Skip to main content
Download PDF

Granulosa cell RNA-Seq insights into senescence and sphingolipid metabolism disorder in PCOS: aspirin as a potential therapeutic drug

Reproductive Biology and Endocrinology volume 23, Article number: 61 (2025) Cite this article

Abstract

Background

Polycystic ovary syndrome (PCOS) is a pivotal cause of anovulatory infertility and the pathogenesis remains elusive. Cellular senescence and sphingolipid metabolism disorder are closely intertwined, and both have been demonstrated present within the granulosa cells of PCOS, while research on the combined impact of senescence and sphingolipids on PCOS-related anovulation is scarce.

Methods

Here, we leveraged four datasets of PCOS and executed differential gene expression analysis, engaged in WGCNA, and harnessed machine learning algorithms—including RF, SVM-RFE, and LASSO—to deeply explore the key genes that interact with senescence and sphingolipid metabolism in granulosa cells of PCOS. These key genes were subjected to further analysis to construct a diagnostic model, forecast immune cell infiltration, and identify potential agents. Additionally, within the testosterone-stimulated granulosa cells, we validated the expression of key genes, confirmed senescence and sphingolipids dysregulation, and evaluated the therapeutic efficacy of the candidate agent.

Results

Our research pinpointed a set of genes (LYN, PLCG2, STAT5B, MMP9, and IL6R) that showed promise as biomarkers for PCOS-related anovulation and the diagnostic nomogram was developed. These biomarkers were linked to various immune cell types infiltration. In testosterone-stimulated granulosa cells, we observed increased expression of these biomarkers, accompanied by signs of senescence and changes in sphingolipids. Importantly, the potential agent aspirin displayed the ability to ameliorate these two processes.

Conclusion

This study highlighted the important value of genes associated with senescence and sphingolipids dysregulation in PCOS. Aspirin targeting senescence could be a promising therapeutic drug for addressing anovulation associated with PCOS.

Introduction

Polycystic ovary syndrome (PCOS) is a common endocrine disorder affecting 6–10% women of reproductive age worldwide, characterized by a triad of symptoms including hyperandrogenism, menstrual irregularities, and the presence of multiple cysts in the ovaries [1]. It is the leading cause of infertility in women and is also associated with an increased risk of metabolic disorders such as insulin resistance, type 2 diabetes, and dyslipidemia [2]. Anovulation stemming from abnormal follicular development is the most critical characteristic of PCOS. The disrupted follicular microenvironment in PCOS is a vital factor of the abnormal follicular development and oocyte maturation [3]. Granulosa cells, the predominant somatic cell population in the follicular microenvironment, normally respond to hormonal signals and support the growth of a dominant follicle, which eventually leads to ovulation [4]. However, this process is impaired in PCOS patients, resulting in the development of numerous small cysts or follicles, which ultimately leads to anovulation or even infertility [5]. Therefore, further study on the dysfunction of granulosa cells is helpful to understand the molecular mechanism of abnormal follicular development in women with PCOS.

Sphingolipids, a significant lipid class, are found in all eukaryotic cells and account for 2–20% of the membrane lipids [6]. Bioactive sphingolipids especially ceramide (Cer), sphingosine-1-phosphate (S1P), ceramide-1-phosphate (C1P), and others [7], play pivotal roles in various biological processes including cell differentiation, proliferation, apoptosis, and migration [8]. Notably, S1P is well acknowledged for its protective role in follicular development by inhibiting granulosa cells apoptosis [9], and has been utilized in follicle cryopreservation and ovarian transplantation [10, 11]. In contrast to S1P, Cer has been reported to inhibit granulosa cell growth and consequent follicular atresia [9, 12]. Therefore, the balance of sphingolipid metabolism, particularly S1P/Cer ratio, is essential for follicular development and ovulation. Nevertheless, in the follicular fluid and granulosa cells derived from PCOS patients, diminished levels of S1P and elevated levels of Cer have been detected [13,14,15].

Cellular senescence is defined by its hallmark features of irreversible cell cycle arrest, resistance to apoptosis, dysregulation of metabolism, and the presence of a senescence-associated secretory phenotype (SASP) [16]. Recent research has revealed the accumulation of senescent granulosa cells in PCOS, the expression levels of senescence markers, including p16, p21, p53, and β-galactosidase, were markedly elevated in the granulosa cells from PCOS patients and PCOS mice model compared to controls [17]. These senescent granulosa cells displayed dysfunctional attributes that substantially hinder folliculogenesis in PCOS [18].

The concomitant presence of sphingolipid metabolism disorder and cellular senescence within PCOS granulosa cells provides a compelling rationale to investigate potential pathophysiological crosstalk between these two processes in mediating follicular developmental impairment. Indeed, emerging evidence has implicated Cer as a potent senescence inducer across various cell types, including vascular endothelial cells [19], neuron cells [20], and skeletal myoblasts [21]. In PCOS granulosa cells, despite the absence of evidence directly substantiating Cer-mediated pro-senescence effects, this bioactive sphingolipid has been demonstrated to elicit oxidative stress [12], mitochondrial dysfunction [22], and DNA damage [23], all of which are critical drivers of cellular senescence [24]. In addition to the pro-senescence effects of certain sphingolipids, senescence has been reported to modulate cellular sphingolipid profiles. Laurila et al. identified marked accumulation of dihydroceramide in senescent murine skeletal muscle tissue [25]. In tumor cells, chemotherapy-induced senescence led to elevated ceramide biosynthesis [26, 27]. These researches demonstrate the significant correlation between sphingolipid metabolic disorders and cellular senescence, which has not been investigated within PCOS granulosa cells to date. Hence, this study is designed to focus on the interplay between cellular senescence and sphingolipid metabolism in PCOS through a comprehensive bioinformatics and machine learning-driven analysis of granulosa cells RNA-Seq from PCOS-associated datasets. The study also seeks to evaluate the diagnostic efficacy of genes implicated in both cellular senescence and sphingolipid metabolism for PCOS-related anovulation, and to identify potential pharmacological interventions.

Materials and methods

Data acquisition and processing

In this study, four transcription profiles (GSE106724, GSE137684, GSE95728, GSE34526) were downloaded from the NCBI GEO datasets (https://www.ncbi.nlm.nih.gov/gd/). Inclusion criteria were as follows: (1) The 2003 Rotterdam criteria served as the foundation for diagnosing PCOS; (2) datasets encompassing data from both PCOS subjects and normal controls; (3) test specimens obtained from datasets originating from human ovarian granulosa cells. The raw data was processed by the ArrayExpress R package (version 1.62.0, https://www.bioconductor.org/packages/2.9/bioc/html/limma.html). Subsequently, the sva R package (version 3.50.0, http://www.bioconductor.org/packages/release/bioc/html/sva.html) was employed to eliminate batch effects and to integrate the expression profile data from datasets GSE106724, GSE137684, and GSE95728. The raw expression data underwent log(x + 1) normalization, and genes expressed as 0 in over 50% of the samples were excluded to form the training set for subsequent analysis. The GSE34526 dataset was utilized as the validation set. Figure 1 shows the process flow of data processing in this research.

Fig. 1
figure 1

The workflow of this study

Full size image

Differential expression genes analysis

The limma R package (version 3.58.1, https://bioconductor.org/packages/release/bioc/html/limma.html) was utilized to perform differential analysis between the PCOS and control groups, and yielded the corresponding P value and log2 fold change (log2FC) for each gene. And P < 0.05 & |log2FC|≥ 0.5 were as the threshold for selecting differentially expressed genes (DEGs).

