Introduction

Polycystic ovary syndrome (PCOS) is a common gynaecological endocrine disorder in women of reproductive age, with a global incidence of approximately 5–18%1. Clinical manifestations include menstrual disorders, infertility, hyperandrogenism, insulin resistance (IR) and obesity2,3. The clinical manifestations of PCOS are highly heterogeneous and complex. Diagnostic criteria vary and treatment is difficult, with serious implications for women’s physical and mental health4. Studies have shown that approximately 50% of women with PCOS are obese5. A meta-analysis of 16 studies found that women with PCOS have a higher risk of metabolic disease6. A prospective study in Spain found that 28% of overweight and obese women of childbearing age had PCOS7. The risk of PCOS increases with increasing BMI. Although the aetiology of PCOS is still unclear, obesity may be the driving factor for PCOS in at-risk individuals, and weight loss is considered an important treatment to improve symptoms in PCOS patients5,8. However, researchers still do not know why some women develop PCOS with obesity and whether some genetic disorders contribute. Therefore, a better understanding of the pathophysiological mechanism of PCOS and active exploration of new molecular mechanisms and potential therapeutic targets for obese patients with PCOS are important. Studies have shown that caffeine intake may reduce the risk of metabolic syndrome and type 2 diabetes mellitus9, but high caffeine intake (more than 200–300 mg/day) may lead to reduced fertility and adverse effects on offspring10. In addition, some clinical case studies have found that higher coffee consumption is associated with a lower risk of polycystic ovary syndrome11, where the polyphenols in coffee have been shown to increase insulin sensitivity, and some studies have demonstrated the ability of caffeine to inhibit inflammatory responses12,13, but high doses of caffeine have been found to inhibit the enzyme CYP19 (aromatase), leading to increased testosterone concentrations that worsen the symptoms of PCOS14.

In recent years, bioinformatics and microarray technology have been widely used to identify potential genes involved in diseases15, providing new methods for disease prevention and treatment. In this study, bioinformatics analysis was used to screen the SLC16A6 gene potentially related to the pathogenesis of obese PCOS patients. This gene belongs to the human solute carrier (SLC) transporter family, which comprises 400 genes and 52 subfamilies16 and has been shown to be involved in metabolic diseases17,18. As most SLCs are involved in the transport of small organic molecules, several drugs targeting SLCs have been approved for marketing19. At present, the study of SLCs has not progressed and no study has shown that the SLC family can be used as a therapeutic target for the treatment of polycystic ovary syndrome. Current research is focusing on the role of SLC transporters in metabolic diseases and determining whether a subtype of the SLC protein family interacts with the endocrine and metabolic problems associated with PCOS. These findings may help to identify or treat this complex reproductive endocrine disorder. Using reverse network pharmacology, this study predicted that coffee would have a potential therapeutic effect on obese PCOS patients. A clinical trial showed that healthy subjects who drank black coffee daily for 2 weeks experienced significant reductions in body weight and BMI. In addition, waist circumference and abdominal fat were reduced20. The main component of coffee is caffeine. Varillas-Delgado et al. reported that caffeine promotes fat oxidation and is involved in lipid metabolism in the body21, and a prospective cohort study confirmed that caffeine can effectively reduce the risk of obesity22. Caffeine reduces the risk of type 2 diabetes by ameliorating disturbances in the central insulin signaling pathway23.

Therefore, this study employed bioinformatics analysis alongside in vivo experiments to demonstrate that SLC16A6 has a regulatory effect on PCOS. Additionally, it suggests that coffee and its potent monomers may offer therapeutic benefits for PCOS by modulating SLC16A6. This research aims to provide valuable life advice for obese women with PCOS and to explore new therapeutic targets for this demographic.

