Long RNA profiles of endometrial extracellular vesicles provide new insights into the pathogenesis of ovarian endometriosis

Background Ovarian endometriosis(OEM) is a highly prevalent condition that significantly affects women’s health. Tissue-derived extracellular vesicles(Ti-EVs) contain multiple molecules that maintain intercellular communication. They participate in pathological processes contributing to the progression of OEM. However, little is known about the roles of long RNAs within Ti-EVs in OEM.

Ti-EVs were extracted from the ectopic endometrium of women with endometriosis(EM) and from normal controls(CTRL). We performed long RNA profiling of the isolated Ti-EVs, followed by weighted gene co-expression network analysis and pathway analysis. Quantitative real-time PCR and immunohistochemistry were used to validate the expression of hub genes and to investigate their association with clinical features.

We discovered that 535 mRNAs, 84 long non-coding RNAs, and 104 circular RNAs were differentially expressed between EM-EVs and CTRL-EVs. The differentially expressed mRNAs were enriched in pathways related to the inflammatory response, negative regulation of interferon-gamma production, cell surface receptor signalling pathways, and sensory perception of pain. Competing endogenous RNA networks were constructed to explore the functions of differentially expressed lncRNAs and circRNAs. Weighted gene co-expression network analysis identified five hub genes (C7, ACTG2, DLK1, HOXC6, and PDLIM3) significantly associated with endometriosis. Quantitative real-time PCR and immunohistochemistry confirmed that these hub genes were consistently upregulated in EM-EVs. Furthermore, the protein expression levels of HOXC6, DLK1, and C7 were correlated with both CA125 levels and disease stage in women with OEM.

Our study provides the first assessment of long RNAs in Ti-EVs derived from ectopic endometrium and identifies several key genes in Ti-EVs that are significantly correlated with OEM. These findings provide novel insights into the pathogenesis of OEM. Similar content being viewed by others

Endometriosis (EM) is a benign gynaecological disease that mainly affects women of reproductive age and is characterised by the presence of endometrial-like tissue outside the uterus [1,2,3]. EM affects approximately 10% of women of reproductive age and leads to symptoms, such as pelvic pain, dysmenorrhoea, dyspareunia, and infertility. Furthermore, it can profoundly impact quality of life and work productivity [4, 5]. Traditionally, EM is defined surgically by the identification of endometrial tissue outside the uterus [6]. However, this narrow anatomical definition fails to elucidate the complete natural history of EM, the broad spectrum of its clinical manifestations, the frequent recurrence of symptoms, the underlying molecular pathophysiology, and responsiveness to current treatment options. EM involves dysregulation of multiple signalling pathways, affecting angiogenesis, attachment, adhesion, invasion, migration, proliferation, apoptosis, and inflammation [7]. EM is increasingly recognised as a multifactorial condition in which genetic factors play a crucial role in its pathogenesis [8]. Extracellular vesicles (EVs) are lipid-bilayer-enclosed particles secreted by cells into the extracellular space that serve as key mediators of intercellular communication. EVs have historically been categorised into microvesicles, exosomes, and apoptotic bodies based on biogenesis, release pathways, size, content, and function [9]. Extracellular RNA (exRNA) comprises RNA species localised outside parental cells. These molecules are typically enclosed within membrane-bound EVs or complexed with proteins and lipids [10]. Tissue-derived EVs (Ti-EVs) play important roles in multiple pathogenic processes, including EM. Ti-EVs, which can be released by virtually all cell types, are significant mediators of intercellular communication and interorgan crosstalk through the transfer of bioactive molecules, including proteins, lipids, and RNAs between cells [11]. For example, microRNAs in Ti-EVs have been implicated in key mechanisms of endometriosis pathophysiology, including the immunomodulation of peritoneal macrophages (pMφ), the proliferation and migration of ectopic lesions, and angiogenesis [12]. Therefore, a better understanding of Ti-EVs may provide valuable targets for disease diagnosis and treatment. In this study, we sought to characterise Ti-EVs and their long RNA cargo from the ectopic endometrium of women with OEM and control individuals. These findings offer novel insights into the development of improved diagnostic and therapeutic strategies for OEM.