WGCNA algorithm identifies module genes related to PCOS

Weighted gene co-expression network analysis (WGCNA) was utilized to uncover gene modules that are highly correlated with sample phenotypes, and to summarize the feature genes within these gene modules. It starts by assuming that gene networks follow a scale-free network model, defining a gene co-expression correlation matrix and an adjacency function for gene network formation, then calculating the dissimilarity coefficient for different nodes, and identifying gene sets associated with phenotypes. We employed the WGCNA R package (version 1.72–5, https://cran.r-project.org/web/packages/WGCNA/index.html) to analyze the matrix data of the top 5000 genes based on absolute median deviation (MAD), using PCOS and control groups as traits, and selected module genes related to PCOS for analysis.

Identifying intertwined senescence-sphingolipid metabolism genes in PCOS

The cellular senescence-related genes (CSGs) were downloaded from the Molecular Signatures Database (MSigDB, https://www.gsea-msigdb.org/gsea/msigdb/index.jsp) and the Human Cellular senescence Genomic Resources (HAGR, http://genomics.senescence.info/genes). The sphingolipid metabolism-related genes (SMGs) were obtained from the MSigDB database and the InnateDB database (http://www.innatedb.com). The genes at the intersection of database CSGs, WGCNA module genes, and DEGs were identified as DEGs-WGCNA-CSGs. Similarly, DEGs-WGCNA-SMGs were determined by intersecting database SMGs, WGCNA module genes, and DEGs. Subsequently, DEGs-WGCNA-CSGs and DEGs-WGCNA-SMGs were filtered with a correlation coefficient threshold of |R|> 0.6 (adjustable) and P < 0.001 to identify senescence-related sphingolipid metabolism genes (intertwined genes) in PCOS. These intertwined genes were further screened using the Wilcoxon test between PCOS and control samples (P < 0.05).

Functional enrichment analysis of intertwined genes in PCOS

We used the clusterProfiler R package (version 4.10.1, http://bioconductor.org/packages/release/bioc/html/clusterProfiler.html) for Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis. An adjusted P < 0.05 indicated statistical significance.

PPI network construction

The protein interaction relationships between the intertwined genes were searched using STRING (https://cn.string-db.org/), with a threshold of combined_score > 0.4 to construct the Protein–Protein Interaction (PPI) network. Subsequently, Cytoscape (version 3.8.2, https://cytoscape.org/) was utilized, which employs four algorithms: Maximum Connectivity Cluster (MCC), Maximum Neighborhood Component (MNC), Degree, and Extracellular Processing Cluster (EPC) to predict and explore the top 30 significant genes within the PPI network. The intersection of genes identified by the four algorithms are considered candidate hub genes.

Machine learning for selecting diagnostic feature genes

Three machine learning algorithms—LASSO, random forest (RF), and support vector machine-recursive feature elimination (SVM-RFE)—were utilized to further screen the potential feature genes for the diagnosis of PCOS. Feature selection is performed through tenfold cross-validation, with genes having a MeanDecreaseGini > 1 considered important variables. The LASSO regression, SVM-RFE, and RF analyses were conducted using the R packages “glmnet version 4.1–8”, “e1071 version 1.7–14”, and “randomForest version 4.7–1.1”. The common genes extracted through the amalgamation of these three machine-learning algorithms were subsequently considered as biomarkers for the prediction of PCOS.

Immune infiltration analysis and its correlation with feature genes

The single sample gene set enrichment analysis (ssGSEA) algorithm of GSVA R package (version 1.50.5, https://www.bioconductor.org/packages/release/bioc/html/GSVA.html) was utilized to calculate the proportion of 28 immune cell types based on the expression levels of samples in the training set. The Wilcoxon test was applied to assess the differences in immune cell distribution between PCOS and control groups. Pearson's correlation was then used to calculate the associations between each immune cell type. Finally, the analysis examined the correlation between the expression levels of feature genes and immune cells.

GSEA for molecular mechanisms

Gene set enrichment analysis (GSEA) is a computational approach designed to determine if a specific set of genes exhibits statistically significant differences across two distinct biological conditions. We employed GSEA to investigate the enrichment of KEGG pathways among the feature genes. The threshold applied were an adjusted P < 0.05 and |NES|> 1, ensuring the identification of significantly enriched pathways.

Diagnostic nomogram visualization

The nomogram for PCOS based on the feature genes was meticulously formulated through the rms R package (version 6.8–0, https://cran.r-project.org/web/packages/rms/index.html). This nomogram is a graphical representation of a regression model, allocating scores according to the regression coefficients of the independent variables. The accuracy of the nomogram's predictions is evaluated through the use of calibration curves, decision curve analysis (DCA), and ROC curves.

Drug prediction

The Drug Signatures Database (DSigDB, https://dsigdb.tanlab.org/DSigDBv1.0/) is a comprehensive online database with information about drugs and their targets. Utilizing this database, we endeavored to predict the drugs that may target the key DEGs in the granulosa cells of PCOS patients and conducted visualization to facilitate subsequent correlation analyses.

Human ovarian granulosa tumor cell line (KGN) culture and treatment

The human ovarian granulosa tumor cell line (KGN) was utilized in this study. This cell line, derived from granulosa cells of an ovarian cancer patient and subsequently immortalized, exhibits the capacity to synthesize steroid hormones and retains the growth characteristics of granulosa cells. These features make it a commonly employed model for investigating the functions and hormonal regulation of granulosa cell [28]. KGN cells were maintained in Dulbecco's Modified Eagle Medium (DMEM, Gibco, USA) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin at 37 °C within a CO2 incubator. To simulate the pre-ovulatory granulosa cells in PCOS, KGN cells were treated with 10 μM testosterone for 24 h. However, this method has its limitations, which will be further discussed in the Discussion section. Furthermore, to ascertain the influence of aspirin, the testosterone-preconditioned KGN cells were incubated with varying concentrations of aspirin (0 μM, 10 μM, 20 μM, 40 μM) for 24 h.

Total RNA extraction, reverse transcription and quantitative real-time PCR

Total RNA was extracted from cells using TRIzol reagent (Takara, Japan). 1 μg total RNA was reverse transcribed in a 20 μL volume by HiScript IV RT SuperMix for qPCR (Vazyme, China). Real-time PCR was performed with Taq Pro Universal SYBR qPCR Master Mix (Vazyme, China). Specific primers used for PCR amplification were synthesized with the sequences as shown in Table 1 The relative expression levels of target genes were analyzed using the 2−ΔΔCt method.

Table 1 Primers used for real-time PCR
Full size table

Protein extraction and western blot analysis

Cells were washed with PBS and lysed in RIPA buffer (Beyotime, China), enriched with phenylmethylsulfonyl fluoride (PMSF, Servicebio, China) and a protease inhibitor cocktail (Servicebio, China), on ice for 10 min followed by centrifugation at 12,000 rpm for 15 min at 4 °C. Proteins were then mixed with SDS-PAGE loading buffer (Servicebio, China), heated at 95 °C for 10 min, and 20 μg protein was subjected to SDS-PAGE. The resolved proteins were transferred onto 0.45 μm PVDF membranes (Millipore, Germany). The membranes were blocked with 5% nonfat milk in Tris-buffered saline-Tween 20 (TBST) for 1 h at room temperature before being incubated with primary antibodies against p16, p21 and GAPDH (all from Proteintech, China) overnight at 4 °C. On the next day, the membranes were washed and incubated with secondary antibodies (Proteintech, China) for 1 h at room temperature. After washing, the immunoreactive blots were detected with an enhanced chemiluminescence (ECL, Vazyme, China) substrate.

