientific basis for targeting the gut microbiota ecosystem as a therapeutic strategy for PCOS.

Given this context, natural plants and their bioactive constituents, particularly those with a history of dietary and medicinal use, have attracted growing interest for their multi-target capacity, high safety, and potential for microbiota modulation. Gynostemma pentaphyllum (G. pentaphyllum), a traditional herb also consumed as a functional food, is a salient example. Its primary bioactive compounds, gypenosides (GPs), have demonstrated efficacy in ameliorating several pathological features of PCOS, including counteracting inflammation,²⁴ oxidative stress,²⁵ dysregulated glucose and lipid metabolism,^26,27 and insulin resistance.²⁸ Notably, emerging evidence suggests that the systemic metabolic benefits of G. pentaphyllum and GPs are closely linked to their ability to remodel the gut ecosystem by enhancing the abundance of potentially beneficial taxa (e.g., Lactococcus spp.²⁹ and Prevotella) and reinforcing the intestinal barrier.³⁰ This mode of action is consistent with other microbiota-directed interventions, such as the Bu Shen Hua Tan formula, which alleviates PCOS by modulating microbial short-chain fatty acids (SCFAs) production and activating PPARγ signaling.³¹ Furthermore, HeQi San has been reported to alleviate chronic inflammation and modulate gut flora in dehydroepiandrosterone (DHEA)-induced PCOS mice,³² and berberine has been shown to regulate gut microbiota and metabolites to improve PCOS.³³ These parallel findings strongly posit gut microbiota modulation as a key mechanism through which G. pentaphyllum confers its therapeutic benefits in PCOS.

This study was therefore designed to elucidate the therapeutic potential of GPs in PCOS through the lens of gut microbiota regulation. We comprehensively assessed the effects of GPs on reproductive, endocrine, and metabolic phenotypes in a PCOS rat model. By integrating 16S rRNA gene sequencing and metagenomic analyses, we investigated the compositional and functional changes in the gut microbiota, as well as the inferred ecological interactions. Notably, while previous studies have implicated gut microbiota dysbiosis in PCOS pathogenesis,¹² direct causal evidence demonstrating that microbiota modulation alone can mediate the therapeutic effects of a natural product in PCOS remains limited. Furthermore, although GPs have been reported to exert anti-inflammatory and antioxidant effects in various disease models,³⁴ their specific mechanism of action in PCOS—particularly whether these effects are mediated via the gut microbiota—has not yet been established. To address this gap, we performed FMT experiments to establish causality and integrated ovarian transcriptomic profiling with gut microbiome analysis to systematically dissect the underlying molecular pathways. Our work provides causal insights into the role of gut microbiota in PCOS pathogenesis and highlights the promise of GPs as a gut-microbiota-targeted dietary intervention for PCOS management.

Results

Network pharmacology prediction

Liquid chromatography-mass spectrometry (LC-MS) analysis identified 20 distinct saponins in Gynostemma pentaphyllum extract. Based on in silico prediction of intestinal absorption and oral bioavailability (using SwissADME), 10 components with high potential bioavailability were prioritized for subsequent network pharmacology analysis (Table S1). Using the SwissTargetPrediction database, 273 potential protein targets were predicted for these bioactive saponins. Notably, no direct targets were identified for 10 components, including malonyl-GP XLIX, ginsenoside Rb1, GP A, and acetyl-GP A, suggesting their bioactivity may be mediated through indirect or non-target-binding mechanisms.

Meanwhile, 1,983 PCOS-associated targets were retrieved from the GeneCard, OMIM, and NCBI databases using the keyword “polycystic ovary syndrome”. A Venn diagram revealed 84 overlapping targets between GPs and PCOS (Figure 1A), representing putative direct targets for the intervention. A protein-protein interaction (PPI) network was constructed for these overlapping targets to further investigate their functional relationships (Figure 1B). To elucidate the biological functions of these 84 common targets, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were performed. GO analysis indicated their strong involvement in cellular responses to external stimuli (e.g., proliferative and survival signals), metabolic reprogramming, and the integrated regulation of hormonal synergy (Figure 1C). KEGG analysis identified 148 significantly enriched pathways, including PI3K-Akt, Ras, FoxO, EGFR, inflammatory response, endocrine resistance, and immune-related pathways (Figure 1D). Of particular relevance, the PI3K/AKT signaling pathway, along with oxidative stress- and inflammation-related pathways, was most significantly enriched.

Figure 1.

Network pharmacological analysis of gypenosides on polycystic ovary syndrome

(A) Venn diagram analysis of gypenosides and polycystic ovary syndrome.

(B) PPI protein interaction network.

(C) GO pathway enrichment analysis.

(D) KEGG pathway enrichment analysis.

(E) Network diagram of herb-compound-target-pathway (GPs represent gypenosides) in the treatment of polycystic ovary syndrome.

Subsequently, a comprehensive “herb-compound-target-pathway” network was constructed using Cytoscape 3.10.3 to visualize the multi-component, multi-target, and multi-pathway interactive characteristics (Figure 1E). The network comprised 305 nodes and 1,331 edges, with red, blue, and green nodes representing the active components of Gynostemma saponins, the signaling pathways, and the potential targets, respectively. Network topology analysis predicted that GP XLIX, GP XXXVI, and ginsenoside Rb3 may serve as the core active components.

GPs ameliorate ovarian dysfunction in PCOS rats

Analysis of the body weight growth curve revealed a significant increase in the body weight of the PCOS model group compared to the control group (p < 0.05). In contrast, treatment with various doses of GPs and metformin (Met) effectively attenuated this weight gain, with all treatment groups exhibiting final body weights significantly lower than the PCOS group (p < 0.05; Figure 2A).

Figure 2.

GPs can alleviate the metabolic disorders and ovarian dysfunction in PCOS rats

(A) Body weight at 4 weeks and 8 weeks. n = 6 biological replicates of each. Mean ± SD. ∗p < 0.05 and ∗∗∗p < 0.001 compared to the NC group by one-way ANOVA.

(B) Representative sections of ovarian tissue stained with H&E (scale bars, 500 μm). ▲, cystic follicles; ★, corpus luteum.

(C) Number of cystic follicles and corpus lutea. n = 6 biological replicates of each. Mean ± SD. ∗p < 0.05, ∗∗p < 0.01, and ∗∗∗p < 0.001 compared to the NC group by one-way ANOVA.

(D) Serum sex hormone levels of rats in each group. Serum testosterone (T), luteinizing hormone (LH), follicle-stimulating hormone (FSH) levels, and the ratio of LH to FSH (LH/FSH). n = 5–6 biological replicates of each. Mean ± SD. ∗∗p < 0.01 and ∗∗∗p < 0.001 compared to the NC group by one-way ANOVA.