Material collection process The inclusion criteria for the EM group were as follows: (1) women aged 25–40 years with regular menstrual cycles; (2) no adenomyosis or other malignant conditions; (3) no history of steroid hormone therapy within 3 months before surgical intervention; and (4) histopathological confirmation of OEM. Disease severity was classified according to the American Society for Reproductive Medicine (ASRM) staging system, with patients categorised into two subgroups: mild endometriosis (Stage I–II) and severe endometriosis (Stage III–IV) [13]. The inclusion criteria for the control group were as follows: (1) women undergoing surgical treatment for cervical intraepithelial neoplasia or hysterectomy for pelvic organ prolapse; and (2) women undergoing hysteroscopy for fallopian tubal diseases without endometriosis lesions. Exclusion criteria included intrauterine device use, active genitourinary infections or sexually transmitted diseases, perimenopausal status, hormone therapy within the preceding year, concurrent malignancies, and coexisting uterine pathologies such as leiomyomas or adenomyosis. All clinical samples in our study were collected during the luteal phase. The entire procedure followed sterile principles. For patients with OEM, endometriotic lesions were surgically obtained using sterile hemostatic forceps. The specimens were rinsed with isotonic sodium chloride solution to remove blood stains and other impurities, then cut into 1 × 0.5 × 0.5 cm³ pieces and aliquoted into cryovials. For normal endometrium (CTRL) from non-endometriotic patients undergoing hysteroscopy, the uterine cavity was gently scraped for 1 week using a curette, washed twice with isotonic sodium chloride solution, and aliquoted into cryopreservation tubes. Isolation of EVs from tissue samples Tissue samples(1 g) were collected from the EM and CTRL groups. The frozen tissue was chopped into 5 mm³ sections on ice using a scalpel. The tissue pieces were transferred into 10 mL of Hanks’ Balanced Salt Solution (HBSS) containing 100µL of collagenase/dispase. The mixture was then incubated at 37 °C for 30 min on a shaking bed. After incubation, the samples were cooled on ice, and a protease inhibitor cocktail was added to prevent protein degradation. The homogenised samples were filtered through a 40 μm mesh filter to remove larger debris. The filtered samples were centrifuged at 300 × g for 5 min, and the pellet was collected (referred to as “T1”). The supernatant from the first centrifugation was transferred to new tubes and centrifuged at 2,000 × g for 10 min, and the pellet was collected (referred to as “T2”). The supernatant from the second centrifugation was transferred to new tubes and centrifuged at 10,000 × g or 15,000 × g for 10 min, and the pellet was collected (referred to as “T3”). The supernatant from the third centrifugation was concentrated to 500 µL using ultrafiltration techniques. The concentrated 500 µL sample was loaded onto a qEV original(70 nm) SEC column. Fractions were eluted with phosphate-buffered saline (PBS) according to the manufacturer’s instructions. EVs were detected in fractions seven and eight, each containing 500 µL. To further purify and concentrate EVs, either ultracentrifugation or ultrafiltration was performed. Ultracentrifugation was performed for 70 min at 110,000 × g (average) at 4 °C (TH-641 rotor, Thermo Fisher; thinwall polypropylene tube, 13.2 mL capacity). The supernatant was removed, and the pellet was resuspended in 100 µL PBS. Ultrafiltration was performed using a 10 kDa molecular weight cut-off protein concentrator (Thermo Fisher, 88516), concentrating the original 1000 µL to 500 µL. Identification of EVs In accordance with the Minimal Information for Studies of Extracellular Vesicles 2018 guidelines, western blotting(WB), transmission electron microscopy (TEM), and nanoparticle tracking analysis (NTA) were used to identify isolated Ti-EVs [14]. WB Isolated EVs were homogenised in RIPA lysis buffer on ice for 30 min, with mixing every 10 min. EV proteins were sedimented by centrifugation, and protein concentration was determined using a bicinchoninic acid (BCA) assay. Twenty micrograms of protein were separated using SDS-PAGE and transferred onto polyvinylidene difluoride membranes. The membranes were incubated overnight at 4 °C with primary antibodies against CD9, Alix, TSG101, and calnexin. After primary incubation, the membranes were incubated with horseradish peroxidase-conjugated secondary antibodies for 1 h at room temperature. Results were visualised using an enhanced chemiluminescence detection reagent. Transmission electron microscopy (TEM) The morphological characteristics of the EVs were analysed using a Hitachi H7650 transmission electron microscope (Hitachi, Tokyo, Japan). Briefly, a 10 µL aliquot of the EV pellet was placed on a copper mesh, incubated at room temperature for 10 min, and then washed with sterile distilled water. Subsequently, the samples were negatively stained with 10 µL of 2% uranyl acetate for 1 