{"paper_id":"5073a627-7e6c-4c32-a3f3-6bad7bace48d","body_text":"Abstract\nBackground\nEndometriosis (EM) is a common gynecological disease. Though m6A RNA methylation has emerged as a key regulator in various physiological and pathological processes, its cell-type-specific function in EM remains unclear.\nMethods\nWe obtain batch and single-cell RNA sequencing (scRNA-seq) data from the Gene Expression Omnibus (GEO) database. The “Seurat” package was used to annotate and label cell types in scRNA-seq data. After identifying m6A-associated modules with weighted gene co-expression network analysis (WGCNA), core biomarkers for EM were filtered and selected by conducting least absolute shrinkage and selection operator (LASSO), support vector machine–recursive feature elimination (SVM–RFE), and eXtreme Gradient Boosting (XGBoost). Subsequently, functional enrichment analysis was performed with the \"clusterProfiler\" package. Immune infiltration and pathway activity were assessed by single-sample gene set enrichment analysis (ssGSEA). Drug prediction was performed using DSigDB in Enrichr tool, and molecular docking was conducted with AutoDock Vina. Finally, quantitative real-time PCR (qRT-PCR) and Western blot were performed to detect gene expression in vitro, and cell migration and invasion were assessed by carrying out wound healing and Transwell assays.\nResults\nThe scRNA-seq analysis revealed altered cellular composition in EM, characterized by increased proportions of epithelial and NK/T cells. m6A regulatory activity was significantly elevated in EM samples, particularly in immune cells, and was associated with neutrophil activation and RNA splicing pathways. Integrated analysis of WGCNA and machine learning identified EPCAM, DST, HSPH1, and NAP1L1 as core m6A-related genes. These four genes were correlated with specific immune cell subsets, including immature dendritic cells and effector memory CD4 + T cells, and were enriched in cell cycle-related oncogenic pathways. Drug prediction revealed that dronabinol and minocycline might be potential therapeutics for EM, with molecular docking confirming favorable binding affinities. Functionally, EPCAM was downregulated in endometriotic 12Z cells, and its overexpression significantly suppressed cell migration.\nConclusions\nThis study revealed that m6A-related dysregulation contributed to EM pathogenesis through dysregulated immune responses and affected cell migration. The identified gene signature and candidate drugs provide novel insights for potential therapeutic strategies.\nSimilar content being viewed by others\nIntroduction\nEndometriosis (EM) is a prevalent gynecological disorder characterized by ectopic growth of endometrial-like tissues outside the uterine cavity [1]. This chronic inflammatory condition is a leading cause of dysmenorrhea, pelvic pain, and infertility [2]. At present, despite its high prevalence, the pathogenesis of EM remains incompletely understood. Moreover, current therapeutic options are often limited by side effects and high recurrence rates [3]. Thus, a more comprehensive understanding of the molecular and cellular mechanisms underlying disease initiation and progression is urgently needed.\nAccumulating evidence highlights the importance of epigenetic regulation, particularly N6-methyladenosine (m6A) RNA modification, in controlling key processes implicated in EM, such as immune responses, cell proliferation, and inflammation [4, 5]. For example, METTL3-mediated m6A methylation, recognized by reader proteins, such as YTHDF2, has been shown to regulate mRNA stability of key signaling molecules involved in cell proliferation, migration, and senescence. These processes are central to the pathogenesis of gynecological disorders [6]. m6A modifications are dynamically deposited, removed, and interpreted by “writers”, “erasers”, and “readers,” respectively, and influence mRNA stability, splicing, translation, and decay [7, 8]. Although dysregulation of the m6A machinery has been linked to various inflammatory and neoplastic diseases, its cell-type-specific role in EM remains largely unexplored [9].