Materials and methods

Bioinformatics

Analysis of differentially expressed genes and ROC curve analysis of obese PCOS patients

The datasets were downloaded from GEO, normalized, and DEGs were extracted with the R package “edgeR”20. Fold changes (FCs) were calculated for individual genes. After searching the GEO database, “GSE10946” and “GSE193812” were selected as the main research datasets. In GSE10946, 5 samples of granulosa cells from obese non-PCOS patients and 7 samples of granulosa cells from obese PCOS patients were used. GSE193812 used abdominal fat samples from 4 obese non-PCOS patients and 4 obese PCOS patients. The R packages “limma”, “edgeR” and “DESeq2” were used to analyse the DEGs in the two datasets, with the criteria set as P < 0.05 and absolute logFC ≥ 1. The intersection of the DEGs was obtained using a Venn diagram. The GSE10946 dataset was used for ROC curve analysis of the intersecting genes using the PROC package.

Analysis of differentially expressed genes and enrichment analysis of the SLC16A6 high- and low-expression groups

The “DESeq2” R package was used to select the samples with high and low SLC16A6 expression for the analysis of differentially expressed genes. After the DEGs were obtained, “clusterProfiler” was used for GO and KEGG enrichment analyses24,25,26.

Network pharmacology

Reverse network pharmacology analysis of SLC16A6

The herb database (http://herb.ac.cn/) was used to search for natural compounds related to SLC16A6 in the database by entering “SLC16A6” and performing target queries.

Target analysis of active coffee ingredients

The active compounds in coffee were obtained by literature searches, the active compounds were added to the herb database and the TCMSP database for queries, and the active compounds were stored with corresponding targets.

PCOS-related targets and potential targets of coffee in the treatment of PCOS

The PCOS targets were obtained from the following two databases: GeneCards27 and OMIM28. The keywords 'polycystic ovarian syndrome’ and ‘PCOS’ were searched in both databases to obtain disease targets. Online Wayne figure tools (https://bioinfogp.cnb.csic.es/tools/venny/) were used to match PCOS disease-related genes and corresponding coffee target intersections to identify potential coffee targets for the treatment of PCOS.

Protein‒protein interactions

Studying the interactions between protein networks can help to identify the core genes involved. A protein–protein interaction (PPI) analysis was performed using STRING, the potential targets obtained from the Venn diagram were entered into the “multiple proteins” analysis in STRING, and the species was restricted to “Homo sapiens”. The default confidence of the platform was used to improve the reliability of the data. The degree values of each node in the PPI network were obtained using the Cytoscape 3.8.0 analysis plug-in, in which the goal of the height value played a key role. Through this step, the core objectives of coffee were obtained.

Network construction

The data on the active components of coffee and potential targets for the treatment of PCOS were collated. The data obtained were imported into Cytoscape 3.8.0 software to visualise the drug-drug-disease-target gene network, and the number of targets corresponding to each drug was calculated.

Enrichment analysis

Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were performed using the DAVID tool29. Biological processes, cellular components, molecular functions and key signalling pathways were identified to explore the core mechanism of coffee and related biological pathways. The functional annotations with P values less than 0.05 in the enrichment results were further analysed. A bubble map of the enrichment results was drawn and three pathways, namely insulin, fat and oocyte development, were extracted for visualisation.

Molecular docking

Caffeine (PubChem CID: 2519) was molecularly docked to the protein SLC16A6 (Uniprot ID: O15403) using AutoDock Vina 1.1.2 software30. Protein models were downloaded from the AlphaFold protein structure database (https://alphafold.ebi.ac.uk/). Protein pre-processing (removal of water molecules and excess ligands, addition of hydrogen atoms) was performed using PyMol 2.4. Chemdraw 20.0 was used for energy minimisation of compounds. AutoDock Tools 1.5.6 was used to generate PDBQT files for docking simulations. The docking box was set to enclose the entire protein structure. Other parameters were left at their default values. The docking results were set to output 9 docking positions. The docking conformation with the lowest binding energy and the highest clustering frequency was considered to be the most likely binding mode between the ligand and the protein. Finally, the docking results were visualised using Pymol 2.4 software.