Cell viability measurements

KGN cells were cultured in 96-well plates overnight. Subsequently, 10 μL of the Cell Counting Kit-8 (CCK-8, Beyotime, China) reagent mixed with 90 μL of DMEM was added to each well. The plates were then incubated at 37 °C for 1 h before the optical density (OD) at 450 nm was measured using a multimode microplate reader (Thermo Scientific, USA).

Senescence associated SA-β-Gal activity

SA-β-Gal staining was conducted with β-Galactosidase Staining Kit (Beyotime, China), following the protocol in reference [29]. The KGN cells were first fixed with 0.5% glutaraldehyde for 15 min, then washed with PBS containing 1 mM MgCl2. The staining procedure continued at 37 °C overnight in a PBS solution including 1 mM MgCl2, 1 mg/mL X-Gal, and 5 mM each of potassium ferricyanide and potassium ferrocyanide. Six representative images from each sample were assessed.

Edu staining experiment

Place KGN cells on a coverslip in a 24-well plate and culture until confluent. Incorporate 10 μM Edu into the medium and incubate for 6 h. Wash with PBS, and fix cells with 4% paraformaldehyde for 30 min. Permeabilize cells with 0.5% Triton X-100 for 10 min. Apply click-chemistry mixture with CuSO4 and a fluorescent dye for 30 min. Use Hoechst to stain cell nuclei for 10 min. Wash and mount coverslip on a slide. Examine Edu-positive cells under a fluorescence microscope. Count Edu-positive and total cells in six fields.

Statistical analysis

All experimental data are the result of three independent experiments. All results are presented as the mean ± standard deviation. GraphPad Prism software (version 9.0, GraphPad Software, USA) was used for statistical analysis. Student’s t-test was utilized for analyzing the difference between the two groups, and one-way analysis of variance (ANOVA) followed by Tukey’s multiple comparison test was used to determine significance across multiple groups. The correlation between the variables was determined using Pearson’s or Spearman’s correlation test. All statistical P values were two-sided, and P < 0.05 was regarded as statistical significance.

Results

GEO expression data processing and analysis of variance

Four datasets (GSE106724, GSE137684, GSE95728, and GSE34526) were retrieved from GEO, with the first three datasets used as training sets and the fourth as a validation set. Gene expression data from the training sets were first subjected to quantile normalization to standardize the datasets. These datasets were then merged, and batch effects were removed to ensure homogeneity. A total of 23 granulosa cell samples from PCOS patients and 15 samples from healthy controls were included. As shown in Fig. 2A, the datasets were initially distinct, with no overlap, but their intersection was used as a unified batch for subsequent analysis after preprocessing. Figure 2B illustrates the principal component analysis (PCA) results for the control and PCOS groups across the datasets. Differential gene expression analysis was performed comparing granulosa cells from PCOS patients with those from control subjects. A total of 963 DEGs were identified, with 797 upregulated and 166 downregulated (Supplementary Material 1). The volcano plot and heatmap of the top 50 DEGs are shown in Fig. 2C–D, respectively.

Fig. 2
figure 2

GEO expression data pre-processing. A, B Principal component analysis (PCA) of the GSE106724, GSE137684, and GSE95728 datasets. The points of the scatter diagrams display the samples based on the top two principal components of gene expression profiles without (A) and with (B) the removal of batch effect. C Volcano map of DEGs between PCOS and control samples, blue dots represent low expression and red dots represent high expression. D Heat map of DEGs between PCOS and control samples

Full size image

Identification of module genes associated with PCOS

To identify the most significant modular genes in granulosa cells of PCOS, we performed WGCNA and used the correlation coefficient method to construct a sample clustering dendrogram (Fig. 3A). Hierarchical clustering was then applied to generate a gene co-expression network. The soft threshold for constructing the scale-free network was determined to be 3, as it was the first power value that resulted in a scale-free topology index of 0.85 (Fig. 3B). We then partitioned the gene expression data and merged modules with high similarity, identifying six distinct modules, each containing at least 50 genes (Fig. 3C). Based on the Pearson correlation heatmaps for each module, the turquoise module, consisting of 2,689 genes (R = 0.68, P = 5e-06), showed the strongest significant correlation (Fig. 3D, Supplementary Material 2). This finding was further supported by scatter plots, which demonstrated a highly significant level of connectivity within the turquoise module (cor = 0.77, P < 1e-200, Fig. 3E).

Fig. 3
figure 3

Weighted gene co-expression network analysis based on DEGs. A Clustered tree diagram of PCOS and its control groups. B The scale-free fit index for various soft-thresholding powers (β) and the mean connectivity for various soft-thresholding powers. C Clustering dendrogram of DEGs with dissimilarity based on the topological overlap, together with assigned module colors. D Heatmap of the correlation between modules and clinical traits, where red represents positive correlation, blue represents negative correlation, and the intensity of the color signifies the strength of the association. E Scatterplot of correlations between gene significance and module membership in turquoise module

Full size image

Identification of genes associated with both the cellular senescence and sphingolipid metabolism in PCOS granulosa cells

By performing an intersection analysis combining DEGs, hub module genes identified through WGCNA, and SMGs, we identified 7 pivotal genes (Supplementary Material 3), referred to as DEGs-WGCNA-SMGs, which were correlated with sphingolipid metabolism processes in PCOS granulosa cells (Fig. 4A). Similarly, by focusing on CSGs, we identified 77 key genes (Supplementary Material 4), termed DEGs-WGCNA-CSGs, which are highly involved in this metabolic pathway within PCOS granulosa cells (Fig. 4B). A Pearson correlation analysis was then performed on the genes from both DEGs-WGCNA-CSGs and DEGs-WGCNA-SMGs, using a threshold of an absolute correlation coefficient |R| > 0.6 and P < 0.001. As a result, a total of 79 genes termed CS-SMGs were identified (Supplementary Material 5), indicating an interplay between cellular senescence and sphingolipid metabolism in PCOS granulosa cells. To further investigate the biological functions of these 79 interconnected genes, GO functional analysis and KEGG pathway analysis were performed, with an adjusted P < 0.05 used to identify significantly enriched terms. The biological processes (BP) included cellular response to biotic stimulus, cellular response to lipopolysaccharide, and cellular senescence. Molecular functions (MF) included cytokine activity and cytokine receptor binding. Cellular components (CC) included azurophil granule lumen, cytoplasmic vesicle lumen, and early endosome (Fig. 4C). KEGG analysis revealed enrichment in the MAPK signaling pathway, FoxO signaling pathway, and other relevant pathways (Fig. 4D). An interaction network for the 79 interconnected genes was constructed using STRING, which included 65 nodes and 366 interaction relationships (Fig. 4E‒F). We then utilized the cytoHubba plugin in Cytoscape to predict and highlight the top 30 pivotal genes in the PPI network. Using four distinct topological analysis algorithms—MCC, MNC, EPC, and Degree—we successfully identified 28 key genes likely to play significant roles in this network (Fig. 4G, Supplementary Material 6).