\nRecent advances in scRNA-seq have improved our ability to dissect cellular heterogeneity in complex tissues [10,11,12]. These technologies enable the identification of rare cell populations, characterization of cell-state transitions, and mapping of cell–cell communication networks at unprecedented resolution [13, 14]. In EM research, such approaches hold promise for analyzing specific cell types and pathways that contribute to lesion formation, immune dysregulation, and tissue invasion [15, 16]. While bulk transcriptomic studies can help identify differentially expressed genes (DEGs) in endometriotic tissues, they often obscure cell-type-specific signals due to tissue-level averaging [17]. Integrating scRNA-seq data with functional genomics approaches, such as co-expression network analysis and pathway enrichment, can help delineate key regulatory modules and uncover potential therapeutic targets underlying EM pathogenesis [18].\nHere, this study aimed to systematically analyze the cell-type-specific regulatory role of m6A RNA modification in the development and progression of EM, along with its underlying molecular mechanisms. By integrating single-cell transcriptomics with multi-omics data, we established a multidimensional research framework encompassing cellular heterogeneity, key gene screening, functional validation, and therapeutic target prediction. The present findings may improve our understanding of the epigenetic transcriptional regulation mechanisms in EM and provide crucial insights for the discovery of novel biomarkers and potential therapeutic strategies, underscoring its potential academic and clinical significance.\nMaterials and methods\nData collection\nGene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/) was accessed to obtain bulk and scRNA-seq data. The bulk RNA-seq data sets GSE51981 (77 EM and 71 control samples), GSE6364 (21 EM and 16 control samples), and GSE7305 (10 EM and 10 control samples) were generated using the GPL570 platform (Affymetrix Human Genome U133 Plus 2.0 Array). The scRNA-seq data set GSE214411 contained 6 patients with EM and 7 healthy controls. The 23 m6A regulators analyzed in this study were obtained from a previously published study [19].\nData preprocessing\nThe scRNA-seq data set GSE214411 was preprocessed to retain cells with mitochondrial gene percentage < 15%, total gene counts between 200 and 100,000, and each gene expressing at least three cells. Data normalization was performed using the SCTransform function, followed by principal component analysis (PCA). The top 20 PCs were used to correct batch effects with the “harmony” R package, and then subjected to UMAP dimensionality reduction (Figure S1). Next, cell clustering was performed using the FindNeighbors and FindClusters functions in “Seurat” R package. Finally, cell types were annotated using marker genes from the CellMarker2.0 database.\nSingle-sample gene set enrichment analysis (ssGSEA)\nThe “GSVA” R package was used to perform ssGSEA. The m6A score was calculated for each sample based on a predefined m6A-related gene set to reflect the relative activation level of the m6A pathway. The DEGs between different m6A score groups were identified using the FindMarkers function. Functional enrichment analysis of Gene Ontology (GO) biological processes was subsequently conducted with the “clusterProfiler” R package. An adjusted p < 0.05 was considered statistically significant.\nWeighted gene co-expression network analysis (WGCNA)\nThree transcriptomic datasets (GSE51981, GSE6364, and GSE7305) were merged, and batch effects were corrected using the ComBat function in the “sva” R package. This yielded a final expression matrix of 108 EM and 97 control samples. WGCNA was performed to identify co-expression modules and potential biomarkers associated with m6A modification phenotype. After sample clustering and outlier removal, a soft-threshold power β = 16 was selected using the pickSoftThreshold function, achieving a scale-free network topology (scale-free R2 > 0.85). The gene expression matrix was transformed into an adjacency matrix and subsequently converted into a TOM. Hierarchical clustering with average linkage was performed based on TOM dissimilarity. Gene modules were detected using the dynamic tree-cutting algorithm (parameters: deepSplit = 2, minModuleSize = 50, and merge cut height = 0.25). Module eigengenes were calculated for each module, retaining those with |GS|> 0.5 and |MM|> 0.5 for module merging and downstream association analysis.\nMachine learning for biomarker screening\nTo identify robust m6A-related biomarkers, the m6A-associated gene set was intersected with DEGs to obtain common genes. Three machine learning algorithms were applied sequentially for feature selection. First, LASSO regression was performed using the “glmnet” R package. Second, RFE was implemented using the “caret” package to identify the feature subset that minimized model error. Third, the XGBoost algorithm was employed via the “xgboost” package to rank feature importance, retaining the top 10 most important genes. Finally, genes identified by all three methods were intersected to define the final set of core m⁶A-related biomarkers.