Experimental verification

Preparation and grouping of the obese PCOS rat model

Thirty-six 5-week-old female Sprague–Dawley rats of SPF grade were randomly selected for the study (weight 160 g-180 g), SD rats were obtained from Beijing Viton Lever Laboratory Animal Co. Ltd, [licence number: SCXK (Beijing) 2016-0006]. Ltd, [licence number: SCXK (Shanghai) 2017-0005]. They were kept in the animal room of the Experimental Centre of Shanghai Changhai Hospital, [Licence No.: SCXK (Shanghai) 2017-0005], in separate cages with free access to food and water. The indoor temperature of the animal room in the experimental centre was 28 ± 1.5 °C, the indoor relative humidity was 40 ± 5%, the day and night light/dark period was 12 h, and the experiment was started after 1 week of acclimatisation feeding. This experiment was reviewed and approved by the Experimental Animal Ethics Committee of the Medical Ethics Committee of Shanghai Changhai Hospital, Ethical Review No. CHEC (A.E) 2022-011, and all experimental procedures complied with the ethical regulations of animal experimentation.

At 5 weeks of age, female Sprague–Dawley rats were randomly divided into PCOS model group rats (model group, n = 22) and control group rats (control group, n = 8). In this study, according to the previous studies, the rats in the model group were administered 1 mg/kg letrozole by gavage daily along with high-fat chow (animal high-fat chow formula energy composition:60.65% fat, 21.22% carbohydrate and 18. 14% proteins from Synergistic Biologicals (XTHF60)), and the rats in the control group were administered 1 mg/kg saline by gavage daily, along with After 28 days of the experiment, the body mass of the rats was weighed, and compared with the control group, the body mass of the model group increased by 20%, which indicated that the obesity model was successful21, and the rats with body mass increased by more than 20% were further grouped, and the rats in the model group were randomly divided into the model group (n = 8), CFYN group (n = 6) and GLP-1 group (n = 8), and the model group was given 1 mg/kg saline by gavage daily. The model group was given 1 mg/kg saline by gavage daily, the CFYN group was given 10 mg/kg caffeine by gavage daily and the GLP-1 group was given 0.63 mg/kg simethicone by gavage once a day and was also fed high fat chow for 21 consecutive days.

Analysis of reproductive and metabolic phenotypes and blood and tissue sampling

From the beginning of the modelling period, the body weight of the rats in each group was weighed weekly and summarised. From the 15th day of modelling, the rats in each group were subjected to vaginal exfoliative cytology at 8:00 am every day until the end of the experiment. After the vaginal scrapings were stained with methylene blue, the estrous cycle of the rats was read and an estrous cycle chart was created to calculate the rate of estrous cycle disruption. At the end of the experiment (week 7), all rats were killed under chloral hydrate anaesthesia and blood and ovarian tissue were collected. Whole blood was centrifuged, serum was collected and stored at − 80 °C, the left ovary was fixed in 4% paraformaldehyde and the right ovary was frozen for mRNA determination.

Morphology of the rat ovary tissue

Ovaries were fixed in 4% paraformaldehyde and dehydrated through various concentrations of ethanol. Representative ovaries were embedded in paraffin, sectioned at 4 μm thickness and stained with haematoxylin and eosin. The fixed sections of rat ovarian tissue were examined under a microscope and photographed, and the number of cystic follicles and corpus luteum were counted.

Rat serum analysis using enzyme-linked immunosorbent assay (ELISA)

Serum levels of follicle-stimulating hormone (FSH), luteinising hormone (LH), fasting insulin (FINS) and fasting blood glucose (FBG) were determined using enzyme-linked immunosorbent assay (ELISA) kits, and the ratio of LH to FSH and the insulin resistance index (HOMA-IR) were calculated. The specific calculation method used was FINS*FBG/22.5. ELISAs were performed according to the instructions of the ELISA detection kits.

RT‒PCR was used to detect the expression of related genes in rat tissues

The tissues to be tested were excised from the frozen ovaries, and RNA was extracted using the TRIzol lysis method (according to the TRIzol instructions from Invitrogen Company) and reverse transcribed into cDNA (according to the RR047A instructions from TAKARA Company). The mRNA expression of the target gene was detected by real-time PCR (according to the instruction manual of TAKARA RR420A). After the reaction, the amplification curve and melting curve of real-time PCR were confirmed, and the data were analysed.

