Letrozole was purchased from Heng Rui Pharmaceutical Company (190605KG, Lianyungang, Jiangsu, China). Mogroside V (50.42% mogroside V), prepared from fresh fruit of Siraitia grosvenorii, was a generous gift from Professor Xingwei Liang, who purchased from Guilin Layn Natural Ingredients Corp. (MOV04-18022504, Guilin, Guangxi, China). Carboxymethylcellulose sodium (CMC) was purchased from Henan Wan Bang Industrial Co., Ltd. (Henan, China). TransScript^® One-Step gDNA Removal and cDNA Synthesis SuperMix Kit (AT311-02) were purchased from TransGen Biotech (Beijing, China). 2 × Universal SYBR Green Fast qPCR Mix kit (RK21203) was purchased from ABclonal Technology Co., Ltd. (Wuhan, China). LDHA (C28H7) rabbit mAb (#3558), hexokinase II (HK2) rabbit mAb (#2867) and PKM2 (D78A4) XP^® mAb (#4053) were purchased from Cell Signaling Technology, Inc. (Danvers, Massachusetts, USA). β-tubulin (No: 66240-1-Ig), horseradish peroxidase-conjugated goat anti-rabbit IgG (H+L) (SA00001-2) and biotin-conjugated affinipure goat anti- rabbit IgG (H+L) (SA00004-2) were purchased from Protein Tech Group Inc. (Chicago, USA). Wright’s–Giemsa Stain solution (G1020) and 20 × Metal Enhanced DAB Substrate Kit (DA1015) were purchased from Beijing Solarbio Science & Technology Co., Ltd (Beijing, China). BCA Protein Assay Kit (CW0014S), eECL Western Blot Kit (CW0049M) and SDS-PAGE Gel Kit (CW0022S) were purchased from Beijing ComWin Biotech Co., Ltd. (Beijing, China). Trizol reagent (15596026 and 15596018) was purchased from Thermo Fisher Scientific (Waltham, USA). Key resources of the study were also summarized in Supplemental Tables S1, S2 . All primers were designed and synthesized by Shanghai Sangon Biotechnology Co., Ltd. (Shanghai, China).

Female Sprague Dawley (SD) rats (n = 30, 5 weeks old, 173.66 ± 7.73 g) were purchased from the Laboratory Animal Center of the University of South China (Hengyang, China). All animals were kept in standard environmental conditions (25°C, 12 h light/dark cycle) with free access to standard food pellets and water. After adaptive feeding for one week, the animals were randomly divided into the control group (n = 10) and the PCOS group (n = 20). The rats from the control group were fed with standard food pellets; the rats from PCOS group were orally administrated with letrozole (1 mg/kg/day, dissolved in 1% (w/w) CMC) (7) and fed with a high-fat diet (HFD, consisting of 61.5% standard food, 12% lard, 5% sucrose, 5% milk powder, 5% peanut, 10% egg, 1% sesame oil, 0.5% salt) from 1^st to 30^th days. From 21^st to 30^th day, the estrous cycle was identified by vaginal smear. According to the estrous cycle, the female rats successively stayed in metestrus and diestrus stages were preliminarily considered as PCOS models. Then, the young-adult PCOS rats were further divided into the PCOS group and the PCOS + MV group (n = 10 for each group). Rats in the PCOS + MV group were administrated through an oral gavage with mogroside V (MV, 600 mg/kg/day, dissolved in 0.9% NaCl) from 31^st to 60^th days, and vaginal cytology was performed from 51^st to 60^th days. At the end of the treatment period, all rats were anesthetized with an intraperitoneal injection of 10% chloral hydrate (0.8 ml/100 g) before sacrifice. All animal experiments were approved by the Animal Ethics Committee of University of South China (No. SYXK2020-0002).

Vaginal smears were performed for 10 consecutive days at 13:00 from the 21^st day of modelling and initiation of MV administration, respectively. Briefly, the vagina was flushed with 25-30 μl saline (0.9% NaCl) for 2~3 times. The vaginal fluid (10-20 μl) was sampled onto the glass slides and air dried at room temperature. Then, the slides were stained with Wright’s–Giemsa Stain solution (G1020, Solarbio, Beijing, China) following the manufacturer’s protocol. Finally, the estrous cycle including proestrus, estrous, metestrus and diestrus phases and changes of vaginal epithelial cells were evaluated according to the predominant cell type (19, 20) under a light microscope (BX43, OLYMPUS, PA, USA).