\nImmune infiltration analysis\nTo investigate the relationship between the identified biomarker genes and immune cell infiltration, we performed ssGSEA to calculate enrichment scores for 28 immune cell types [20], followed by correlation analysis between the expression of feature genes and each immune cell score. Pathway activity was assessed using gene sets from the h.all.v2025.1.Hs.symbols.gmt collection (MSigDB, GSEA website), which includes canonical biological pathways and processes. Pathway activity scores were computed using ssGSEA, and correlations between the feature gene expression and pathway scores were calculated by performing Pearson correlation analysis.\nDrug prediction and molecular docking\nDrug candidates were predicted using the DSigDB database via the “Enrichr” R package, which links gene expression signatures to drug responses. To evaluate potential interactions between the selected compounds and target proteins, molecular docking was conducted. The 3D structures of candidate drug molecules were retrieved from PubChem and subjected to energy minimization in Chem3D using the MM2 force field to obtain stable conformations. Crystal structures of target proteins were downloaded from the Protein Data Bank (PDB) and preprocessed in PyMOL to remove water molecules and co-crystallized ligands. Hydrogen atoms were added and polar residues were assigned using AutoDockTools. Ligands were prepared by adding hydrogen atoms and defining rotatable bonds. Semi-flexible docking simulations were performed using AutoDock Vina, with a grid box centered on the putative active site of each protein. Docking outputs were analyzed for binding modes and interaction patterns, and representative conformations were visualized in PyMOL.\nCell culture and transfection\nDMEM/F12 medium added with 10% FBS and 1% penicillin–streptomycin was employed to cultivate immortalized human EM epithelial cell line 12Z (catalog no. Eou84715, Beijing Yo-Sci Biotechnology Co., Ltd, Beijing, China). Primary human endometrial epithelial cells (hEECs, catalog no. Delf-10667, Hefei Whole Biology Co., Ltd, Hefei, China) were maintained in specific epithelial cell growth medium, following the manuals. All cells were incubated in a humidified environment with 5% CO₂ at 37 °C. To investigate the functional role of EPCAM, an EPCAM overexpression plasmid (oe-EPCAM) and a negative control plasmid (oe-NC) were constructed using pLVX-3 × FLAG-Puro vector (Shanghai Newpu Biotech, China). All constructs were sequence-verified prior to transfection with Lipofectamine 3000 transfection kit (Invitrogen). At 48 h post-transfection, 12Z cells were harvested for qPCR and functional assays to confirm overexpression efficiency.\nQuantitative real-time PCR (qRT-PCR)\nUsing RNA Extraction Reagent (R0017S, Beyotime, Shanghai, China), we extracted total RNA and then detected RNA concentration with NanoDrop. Samples were stored at − 80 ℃ until use. The cDNA was synthesized using the BeyoRT™ II First Strand cDNA Synthesis Kit (D7168S, Beyotime, China). qPCR was performed on a CFX96 Real-Time PCR Detection System (Bio-Rad) with BeyoFast™ SYBR Green qPCR Mix (2X) (D7260, Beyotime). Applying the 2⁻ΔΔCt method, relative mRNA expressions of target genes were calculated, with GAPDH as an internal reference gene. Each experiment was performed in triplicate. Specific primer sequence is listed in Table 1.\nWound healing and Transwell assay\nFor migration assessment, a wound healing assay was performed. Briefly, transfected 12Z cells were seeded into 6-well plates and grown to 90% confluence. A uniform scratch was created using a 200 μL pipette tip, and detached cells were washed away with PBS. Images were captured at 0 h and 48 h post-scratching under an inverted microscope. The wound closure area was quantified using ImageJ software. For the Transwell assay, the transfected 12Z cells suspended in serum-free medium (300 μL) were seeded into the upper chamber of the Transwell (8 μm pore size, Corning), while the lower chamber contained 700 μL of medium with 10% FBS. After 48-h incubation, migrated cells were fixed with 4% paraformaldehyde, colored with 0.5% crystal violet for 10 min, and imaged under a microscope. The migrated cells were quantified from five random fields per chamber.