Statistical analysis

Means ± standard deviations were used for statistical descriptions. ANOVA was used to compare more than two groups of normally distributed data. The Wilcoxon rank sum test was used to compare sequencing data that were not normally distributed. All statistical tests were two-tailed. When P < 0.05, the difference was considered statistically significant.

In the Methods section, the inclusion of live animals is reported according to the ARRIVE guidelines (PLoS Bio 8(6), e1000412,2010): (1) identification of the institutional and/or licensing committee that approved the experiments, including all relevant details; (2) confirmation that all experiments were performed in accordance with relevant guidelines and regulations.

Results

Bioinformatics

From the GSE193812 dataset, a total of 472 differentially expressed genes in adipose tissue were identified, as shown in Fig. 1A (Table S1). In the GSE10946 dataset, a total of 61 DEGs were found in the granule cell cluster from obese or non-obese PCOS subjects, as shown in Fig. 1B (Table S2). The common differentially expressed gene was SLC16A6 (Fig. 1C). The GSE10946 dataset was selected to verify the expression of SLC16A6 and perform a ROC curve analysis. SLC16A6 was significantly increased in the granulosa cells of the PCOS group (P < 0.05) (Fig. 1D). The AUC was 0.866, indicating that SLC16A6 expression in the granulosa cells of the PCOS group was strongly reliable (Fig. 1E). A total of 356 DEGs were detected between the SLC16A6 high- and low-expression groups (Fig. 1F). GO and KEGG analyses were performed on 356 DEGs, and a total of 380 significant GO terms were identified (Table S3). The top 30 GO terms are shown in Fig. 1G. These pathways included “insulin-like growth factor receptor binding”, “insulin-like growth factor II binding”, insulin-like growth factor I binding, estradiol 17-beta-dehydrogenase activity, steroid dehydrogenase activity, etc., and “sex hormone-related pathways” (Fig. 1G). Twenty significantly different pathways were identified according to the KEGG enrichment results (Table S4): “TGF-beta signaling pathway”, “ovarian steroidogenesis”, “proline signaling pathway”, “regulation of lipolysis in” “adipocytes”, “type II diabetes mellitus” and other inflammation-, intraovarian sterol-, and diabetes-related pathways (Fig. 1H).

Fig. 1
figure 1

Bioinformatics analysis. (A) Differentially expressed genes in adipose tissue. (B) Differentially expressed genes in granular cells. (C) Shared differentially expressed genes in adipose tissue and granulosa cells. (D) Differential expression of the SCL16A6 gene in granulosa cells. (E) ROC curve of SCL16A6 expression in granulosa cells. (F) Differentially expressed genes between the SCL16A6 high- and low-expression groups. (G) GO enrichment analysis of the top 30 genes. (H) KEGG analysis of the top 20 differentially expressed genes.

Full size image

Network pharmacology

A total of 15 corresponding components associated with SLC16A6 (Table S5), which contained three coffee-related substances, namely, “coffee acid”, “caffeic”, and “trans-caffeic acid”, were obtained from the herb database. Seven effective substances in coffee were collected, namely, “coffee acid”, “caffeic”, “trans-caffeic acid”, “cafestol”, “trigonelline”, “caffeine”, and “chlorogenic acid”. A total of 147 targets corresponding to 7 substances were identified (Table S6) (Fig. 2A). The database revealed 3240 PCOS genes and 73 intersecting genes between active coffee ingredients and PCOS (Table S7) (Fig. 2B). The PPI network diagram of the intersecting genes was constructed (Fig. 2C), and the core targets of the PPI network diagram were analyzed. The top three targets were “INS”, “TP53”, and “TNF” (Fig. 2D). A network diagram of coffee components and therapeutic targets for PCOS treatment was constructed (Fig. 2E), and the corresponding targets of each component were statistically analyzed. Forty-six targets corresponding to caffeine were identified (Fig. 2F), which was far more than the number of other components, indicating its core role. KEGG and GO analyses were performed on the 73 intersecting genes. Figure 3A shows the involvement of genes in each pathway, and the INS genes were involved in multiple pathways. Figure 3B shows the top 30 GO terms, including “regulation of insulin secretion”, “glucose transmembrane transporter activity”, “D-glucose transmembrane transporter” activity” and “other insulin or glucose-related pathways”. After the KEGG analysis, the intersecting gene enrichment pathways were closely related to fat metabolism, insulin, and oocyte development, as shown in Fig. 3C.