(E) Representative vaginal smears during each estrous cycle during the experiment (scale bars, 100 μm). Pre-estrus, estrus, post-estrus, and inter-estrus period.

(F) Changes in estrous cycles of rats in each group after treatment.

We next evaluated the histopathological changes in the ovarian tissues. Compared with the controls, the PCOS group exhibited abnormal ovarian morphology, featuring a thickened ovarian cortex, a reduced number of growing follicles, an increased number of cystic follicles, and absent or scarce corpora lutea. Conversely, administration of GPs across different doses and Met resulted in notable improvements, manifested as a thinner ovarian cortex, an increased population of growing follicles, reduced cystic dilation, and a restored number of corpora lutea (Figures 2B and 2C). Of note, the ovarian morphology in the high-dose GPs group was most comparable to that of the control group.

Assessment of serum sex hormone profiles showed that the PCOS group had significantly decreased follicle-stimulating hormone (FSH) levels and significantly elevated testosterone (T), luteinizing hormone (LH), and LH/FSH ratios compared to the control group (p < 0.05). Following intervention with GPs and Met, these aberrations were significantly reversed, as evidenced by increased FSH levels and decreased T, LH levels, and LH/FSH ratios relative to the PCOS group. The most substantial effects were observed in the high-dose GPs group (p < 0.05; Figure 2D).

To further assess ovarian function, we monitored the estrous cycle. Control rats displayed regular cyclicity, whereas PCOS rats exhibited severe cyclicity disruption, frequently arrested in the diestrus phase, indicating ovulatory dysfunction. Treatment with GPs and Met markedly improved estrous cycle regularity (Figures 2E and 2F), demonstrating the efficacy of GPs in countering PCOS-induced estrous cycle abnormalities.

Impact of GPs on pathway expression in PCOS rats

In this part, we first identified differentially expressed genes (DEGs) between the control and PCOS model groups (Figure 3A), as well as between the PCOS and GPs-treated groups (Figure 3B). The intersection of these DEGs yielded a set of genes specifically modulated by GP intervention (Figure 3C). A PPI network was constructed for these overlapping targets to further investigate their functional relationships (Figure 3D). GO functional enrichment analysis revealed that these key genes were significantly enriched in biological processes such as postsynaptic signaling, synaptic regulation mediated by the EGFR pathway, and extracellular matrix interactions (Figure 3E). KEGG pathway analysis further indicated significant regulation of pathways related to inflammatory response, oxidative stress, and cell proliferation/differentiation, as well as the PPAR signaling pathway, following GP treatment (Figure 3F).

Figure 3.

Transcriptome sequencing analysis of GPs in PCOS rats

(A) Volcano plot of differentially expressed genes between the control group and the PCOS group.

(B) Volcano plot of differentially expressed genes between the PCOS group and the GPs-treatment group.

(C) Intersection Venn diagram of differentially expressed genes.

(D) PPI protein interaction network.

(E) GO enrichment analysis.

(F) KEGG pathway enrichment analysis.

To experimentally validate the anti-inflammatory and antioxidant effects suggested by the transcriptome data, we quantified relevant serum biomarkers. As shown in Figure 4, GP treatment significantly lowered the levels of the pro-inflammatory cytokines IL-1β and IL-6. Concurrently, GP administration markedly enhanced the activity of the antioxidant enzyme superoxide dismutase (SOD) and reduced the level of the lipid peroxidation product malondialdehyde (MDA) (p < 0.05).

Figure 4.

The inhibitory effect of GPs on the inflammatory response and oxidative stress in PCOS rats

(A) Levels of IL-1β and IL-6. n = 6 biological replicates of each. Mean ± SD. ∗p < 0.05, ∗∗p < 0.01, and ∗∗∗p < 0.001 compared to the NC group by one-way ANOVA.

(B) Levels of MDA and activity of SOD. n = 6 biological replicates of each. Mean ± SD. ∗p < 0.05, ∗∗p < 0.01, and ∗∗∗p < 0.001 compared to the NC group by one-way ANOVA.

The therapeutic effects of GPs on PCOS are mediated by remodeling gut microbiota

During model development, we observed increased stool volume and abnormal morphology in PCOS model rats, suggesting potential intestinal flora imbalance that might affect intestinal motility and lead to excretory dysfunction. We then evaluated the intestinal safety and effects of GPs. Histopathological analysis revealed no significant structural abnormalities in intestinal tissues of the GPs-treated group compared to the PCOS group, while the Met group exhibited discernible structural alterations (Figure 5A). These results indicate good intestinal safety of GPs at therapeutic doses, in contrast to Met, which despite alleviating PCOS symptoms, may cause intestinal damage.

Figure 5.

Safety of GPs saponins on the intestine and their ability to regulate the structure of the intestinal microbiota and improve PCOS

(A) Representative sections of intestinal tissue stained with H&E (scale bars, 50 μm).

(B) ASV release curve and ASV Rank curve.

(C) Venn diagram analysis of different groups.

(D and E) Alpha diversity of the intestinal microbiota, represented by Shannon, Simpson, Chao1, and ACE rarefaction indices. n = 4 biological replicates of each.

(F) β-diversity based on principal coordinates analysis (PCoA) and non-metric multidimensional scaling analysis (NMDS)

To test the hypothesis that GPs alleviate PCOS through gut microbiota modulation, we conducted systematic 16S rRNA gene sequencing analysis of gut microbiota in PCOS rats. Rank-abundance curves indicated uniform species distribution and high richness across groups, while the ASV rarefaction curve plateaued with increasing sequencing depth (Figure 5B), confirming adequate sequencing depth and data reliability. ASV analysis identified 8 unique ASVs in the PCOS group (Figure 5C), whereas no unique ASVs were detected in the GPs-treated group, suggesting PCOS-induced alterations in gut microbiota composition. Alpha diversity analysis detected no statistically significant differences in Shannon, Simpson, Chao1, or ACE indices among the three groups (p > 0.05; Figures 5D and 5E). β-Diversity analysis (PCoA and non-metric multidimensional scaling [NMDS]) demonstrated clear separation between PCOS and control/GPs groups (Figure 5F), indicating that GPs can reverse PCOS-induced gut microbiota dysbiosis.

Analysis of microbial composition revealed significant changes at multiple taxonomic levels. At the phylum level, GPs treatment increased the relative abundance of Firmicutes while decreasing Actinobacteriota (Figure 6A). At the genus level, GPs treatment reversed the PCOS-induced reduction in Romboutsia and Lactobacillus (Figure 6B); note: species-level assignment based on V3–V4 16S rRNA amplicon sequencing is tentative; all conclusions are primarily drawn at the genus level.

Figure 6.

Gynostemma pentaphyllum saponins can regulate the structure and function of the intestinal microbiota and improve PCOS

(A) Comparison of the composition of the intestinal microbial community at the phylum level. n = 4 biological replicates of each.