min. After removing excess liquid with filter paper, the grid was dried under an incandescent lamp for 2 min before imaging at an accelerating voltage of 80 kV. NTA The concentration, size distribution, and particle movement of the isolated EVs were determined by NTA using a ZetaView PMX 110 particle tracker (Particle Metrix, Germany). Briefly, EV samples were diluted in PBS to a concentration within the optimal measuring range of 1 × 10⁷ to 1 × 10⁹ particles/mL. Measurements were performed using a 405 nm laser, and the scattered light from individual particles was tracked and analysed using the instrument software to determine particle concentration and size. RNA-sequencing workflow RNA sequencing was performed using an Illumina HiSeq 3000 platform. Sequencing libraries were prepared using a Total RNA-Seq Library Prep Kit(Illumina). The reads were first mapped to the latest University of California, Santa Cruz genome browser transcript set using Bowtie2(version 2.1.0), and gene expression levels were estimated using RSEM(v1.2.15). For lncRNA expression analysis, we used the transcript set from Lncipedia (www.lncipedia.org). The trimmed mean of M-values method was used to normalise gene expression. For circRNA expression analysis, reads were mapped to the human genome using STAR, and DCC was used to annotate and quantify circRNA expression levels. Trimmed mean of M-values was used for normalisation. The sequencing depth was set at approximately 50 million reads per sample to ensure sensitive detection of low-abundance exosomal transcripts. Whole-transcriptome high-throughput sequencing and bioinformatics Whole-transcriptome high-throughput sequencing was performed for the EM and CTRL groups. Pathway enrichment analysis was conducted using Gene Ontology (GO) and Hallmark databases. The GSE120103 and GSE141549 datasets were downloaded from the Gene Expression Omnibus (GEO) database. The GSE120103 dataset contains 18 endometriotic lesions from women with OEM and 18 normal endometrium samples. GSE141549 contains the expression matrix and clinical information for 198 endometriotic lesions. The final RNA-Seq expression matrix was organised with column names representing the sample IDs and row names representing the gene symbols. We applied the Benjamini-Hochberg (BH) procedure to control the false discovery rate (FDR) at a level of 0.05. Differential gene screening was performed using the edgeR package, with a threshold of |log2FC| ≥ 0.5 and FDR < 0.05. However, given the limited sample size and low abundance of circRNAs in Ti-EVs, we used a P-value < 0.2 without correction to avoid excluding potentially meaningful signals. Pathway enrichment analyses (GO, KEGG, and GSEA) were performed using the R package ‘clusterProfiler’, and significant GO terms were visualised using ‘Goplot’. Network analysis of differentially expressed mRNAs(DEmRNAs) was conducted using NetworkAnalyst. Weighted Gene Co-expression Network Analysis (WGCNA) To explore gene co-expression patterns, WGCNA was used. The GSE120103 dataset was downloaded as a matrix file. Raw CEL files for GSE141549 were downloaded and processed using the affy package in R. Background correction and normalisation were performed using the Robust Multi-array Average algorithm. Probes without corresponding gene symbols were removed. When multiple probes mapped to the same gene, the probe with the highest mean expression was retained as the representative of that gene. Batch correction was not applied to GSE120103. ComBat batch correction was applied using the sva package to remove potential batch effects from GSE141549. Outlier samples were removed by hierarchical clustering, and the soft-thresholding power was selected to achieve a scale-free topology (R2 > 0.85). Module preservation analysis was performed to evaluate the stability of the identified modules across datasets. Subsequently, the adjacency matrix was transformed into a Topological Overlap Matrix to measure network interconnectedness. A hierarchical clustering function was used to group genes with similar expression profiles into distinct modules. The derived gene modules were subsequently associated with clinical features by assessing both gene significance (GS) and module membership (MM). Genes with MM > 0.8 and gene significance > 0.2 were considered candidate hub genes within each module. To ensure cross-dataset consistency, we identified the intersection of candidate hub genes derived from corresponding modules in the two datasets. EV RNA extraction Total RNA was extracted and purified from the EVs using the miRNeasy Mini Kit (catalogue #217004, Qiagen, Düsseldorf, Germany) according to the manufacturer’s instructions. A NanoDrop was used to assess RNA concentration and purity. RNA concentration was measured using 1 µL of sample, with DEPC-treated water as the blank control. RNA samples with an A260/A280 ratio of approximately 2.0 were considered high quality. Samples with ratios below 1.8 or above 2.2 were considered potentially contaminated and were not used for