Fig. 2
figure 2

Network pharmacology of coffee. (A) Target map of each component of coffee. (B) Related targets of coffee in the treatment of PCOS. (C) PPI network of intersecting targets. (D) HUBBA map of core targets. (E) Therapeutic target–active component network diagram. (F) Ranking of component targets.

Full size image
Fig. 3
figure 3

GO/KEGG analyses of the intersecting genes. (A) Sankey diagram of pathway targets. (B) Target GO map. (C) KEGG map of action targets.

Full size image

Molecular docking

Molecular docking was performed to verify the affinity between the targets of caffeine and SLC16A6, and the correspondence between targets and components is shown in Fig. 4. The SLC16A6 protein sequence structure was searched in the Alphafold2 database and pre-processed for all-atom docking with the molecular conformation of the energetically minimised component. In docking, the lower the binding energy, the more stable the binding.

Fig. 4
figure 4

The results showed that four sites, THR-305, PHE-308, TYR-39 and GLU-159, had high docking activity with caffeine. The main interaction modes were Carbon-Hydrogen Bond, Fi-Fi Stacked and Fi-Alkyl.

Full size image

The best conformations of active compounds docked to target molecules are shown in Fig. 4. The results showed that four sites, THR-305, PHE-308, TYR-39 and GLU-159, had higher docking activities with caffeine and their interaction modes were mainly carbon-hydrogen bond, fi-fi stacked and fi-alkyl, suggesting that caffeine may act as an endogenous ligand and specifically bind to the corresponding target proteins in the body to play a role in lowering blood pressure. It is suggested that caffeine may act as an endogenous ligand, specifically binding to the corresponding target protein in the body and exerting a blood glucose-lowering effect.

Experimental verification

Reproductive and metabolic phenotypes of obese PCOS rats

In this study, the obese PCOS group was confirmed by changes in body weight. At the end of the induction period (4 weeks), the body weight of the model group increased significantly (339.5 ± 29.8 g) compared to the control group (234.7 ± 11.8 g) (Fig. 5A). Compared with the control group (35.04%), the weight change in the model group was significantly greater (95.35%) (Fig. 5B). The vaginal images of the control group suggested that the control group had a regular, complete estrous cycle of 4 to 5 days, whereas the vaginal smear of the model group showed aperiorism, mainly in the interestrus period (Fig. 5C,D).

Fig. 5
figure 5

The establishment of an obese PCOS rat model was determined by weight changes. (A) Body weight and (B) body weight change (% initial body weight). The data are presented as the means ± SDs (significance level * P < 0.05). Estrous cycle and ovarian morphological changes in obese PCOS rats. (C, D) Estrous cycles of two representative rats: E (estrus), M (metestrus), D (diestrus), and P (proestrus). (E) Photographs of representative ovarian morphology.

Full size image

HE-stained sections from the control group showed normal ovarian morphology. Microscopy revealed that the ovaries of the control group had multiple corpora lutea and follicles at various stages of development, and multilayered granulosa cells were neatly arranged. The HE-stained sections of the ovaries in the model group showed obvious polycystic morphology with typical polycystic pathological changes under the microscope, with a large number of dilated cystic follicles, no or little luteal tissue, corona radiata and oocytes, and a significant reduction in the number of granulosa cell layers (Fig. 5E).

In conclusion, rats subjected to long-term intragastric administration of letrozole and a high-fat diet (model group) exhibited obvious obesity-like PCOS changes (Fig. 6).