\nWestern blot\nThe total protein samples from the transfected 12Z cells using the RIPA lysis assay kit (P0013B, Beyotime) and the concentration was tested. Thereafter, the protein samples were separated in SDS-PAGE separation gel and transferred to the PVDF membranes, which were blocked in 5% skimmed milk to prevent non-specific binding. The primary antibodies against EPCAM (ab225382, Abcam, Cambridge, UK) and housekeeping control GAPDH (ab181602, Abcam) were applied to incubated with the membrane at 4 ℃ overnight, and then the HRP-conjugated secondary antibody was further reacted with the membrane for 1-h incubation at ambient temperature. The membrane containing the antibodies were finally visualized using the ECL visualization kit (P0018S, Beyotime) and the ChemiDoc Imaging System (Bio-Rad, Hercules, CA, USA). The density of the protein band was calculated using ImageJ 1.40 (NIH, Bethesda, MD, USA).\nStatistical analysis\nTwo-group comparison was conducted using Student’s t test, while one/two-way ANOVA followed by Tukey’s post hoc test was used for comparisons across multiple groups. Statistical analysis was conducted in R software (4.2.1) and GraphPad Prism 9.0. A p < 0.05 was regarded as statistically significant.\nResults\nSingle-cell landscape analysis\nAfter quality control, dimensionality reduction and clustering of scRNA-seq data from patient and control samples, a total of 12 clusters (0–11) containing 125,899 cells were obtained (Fig. 1A). Based on canonical marker gene expression, cells were annotated into seven cell types: smooth muscle cells, epithelial cells, NK/T cells, endothelial cells, macrophages, ciliated cells, and mast cells (Fig. 1B, C). Cell composition analysis revealed that the proportions of epithelial and NK/T cells in EM samples were higher than in controls (Fig. 1D).\nM 6 A dysregulation in EM involved specific immune cells and splicing pathways\nTo assess the role of m6A modification in EM, we applied the AUCell algorithm to score the enrichment of a curated m6A-regulated gene set in each single cell based on ranked gene expression. It was observed that m6A scores were significantly elevated in EM cells compared to controls (Fig. 2A, p < 0.0001). Cell-type-specific analysis revealed that NK/T cells, macrophages, ciliated cells, and mast cells exhibited the highest m6A scores (Fig. 2B, p < 0.0001). Cells were subsequently classified into high- and low-m6A groups based on the median score. Differential gene expression analysis identified a set of significantly upregulated and downregulated DEGs in the high-m6A score group, including HNRNPC, WTAP, ZC3H13, MT2A, MT1G (Fig. 2C). Furthermore, enrichment analysis further revealed that these DEGs were primarily enriched in pathways, such as neutrophil activation and RNA splicing (Fig. 2D).\nWGCNA identified the MEblack module as closely associated with the m 6 A-related model\nTo identify gene modules associated with m6A modification in EM, we performed WGCNA on the batch-corrected transcriptomic data set. A soft-thresholding power of β = 16 was selected to achieve a scale-free topology, yielding 10 co-expression modules (Fig. 3A, B). To link these modules to m6A-related gene signature, we calculated m6A-related gene set enrichment scores for each sample using ssGSEA and used these scores as a trait in module–trait association analysis. The resulting heatmap revealed that the MEblack module exhibited the strongest association with this m6A signature (Fig. 3C, r = 0.9, p < 0.0001). Within the MEblack module, 2440 genes with |GS|> 0.5 and |MM|> 0.5 were identified as closely related to m6A regulation (Fig. 3D).\nMachine learning revealed a key gene signature in EM\nFrom the intersection of WGCNA-derived m6A-associated genes and DEGs, 38 overlapping genes were obtained. LASSO regression selected a set of genes at lambda.1se = 0.0244 (Fig. 4A). RFE analysis indicated that 31 features minimized model error (Fig. 4B), and XGBoost feature importance ranking identified the top 10 genes (Fig. 4C). The intersection of genes selected by all three methods yielded four key genes, namely, EPCAM, DST, HSPH1, and NAP1L1 (Fig. 4D). Further analysis of the expression of these hub genes in EM and normal samples showed that EPCAM, DST, and HSPH1 were significantly downregulated in disease samples (Fig. 4E, p < 0.01).