(B) Comparison of the composition of the intestinal microbial community at the genus level. n = 4 biological replicates of each.

(C) Functional prediction analysis. n = 4 biological replicates of each.

Functional predictions further support the association of GPs intervention with pathways related to immunity, oxidative stress, metabolism, and growth (Figure 6C).

Antibiotic depletion and FMT confirm the causal role of gut microbiota in GPs-mediated PCOS amelioration

To verify whether the therapeutic efficacy of GPs depends on modulation of the intestinal microbiota, we first employed a broad-spectrum antibiotic mixture to eliminate the gut microbiota in PCOS rats. Recipient rats that underwent antibiotic pre-treatment and were transplanted with a PCOS model microbiota still exhibited PCOS-related phenotypes: serum testosterone and LH levels remained elevated compared to the PCOS model group, the LH/FSH ratio failed to normalize (p < 0.05), and ovarian histology continued to show abundant cystic follicles with scant corpus luteum (Figures 7A–7C).

Figure 7.

Effect of antibiotic-induced intestinal microbiota depletion on the efficacy of GPs in treating PCOS

(A) Representative images of ovarian tissue H&E staining (scale bars, 1 mm), with cystic follicles (▲) and corpus lutea (★) labeled on the plot.

(B) Quantitative analysis results of cystic follicles and corpus lutea. n = 6 biological replicates of each. Mean ± SD. ∗p < 0.05, ∗∗∗p < 0.001, and ns compared to the NC group by unpaired Student’s t test.

(C) Serum levels of testosterone (T), luteinizing hormone (LH), follicle-stimulating hormone (FSH), and the LH/FSH ratio in rats across groups. n = 6 biological replicates of each. Mean ± SD. ∗p < 0.05, ∗∗∗p < 0.001 and ns compared to the NC group by one-way ANOVA.

(D) Estrous cycle determined by vaginal exfoliative cell Papanicolaou staining in each group.

We subsequently conducted an FMT experiment to determine whether the microbial community derived from GPs-treated donors could replicate the therapeutic effects of GPs. Hormone level results demonstrated that compared with recipients receiving untreated PCOS donor microbiota (FMT_PCOS group), PCOS recipient rats receiving GPs-treated donor microbiota (FMT_GPs group) exhibited significantly reduced serum testosterone and LH levels, as well as restored LH/FSH ratios (Figure 7C). Notably, the degree of improvement in the FMT_GPs group was comparable to that observed in the original GPs treatment group. Ovarian tissue hematoxylin and eosin (H&E) staining (Figure 7A) further revealed that FMT_PCOS group rats showed a significant increase in cystic follicles and a decrease in corpus lutea, whereas FMT_GPs group rats exhibited a reduction in cystic follicles and improved corpus luteum formation (Figure 7B). Additionally, dynamic observation of vaginal exfoliative cell Papanicolaou staining (Figure 7D) indicated that FMT_PCOS group rats exhibited estrous cycle disturbances with persistent diestrus, whereas FMT_GPs group rats showed significant restoration of their menstrual cycles, demonstrating that the GPs-reprogrammed microbiota could restore ovarian periodicity in recipient rats.

To further elucidate the molecular and microbiomic changes in recipient rats following FMT, we conducted a combined analysis of ovarian transcriptome sequencing and intestinal microbial 16S rRNA sequencing between the FMT_GPs group and the FMT_PCOS group (Figure 8). The ovarian transcriptome sequencing results further revealed its underlying molecular mechanisms. As shown in Figures 8A–8E, the transcriptomic analysis demonstrated intergroup differences in DEGs, PPI networks, GO enrichment, and KEGG pathway enrichment. KEGG enrichment analysis revealed significant enrichment in circadian rhythms and their associated pathways, cyclic AMP (cAMP) signaling, ErbB signaling, and neuroactive ligand-receptor interactions.

Figure 8.

Joint analysis of ovarian transcriptome and gut microbiota 16S rRNA sequencing in GP-treated PCOS rats

(A) Volcano plot of differentially expressed genes in ovarian tissue.

(B) Protein-protein interaction (PPI) network diagram of differentially expressed genes.

(C) Expression heatmap of key differentially expressed genes.

(D) GO enrichment analysis results of differentially expressed genes.

(E) KEGG pathway enrichment analysis results of differentially expressed genes.

(F) Representative images of intestinal tissue H&E staining (scale bars, 100 μm); the triangular symbol (▲) indicates mild intestinal injury in the FMT_pcos group.

(G) Venn diagrams for the indicated groups.

(H) Sample rarefaction curve and rank-abundance curve.

(I) Intestinal microbiota β-diversity based on principal component analysis (PCoA) and non-metric multidimensional scaling (NMDS).

(J) Bar plot of microbial composition at the genus level for differentially abundant microorganisms.

(K) LEfSe analysis results (LDA score distribution)showing differentially abundant genera between groups.

(L) Relative abundance of intestinal microbiota KEGG functional pathways predicted by PICRUSt2.

Intestinal H&E staining revealed mild structural damage to the intestinal epithelium in the FMT_PCOS group (Figure 8F, indicated by ▲), suggesting persistent mild impairment of the intestinal barrier in these recipients, likely attributable to microbial dysbiosis. ASV Venn diagrams and rarefaction curves demonstrated adequate sequencing depth and distinct ASV compositions between groups (Figures 8G and 8H), confirming the reliability of the sequencing data and the intergroup differences in microbial composition. β-Diversity analysis revealed distinct microbial community structures between the two groups, further confirming significant differences in the overall gut microbiota composition between the recipient rats (Figure 8I). Differential microbial bar plots and LEfSe analyses identified key taxa with intergroup differences (e.g., Lactobacillus) (Figures 8J and 8K). Functional prediction using PICRUSt2 revealed that the reconstructed gut microbiota exhibits significant enrichment in pathways such as carbohydrate metabolism, cysteine/methionine metabolism (associated with antioxidant activity), the two-component system, and quorum sensing (Figure 8L).

Discussion

In this study, we demonstrated that GPs, the primary bioactive saponins from Gynostemma pentaphyllum, exert significant therapeutic effects in a letrozole-induced rat model of PCOS. Our comprehensive investigation, integrating network pharmacology, ovarian transcriptomics, gut microbiome 16S rRNA sequencing, and FMT, revealed that GPs ameliorate PCOS through a multi-targeted mechanism centered on the “gut microbiota-oxidative stress-ovary” axis.