further experiments. Reverse transcription was performed immediately for samples that met the quality criteria. Remaining RNA samples were stored at − 80 °C for long-term preservation. Quantitative Real-Time PCR (qRT-PCR) cDNA was synthesised using PrimeScript RT Master Mix. qRT-PCR assays were performed using the Green Premix Ex Taq kit. GAPDH was used as the reference gene for mRNA expression. Relative gene expression was quantified using the 2−ΔΔCt method, and each reaction was performed in triplicate. The primers used are listed in Table S1. Immunohistochemical staining Tissue sections were dewaxed in xylene I and II for 15 min each with gentle agitation, rehydrated using a graded ethanol series (100% ethanol I/II, 90%, 85%, and 75% for 5 min each), and rinsed in double-distilled water. After washing in PBS (3 × 5 min with agitation), antigen retrieval was performed using citrate buffer (pH 6.0) with high-pressure heating for 3 min, followed by cooling and washing with PBS. Endogenous peroxidase activity was blocked by incubation with 3% hydrogen peroxide for 25 min at room temperature. Non-specific binding was minimised by incubation with 5% BSA for 40 min at room temperature. Primary antibodies were applied and incubated overnight at 4 °C in a humidified chamber (12–16 h), followed by PBS washing with PBS and incubation with the appropriate secondary antibodies for 60 min at room temperature. Chromogenic detection was performed using freshly prepared DAB substrate for 3–10 min, followed by haematoxylin counterstaining (2–3 min), differentiation in 1% acid alcohol, and nuclear staining in aqueous ammonia. Sections were then dehydrated using graded ethanol concentrations, cleared in xylene, and mounted with neutral resin. Quantitative analysis of staining intensity and positive area was conducted using ImageJ software, and immunohistochemical scores were compared across experimental groups. Evaluation of immunohistochemical results Immunohistochemical results were assessed by examining five randomly selected fields per section at 400× magnification. Brown or yellow granular cytoplasmic staining was considered positive. Quantitative analysis was conducted using ImageJ by measuring the integrated density of the positively stained area and normalising it to the corresponding area to calculate the average optical density (integrated density /area). This standardised quantification method enabled objective, reproducible, and unbiased comparison of protein expression levels across groups, while ensuring compatibility with downstream statistical analyses. Field selection was performed by a researcher, and all slides were coded to ensure that the researcher was blinded to the clinical information during selection. Clinical data collection Basic clinical information was collected from all enrolled participants, including age, weight, height, body mass index (BMI), endometriosis stage, presence of pelvic pain, presence of infertility, and serum CA125 levels. Statistical analysis SPSS 26.0 and GraphPad Prism 9.0 were used for data processing and statistical analysis. Normally distributed measurement data were expressed as the mean ± standard deviation, and comparisons between groups were conducted using independent samples t tests. Measurements that were not normally distributed were presented as medians (interquartile ranges), and comparisons between groups were conducted using the non-parametric Mann–Whitney U test. Correlation analyses were conducted according to data distribution: Pearson’s correlation was applied to normally distributed variables, Spearman’s correlation to non-normally distributed variables, and Kendall’s tau correlation to relations between continuous and ordinal variables. To account for multiple testing, Fisher’s Z method was applied, and correlations with an FDR < 0.05 were considered statistically significant.

Characterisation of EVs isolated from endometrium As shown in Fig. 1A and B, endometrial Ti-EVs were isolated from clinical samples and used for whole-transcriptome high-throughput sequencing. EVs were visualised using TEM. Figure 1C shows small spherical vesicles with diameters < 200 nm in samples obtained from women with OEM. These results indicate that the purity of the isolated EVs met the MISEV2018 criteria. NTA was used for single-particle tracking, and Fig. 1D shows that EV diameters ranged from 100 to 150 nm. Western blot analysis demonstrated that EV marker proteins (CD9, Alix, and TSG101) were highly expressed in the isolated EVs, whereas the negative marker calnexin was not detected (Fig. 1E). CTRL and EM EVs exhibit distinct mRNA profiles We quantified 13,198 mRNAs in the ectopic and control endometrium. A total of 535 DEmRNAs were identified in Ti-EVs, including 447 upregulated genes and 88 downregulated genes, in ectopic endometrial samples (EM) compared with those in normal endometrial samples (CTRL) (Fig. 2A). According to GO biological process (GO-BP) enrichment analysis, EM-EVs were significantly enriched with inflammatory