Caffeine treatment restored the estrous cycle in obese PCOS rats

Vaginal exfoliative cytology was performed on the four groups of rats and continuous observations were made until the end of the experiment. The results of the vaginal smears showed that all rats in the control group had a regular and complete estrous cycle lasting 4–5 days. In the model group, the estrous cycle was disrupted in 6 rats and the normal rate of the estrous cycle was only 25%, mainly manifested as a prolongation of the interestrus. The CFYN and GLP-1 groups were similar to the control group. Five rats in the CFYN group had a normal estrous cycle after the experiment, and the normal rate of the estrous cycle was 83%. Six rats in the GLP-1 group had normal estrous cycles at the end of the experiment, and the percentage of normal estrous cycles was 75% (Fig. 7A). Based on histological HE staining, the study showed that the model group still had significant PCOS-like changes in the ovaries, and the CFYN and GLP-1 groups showed reduced numbers of cystic follicles, increased numbers of corpus lutea, and improvements in polycystic changes (Fig. 7B).

Fig. 7
figure 6

Caffeine treatment and the improvement of metabolism in obese PCOS rats. (A) Body weights, (B) FBG levels, (C) FINS levels, (D) GLP-1 levels, and (E) HOMA-IR. Caffeine treatment improved hormone levels in obese PCOS rats. (F) T levels, (G) LH levels, and (H) FSH levels. The data are presented as the means ± SDs (significance level: * model vs. CON, P < 0.05; #CFYN group and GLP-1 group compared with the model group, P < 0.05).

Full size image

Effects of caffeine on body weight and endocrine metabolism in obese PCOS rats

In this study, the successful establishment of an obese PCOS model was confirmed by changes in body weight. After the establishment of the model (after 4 weeks), the obese PCOS rats were randomly divided into three groups: the untreated obese PCOS group (model group), the caffeine gavage treatment group (CFYN group) and the oral semaglutide gavage treatment group (GLP-1 group). After 3 weeks of treatment, no significant difference in body weight was observed between the CFYN group (373.3 ± 66.2 g) and the model group (401.5 ± 31.7 g), whereas the GLP-1 group (281.8 ± 36.3 g) and the model group (401.5 ± 31.7 g) showed a significant reduction in body weight (Fig. 6A).

Fig. 6
figure 7

Estrous cycle and ovarian morphological changes in obese PCOS rats treated with CFYN. (A) Estrous cycles of four representative rats: E (estrus), M (metestrus), D (diestrus), and P (proestrus). (B) Photographs of representative ovarian morphology.

Full size image

Serum samples were collected after a 12-h fast. GLP-1, fasting plasma glucose and fasting insulin levels were compared between the model, CFYN, GLP-1 and control groups, and the corresponding insulin resistance index was calculated. A significant difference was observed between the model and control groups (P < 0.05). The FINS level of the CFYN group was significantly improved compared to the model group (P < 0.05), and no significant difference was observed between the CFYN group and the GLP-1 treatment group (Fig. 6B–E).

In this study, serum testosterone, FSH and LH levels were analysed in each group (Fig. 6F–H). The data of this study showed that the testosterone level of the model group (11.74 ± 1.2 nmol/L) was significantly higher than that of the control group (7.23 ± 1.35 nmol/L) (P < 0.05), showing the typical hyperandrogenemic features of PCOS. No significant difference in testosterone levels was observed between the CFYN group and the GLP-1 group. (P < 0.05) (Fig. 6F). The FSH level in the model group (5.9 ± 0.84 U/l) was significantly higher than that in the control group (2.79 ± 0.38 U/l) (P < 0.05), and a significant difference was observed between the CFYN group and the model group (P < 0.05); however, no significant difference was observed between the CFYN group and the GLP-1 group (Fig. 6G). Furthermore, the LH level in the model group (8.69 ± 1.49 mIU/ml) was significantly higher than that in the control group (3.69 ± 1.27 mIU/ml) (P < 0.05), which is a typical feature of PCOS. A significant difference was observed between the CFYN group and the model group (P < 0.05), but no significant difference was detected between the CFYN group and the GLP-1 group (Fig. 6H), indicating that CFYN could significantly restore the ovarian function of PCOS rats and improve their symptoms.