Fig. 4
figure 4

Screening of genes associated with both cellular senescence and sphingolipid metabolism. A Venn diagram for intersection genes shared by WGCNA, DEGs and SMGs, and these genes were identified as DEGs-WGCNA-SMGs. B Venn diagram for intersection genes shared by WGCNA, DEGs and CSGs, and these genes were named DEGs-WGCNA-CSGs. C Dotplot for GO functional analysis and (D) dendrogram for KEGG pathway analysis of 79 DEGs associated with both cellular senescence and sphingolipid metabolism. E, F PPI network of these 79 DEGs. Threshold: adjusted P < 0.05. G Venn diagram for screening intersection genes based on four topological analysis algorithms—MCC, MNC, EPC, and Degree

Full size image

Identification of diagnostic biomarkers for PCOS via machine learning

The LASSO, SVM-RFE, and RF algorithms were employed to further screen diagnostic genes associated with cellular senescence and sphingolipid metabolism in PCOS granulosa cells. Using the LASSO algorithm, we applied ten-fold cross-validation to select the optimal criteria with high accuracy, ultimately identifying 6 relevant feature genes (Fig. 5A‒B). For the SVM-RFE algorithm, the error rate was minimized when the number of features was set to 17 (Fig. 5C‒D). Under the criterion of Mean Decrease Gini > 1, the RF algorithm identified 7 important feature genes (Fig. 5E‒F). By intersecting the results from the LASSO, SVM-RFE, and RF algorithms, we identified five common feature genes: LYN, PLCG2, STAT5B, MMP9, and IL6R. Their interrelationships were visualized using a Venn diagram (Fig. 5G). Furthermore, the ROC curve analysis confirmed the significant predictive value of these feature genes (Fig. 5H‒I).

Fig. 5
figure 5

Identification of potential diagnostic biomarkers. A Ten-fold cross-validation of tuning parameter selection in the LASSO model. Each curve corresponds to one gene. B LASSO coefficient profile. The solid vertical line indicates the standard error of the partial likelihood deviation. The dashed vertical line is drawn at the optimal λ value. C, D SVM-RFE algorithm for validating biomarker gene expression. E RF plot for the relationship between the number of trees and the error rate. F Arrangement based on the relative importance of genes. G Venn diagram of intersected genes identified by LASSO, SVM-RFE, and RF algorithms. H, I Diagnostic efficacy of the ROC curves of diagnostic biomarkers in training set and validation set, respectively

Full size image

Construction and validation of diagnostic nomogram for PCOS

To evaluate the collective diagnostic efficacy of LYN, PLCG2, STAT5B, MMP9, and IL6R for PCOS in clinical practice, nomogram plots were constructed (Fig. 6A). These plots allow for the calculation of patient scores based on the expression profiles of the selected signature genes, facilitating the estimation of the likelihood of PCOS presence in patients. The calibration curves revealed a minimal discrepancy between the actual and predicted risk of PCOS occurrence (Fig. 6B), while the decision curve analysis (DCA) demonstrated that the nomogram curve outperformed the intervention-none reference line (Fig. 6C). Both results highlighted the higher net benefit and accuracy of the nomogram in diagnosing PCOS. The AUC value (AUC = 0.904) in the ROC curves exceeded 0.85, further confirming the reliability and accuracy of the prediction model (Fig. 6D).

Fig. 6
figure 6

Construction and validation of a nomogram model. A Nomogram of diagnostic biomarkers in the training group. B Calibration curve to assess the predictive power of the nomogram model (C) DCA curve to evaluate the clinical application value of nomogram model. D ROC curve for the nomogram model

Full size image

Single-gene GSEA of feature genes

Using the KEGG gene sets as a reference, we performed GSEA to identify the KEGG signaling pathways associated with the five feature genes, applying a threshold of adjusted P < 0.05 and |NES| > 1. The top 10 most significant pathways for each feature gene are shown in Fig. 7, providing insights into the biological processes potentially regulated by these genes.

Fig. 7
figure 7

GSEA reveals the KEGG enrichment pathways of the 5 diagnostic genes

Full size image

Detection of immunological features in PCOS

Follicular development is a highly intricate process that is not only regulated by hormonal signals but also influenced by a complex interplay of immune cells. The follicular environment is infiltrated by various immune cells, such as T cells, B cells, macrophages, and dendritic cells, which secrete cytokines and growth factors that affect follicular growth, selection, and atresia [30]. In this study, we analyzed the infiltration of immune cells associated with the five feature genes in PCOS. Using expression profile data from the training set, we applied the ssGSEA algorithm to determine the distribution levels of 28 distinct immune cell types between the PCOS and normal groups. A total of 22 immune cell types exhibited significant differences in their distribution (Fig. 8A‒B). Furthermore, we calculated the correlations between the five feature genes and the 28 immune cell types, revealing the complex relationships between gene expression profiles and immune cell distributions in the context of PCOS (Fig. 8C).

Fig. 8
figure 8

Evaluation of immune cell infiltration. A Heatmap of immune cells infiltration in PCOS samples and control samples based on ssGSEA algorithm. B Boxplot for relative percentage of immune cells in each sample. (C) Heatmap for relationships between 5 signature genes and immune cells. * P < 0.05, ** P < 0.01, *** P < 0.001. ns, no statistically significant difference

Full size image

Identification of candidate small-molecular compounds for PCOS treatment

To further investigate potential small molecule compounds that may reverse the gene expression pattern associated with PCOS and exert therapeutic effects, the 5 diagnostic genes in PCOS were imported into the DSigDB database. The top 10 compounds including GLYCOPROTEIN (BOSS), methotrexate (BOSS), glutathione (CTD 00006035), raloxifene (CTD 00007367), benzene (CTD 00005481), aspirin (CTD 00005447), 7,8-Benzoflavone (CTD 00000606), 170449–18-0 (CTD 00003361), hexachlorobenzene (CTD 00006091), and dexamethasone (CTD 00005779), were identified as potential agents for PCOS treatment (Fig. 9).

Fig. 9
figure 9

Barplot for the relationship of candidate small-molecular compounds and the 5 diagnostic genes

Full size image

Validation of the expression of diagnostic genes and the effects of potential drugs

Hyperandrogenism is a common phenomenon in women with PCOS and is a primary contributor to the granulosa cells dysfunction [31]. Therefore, to determine the mRNA levels of the 5 diagnostic genes in granulosa cells of PCOS, we exposed the KGN cell line (a model for granulosa cells) to testosterone. As a result, the mRNA expression of these genes upregulated (Fig. 10A). We further investigated the impact of testosterone on senescence and sphingolipid metabolism in KGN cells. The testosterone-stimulated KGN cells displayed an elevated rate of SA-β-Gal positivity (Fig. 10B), a common indicator of cellular senescence, along with upregulated expression of the senescence-associated proteins p21 and p16 (Fig. 10C). This was coupled with a noticeable decrease in cell proliferation, as evidenced by the diminished Edu staining (Fig. 10D) and the corresponding cell proliferation curve (Fig. 10E). Moreover, testosterone treatment led to an increase in intracellular Cer level within the KGN cells (Fig. 10F).