The phenotypic findings demonstrate that GPs significantly alleviated abnormal weight gain, improved ovarian histopathology, regulated serum sex hormone levels, and restored estrous cyclicity in PCOS rats. This finding is consistent with the clinical profile of PCOS, which is often characterized by metabolic abnormalities and an elevated risk of obesity.³⁵ Hyperandrogenism is a core pathological feature of PCOS; it not only promotes antral follicle accumulation but also suppresses FSH levels and elevates the LH/FSH ratio via negative feedback on pituitary function, the latter being a key diagnostic marker for PCOS.³⁶ Thus, the ability of GPs to significantly reduce T levels and normalize the LH/FSH ratio underscores their potential to fundamentally correct the endocrine imbalance in PCOS. This finding is supported by the work of Yang et al.,³⁷ who identified that GP XVII (GP-17), a specific saponin from Gynostemma pentaphyllum, can function as a phytoestrogen and modulate hormonal homeostasis via estrogen receptor (ER)-mediated pathways, providing direct molecular evidence for the reproductive endocrine-regulatory potential of GPs.

As natural bioactive compounds, GPs have attracted considerable research interest due to their diverse biological activities. Beyond their documented anti-inflammatory effects,³⁸ they exhibit potential in areas such as anti-tumor activity,³⁹ lipid metabolism regulation,⁴⁰ and immune modulation,⁴¹ alongside a favorable safety profile.⁴² Particularly relevant to PCOS, Akter et al.⁴³ reported that the anti-obesity effect of Gynostemma involves the PPARγ/CEBPα signaling pathway, and obesity is closely linked to the metabolic disturbances in PCOS.³⁵ Furthermore, dietary intervention with GPs has been shown to modulate the gut microbiota in mice fed a high-fat diet,⁴⁴ offering a clue that their benefits in PCOS may involve the gut microbiota-oxidative stress-ovary axis (Figure S2). Previous studies also highlight the positive roles of GPs in enhancing the efficacy of gastric cancer immunotherapy,⁴⁵ suppressing inflammatory responses,⁴⁶ and mitigating septic acute lung⁴⁷ and intestinal injury.⁴⁸ These pleiotropic effects are mediated through the regulation of multiple signaling pathways (e.g., PI3K/AKT/mTOR, NF-κB/MAPKs/AP-1, and Sirt1/Nrf2) and biological processes, including antioxidative stress, anti-inflammation, and anti-apoptosis. In a related context, Su et al.³⁰ found that Gynostemma pentaphyllum polysaccharides can modulate the oxidative stress response by enriching anti-inflammatory gut microbiota, suggesting that the antioxidant capacity of Gynostemma components may also be crucial for protecting ovarian function.

The transcriptomic and biochemical findings revealed that GPs modulated ovarian gene expression related to inflammatory response, oxidative stress, and cellular functions. KEGG pathway analysis indicated significant regulation of pathways related to inflammation, oxidative stress, and the PPAR signaling pathway, which is directly linked to insulin sensitivity and metabolic regulation. Consistently, Bai et al.⁴⁹ and Li et al.⁵⁰ have reported that alleviating inflammation and oxidative stress effectively mitigate PCOS pathology. The anti-inflammatory and antioxidant properties of GPs are not unique to PCOS but have also been documented in other pathological conditions, including atherosclerosis,³⁴ cerebral ischemic injury,⁵¹ and psoriasis,⁵² wherein they exert protective effects by regulating downstream pathways. Direct evidence from animal models confirms the potent antioxidant function of GPs, demonstrated by their ability to significantly enhance serum total antioxidant capacity (T-AOC) and SOD activity while reducing MDA content.⁵³ Furthermore, the aglycone metabolites of GPs have been shown to modulate gut microbiota, thereby ameliorating hyperlipidemia and hepatic steatosis—effects closely linked to the attenuation of oxidative stress and inflammation.²⁵ Our serum biochemical results robustly corroborate the transcriptomic findings, collectively indicating that the therapeutic effect of GPs on PCOS is likely mediated through the modulation of anti-inflammatory and antioxidant signaling pathways.

The 16S rRNA sequencing analysis revealed that GPs treatment reshaped gut microbiota by increasing the relative abundance of Firmicutes while decreasing Actinobacteriota at the phylum level and reversed the PCOS-induced reduction in Romboutsia and Lactobacillus at the genus level. This regulation is consistent with the findings of Zhang et al.,⁵⁴ who reported that modulation of these phyla is associated with improved PCOS outcomes. This is particularly significant as Liu et al.⁵⁵ demonstrated that increasing Lactobacillus abundance improves PCOS. Literature evidence supports the functional relevance of these microbial changes: Lactobacillus regulation can suppress inflammation through reduction of chronic inflammation and control of systemic immune responses⁵⁶ and reduce blood glucose by increasing glucose tolerance and alleviating oxidative stress in diabetic mice,⁵⁷ while oral administration of related strains improves depression by producing ergothioneine, which prevents stress-induced depressive behaviors and exerts anti-inflammatory effects.⁵⁸ Furthermore, the increased abundance of Romboutsia and Lactobacillus following GPs treatment is noteworthy, given their documented benefits in metabolic diseases⁵⁹ and renal injury.⁶⁰

Notably, the changes in Lactobacillus observed in our study present an interesting parallel with findings from clinical research: Kuang et al.⁶¹ demonstrated that probiotic Lactobacillus acidophilus JYLA-126 supplementation in PCOS patients improved metabolic outcomes, optimized hormonal profiles (including LH and anti-Müllerian hormone (AMH)), enhanced gut microbiota diversity, and alleviated Met-induced gastrointestinal side effects, highlighting the translational potential of Lactobacillus-based interventions. Similarly, Liu et al.⁶² reported a reduced abundance of beneficial bacteria such as Akkermansia in PCOS patients. These studies collectively underscore the central importance of specific gut microbiota in PCOS treatment. As gut microbiota dysbiosis directly impacts insulin resistance, hormonal balance, and systemic inflammation—core pathological features of PCOS⁶³—the ability of GPs to modulate these processes through the gut microbiota-oxidative stress-ovary axis is of particular significance.