\nBiomarker genes correlated with immune and pathway alterations in EM\nTo characterize the functional and immunological context of the key genes (EPCAM, DST, HSPH1, NAP1L1), we performed correlation analyses between their expression, immune cell infiltration, and pathway activities. The expression of these genes was negatively correlated with MDSCs and CD56dim natural killer cells, and positively associated with immature dendritic cells and effector memory CD4 T cells (Fig. 5A). Pathway analysis revealed that the genes were positively linked to oncogenic and cell cycle-related pathways, including ANDROGEN_RESPONSE, MITOTIC_SPINDLE, and G2M_CHECKPOINT. In contrast, they were negatively correlated with tumor-suppressive and differentiation-related pathways, such as P53_PATHWAY, MYOGENESIS, and KRAS_SIGNALING_DN (Fig. 5B).\nDronabinol and minocycline identified as candidate therapeutics targeting EM\nDrug prediction identified several candidate compounds associated with the expression of EPCAM, DST, HSPH1, and NAP1L1, with dronabinol and minocycline emerging as top candidates. Molecular docking revealed stable binding conformations between the selected drugs and their target proteins. Dronabinol showed favorable binding affinities to both EPCAM and HSPH1 proteins, forming hydrophobic interactions and hydrogen bonds with key residues. Minocycline also exhibited strong binding to HSPH1 protein, with additional polar contacts within the binding pocket. The predicted binding modes and interaction profiles are visualized in Fig. 6, and detailed docking and residue-level interaction information is provided in Table 2.\nEPCAM suppressed cell migration in EM\nTo investigate the potential roles of dysregulated genes in EM pathogenesis, we compared the expression of candidate genes hEECs and 12Z cells in vitro. We found that EPCAM, DST, and HSPH1 were significantly downregulated in 12Z cells than in hEECs (Fig. 7A). Given that EPCAM exhibited the lowest expression levels among the four genes in 12Z cells, EPCAM was selected for subsequent functional validation. Following successful establishment of the EPCAM overexpression model, both qPCR and Western blot analyses confirmed a significant upregulation in EPCAM expression in 12Z cells (Fig. 7B, C). Functional assays demonstrated that overexpression of EPCAM markedly suppressed the migration of 12Z cells (Fig. 7D, E), suggesting its potential role in inhibiting aberrant cellular motility in EM. These results showed that downregulation of EPCAM may enhance the migratory phenotype of ectopic endometrial cells, highlighting its potential role as a suppressor of progression in EM.\nDiscussion\nEM has long been viewed through the lens of inflammation and hormonal dysregulation, yet the molecular logic underlying cellular identity shifts and ectopic survival remains elusive [21, 22]. In this study, our multi-omics approach revealed a dysregulated cellular landscape in endometriotic lesions, highlighted the critical role of m6A RNA modification across immune and stromal compartments, and identified EPCAM, DST, HSPH1, and NAP1L1 as a core gene set functionally linked to disease progression. Notably, our functional assay in 12Z cells revealed that EPCAM may serve as a suppressor of migration in endometrial epithelial cells, providing new insight into the molecular drivers of ectopic implantation.\nA key finding of this study was altered cellular composition in EM, particularly increased proportions of epithelial and NK/T cells. The expansion of epithelial cells aligns with the pathological hallmark of ectopic glandular structures [23], while the increase of NK/T cells suggests a shift toward a pro-inflammatory and immune-dysregulated microenvironment [24]. This observation is consistent with growing evidence that EM is not merely an inflammatory condition but also an immune-evasion disorder, in which aberrant immune cell infiltration fails to clear ectopic tissue [25]. Our finding that NK/T cells exhibited an elevated m6A regulatory state further implicated RNA methylation in modulating immune responses in EM, potentially affecting cytokine production, cytotoxicity, or T-cell differentiation [26]. These processes are known to be regulated by m6A in other contexts. Notably, genes comprising such a high m6A signature were enriched not only in immune activation pathways such as neutrophil activated response but also in post-transcriptional regulatory processes, particularly RNA splicing. This dual enrichment underscores the pleiotropic role of m6A in simultaneously shaping both the functional state of immune cells and transcriptomic diversity of stromal and epithelial populations, potentially through mechanisms, such as alternative splicing, mRNA stability, and nuclear export [27].