Fig. 10
figure 10

The key genes and potential drugs vValidation. KGN cells were initially subjected to a 24-h pre-treatment with 10 μM testosterone, after which we assessed: (A) the mRNA levels of 5 diagnostic genes, (B) SA-β-Gal activity, (C) the protein expression levels of p16 and p21, and (D) Edu immunofluorescence staining. E The proliferation curves of KGN cells with and without 10 μM testosterone intervention, across a 5-day period. F The intracellular ceramides levels of KGN cells with or without 24-h 10 μM testosterone. G KGN cells, preconditioned with testosterone, were treated with a range of aspirin concentrations (0 μM, 10 μM, 20 μM, 40 μM) for 24 h, after which the protein levels of p16 and p21 were quantified. H The activity of SA-β-Gal, (I) Edu immunofluorescence staining, (J) cell proliferation rates, and (K) intracellular ceramide levels were additionally assessed in KGN cells treated with either 0 or 20 μM aspirin. * P < 0.05, ** P < 0.01, *** P < 0.001. ns, no statistically significant difference

Full size image

The potential of aspirin to alleviate the senescence and sphingolipids dysregulation in granulosa cells of PCOS was further assessed in this study. We treated KGN cells, which had been pre-exposed to testosterone, with different concentrations of aspirin for 24 h. The protein levels of p21 and p16 significantly reduced when aspirin concentrations exceeded 20 μM (Fig. 10G), and a decrease in SA-β-Gal positivity was observed with 20 μM aspirin treatment in KGN cells (Fig. 10H), suggesting that aspirin can mitigate testosterone-induced senescence in granulosa cells. Additionally, aspirin treatment enhanced the intensity of Edu staining and resulted in relatively accelerated cell proliferation (Fig. 10I-JJ). In parallel, the intracellular level of ceramide was significantly reduced in KGN cells following aspirin treatment (Fig. 10K).

These results suggest that alterations in 5 diagnostic genes associated with cellular senescence and sphingolipid metabolism may be present in the granulosa cells of PCOS patients, and aspirin could potentially serve as a therapeutic agent for infertility related to PCOS.

Discussion

PCOS is associated with 80% of anovulatory infertility [32], while the pathophysiology of PCOS-related anovulation remains unclear. Additionally, PCOS is linked to adverse pregnancy outcomes, including gestational diabetes, hypertensive disorders, and preterm birth [33]. Granulosa cells, being essential in follicular development, exhibit dysfunction which may contribute to the anovulation in PCOS [4]. In recent research, the interplay between cellular senescence and sphingolipid metabolism, recognized as potential contributors to granulosa cell dysfunction, has garnered attention. Therefore, we dissected the pathological mechanisms underlying granulosa cells dysfunction in PCOS by integrating insights from both cellular senescence and sphingolipid metabolism disorders, and explored new biomarkers for PCOS-related anovulation and potential therapeutic strategies based on bioinformatics analysis combining machine learning.

In the pursuit of diagnostic biomarkers, we initially identified 79 DEGs linking cellular senescence and sphingolipid metabolism in PCOS granulosa cells. A number of senescence-associated genes were found upregulated, among which CDKN1B, IL6, and TNF-α have been previously validated in PCOS [34, 35]. The functional enrichment analysis confirmed the activation of cellular senescence signaling pathways, as well as senescence-related MAPK and NF-κB signaling pathways. We further substantiated the presence of senescence-associated phenotypes in testosterone-stimulated KGN cells by heightened SA-β-Gal levels, upregulated expression of p21 and p16, and diminished Edu levels. Regarding sphingolipid metabolism, we noted that CerS4 (catalyzing ceramide synthesis) was significantly upregulated, while SGMS1 (facilitating ceramide degradation) was downregulated. This shift in enzyme expression patterns suggests a sphingolipid metabolism towards ceramide accumulation in PCOS granulosa cells. And we confirmed the increased ceramide levels in testosterone-stimulated KGN cells. These findings established the co-existence of cellular senescence and sphingolipid metabolism disorders in PCOS granulosa cells. Given that TNF-α and IL6 have been demonstrated to enhance ceramide production [36, 37], we propose that senescent granulosa cells may contribute to ceramide accumulation through autocrine or paracrine secretion of TNF-α and IL6. This process could lead to the decreased S1P/Cer ratio and consequently abnormal follicular development in PCOS. Future studies will be dedicated to validating this hypothesis. To further screen for diagnostic biomarkers in PCOS-related anovulation, we utilized four topological analysis algorithms and filtered out 28 hub genes with pivotal roles in the PPI network from 79 DEGs. Ultimately, the 5 most critical genes: LYN, PLCG2, STAT5B, MMP9, and IL6R, were selected based on machine learning.

To explore the biological functions mediated by these 5 diagnostic genes, we further conducted GSEA, and uncovered their significant involvement in immune cell-associated pathways including T cells, B cells, natural killer (NK) cells, as well as pathways related to chemokine and cytokine signaling. These findings suggest that these genes are potent regulators of the follicular immune microenvironment in PCOS. This hypothesis was further verified by the immune infiltration landscape based on these 5 genes in PCOS that they regulated the infiltration of up to 22 distinct immune cell types. Among these genes, MMP9 expression was found negatively correlated with monocyte and M1 macrophages, while positively correlated with M0 macrophages in PCOS patients [38]. Additionally, a cross-sectional study revealed that MMP9 levels significantly increased in obese PCOS patients and positively correlated with the duration of infertility [39]. Beyond MMP9, PLCG2 and IL6 also have been implicated in modulating immune cell infiltration in PCOS based on bioinformation analysis [40, 41]. Despite the lack of direct research exploring the influence of LYN on immune cells, Shen and colleagues have found that LYN was associated with insulin resistance in PCOS patients, and the blockade of LYN improved ovarian dysfunction in PCOS mice [42]. Taking these together, these 5 diagnostic genes associated with both cellular senescence and sphingolipid metabolism play important roles in regulating the follicular immune microenvironment in PCOS.

In addition, we identified potential therapeutic pharmaceuticals of PCOS-related anovulation based on these 5 genes. Among the candidate pharmaceuticals, we noted aspirin, a widely used nonsteroidal anti-inflammatory drug (NSAID). Aspirin exerts primary effects through irreversible inhibition of cyclooxygenase (COX), suppressing thromboxane A2 synthesis (antiplatelet action) and prostaglandin production (anti-inflammatory effect). The anti-inflammatory property intersects with PCOS therapeutic target: ovarian dysfunction mediated by low-grade inflammation. A clinical trial has demonstrated that low-dose aspirin (75‒100 mg/day) mitigated the risk of ovarian hyperstimulation syndrome (OHSS) in PCOS patients undergoing in vitro fertilization-embryo transfer (IVF-ET) [43]. Additionally, low-dose aspirin supplement enhanced the efficacy of ovulation-inducing agents such as letrozole and tamoxifen, improving the ovulation rate and pregnancy rates in anovulatory PCOS patients [44, 45]. Emerging evidence suggests secondary insulin-sensitizing potential of aspirin, which may help ameliorate the metabolic dysregulation of PCOS patients [46]. However, high-dose aspirin (> 300 mg/day) completely blocking COX-2 may paradoxically impair ovulation by disrupting follicular prostaglandin gradients [47]. And pharmacodynamic interference of aspirin with spironolactone via competitive albumin binding, potentially reducing the anti-androgenic efficacy of spironolactone [48]. The dose-dependent gastrointestinal toxicity of aspirin also requires vigilant attention [49]. Therefore, it appears that low-dose aspirin may be more suitable for PCOS management. In our in vitro validation, low-dose aspirin decreased senescence and ceramide levels in testosterone-stimulated KGN cells. Similarly, aspirin retarding senescence has also been observed in preeclampsia, age-related hearing loss and Alzheimer's disease [50,51,52]. Furthermore, Knapp et al. uncovered that aspirin regulated sphingolipid metabolism in dose-dependent manner, with low-dose aspirin (75 mg/day) increasing plasma S1P levels of volunteers, while high-dose (300 mg/day) decreasing S1P levels [53]. These studies combined with our results elucidated that aspirin might improve ovarian dysfunction through mitigating senescence and reversing sphingolipid metabolism disorder in granulosa cells, thereby supporting the clinical application of low-dose aspirin to treat PCOS-related anovulation.