The FMT experiments in our study established causality: PCOS-donor microbiota induced PCOS-like traits in normal rats, whereas GPs-conditioned microbiota alleviated PCOS phenotypes in recipients. These findings are consistent with those reported by Wang et al.⁶⁴ regarding how mulberry branch alkaloids ameliorate the phenotypes of DHEA-induced PCOS rats, who explicitly identified gut microbiota remodeling and alterations in metabolite profiles as key mediators of improved reproductive endocrine function. Additionally, Huang et al.¹⁴ demonstrated that transplantation of fecal microbiota from PCOS patients into germ-free mice induced ovarian dysfunction, lipid metabolism disorders, insulin resistance, and obesity-like phenotypes in recipient mice, whereas transplantation of healthy control microbiota did not elicit such changes. The results of this study provide bidirectional validation of these findings: on one hand, PCOS-associated microbiota can transmit pathological phenotypes to healthy recipients; on the other hand, microbiota reconstructed by GPs can cf. therapeutic effects to PCOS recipients, collectively establishing a causal relationship between microbiota and PCOS from both perspectives. The transcriptomic alterations in FMT recipients align with the “gut-ovarian axis” hypothesis proposed by Zhao et al.,⁶⁵ which posits that gut microbiota dysbiosis can influence the pathogenesis and progression of PCOS through multiple mechanisms, including immunomodulation, metabolic disturbances, and hormonal imbalances. The enrichment of beneficial bacterial genus such as Lactobacillus in GPs-reprogrammed microbiota may mediate the protective effects, as these probiotics have been previously reported to exhibit beneficial functions, including ameliorating metabolic disorders, modulating inflammatory responses, and enhancing intestinal barrier function.⁵⁶ Notably, L. reuteri has been demonstrated to improve metabolic abnormalities induced by circadian rhythm disturbances associated with PCOS.⁶⁶ Additionally, Romboutsia, a genus with anti-inflammatory and metabolic regulatory potential,^59,67 may contribute to improving the chronic inflammatory state linked to PCOS by enhancing host immune responses and energy metabolism through increased abundance.

In conclusion, the above results demonstrate from both host ovarian transcriptional regulation and intestinal microecological perspectives that the reconstructed gut microbiota following GP intervention is sufficient to independently improve ovarian function in PCOS rats by modulating pathways related to inflammation and oxidative stress. The reconstituted gut microbiota not only alters the composition of recipient microbiota but also enhances ovarian function by regulating key signaling pathways, thereby establishing the causal mediating role of the “gut microbiota-ovarian axis” in GP therapy for PCOS. This study provides direct evidence from a causal perspective regarding the involvement of gut microbiota in PCOS pathogenesis and offers a theoretical foundation for clinical applications of microbiota-targeted treatment strategies (e.g., probiotics, prebiotics, and FMT).

Limitations of the study

Several limitations should be acknowledged. The sample size is modest (n = 5–6 per group), yet the consistent and statistically significant improvements across multiple parameters, together with causal evidence from FMT experiments, support the validity of our conclusions. The current conclusions, based primarily on a rat model, require clinical validation in PCOS patients. Future studies with larger cohorts are warranted to further validate the therapeutic potential of GPs. Additionally, while our transcriptomic analysis identified key pathways regulated by GPs, we did not perform qPCR validation for selected RNA sequencing (RNA-seq) targets. However, our mechanistic interpretations are primarily drawn from pathway-level enrichment analyses rather than individual gene-level claims, and the key functional conclusions are independently validated at the protein/functional level by serum biomarker measurements (IL-1β, IL-6, MDA, and SOD). The absence of targeted metabolomics data (e.g., short-chain fatty acids, bile acids, and tryptophan metabolites) is another limitation. Future research should focus on conducting large-scale clinical trials to verify the efficacy and safety of GPs in PCOS patients; applying metabolomics to clarify the interactions between GPs and specific gut microbiota, and the roles of microbial metabolites in PCOS pathophysiology; and exploring combination therapies with existing PCOS treatments, which holds significant clinical translational value. It should be noted that letrozole itself may directly affect gut microbiota composition, as previously reported in rodent models.⁶⁸ Therefore, the observed dysbiosis in the PCOS model group may partly result from letrozole treatment rather than PCOS per se. Nevertheless, the reversal of these microbial changes by GPs suggests a therapeutic modulation regardless of the initial trigger. Furthermore, the precise molecular pathways through which GPs regulate gut microbiota and influence ovarian function need further elucidation. The mechanisms by which gut microbiota regulate metabolism and reproductive function are highly complex, involving synergistic interactions among various metabolites such as short-chain fatty acids, bile acids, and lipopolysaccharides. Further research is required to elucidate which specific metabolites are targeted by GPs for their therapeutic effects on PCOS and how different microbial communities collaboratively regulate these metabolic networks. As a female-specific disorder was modeled, only female rats were used; therefore, the findings cannot be generalized to males, and sex-specific microbiome effects were not assessed.

Resource availability

Lead contact

Requests for further information and resources should be directed to and will be fulfilled by the lead contact, Huali Huang (hualihuang999@126.com).

Materials availability

• •Gypenosides (GPs) used in this study were purchased from Baist Biological Technology Co., Ltd. (Xi’an, China). Letrozole, Met hydrochloride, and other chemical reagents were obtained from commercial sources as listed in the key resources table.
• •Fecal microbiota suspensions used for transplantation were freshly prepared for each experiment as described in the method details section.
• •This study did not generate new unique materials (e.g., new cell lines, antibodies, or recombinant proteins). All materials used in this study are commercially available or can be prepared following the protocols described herein.

Data and code availability

• •The 16S rRNA gene sequencing data and ovarian RNA sequencing data generated in this study have been deposited in the NCBI Sequence Read Archive (SRA) and are publicly available as of the date of publication. Accession numbers are listed in the key resources table.
• •This paper does not report any original code. All bioinformatics analyses were performed using publicly available software packages and R scripts, which are cited in the method details and key resources table.
• •Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Acknowledgments

This work was supported by Middle/Young aged Teachers’ Research Ability Improvement Project of Guangxi Higher Education (grant no. 2024KY0103), Self-funded research project of Health Commission of Guangxi Zhuang Autonomous Region (grant nos. Z-A20231144, Z-A20240991, Z-A20240397, and Z-A20250366), Self-funded research project of Administration of Traditional Chinese Medicine of Guangxi Zhuang Autonomous Region (grant nos. GXZYA20240466, GXZYA20240462, GXZYA20240929, and GXZYA20250906), the key clinical specialist pathology unit program of Guangxi Zhuang Autonomous Region (grant no. 2023QZD01), and the Key Research and Development Project of Nanning Science and Technology Bureau (grant no. 20253036-2).

Author contributions

S.Li, conceptualization, methodology, formal analysis, investigation, and writing; J.W., conceptualization, methodology, formal analysis, investigation, and writing; Z.L., investigation and writing; J.L., investigation and writing; H.J., investigation and writing; L.L., conceptualization, validation, writing – original draft, and supervision; Y.L., methodology and formal analysis; S.Luo, investigation and writing; H.H., conceptualization, validation, writing – original draft, and supervision.