Following 12 h of fasting, blood collected from the tail vein was used to measure fasting blood glucose (FBG) with a Roche blood glucose meter (Roche diagnostic GmbH, Germany). All rats were weighed and anesthetized with 20% urethane, and the whole blood sample was collected by abdominal aortic puncture immediately. The serum was further isolated by centrifuging (2000 g) for 10 min at 4 °C. Finally, the levels of serum luteinizing hormone (LH), follicle-stimulating hormone (FSH), total testosterone and insulin were determined with radioimmunoassay (Beijing North Institute of Biotechnology Co., Ltd., Beijing, China). In brief, the serum samples (FSH: 200 μL, LH: 200 μL, testosterone: 50 μl, Insulin: 100 μL) were incubated with corresponding labeled antibodies (FSH and LH: 4 °C for 20 h, testosterone: 37 °C for 1 h, insulin: 37 °C for 2 h). Then, an immune separation agent was added to remove the uncombined antigens by mixing and centrifuging at 3800 rpm for 15 mins. After that, the radioactivity of the compound was detected to measure the corresponding hormone levels. Formula (FBG × insulin level)/22.5 was used to detect homeostasis model assessment of insulin resistance (HOMA-IR) (21).

The left ovaries of rats from each group were fixed in 4% paraformaldehyde and embedded in paraffin after serial dehydration, then were sequentially sectioned with a thickness of 6 μm and stained with hematoxylin and eosin (H&E). The morphological changes of the ovarian tissues were checked and photographed using a light microscope (BX 43; Olympus, PA, USA). In addition, the number of corpora lutea was counted in one representative section taken from the middle of the ovary under 10 × or 20 × magnification according to previous studies (22, 23).

The right ovaries of rats from each group were sent to Shanghai Applied Protein Technology Co., Ltd. (Shanghai, China) to analyze the targeted energy metabolomics via LC-MS/MS. In brief, 100 mg ovarian tissues, 200 μl of ultra-pure water and 800 μl of methanol/acetonitrile (1:1, v/v) were first added into the Eppendorf tubes. Then, the samples were grounded by ultrasonic homogenizers, incubated at 20 °C for 60 min, and centrifuged at 13, 000 rpm for 15 min at 4 °C. The supernatants were harvested and dried with nitrogen, and the lyophilized powder was stored at -80 °C. The lyophilized samples were reconstituted by dissolving with 100 μl solvent mixture containing water/acetonitrile (1:1, v/v). The samples were vortexed for 1 min and centrifuged at 14, 000 rpm for 15 min at 4 °C. Finally, the supernatants were collected for LC-MS/MS analysis. A hierarchical clustering heatmap was produced to visualize the metabolites expression level using the OmicStudio tools. (https://www.omicstudio.cn/tool/4).

RNA-seq library construction and RNA sequencing of ovaries (n = 4 rats per group) were performed by the Novogene Corporation (Beijing, China). Total RNA (3 μg) was used as the initial RNA for database construction. The NEBNext^® Ultra TM RNA Library Prep Kit (Illumina, USA) was employed to build the library. The Illumina platform was employed to sequence these libraries using a 150 bp paired-end strategy. Reads with adapter, reads with N (N means the base information cannot be determined) and low-quality reads (reads with Qphred ≤ 20 bases accounting for more than 50% of the read length) were mainly removed and filtered from the original data. The construction of the index of the reference genome and the comparison of clean reads at the mating end with the reference genome were both performed with HISA T2 v2.0.5.

Feature Counts were used to calculate the readings mapped to each gene (24). The fragments per kilobase of transcript per million fragments mapped (FPKM) was then calculated based on the length of the gene and the readings mapped to that gene were calculated. Differential expression analysis of two conditions/groups (three biological replicates per condition) was performed using the DESeq2 R package (1.16.1). The resulting P-values were adjusted using Benjamini and Hochberg’s approach to control the false discovery rate (25). Genes with an adjusted P-value (Padj) < 0.05 were assigned as differentially expressed genes (DEGs).

GO enrichment analysis of DEGs was implemented by the cluster-Profiler R package. GO terms with a corrected P-value less than 0.05 were considered to be significantly enriched by DEGs. ClusterProfiler R package was used to test the statistical enrichment of DEGs in KEGG pathways (http://www.genome.jp/kegg/).