Effect of caffeine on the expression of the SLC16A6 transporter gene in obese PCOS rats

RT-PCR was used to determine the gene expression of the SLC16A6 transporter in different groups of rats. Compared with the normal control group, the expression of the SLC16A6 gene was significantly decreased in the model group, and no significant differences were observed between the control, CFYN and GLP-1 groups. However, significant differences were observed between the CON, CFYN and GLP-1 groups and the model group (P < 0.05) (Fig. 8).

Fig. 8
figure 8

Caffeine treatment affects SLC16A6 gene expression in obese PCOS rats. The data are presented as the means ± SDs (significance level *P < 0.05).

Full size image

Discussion

PCOS is a common heterogeneous endocrine metabolic disorder in women of reproductive age, the pathogenesis and pathological features of which are not fully understood31, but recent studies have shown that PCOS patients with obesity are more likely to have reproductive and metabolic abnormalities, and that obesity can exacerbate the reproductive and metabolic abnormalities in PCOS patients32,33, Excessive weight gain can lead to insulin resistance and other hormonal disturbances, and obesity and elevated androgen levels can also affect female reproductive function34,35. In addition, studies have shown that obesity and increased insulin resistance are potentially modifiable risk factors for high risk of female reproductive dysfunction, and addressing these risk factors may help to alleviate or treat female reproductive dysfunction36. According to the treatment recommendations of the International Evidence-Based Guidelines for PCOS, lifestyle interventions including diet, exercise and behaviour are recommended as first-line treatment for obese PCOS patients37. However, simple dietary and lifestyle changes have relatively high loss rates in most cases, and it is difficult for obese PCOS patients to adhere to them effectively to achieve weight control; therefore, it is important to use medication-assisted therapy to help obese PCOS patients control their weight and improve their endocrine levels.

SLC16 consists of 14 members of the monocarboxylic acid transporter (MCT) family of proteins, which play a critical role in the transport of essential cellular nutrients, as well as cellular metabolism and pH regulation38. Transgenic animal models have shown that SLC transporter proteins are involved in many important metabolic processes, including nutrient supply, metabolic conversion, energy homeostasis, etc.39. Current evidence suggests that SLC16 family plays a critical role in lipid metabolism and insulin sensitivity. Deletion of SLC16a13 lead to reduced hepatic lipid accumulation when fed a high-fat diet. SLC16A1-4 are in involved in the transport of monocarboxylates like lactate and pyruvate, which are important for cellular metabolism. As a member of SLC16 family, it has been found that the absence of SLC16A6 expression in liver tissue is closely associated with the development of type 2 diabetes, suggesting that the pathway of action of SLC16A6 transporter proteins is related to glucose and lipid metabolism40. The key factor in the treatment of obese PCOS patients is controlling body weight, and losing weight can improve follicular development. The expression of SLC16A6 transporter is closely related to glucose and lipid metabolism and the transport and excretion function of ketone body41. According to DEGs in adipose tissue and granulosa cells, only SLC16A6 was downregulated in both adipose tissue and granule cell. In addition, A total of 15 corresponding components associated with SLC16A6 (Table S5), which contained three coffee-related substances, namely, “coffee acid”, “caffeic”, and “trans-caffeic acid”, were obtained from the herb database. Combined with gene expression of SLC16A6, SLC16A6 possibly plays a critical role in the development of obese PCOS.

Results obtained by screening for DEGs in PCOS samples from obese or non-obese PCOS subjects were found that the expression of SLC16A6 is elevated in both adipose and granulosa cells of obese patients with PCOS, and considering the physiological function of SLC16A6, this elevation may be physiological compensatory and needs to be further explored. And after validation in animal experiments, the present study confirmed the correlation between SLC16A6 expression and obese PCOS. Reduced expression of SLC16A6 was found in the ovaries of obese PCOS rats, which may be related to the impaired glucose-lipid metabolism pathway in obese PCOS, and it was hypothesised that reduced expression of the SLC16A6 transport protein is one of the causes of insulin resistance in obese PCOS. The expression level of SLC16A6 could be restored by caffeine treatment, suggesting that there is also a correlation between the expression of the SLC16A6 transporter protein and the alteration in reproductive function.