Declaration of interests

The authors declare no competing interests.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Chemicals, peptides, and recombinant proteins
Gypenosides (GPs)	Baist Biological Technology Co., Ltd. (Xi’an, China)	Cat #BST4-05-100
Letrozole	MedChemExpress (Monmouth Junction, NJ, USA)	Cat #HY-14248
Metformin hydrochloride	Sigma-Aldrich (St. Louis, MO, USA)	Cat #PHR1084
Ampicillin sodium	Sigma-Aldrich	Cat #PHR1424
Neomycin sulfate	Sigma-Aldrich	Cat #PHR1491
Metronidazole hydrochloride	MedChemExpress	Cat #HY-B0318A
Vancomycin	Sigma-Aldrich	Cat #ST9H9BC19E08
Sodium carboxymethyl cellulose (CMC-Na)	Aladdin (Shanghai, China)	Cat #C1440802
Critical commercial assays
Rat Total Testosterone (T) ELISA Kit	Elabscience (Wuhan, China)	Cat #E-OSEL-R0003
Rat Follicle-Stimulating Hormone (FSH) ELISA Kit	Elabscience	Cat #E-EL-R0391
Rat Luteinizing Hormone (LH) ELISA Kit	Elabscience	Cat #E-EL-R0026
Rat IL-1β ELISA Kit	Abcam (Cambridge, UK)	Cat #ab255730
Rat IL-6 ELISA Kit	Abcam	Cat #ab234570
Malondialdehyde (MDA) Assay Kit	Beyotime Biotechnology (Shanghai, China)	Cat #S0131
Superoxide Dismutase (SOD) Activity Assay Kit	Beyotime Biotechnology	Cat #S0101
Fecal Genomic DNA Extraction Kit	TIANGEN Biotech (Beijing, China)	Cat #DP328
NEBNext Ultra II DNA Library Prep Kit	New England Biolabs (Ipswich, MA, USA)	Cat #E7645
NEBNext Ultra II Directional RNA Library Prep Kit	New England Biolabs	Cat #E7760
Deposited data
RNA-Seq analysis of ovarian transcriptome in PCOS rats treated with gypenosides	This paper	NCBI SRA: PRJNA1374190
16S rRNA gene sequencing of the gut microbiota in PCOS rats following gypenosides intervention	This paper	NCBI SRA: PRJNA1376439
RNA-Seq analysis of ovarian transcriptome in recipient rats after fecal microbiota transplantation from GPs-treated or PCOS-model donors	This paper	NCBI SRA: PRJNA1479305
16S rRNA gene sequencing of gut microbiota in recipient rats after fecal microbiota transplantation from GPs-treated or PCOS-model donors	This paper	NCBI SRA: PRJNA1478769
Experimental models: Organisms/strains
Rat: Sprague-Dawley (female, 6 weeks old, 180 ± 20 g) — PCOS model experiment	Experimental Animal Center of Guangxi Medical University (Nanning, China)	Approval No. 202503012
Rat: Sprague-Dawley (female, 6 weeks old, 180 ± 20 g) — FMT experiment	Experimental Animal Center of Guangxi Medical University (Nanning, China)	Approval No. 202601008
Oligonucleotides
16S rRNA V3–V4 forward primer: 5′-CCTAYGGGRBGCASCAG-3′	This paper (Synthesized by Sangon Biotech, Shanghai, China)	N/A
16S rRNA V3–V4 reverse primer: 5′-GGACTACNNGGGTATCTAAT-3′	This paper (Synthesized by Sangon Biotech, Shanghai, China)	N/A
Software and algorithms
bioinformatics.com.cn online platform	bioinformatics.com.cn	https://www.bioinformatics.com.cn
GraphPad Prism (v10.6.0)	GraphPad Software, USA	RRID:SCR_002798
QIIME2 (v2023.5)	QIIME2 development team	https://qiime2.org
DADA2 (v1.28.0)	Bioconductor	https://bioconductor.org/packages/dada2
PICRUSt2 (v2.5.0)	picrust2 development team	https://picrust2.github.io
HISAT2 (v2.2.1)	Daehwan Kim lab	http://daehwankimlab.github.io/hisat2
featureCounts (v2.0.1)	Subread package	http://subread.sourceforge.net
Cytoscape (v3.10.3)	Cytoscape Consortium	RRID:SCR_003032
Compound Discoverer (v3.3)	Thermo Fisher Scientific	https://www.thermofisher.com
fastp (v0.23.1)	Open source	https://github.com/OpenGene/fastp
LEfSe	Huttenhower lab	https://github.com/SegataLab/lefse
Other
Waters ACQUITY UPLC BEH C18 column (1.7 μm, 2.1 × 100 mm)	Waters Corporation	Cat #186002350
Q-Exactive Plus mass spectrometer	Thermo Fisher Scientific	Model: Q-Exactive Plus
Illumina NovaSeq 6000 platform	Illumina, USA	N/A
Olympus BX53 microscope	Olympus, Japan	Cat #BX53
SwissTargetPrediction database	Swiss Institute of Bioinformatics	https://www.swisstargetprediction.ch
STRING database (v11.5)	EMBL	https://string-db.org; RRID:SCR_005223
SILVA database (release 138)	Max Planck Institute	https://www.arb-silva.de
GeneCards database	Weizmann Institute of Science	https://www.genecards.org
OMIM database	Johns Hopkins University	https://www.omim.org
NCBI Gene database	National Center for Biotechnology Information	https://www.ncbi.nlm.nih.gov/gene
Traditional Chinese Medicine Systems Pharmacology Database (TCMSP)	Northwest A&F University	https://tcmsp-e.com
Waters ACQUITY UPLC BEH C18 column (1.7 μm, 2.1 × 100 mm)	Waters Corporation	Cat #186002350
Q-Exactive Plus mass spectrometer	Thermo Fisher Scientific	Model: Q-Exactive Plus
Illumina NovaSeq 6000 platform	Illumina, USA	N/A
SwissTargetPrediction database	Swiss Institute of Bioinformatics	https://www.swisstargetprediction.ch

Experimental model and study participant details

Animals

All animal experiments were conducted in accordance with the ARRIVE guidelines (https://arriveguidelines.org) and approved by the Animal Ethics Committee of Guangxi Medical University (Approval No. 202503012). A total of 36 female Sprague-Dawley rats (6 weeks old, weighing 180 ± 20 g at the start of the experiment) were obtained from the Experimental Animal Center of Guangxi Medical University (Nanning, China). Rats were housed under standard laboratory conditions with controlled temperature (22 ± 2 °C), relative humidity (40–60%), and a 12-h light/12-h dark cycle. Animals were housed in groups of three per cage with free access to standard rodent chow and autoclaved drinking water. All rats were acclimatized to housing conditions for one week prior to any experimental procedures. Female rats were selected because PCOS is a female-specific endocrine disorder, and the letrozole-induced rat model is a well-established preclinical model that recapitulates key features of human PCOS, including hormonal imbalances, ovarian morphological changes, and estrous cycle disruption. No human participants, human tissues, cell lines, or clinical samples were used in this study.