Total mRNA was extracted from ovarian tissues with Trizol reagent according to the manufacturer’s procedure. RNA (2 μg) was reversely transcribed into cDNA by using TransScript^® One-Step gDNA Removal and cDNA Synthesis SuperMix Kit. RT-qPCR was performed in 10 μl volume using 2 × Universal SYBR Green Fast qPCR Mix kit and Applied Biosystems QuantStudio 3 (Thermo Fisher Scientific). Gapdh was used as the reference and gene expression levels were calculated using the 2^-ΔΔCt method. The mRNA expression levels were calculated as relative expression to Gapdh. All the primers were summarized in Table 1 , and the Ct values were summarized in Supplemental Table S3 .

Immunohistochemical analysis was conducted in accordance with a previous study with minor modification (21). Immunohistochemistry was performed on formalin-fixed and paraffin-embedded specimens with primary antibodies including polyclonal LDHA antibodies (1:300 dilution), HK2 antibodies (1:400 dilution) and PKM2 antibodies (1:500 dilution), respectively. The immunoreactive signals of proteins were visualized with 20 × Metal Enhanced DAB Substrate Kit. Finally, sections were observed and photographed with an Olympus DP70 digital camera mounted on a Leica DMR microscope.

Western blotting (WB) assay was performed as described previously (26). Briefly, the protein concentration was detected by a BCA Protein Assay Kit. Briefly, 40 μg protein samples were separated with 10% SDS-PAGE and transferred to polyvinylidene fluoride membranes (Meck Millipore, Merck & Co., Inc., NJ, USA). Then, transferred membranes were blocked by 5% skim milk for 1 h at room temperature. The primary rabbit monoclonal antibodies including LDHA (1:1000 dilution), HK2 (1:1000 dilution), PKM2 (1:1000 dilution), and β-tubulin (1:3000 dilution) were respectively incubated overnight at 4°C, and the incubation of a horseradish peroxidase-conjugated goat anti-rabbit IgG (1:5000 dilution) was performed for 1 h at room temperature. β-tubulin was detected as a loading control. Finally, eECL was added, and the Tanon-5500 Chemiluminescence Imaging System was used to detect the chemiluminescence of protein bands. Finally, blotting images were quantified using Image J analysis software (JAVA image processing program, NIH, Bethesda, USA).

Bioinformatics analysis was performed with the OmicStudio tools (https://www.omicstudio.cn/tool). For all the data results, intra-assay coefficient of variation was < 10%, and inter-assay coefficient of variation was < 15%. All data were shown as mean ± standard error of the mean (SEM) unless otherwise stated. Differences between multiple groups were analyzed by one-way analysis of variance (ANOVA) followed by Bonferroni’s multiple comparison test. All statistical analyses were performed using GraphPad Prism 9 (GraphPad Prism, USA). A p-value less than 0.05 was considered statistically significant.

To identify the effects of MV, young-adult PCOS rat model was established, and the experimental design was shown in Figure 1A . There was a remarkable increase in body weight of the rats after the treatment of HFD and letrozole ( Figure 1B and Table 2 ), and this increase was significantly reduced after MV administration ( Figure 1C and Table 2 ). Estrous disorder is one of the main characteristics of PCOS rats (27, 28), so the effects of MV on the estrous cycle of young-adult PCOS rats were conducted. In the control group, the rats had a regular estrous cycle (4-5 days) including proestrus, estrous, metestrus and diestrus phases sequentially ( Figures 1D, E ), which is consistent with previous studies (19, 20, 29). In the proestrus phase, there were mainly small and round nucleated epithelial cells with greyish red nucleus and greyish cytoplasm (black arrows point), which were relatively uniform in appearance and size. Estrous phase was predominantly characterized with flat and anucleated keratinized epithelial cells showing as blue or purple-blue stacks or layers (red arrows point). However, in the metestrus phase, neutrophils and nuclear epithelial cells were the predominant compositions, and anucleated keratinized epithelial cells were occasionally observed. In the diestrus phase, there were almost full of neutrophils ( Figure 1D , blue arrows point). Compared with the controls group, young-adult PCOS rats showed disrupted estrous cycles with prolonged diestrus but shortened proestrus and estrous phases ( Figure 1F ). Intriguingly, the irregular estrous cycle was gradually returned to be normal after MV administration ( Figures 1D–F ).