\nBy combining three machine learning algorithms, four core genes (EPCAM, DST, HSPH1, and NAP1L1) involved in the pathogenesis of EM were identified. Among these, EPCAM, DST, and HSPH1 were significantly downregulated in endometriotic cells, suggesting a loss of epithelial integrity and differentiation [28]. EPCAM, which encodes epithelial cell adhesion molecule, is a key mediator of homophilic epithelial adhesion [29]; DST (dystonin) anchors the cytoskeleton to cell junctions [30]; and HSPH1 (encoding a heat shock protein) maintains proteostasis under stress [31]. Combined suppression of these genes may destabilize cell–cell contacts, increase cytoskeletal plasticity, and promote a detached, migratory phenotype, which are the hallmarks of retrograde dissemination and ectopic implantation [32]. Importantly, our functional validation in 12Z cells demonstrated that EPCAM overexpression can suppress cell migration, consistent with studies in carcinomas, where loss of EPCAM is associated with metastasis [32]. Meanwhile, our finding revealed that EPCAM was downregulated in endometriotic epithelial cells and suppressed migration, a phenotype potentially associated with dysfunction of E-cadherin and N-cadherin [33]. However, this may reflect dynamic and context-dependent roles of EPCAM during disease progression [34]. Its initial downregulation may facilitate epithelial dissociation and invasion, akin to early collective cancer cell dissemination [35]. In EM, reduced EPCAM may facilitate endometrial cell detachment and dissemination, enabling ectopic implantation. Moreover, its expression is positively correlated with immature dendritic cells and effector memory CD4 T cells, and negatively with MDSCs and CD56ᵈⁱᵐ NK cells, suggesting that EPCAM expression may be associated with an immune contexture that could support anti-inflammatory or regulatory responses [36]. In summary, EPCAM, DST, HSPH1, and NAP1L1 may constitute a functionally synergistic network regulated by m6A. In this study, elevated m6A signature can cause synchronized downregulation of these genes by modulating their mRNA stability, splicing, or translational efficiency. Specifically, reduced expression of EPCAM and DST directly disrupted intercellular junctions and stable anchoring of the cytoskeleton. Furthermore, the downregulation of HSPH1 may impair protein homeostasis and stress tolerance in cells within an ectopic environment, while altered NAP1L1 expression may further affect chromatin remodeling and cell cycle progression. Thus, collective dysfunction of this network collectively undermines epithelial barrier integrity, enhances cellular migration and invasive capacity, and may reshape local immune microenvironment through altered expression or secretory profile of cell surface molecules. Therefore, dysregulation of m6A modification emerges as a key upstream mechanism driving the dysfunction of this synergistic network, contributing to the pathological features of endometriosis.\nFinally, drug enrichment analysis and in silico molecular docking identified dronabinol and minocycline as potential therapeutic candidates. The modeling predicted favorable binding affinities between dronabinol and EPCAM, and between minocycline and HSPH1, suggesting that these compounds could potentially target the identified core gene network. These findings, though preliminary, open new avenues for repurposing anti-inflammatory agents in EM treatment and warrant further experimental investigation.\nSeveral limitations should be acknowledged. First, the scRNA-seq data in this study were derived primarily from ectopic endometrial tissue, with insufficient representation of key lesion subtypes, such as deep infiltrating EM. This limitation restricted the generalizability of our conclusions. To address this, future work will expand the sample size and incorporate diverse lesion subtypes. By integrating spatial transcriptomics technology, we plan to systematically elucidate the association between lesion location and molecular characteristics. Second, the functional validation of EPCAM was conducted solely in vitro using cell lines, and its role has not yet been confirmed in primary cells or in vivo models. Furthermore, there was a lack of in-depth analysis of its downstream signaling pathways. Future studies could utilize primary ectopic endometrial epithelial cells, organoid models, and mouse endometriosis models. Through in vivo experiments and integration of proteomics and phosphoproteomics, specific pathways through which EPCAM regulated cell migration can be elucidated. Finally, the drug predictions and molecular docking results presented here were based entirely on computational simulations and lacked validation through in vitro or in vivo pharmacodynamic experiments. Subsequent studies will test candidate compounds (dronabinol and minocycline) in cell models and mouse disease models to evaluate their effects on lesion growth, core gene expression, and cell migration, followed by performing target validation.