Method details

UPLC-Q-TOF/MS analysis of gypenosides

For chemical characterization of GPs, 20 mg of the GPs powder was accurately weighed and transferred into a 1.5 mL centrifuge tube. Then, 1 mL of methanol-water mixture (70:30, v/v) was added. The mixture was subjected to ultrasonic extraction at room temperature for 30 minutes, followed by centrifugation at 10,000 × g for 5 minutes. The resulting supernatant was filtered through a 0.22 μm nylon membrane prior to analysis.

The analysis was conducted on a Thermo Fisher Scientific Q-Exactive Plus system equipped with a Waters ACQUITY UPLC BEH C18 column (1.7 μm, 2.1 × 100 mm). The mobile phase consisted of 0.1% formic acid in water (solvent A) and 0.1% formic acid in acetonitrile (solvent B) under a gradient elution program. The flow rate was set at 0.4 mL/min with an injection volume of 5 μL. For mass spectrometric detection, electrospray ionization (ESI) was operated in positive and negative switching mode. Full MS scans in the range of m/z 100–1200 were acquired at a resolution of 70,000. Data-dependent acquisition (DDA) mode was employed for MS/MS scans at a resolution of 35,000, utilizing stepped collision energies. Data processing, including compound identification based on primary and secondary mass spectral data, was performed using the TCM workflow within Compound Discoverer software (Thermo Fisher Scientific), querying ChemSpider, mzVault, and mzCloud databases. The identity of the major gypenosides was confirmed by comparison with authentic reference standards (purity > 98%) purchased from Chengdu Chen Shu-kun Biotechnology Co., Ltd. (Chengdu, China).

Network pharmacology analysis

Target prediction was performed for the 20 gypenoside active ingredients identified by mass spectrometry using the SwissTargetPrediction database (https://www.swisstargetprediction.ch). Targets were supplemented and cross-verified via the Traditional Chinese Medicine Systems Pharmacology Database and Analysis Platform (TCMSP, https://tcmsp-e.com). Potential therapeutic targets for PCOS were retrieved from GeneCards (https://www.genecards.org), OMIM (https://www.omim.org), and NCBI (https://www.ncbi.nlm.nih.gov) databases using “Polycystic Ovary Syndrome” as the keyword. Only targets with a Relevance score greater than the median of the retrieval results in GeneCards were included. Targets from these databases were merged and deduplicated to establish the final PCOS disease target set.

The potential targets of gypenosides and PCOS disease targets were intersected to identify common targets using the bioinformatics.com.cn online platform (https://www.bioinformatics.com.cn). A protein-protein interaction (PPI) network of intersecting targets was constructed using the STRING database (https://string-db.org) with species set to “Homo sapiens” and a minimum interaction confidence score of 0.4. Gene Ontology (GO) functional enrichment and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were performed using the same online platform (pvalueCutoff = 0.05, qvalueCutoff = 0.2, Benjamini-Hochberg adjustment), with results visualized through the platform’s built-in tools. A “herb-compound-target-pathway” multi-level interaction network was constructed and visualized using Cytoscape software (version 3.10.3).

Letrozole-induced PCOS rat model and drug administration

Female Sprague-Dawley rats (6 weeks old, 180 ± 20 g) were randomly divided into 6 groups (n = 6 per group, Figure S1).

Group	Induction (Day 1–21)	Treatment (Day 22–49)
(I) Control	0.5% CMC (oral gavage)	0.5% CMC (oral gavage)
(II) PCOS model	Letrozole (1 mg/kg, oral gavage)	0.5% CMC (oral gavage)
(III) Metformin	Letrozole (1 mg/kg, oral gavage)	Metformin (200 mg/kg, oral gavage)
(IV) GPs low-dose	Letrozole (1 mg/kg, oral gavage)	GPs (75 mg/kg, oral gavage)
(V) GPs mid-dose	Letrozole (1 mg/kg, oral gavage)	GPs (150 mg/kg, oral gavage)
(VI) GPs high-dose	Letrozole (1 mg/kg, oral gavage)	GPs (300 mg/kg, oral gavage)

Letrozole (1 mg/kg) was dissolved in 0.5% CMC and administered daily by oral gavage for 21 consecutive days to induce the PCOS model, as previously established. After model induction, GPs (75, 150, or 300 mg/kg) or metformin (200 mg/kg) were administered daily for 28 consecutive days. All substances were dissolved in 0.5% CMC, prepared fresh daily, and the gavage volume was adjusted weekly based on body weight. The GPs doses were selected based on preliminary experiments and previous reports showing efficacy in metabolic disorder models without observable toxicity. Metformin was chosen as the positive control because it is a first-line insulin-sensitizing agent commonly used to improve metabolic and reproductive outcomes in PCOS patients.

Antibiotic treatment and fecal microbiota transplantation (FMT)

To investigate whether the therapeutic effects of GPs depend on gut microbiota, separate cohorts of rats were used for antibiotic-mediated microbiota depletion and FMT. This experiment was conducted under the approval of the Animal Ethics Committee of Guangxi Medical University (Approval No. 202601008).

Donor preparation

Three groups of donor rats (n = 6 per group) were established following the same protocol: control group (normal rats), PCOS model group, and GPs treatment group (300 mg/kg). Fresh fecal samples were collected from donor rats at the end of treatment, pooled within each group, and homogenized in sterile PBS (1:10 w/v). The homogenate was centrifuged at 800 × g for 3 minutes to remove debris, and the resulting bacterial suspension was used immediately for transplantation.

Antibiotic depletion

Recipient rats received a broad-spectrum antibiotic cocktail in drinking water for 7 consecutive days, consisting of ampicillin sodium (1 mg/mL), neomycin sulfate (1 mg/mL), metronidazole hydrochloride (1 mg/mL), and vancomycin (0.5 mg/mL). Fresh antibiotic solutions were prepared every 48 hours.

FMT procedure

A cohort of recipient rats (n = 12) was divided into two groups (n = 6 each): normal recipients and PCOS recipients. All recipients were pretreated with the antibiotic cocktail for 7 days. Subsequently, normal rats received oral gavage of 1 mL per 100 g body weight of fecal suspension from PCOS model donors (FMT_PCOS group), while PCOS rats received fecal suspension from GPs-treated donors (FMT_GPs group), once daily for 7 consecutive days. All recipient rats were then housed for an additional 21 days without treatment. Fecal samples were collected to verify microbiota engraftment by 16S rRNA sequencing.

Estrous cycle observation

Estrous cycles were monitored daily for 8 consecutive days at weeks 4 and 8 of the experimental period. Vaginal smears were obtained at a fixed time (9:00–10:00 AM) by gently introducing a small amount of physiological saline (0.9% NaCl) with a dropper. The lavage fluid was collected, smeared onto glass slides, and air-dried. Smears were stained using the Papanicolaou staining method and examined under a light microscope (Olympus BX53, Japan). Estrous cycle stages (proestrus, estrus, metestrus, and diestrus) were determined based on the predominant cell type population according to standard cytological criteria.