To reveal the underpinning mechanism of protective effects of MV administration on dysfunctional estrous cycle, we further evaluated the effects of MV treatment on the hormone levels, ovary weight and ovary morphology of young-adult PCOS rats. There were no significant differences in FBG among control, PCOS and PCOS-MV groups ( Figure 2A and Table 2 ). Compared with the control group, though obvious increase was found in the serum level of insulin, HOMA-IR and testosterone of PCOS group, MV treatment only overtly reduced the level of testosterone in the PCOS-MV group, even there was downtrend without statistical significance found in insulin level and HOMA-IR ( Figures 2B-D and Table 2 ). In addition, there were no statistical differences in the serum levels of LH, FSH and LH/FSH between control vs PCOS and PCOS vs PCOS-MV groups ( Figures 2E–G ). The ovary weight of PCOS group significantly increased, while that of PCOS-MV group apparently decreased ( Figure 2H ). Moreover, to check the improvement of MV on the folliculogenesis and ovulation of young-adult PCOS rats, we analyzed histological morphology of ovary and counted the number of corpora lutea. Compared with the control group, multiple follicles with cystic expansion presented vacuolated and disorganized structure and only a few corpora lutea were found in the PCOS group. However, follicles at different stages and obviously increased numbers of corpus luteum were observed in the control and PCOS-MV groups ( Figures 2I–K ). Interestingly, the MV administration recovered the thickness of granulosa cell layers ( Figures 2I–K ).

Given the key roles of glycolysis in the development and mature of follicles (10), we mainly focused on the main metabolites of glycolysis including D-Glucose 6-phosphate, pyruvate, lactate, ATP and GTP ( Figure 3A , red fonts). Compared with the control group, the production of D-Glucose 6-phosphate, lactate, ATP and GTP in the PCOS group was significantly down-regulated excepting an obvious up-regulation in the production of pyruvate ( Figures 3A–F ). MV administration markedly increased the levels of D-Glucose 6-phosphate, lactate, ATP and GTP, and inversely decreased the pyruvate level ( Figures 3A–F ).

To further explore the molecular mechanisms underlying the recovery of MV administration on the development and energy metabolism of ovaries, we performed transcriptome sequencing (RNA-seq) of ovarian tissues. There were respective 828 upregulated and 1359 downregulated genes in the PCOS group when compared with the control group (26). Compared to the PCOS group, 969 and 819 genes were respectively upregulated and downregulated in rats from the PCOS-MV group ( Figure 4A ), among which 583 of the 828 genes upregulated genes and 676 of 1359 downregulated in the PCOS group were significantly reversed after MV administration ( Figure 4B ). These reversed genes are considered as MV-responsive genes ( Figure 4B ). In addition, the DEGs were involved in metabolic pathways, such as glycolysis, PI3K-Akt signaling pathway and peroxisome proliferator-activated receptor signaling pathway ( Figure 4C ). Notably, MV administration restored these metabolism-related biological processes, which were dysregulated in young-adult PCOS rats ( Figure 4D ). The expression levels of several genes related to glycolysis such as LDHA, HK2 and PKM exhibited a reversed phase between the PCOS and PCOS-MV groups ( Figure 4D ). These results led to the speculation that MV may improve the glycolysis in the ovaries of young-adult PCOS rats.

ATP production originally relies on the levels of glycolytic rate-limiting enzymes, which play indispensable roles in controlling glycolysis rate (30). LDHA, HK2 and PKM2 are key rate-limiting enzymes of glycolysis (31–33), which regulates the energy supply of oocytes and GCs (34). Given the above-mentioned transcriptome profiling that shows the reversed expression profile of LDHA, HK2 and PKM between the PCOS and PCOS-MV groups ( Figure 4D ), we further detected the expression levels of the three glycolytic rate-limiting enzymes. Compared with the control group, the mRNA expression levels of Ldha, Pkm2 and Hk2 were significantly down-regulated, MV administration obviously increased the mRNA expression of Ldha, Pkm2 and Hk2 ( Figures 5A–C ). WB analysis presented the expression of LDHA, HK2 and PKM2 were significantly down-regulated in the ovarian tissues of young-adult PCOS rats, which was significantly up-regulated after MV administration ( Figures 5D–G ). Immunohistochemistry (IHC) analysis further showed that LDHA, HK2 and PKM2 markedly decreased in the PCOS group as compared with the control group and recovered in the PCOS-MV group ( Figure 5H ).