\nConclusion\nTo conclude, this study revealed the m6A RNA modification as a central node linking epithelial dysfunction and immune dysregulation in EM. The identification of EPCAM as a functional suppressor of invasion provided a possible mechanistic explanation for ectopic implantation and highlighted a promising target for therapeutic intervention in EM.\nData availability\nThe datasets generated and/or analyzed during the current study are available in the [GSE51981] repository, [https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE51981], [GSE6364] repository, [https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE6364] and [GSE7305] repository, [https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE7305].\nAbbreviations\n- scRNA-seq:\n-\nSingle-cell RNA sequencing\n- EM:\n-\nEndometriosis\n- m6A:\n-\nN6-methyladenosine\n- AUCell:\n-\nArea under the cell enrichment curve\n- DEGs:\n-\nDifferentially expressed genes\n- WGCNA:\n-\nWeighted gene co-expression network analysis\n- ssGSEA:\n-\nSingle-sample gene set enrichment analysis\n- GS:\n-\nGene significance\n- MM:\n-\nModule membership\n- LASSO:\n-\nLeast absolute shrinkage and selection operator\n- EPCAM:\n-\nEpithelial cell adhesion molecule\n- hEECs:\n-\nHuman endometrial epithelial cells\nReferences\nLeone Roberti Maggiore U, et al. Epidemiology of infertility in women with endometriosis. Best Pract Res Clin Obstet Gynaecol. 2024;92:102454.\nCrump J, Suker A, White L. Endometriosis: a review of recent evidence and guidelines. Aust J Gen Pract. 2024;53(1–2):11–8.\nAdilbayeva A, Kunz J. Pathogenesis of Endometriosis and Endometriosis-Associated Cancers. Int J Mol Sci. 2024;25(14):7624.\nJiang L, et al. Exploring diagnostic m6A regulators in endometriosis. Aging Albany NY. 2020;12(24):25916–38.\nHong H, et al. VIRMA promotes neuron apoptosis via inducing m6A methylation of STK10 in spinal cord injury animal models. CNS Neurosci Ther. 2024;30(3):e14453.\nWang X, et al. METTL3-mediated m6A modification of SIRT1 mRNA inhibits progression of endometriosis by cellular senescence enhancing. J Transl Med. 2023;21(1):407.\nChen J, et al. Role of m6A modification in female infertility and reproductive system diseases. Int J Biol Sci. 2022;18(9):3592–604.\nZhang B, et al. Interplay of RNA m(6)A modification-related geneset in pan-cancer. Biomedicines. 2024;12(10):2211.\nHuang E, Chen L. RNA N(6)-methyladenosine modification in female reproductive biology and pathophysiology. Cell Commun Signal. 2023;21(1):53.\nYan J, et al. Single-cell analysis reveals insights into epithelial abnormalities in ovarian endometriosis. Cell Rep. 2024;43(3):113716.\nZhang L, et al. Two-Stage Multi-View Graph Spectural Clustering for Single-Cell RNASeq Data. Curr Genomics. 2025;26:1.\nYe L, et al. Unraveling the complex pathophysiology of white matter hemorrhage in intracerebral stroke: a single-cell RNA sequencing approach. CNS Neurosci Ther. 2024;30(3):e14652.\nShih AJ, et al. Single-cell analysis of menstrual endometrial tissues defines phenotypes associated with endometriosis. BMC Med. 2022;20(1):315.\nChen M, et al. Integrated analysis of single-cell and bulk transcriptomic data reveals altered cellular composition and predictive cell types in ectopic endometriosis. Front Med. 2025;12:1641982.\nZhu S, et al. The heterogeneity of fibrosis and angiogenesis in endometriosis revealed by single-cell RNA-sequencing. Biochim Biophys Acta Mol Basis Dis. 2023;1869(2):166602.\nLi Q, et al. The pathogenesis of endometriosis and adenomyosis: insights from single-cell RNA sequencing†. Biol Reprod. 2024;110(5):854–65.\nCollins KE, et al. Transcriptomic analyses of ovarian clear-cell carcinoma with concurrent endometriosis. Front Endocrinol (Lausanne). 2023;14:1162786.\nZhang H, et al. Integration of single cell and bulk RNA-sequencing reveals key genes and immune cell infiltration to construct a predictive model and identify drug targets in endometriosis. J Inflamm Res. 2025;18:2783–804.