Serum hormone and biochemical measurements

At the end of the experiment, rats were anesthetized. Blood samples were collected via abdominal aortic puncture and allowed to clot at room temperature for 2 hours, then centrifuged at 2500 × g for 15 minutes at 4 °C. Serum was collected and stored at –80 °C.

Serum concentrations of total testosterone (T), follicle-stimulating hormone (FSH), luteinizing hormone (LH), interleukin-1β (IL-1β), and interleukin-6 (IL-6) were measured using commercial ELISA kits according to manufacturers’ instructions. Serum malondialdehyde (MDA) levels and superoxide dismutase (SOD) activity were assessed using biochemical assay kits. All ELISA assays and biochemical measurements were performed in duplicate, and absorbance was measured using a microplate reader (BioTek Epoch, USA).

Histopathological analysis

Ovarian and intestinal tissues were fixed in 10% neutral buffered formalin for 48 hours at room temperature, dehydrated through graded ethanol, cleared in xylene, and embedded in paraffin wax. Serial sections (5 μm thickness) were cut using a microtome (Leica RM2235, Germany). Sections were deparaffinized in xylene, rehydrated through graded ethanol, and stained with hematoxylin and eosin (H&E). Stained sections were examined under an optical microscope (Olympus BX53, Japan). Ovarian cystic follicles and corpora lutea were quantified from at least three non-overlapping fields per section by two independent observers blinded to group allocation.

RNA sequencing and bioinformatics analysis of ovarian tissue

RNA extraction and library preparation

Total RNA was extracted from rat ovarian tissues (left ovary) using TRIzol Reagent according to the manufacturer’s protocol. RNA quality was assessed using a Qubit 4.0 Fluorometer and Qsep400 Bioanalyzer. RNA samples with RIN ≥ 7.0 were used for library construction. Libraries were constructed using the NEBNext Ultra II Directional RNA Library Prep Kit for Illumina, which includes mRNA enrichment via poly(A) capture. Libraries were quantified and assessed for insert size distribution using an Agilent 2100 Bioanalyzer, then sequenced on an Illumina NovaSeq 6000 platform to generate 150 bp paired-end reads.

Data processing

Raw sequencing reads were processed with fastp software (v0.23.1) to remove adapters and low-quality sequences (quality score < 20, length < 50 bp). Clean reads were aligned to the Rattus norvegicus reference genome (Rnor_6.0) using HISAT2 (v2.2.1). Gene expression levels were quantified as FPKM using featureCounts (v2.0.1).

Differential expression analysis

Differential gene expression analysis was performed with the “limma” package (v3.50.0) in R (v4.5.0). Genes with |Log₂FC| ≥ 0.5 and adjusted p-value < 0.05 were considered statistically significant DEGs.

Functional enrichment

Functional enrichment: GO and KEGG pathway enrichment analyses were conducted using the bioinformatics.com.cn online platform (https://www.bioinformatics.com.cn) with p-value cutoff of 0.05. PPI networks were constructed using the STRING database and visualized with Cytoscape (v3.10.3).

16S rRNA gene sequencing and microbial community analysis

DNA extraction and PCR amplification

Total genomic DNA was extracted from rat fecal samples using the Fecal Genomic DNA Extraction Kit (TIANGEN, DP328). The V3–V4 hypervariable region of the bacterial 16S rRNA gene was amplified using primers 341F (5′-CCTAYGGGRBGCASCAG-3′) and 806R (5′-GGACTACNNGGGTATCTAAT-3′). PCR reactions (25 μL) contained 12.5 μL of Phusion High-Fidelity PCR Master Mix, 1 μL each of forward and reverse primers (10 μM), 1 μL of template DNA (10 ng/μL), and 9.5 μL of nuclease-free water. Thermal cycling conditions: 98 °C for 1 min; 30 cycles of 98 °C for 10 s, 50 °C for 30 s, and 72 °C for 30 s; and final extension at 72 °C for 5 min. PCR products were verified by 2% agarose gel electrophoresis and purified using the Qiagen Gel Extraction Kit. Libraries were constructed using the NEBNext Ultra II DNA Library Prep Kit and sequenced on an Illumina NovaSeq 6000 platform to generate 250 bp paired-end reads.

Bioinformatics analysis

Raw reads were processed using DADA2 in QIIME2 (v2023.5) to generate an amplicon sequence variant (ASV) table. Taxonomic assignment was performed against the SILVA database (release 138). Alpha diversity (Chao1, Observed ASVs, Shannon, Simpson, Abundance-based Coverage Estimator (ACE)) and beta diversity (Bray-Curtis distance, PCoA, NMDS) were calculated. PERMANOVA was performed with 999 permutations to assess group differences. Functional potential was predicted using PICRUSt2 (v2.5.0) mapped to KEGG pathways. Linear Discriminant Analysis Effect Size (LEfSe) analysis was performed with an LDA score threshold > 2.0 to identify differentially abundant taxa.

Sample collection and processing

At the end of the experiment, rats were deeply anesthetized and euthanized via exsanguination. Fecal samples were collected from the colon, snap-frozen in liquid nitrogen, and stored at –80 °C. Ovaries and intestinal tissues were dissected; the right ovary and intestinal segments were fixed in 10% neutral buffered formalin for histology, while the left ovary was snap-frozen and stored at –80 °C for transcriptomic analysis.

Quantification and statistical analysis

Statistical methods

All data are presented as the mean ± standard deviation (SD) unless otherwise specified. For data with error bars, SD was used for serum biochemical measurements, and SEM was used for body weight and histomorphometric analyses, as indicated in each figure legend. Detailed statistical information, including the specific tests used and exact p-values, is provided in each figure legend and in the Results section where applicable.

Statistical tests

All statistical analyses were performed using GraphPad Prism (v10.6.0). Normality was assessed using the Shapiro-Wilk test (p > 0.05 considered normal). For two-group comparisons, an F-test was conducted for homogeneity of variance; if met, an unpaired Student’s t-test was applied; otherwise, Welch’s t-test was used. For multi-group comparisons, one-way ANOVA was followed by Tukey’s HSD test (equal variances) or Dunnett’s test (against control). For non-normal data, the Kruskal-Wallis test followed by Dunn’s post-hoc test was used. A p-value < 0.05 was considered statistically significant. Significance symbols: ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

Definition of replicates

• •Biological replicates (n): Represent the number of independent animals per group (n = 5–6).
• •Technical replicates: ELISA and biochemical assays were performed in duplicate; sequencing samples were processed individually without pooling.

Software

See key resources table for all software versions and sources.

Additional resources

Not applicable. This study did not involve clinical trials, and no additional resources beyond those described above were required.