There were some limitations to this study. First, this study was only carried out from the perspective of gene transcriptome, and multi-omics and mechanistic studies were not performed. Second, the granulosa cell model used for our in vitro validation was the KGN cell line, which is derived from a human ovarian granulosa cell tumor. This origin means that KGN cells cannot fully represent the normal physiological state of granulosa cells [54]. Moreover, KGN cells primarily reflect the post-ovulatory state of granulosa cells, which may significantly differ from the pre-ovulatory state in PCOS [54]. As with all in vitro models, KGN cells do not fully replicate the complex in vivo environment of the ovary [55]. Thus, further validation of our results through in vivo experiments and clinical practice is needed.

Conclusions

In summary, this study dissected the complex interactions between cellular senescence and sphingolipids dysregulation in granulosa cells of PCOS patients, leading to the identification of 5 diagnostic genes (LYN, PLCG2, STAT5B, MMP9, and IL6R) for PCOS-related anovulation. These pivotal genes orchestrate inflammatory pathways and modulate the infiltration of diverse immune cell populations within PCOS. Moreover, aspirin, a candidate drug pinpointed through the analysis of these genes, demonstrated the capacity to mitigate senescence and sphingolipids disorders in granulosa cells of PCOS, positioning it as a promising treatment for PCOS-related anovulation (Fig. 11). These findings contribute to our understanding of anovulation in PCOS, opening new avenues for targeted interventions in fertility treatment. Future research should focus on elucidating precise molecular mechanisms, and evaluating long-term efficacy and safety of aspirin in fertility treatment of PCOS patients.

Fig. 11
figure 11

The influence of external stimuli accelerates senescence in granulosa cells, causing the dysregulation of sphingolipid metabolism and a subsequent shift in the sphingolipid profile of follicular fluid. These metabolic derangements may lead to compromised follicular maturation or even follicular atresia, eventually contributing to the development of PCOS

Full size image

Data availability

No datasets were generated or analysed during the current study.

Abbreviations

C1P:

Ceramide-1-phosphate

Cer:

Ceramide

CSGs:

Cellular senescence-related genes

DCA:

Decision curve analysis

DEGs:

Differentially expressed genes

EPC:

Extracellular Processing Cluster

GO:

Gene Ontology

GSEA:

Gene set enrichment analysis

IVF-ET:

In vitro fertilization-embryo transfer

KEGG:

Kyoto Encyclopedia of Genes and Genomes

MCC:

Maximum Connectivity Cluster

MNC:

Maximum Neighborhood Component

NSAID:

Nonsteroidal anti-inflammatory drug

OHSS:

Ovarian hyperstimulation syndrome

PCOS:

Polycystic ovary syndrome

PPI:

Protein–Protein Interaction

RF:

Random forest

SASP:

Senescence-associated secretory phenotype

S1P:

Sphingosine-1-phosphate

SMGs:

Sphingolipid metabolism-related genes

SVM-RFE:

Support vector machine-recursive feature elimination

WGCNA:

Weighted Gene Co-expression Network Analysis

References

  1. Revised 2003 consensus on diagnostic criteria and long-term health risks related to polycystic ovary syndrome. Fertil Steril 2004, 81(1):19–25.

  2. Rudnicka E, Suchta K, Grymowicz M, Calik-Ksepka A, Smolarczyk K, Duszewska AM, Smolarczyk R, Meczekalski B. Chronic low grade inflammation in pathogenesis of PCOS. Int J Mol Sci. 2021;22(7):3789.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  3. Zhang CH, Liu XY, Wang J. Essential role of granulosa cell glucose and lipid metabolism on oocytes and the potential metabolic imbalance in polycystic ovary syndrome. Int J Mol Sci. 2023;24(22):16247.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  4. Gougeon A. Regulation of ovarian follicular development in primates: facts and hypotheses. Endocr Rev. 1996;17(2):121–55.

    Article  CAS  PubMed  Google Scholar 

  5. Vatier C, Christin-Maitre S. Epigenetic/circadian clocks and PCOS. Hum Reprod. 2024;39(6):1167–75.

    Article  CAS  PubMed  Google Scholar 

  6. Kuo A, Hla T. Regulation of cellular and systemic sphingolipid homeostasis. Nat Rev Mol Cell Biol. 2024;25(10):802–21.

    Article  CAS  PubMed  Google Scholar 

  7. Hannun YA, Obeid LM. Sphingolipids and their metabolism in physiology and disease. Nat Rev Mol Cell Biol. 2018;19(3):175–91.

    Article  CAS  PubMed  Google Scholar 

  8. Yan K, Zhang W, Song H, Xu X. Sphingolipid metabolism and regulated cell death in malignant melanoma. Apoptosis. 2024;29(11–12):1860–78.

    Article  CAS  PubMed  Google Scholar 

  9. Hernández-Coronado CG, Guzmán A, Espinosa-Cervantes R, Romano MC, Verde-Calvo JR, Rosales-Torres AM. Sphingosine-1-phosphate and ceramide are associated with health and atresia of bovine ovarian antral follicles. Animal. 2015;9(2):308–12.

    Article  PubMed  Google Scholar 

  10. Meng YY, Xu ZH, Wu FF, Chen WM, Xie SS, Liu J, Huang XF, Zhou Y. Sphingosine-1-phosphate suppresses cyclophosphamide induced follicle apoptosis in human fetal ovarian xenografts in nude mice. Fertil Steril. 2014;102(3):871–U327.

    Article  CAS  PubMed  Google Scholar 

  11. Guzel Y, Bildik G, Dilege E, Oktem O. Sphingosine-1-phosphate reduces atresia of primordial follicles occurring during slow-freezing and thawing of human ovarian cortical strips. Mol Reprod Dev. 2018;85(11):858–64.

    Article  CAS  PubMed  Google Scholar 

  12. Guo X, Zhu Y, Guo L, Qi Y, Liu X, Wang J, Zhang J, Cui L, Shi Y, Wang Q, et al. BCAA insufficiency leads to premature ovarian insufficiency via ceramide-induced elevation of ROS. EMBO Mol Med. 2023;15(4):e17450.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  13. Liu LY, Yin TL, Chen Y, Li YH, Yin L, Ding JL, Yang J, Feng HL. Follicular dynamics of glycerophospholipid and sphingolipid metabolisms in polycystic ovary syndrome patients. J Steroid Biochem Mol Biol. 2019;185:142–9.

    Article  CAS  PubMed  Google Scholar 

  14. Ding Y, Jiang Y, Zhu M, Zhu Q, He Y, Lu Y, Wang Y, Qi J, Feng Y, Huang R, et al. Follicular fluid lipidomic profiling reveals potential biomarkers of polycystic ovary syndrome: a pilot study. Front Endocrinol (Lausanne). 2022;13:960274.

    Article  PubMed  Google Scholar 

  15. Cheng D, Zheng B, Sheng Y, Zeng Z, Mo Z. The roles of autophagy in the genesis and development of polycystic ovary syndrome. Reprod Sci. 2023;30(10):2920–31.