\nLi S, et al. Landscape analysis of m6A modification regulators related biological functions and immune characteristics in myasthenia gravis. J Transl Med. 2023;21(1):166.\nCharoentong P, et al. Pan-cancer immunogenomic analyses reveal genotype-immunophenotype relationships and predictors of response to checkpoint blockade. Cell Rep. 2017;18(1):248–62.\nSteinbuch SC, et al. Endometriosis-Associated Ovarian Cancer: from Molecular Pathologies to Clinical Relevance. Int J Mol Sci. 2024;25(8):4306.\nChen Y, et al. Roles of the adaptor protein tumor necrosis factor receptor type 1-associated death domain protein (TRADD) in human diseases. Biomed Pharmacother. 2022;153:113467.\nSingh P, Metkari SM, Bhartiya D. Mice uterine stem cells are affected by neonatal endocrine disruption & initiate uteropathies in adult life independent of circulatory ovarian hormones. Stem Cell Rev Rep. 2022;18(5):1686–701.\nReis JL, et al. The role of NK and T cells in endometriosis. Int J Mol Sci. 2024;25(18):10141.\nMariadas H, Chen JH, Chen KH. The molecular and cellular mechanisms of endometriosis: from basic pathophysiology to clinical implications. Int J Mol Sci. 2025;26(6):2458.\nQin X, et al. RNA modifications in cancer immune therapy: regulators of immune cells and immune checkpoints. Front Immunol. 2024;15:1463847.\nYang F, et al. Regulation of RNA splicing in endometrial tissue and its association with endometriosis. iScience. 2025;28(9):113207.\nGires O, et al. Expression and function of epithelial cell adhesion molecule EpCAM: Where are we after 40 years? Cancer Metastasis Rev. 2020;39(3):969–87.\nBrown TC, Sankpal NV, Gillanders WE. Functional implications of the dynamic regulation of EpCAM during epithelial-to-mesenchymal transition. Biomolecules. 2021;11(7):956.\nNanavati BN, et al. The desmosome-intermediate filament system facilitates mechanotransduction at adherens junctions for epithelial homeostasis. Curr Biol. 2024;34(17):4081-4090.e5.\nHuang T, et al. Heat shock protein family H member 1 HSPH1 expression correlates with progression and prognosis of hepatocellular carcinoma. Sci Rep. 2025;15(1):31151.\nWen KC, et al. The role of EpCAM in tumor progression and the clinical prognosis of endometrial carcinoma. Gynecol Oncol. 2018;148(2):383–92.\nLiu D, et al. Increased expression of epithelial cell adhesion molecule and its possible role in epithelial-mesenchymal transition in endometriosis. J Obstet Gynaecol Res. 2020;46(10):2066–75.\nSpooner-Harris M, et al. A re-appraisal of mesenchymal-epithelial transition (MET) in endometrial epithelial remodeling. Cell Tissue Res. 2023;391(2):393–408.\nWang Z, et al. Correction: CADM2 participates in endometriosis development by influencing the epithelial-mesenchymal transition. Reprod Sci. 2024;31(7):2135.\nFukaya T, et al. Intercellular adhesion molecule-1 and hepatocyte growth factor in human endometriosis: original investigation and a review of literature. Gynecol Obstet Invest. 1999;47(Suppl 1):11–6 (discussion 16-7).\nAcknowledgements\nNot applicable.\nFunding\nThis work was financially supported by the National Natural Science Foundation of China (Grant No. 81902334) and the Jiangsu Provincial Health Commission General Program (Grant No. H2023007).\nAuthor information\nAuthors and Affiliations\nContributions\nYehua Jing designed this study. Qianqian Ma acquired the data. Xingwei Wang interpreted the data. Yehua Jing drafted the manuscript. Qianqian Ma and Xingwei Wang revised the manuscript. All authors read and approved this manuscript.\nCorresponding authors\nEthics declarations\nEthics approval and consent to participate\nEthical approval was not required for this study, because it did not involve any human experiments.\nConsent for publication\nNot applicable.\nCompeting interests\nThe authors declare no competing interests.\nAdditional information\nPublisher's Note\nSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\nSupplementary Information\nRights and permissions\nOpen 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/.\nAbout this article\nCite this article\nWang, X., Ma, Q. & Jing, Y. Integrated single-cell and transcriptomic analyses uncover m6A-related molecular alterations in endometriosis. Eur J Med Res 31, 600 (2026). https://doi.org/10.1186/s40001-026-04150-0\nReceived:\nAccepted:\nPublished:\nVersion of record:\nDOI: https://doi.org/10.1186/s40001-026-04150-0","source_license":"CC0","license_restricted":false}