    Article  PubMed  Google Scholar 

  16. Gorgoulis V, Adams PD, Alimonti A, Bennett DC, Bischof O, Bishop C, Campisi J, Collado M, Evangelou K, Ferbeyre G, et al. Cellular senescence: defining a path forward. Cell. 2019;179(4):813–27.

    Article  CAS  PubMed  Google Scholar 

  17. Tanaka T, Urata Y, Harada M, Kunitomi C, Kusamoto A, Koike H, Xu ZX, Sakaguchi N, Tsuchida C, Komura A, et al. Cellular senescence of granulosa cells in the pathogenesis of polycystic ovary syndrome. Mol Human Reprod. 2024;30(5):015.

    Article  Google Scholar 

  18. Prabhu NB, Adiga D, Kabekkodu SP, Bhat SK, Satyamoorthy K, Rai PS. Bisphenol A exposure modulates reproductive and endocrine system, mitochondrial function and cellular senescence in female adult rats: a hallmarks of polycystic ovarian syndrome phenotype. Environ Toxicol Pharmacol. 2022;96:104010.

    Article  CAS  PubMed  Google Scholar 

  19. Venable ME, Yin X. Ceramide induces endothelial cell senescence. Cell Biochem Funct. 2009;27(8):547–51.

    Article  CAS  PubMed  Google Scholar 

  20. Wu B, Xiao Q, Zhu L, Tang H, Peng W. Icariin targets p53 to protect against ceramide-induced neuronal senescence: implication in Alzheimer’s disease. Free Radic Biol Med. 2024;224:204–19.

    Article  CAS  PubMed  Google Scholar 

  21. Jadhav KS, Dungan CM, Williamson DL. Metformin limits ceramide-induced senescence in C2C12 myoblasts. Mech Ageing Dev. 2013;134(11–12):548–59.

    Article  CAS  PubMed  Google Scholar 

  22. Itami N, Shirasuna K, Kuwayama T, Iwata H. Palmitic acid induces ceramide accumulation, mitochondrial protein hyperacetylation, and mitochondrial dysfunction in porcine oocytes. Biol Reprod. 2018;98(5):644–53.

    Article  PubMed  Google Scholar 

  23. Kwun C, Patel A, Pletcher S, Lyons B, Abdelrahim M, Nicholson D, Morris E, Salata K, Francis GL. Ceramide increases steroid hormone production in MA-10 Leydig cells. Steroids. 1999;64(8):499–509.

    Article  CAS  PubMed  Google Scholar 

  24. Khosla S, Farr JN, Tchkonia T, Kirkland JL. The role of cellular senescence in ageing and endocrine disease. Nat Rev Endocrinol. 2020;16(5):263–75.

    Article  CAS  PubMed  Google Scholar 

  25. Laurila PP, Wohlwend M, de Lima TI, Luan PL, Herzig S, Zanou N, Crisol B, Bou-Sleiman M, Porcu E, Gallart-Ayala H, et al. Sphingolipids accumulate in aged muscle, and their reduction counteracts sarcopenia. Nature Aging. 2022;2(12):1159-+.

    Article  CAS  PubMed  Google Scholar 

  26. Millner A, Lizardo DY, Atilla-Gokcumen CE. Untargeted lipidomics highlight the depletion of deoxyceramides during therapy-induced senescence. Proteomics. 2020;20(10):e2000013.

    Article  PubMed  Google Scholar 

  27. Modrak DE, Leon E, Goldenberg DM, Gold DV. Ceramide regulates gemcitabine-induced senescence and apoptosis in human pancreatic cancer cell lines. Mol Cancer Res. 2009;7(6):890–6.

    Article  CAS  PubMed  Google Scholar 

  28. Ma Y, Zheng L, Wang Y, Gao Y, Xu Y. Arachidonic acid in follicular fluid of PCOS induces oxidative stress in a human ovarian granulosa tumor cell line (KGN) and upregulates GDF15 expression as a response. Front Endocrinol (Lausanne). 2022;13:865748.

    Article  PubMed  Google Scholar 

  29. Li F, Huangyang P, Burrows M, Guo K, Riscal R, Godfrey J, Lee KE, Lin N, Lee P, Blair IA, et al. FBP1 loss disrupts liver metabolism and promotes tumorigenesis through a hepatic stellate cell senescence secretome. Nat Cell Biol. 2020;22(6):728–39.

    Article  PubMed  PubMed Central  Google Scholar 

  30. Xiang Y, Wang H, Ding H, Xu T, Liu X, Huang Z, Wu H, Ge H. Hyperandrogenism drives ovarian inflammation and pyroptosis: a possible pathogenesis of PCOS follicular dysplasia. Int Immunopharmacol. 2023;125(Pt A):111141.

    Article  CAS  PubMed  Google Scholar 

  31. Rosenfield RL, Ehrmann DA. The pathogenesis of Polycystic Ovary Syndrome (PCOS): the hypothesis of PCOS as functional ovarian hyperandrogenism revisited. Endocr Rev. 2016;37(5):467–520.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  32. Rababa’h AM, Matani BR, Yehya A. An update of polycystic ovary syndrome: causes and therapeutics options. Heliyon. 2022;8(10):e11010.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  33. Marinkovic-Radosevic J, Cigrovski Berkovic M, Kruezi E, Bilic-Curcic I, Mrzljak A. Exploring new treatment options for polycystic ovary syndrome: Review of a novel antidiabetic agent SGLT2 inhibitor. World J Diabetes. 2021;12(7):932–8.

    Article  PubMed  PubMed Central  Google Scholar 

  34. Wang BJ, Hao MM, Yang QL, Li J, Guo YH. Follicular fluid soluble receptor for advanced glycation endproducts (sRAGE): a potential protective role in polycystic ovary syndrome. J Assist Reprod Genet. 2016;33(7):959–65.

    Article  PubMed  PubMed Central  Google Scholar 

  35. Devillers MM, François CM, Chester M, Corre R, Cluzet V, Giton F, Cohen-Tannoudji J, Guigon CJ. Androgen receptor signaling regulates follicular growth and steroidogenesis in interaction with gonadotropins in the ovary during mini-puberty in mice. Front Endocrinol (Lausanne). 2023;14:1130681.

    Article  PubMed  Google Scholar 

  36. Hernández-Corbacho MJ, Canals D, Adada MM, Liu ML, Senkal CE, Yi JK, Mao CG, Luberto C, Hannun YA, Obeid LM. Tumor Necrosis Factor-α (TNF α)-induced Ceramide Generation via Ceramide Synthases Regulates Loss of Focal Adhesion Kinase (FAK) and Programmed Cell Death. J Biol Chem. 2015;290(42):25356–73.

    Article  PubMed  PubMed Central  Google Scholar 

  37. Dorweiler TF, Singh A, Ganju A, Lydic TA, Glazer LC, Kolesnick RN, Busik JV. Diabetic retinopathy is a ceramidopathy reversible by anti-ceramide immunotherapy. Cell Metabolism. 2024;36(7):1521–1533.e5.

    Article  CAS  PubMed  Google Scholar 

  38. Zhang W, Wu Y, Yuan Y, Wang L, Yu B, Li X, Yao Z, Liang B. Identification of key biomarkers for predicting atherosclerosis progression in polycystic ovary syndrome via bioinformatics analysis and machine learning. Comput Biol Med. 2024;183:109239.

    Article  CAS  PubMed  Google Scholar 

  39. Sylus AM, Nandeesha H, Chitra T. Matrix metalloproteinase-9 increases and Interleukin-10 reduces with increase in body mass index in polycystic ovary syndrome: a cross-sectional study. Int J Reprod Biomed. 2020;18(8):605–10.

    CAS  PubMed  PubMed Central  Google Scholar 

  40. Gao MG, Liu XH, Du MX, Gu H, Xu H, Zhong XM. Identification of immune cell infiltration and effective biomarkers of polycystic ovary syndrome by bioinformatics analysis. Bmc Preg Child. 2023;23(1):377.

    Article  CAS  Google Scholar 

  41. Qu J, Li B, Qiu M, Wang J, Chen Z, Li K, Teng X. Discovery of immune-related diagnostic biomarkers and construction of diagnostic model in varies polycystic ovary syndrome. Arch Gynecol Obstet. 2022;306(5):1607–15.

    Article  CAS  PubMed  Google Scholar 

  42. Shen H, Xu X, Fu Z, Xu C, Wang Y. The interactions of CAP and LYN with the insulin signaling transducer CBL play an important role in polycystic ovary syndrome. Metabolism. 2022;131:155164.

    Article  CAS  PubMed  Google Scholar 

  43. Jahromi BN, Zolghadri J, Rahmani E, Alipour S, Anvar Z, Zarei A, Keramati P. Effect of low-dose aspirin on the development of ovarian hyperstimulation syndrome and outcomes of assisted reproductive techniques in the women with PCOS, a randomized double-blinded clinical trial. Taiwan J Obstet Gynecol. 2019;58(2):255–60.

    Article  Google Scholar 

  44. Yu QJ, Wang ZH, Su FF, Wang M. Effectiveness and safety of aspirin combined with letrozole in the treatment of polycystic ovary syndrome: a systematic review and meta-analysis. Ann Palliative Med. 2021;10(4):4632–41.

    Article  Google Scholar 

  45. Aref NK, Ahmed WAS, Ahmed MR, Sedik WF. A new look at low-dose aspirin: Co-administration with tamoxifen in ovulation induction in anovulatory PCOS women. J Gynecol Obstet Hum Reprod. 2019;48(8):673–5.

    Article  PubMed  Google Scholar 

  46. Carvalho-Filho MA, Ropelle ER, Pauli RJ, Cintra DE, Tsukumo DM, Silveira LR, Curi R, Carvalheira JB, Velloso LA, Saad MJ. Aspirin attenuates insulin resistance in muscle of diet-induced obese rats by inhibiting inducible nitric oxide synthase production and S-nitrosylation of IRbeta/IRS-1 and Akt. Diabetologia. 2009;52(11):2425–34.

    Article  CAS  PubMed  Google Scholar 

  47. Greenway FL, Swerdloff RS. The effect of aspirin (prostaglandin synthetase inhibitor) on ovulation. Fertil Steril. 1978;30(3):364–5.

    Article  CAS  PubMed  Google Scholar 

  48. Tweeddale MG, Ogilvie RI. Antagonism of spironolactone-induced natriuresis by aspirin in man. N Engl J Med. 1973;289(4):198–200.

    Article  CAS  PubMed  Google Scholar 

  49. Lavie CJ, Howden CW, Scheiman J, Tursi J. Upper gastrointestinal toxicity associated with long-term aspirin therapy: consequences and prevention. Curr Probl Cardiol. 2017;42(5):146–64.

    Article  PubMed  Google Scholar 

  50. Clark DPQ, Zhou Z, Hussain SM, Tran C, Britt C, Storey E, Lowthian JA, Shah RC, Dillon H, Wolfe R, et al. Low-dose aspirin and progression of age-related hearing loss: a secondary analysis of the ASPREE randomized clinical trial. JAMA Netw Open. 2024;7(7):e2424373.

    Article  PubMed  PubMed Central  Google Scholar 

  51. Casoli T, Balietti M, Giorgetti B, Solazzi M, Scarpino O, Fattoretti P. Platelets in Alzheimer’s disease-associated cellular senescence and inflammation. Curr Pharm Des. 2013;19(9):1727–38.

    CAS  PubMed  Google Scholar 

  52. Sugulle M, Fiska BS, Jacobsen DP, Fjeldstad HE, Staff AC. Placental senescence and the two-stage model of preeclampsia. Am J Reprod Immunol. 2024;92(1):e13904.

    Article  CAS  PubMed  Google Scholar 

  53. Knapp M, Lisowska A, Knapp P, Baranowski M. Dose-dependent effect of aspirin on the level of sphingolipids in human blood. Adv Med Sci. 2013;58(2):274–81.

    Article  CAS  PubMed  Google Scholar 

  54. Sudhakaran G, Babu SR, Mahendra H, Arockiaraj J. Updated experimental cellular models to study polycystic ovarian syndrome. Life Sci. 2023;322:121672.

    Article  CAS  PubMed  Google Scholar 

  55. Indran IR, Lee BH, Yong EL. Cellular and animal studies: insights into pathophysiology and therapy of PCOS. Best Pract Res Clin Obstet Gynaecol. 2016;37:12–24.

    Article  PubMed  Google Scholar 

Download references

Acknowledgements

Not applicable.

Clinical trial number

Not applicable.

Funding

This work was supported by grants from the Natural Science Foundation of Jiangsu Province (BK20210011), Zhongda Hospital Affiliated to Southeast University, Jiangsu Province High-Level Hospital (2023GSPKY11, GSP-LCYJFH01, CZXM-GSP-KY), Zhongda Hospital Affiliated to Southeast University, Jiangsu Province High-Level Hospital Pairing Assistance Construction Funds (zdyyxy07), National Clinical Key Discipline Construction Funds (czxm-zk-40), and National Natural Science Foundation of China (82372126, 8207071577, 82301900).

Author information

Author notes

  1. Weiwei Shi and Hao Lin contributed equally to this work.

Authors and Affiliations

  1. Department of Obstetrics and Gynecology, Zhongda Hospital Affiliated to Southeast University, Nanjing, 210009, Jiangsu, China

    Weiwei Shi, Wu Di & Yang Shen

  2. Department of Medicine, Southeast University, Nanjing, 210009, Jiangsu, China

    Hao Lin & Yang Shen

  3. Department of Clinical Science and Research, Zhongda Hospital Affiliated to Southeast University, Nanjing, 21009, Jiangsu, China

    Hao Lin

  4. Key Laboratory of Innovative Applications of Bioresources and Functional Molecules of Jiangsu Province, College of Life Science and Chemistry, Jiangsu Second Normal University, Nanjing, 210013, Jiangsu, China

    Cong He

Authors
  1. Weiwei Shi
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Hao Lin
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Wu Di
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. Cong He
    View author publications

    You can also search for this author inPubMed Google Scholar

  5. Yang Shen
    View author publications

    You can also search for this author inPubMed Google Scholar

Contributions

W.S. contributed to the conception and design of the study, performed the experimental study and the statistical analysis; H.L. wrote the first draft of the manuscript; D.W. and C.H. wrote sections of the manuscript. Y.S. supervised the study and are the corresponding authors for this manuscript.

Corresponding author

Correspondence to Yang Shen.

Ethics declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Shi, W., Lin, H., Di, W. et al. Granulosa cell RNA-Seq insights into senescence and sphingolipid metabolism disorder in PCOS: aspirin as a potential therapeutic drug. Reprod Biol Endocrinol 23, 61 (2025). https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12958-025-01396-x

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12958-025-01396-x

Keywords

Reproductive Biology and Endocrinology

ISSN: 1477-7827

Contact us