response, immunoglobulin production involved in immunoglobulin-mediated immune response, negative regulation of interferon-gamma production, cell surface receptor signalling pathway, and sensory perception of pain. Most DEmRNAs within these GO-BP terms were upregulated (Fig. 2B). GO molecular function analysis indicated enrichment of DEmRNAs in carbohydrate binding, transmembrane signalling receptor activity, endochitinase activity, RNA polymerase II distal enhancer sequence-specific DNA binding, and N, N-dimethylaniline monooxygenase activity, which were generally upregulated (Fig. 2C). GO cellular component analysis showed enrichment in integral component of the plasma membrane, proteinaceous extracellular matrix, external side of the plasma membrane, extracellular space, and extracellular region, and these DEmRNAs were generally upregulated (Fig. 2D). These findings indicate an imbalance in the immune environment of EM tissues. GSEA was performed to identify biological pathways enriched in EM-EVs compared with those in CTRL-EVs. Upregulated pathways included cytokine-cytokine receptor interaction, chemokine signalling pathway, tuberculosis, phagosome, and osteoclast differentiation (Fig. 2E). Network-based analysis was performed to identify hub genes in EM-EVs. In the constructed network, key genes were predominantly associated with inflammatory mediators (Fig. 2F). Significantly upregulated genes included IL6, IL10, CD40, and FCGR1A. IL6, produced by B lymphocytes, has been shown to modulate immune cells, including CD4+ T cells, and is closely associated with endometriois. IL10 has been reported to stimulate proliferation and invasion of endometrial stromal cells in vitro and growth of ectopic lesions in vivo. FCGR1A (CD64), a transmembrane glycoprotein receptor that functions alongside CD32 and CD16, plays an important role in both innate and adaptive immune responses. Eva Vargas reported that CD40 is dysregulated in EM sibling disorders. Most of these genes have not been previously reported in this context, and further studies are required to validate these findings. Differentially expressed lncRNAs and functional enrichment analysis of lncRNA-related target genes In the identification of lncRNAs, 84 lncRNAs were differentially expressed. Among these, 71 lncRNAs were upregulated in EM-EVs and 13 lncRNAs were downregulated (P 1.5) (Fig. 3A). Because lncRNAs exert regulatory effects through neighbouring target genes (cis-regulation) or distant target genes (trans-regulation), functional enrichment analyses were performed for both cis-regulated and trans-regulated target genes (Fig. 3B). GO-BP analysis showed that the target genes of DElncRNAs were significantly enriched in translational initiation, SRP-dependent cotranslational protein targeting to membrane, and viral transcription. Enriched GO cellular component terms included cytosol and focal adhesion. Enriched GO molecular function terms included RNA binding, zinc ion binding, and structural constituent of ribosome. KEGG pathway analysis revealed enrichment in ribosome, lysosome, and phagosome pathways. Competing endogenous RNA (ceRNA) represents an important regulatory mechanism of lncRNAs. A ceRNA network was constructed based on lncRNA–miRNA and miRNA–mRNA interactions (Fig. 3C). The Mircode database was used to predict lncRNA–miRNA interactions, and miRDB, miRTarBase, and TargetScan were used to predict miRNA–mRNA interactions. The lncRNA-miRNA-mRNA ceRNA network comprised 12 downregulated miRNAs, 28 upregulated lncRNAs, and 151 upregulated mRNAs. GO enrichment analysis of genes within the ceRNA network demonstrated pathways overlapping with those identified in cis- and trans-regulation analyses, including translational initiation, focal adhesion, and SRP-dependent co-translational protein targeting to membrane (Fig. 3D). These overlapping pathways may play important roles in the pathogenesis of OEM. Differentially expressed circRNAs and functional enrichment analysis of circRNA-related target genes A total of 104 differentially expressed circular RNAs were selected to visualise their chromosomal locations and expression patterns (Fig. 4A). DEcircRNAs in EM-EVs and CTRL-EVs were identified using a heatmap (Fig. 4B). The volcano plot (Fig. 4C) illustrates representative DEcircRNAs. All DEcircRNAs were upregulated. To further elucidate the functions and biological processes associated with DEcircRNAs, GO and KEGG enrichment analyses were conducted. The top three enriched GO-BP terms were negative regulation of cellular component organisation, synaptic vesicle budding from the presynaptic endocytic zone membrane, and regulation of histone ubiquitination. KEGG pathway enrichment analysis indicated that the most significant pathways were glutamatergic synapse, microRNAs in cancer, and Fc gamma R-mediated phagocytosis (Fig. 4D). The MiRanda program was used to predict circRNA-miRNA interactions, whereas miRDB, miRTarBase, and TargetScan were used to predict miRNA–mRNA interactions. Only miRNA-mRNA interactions identified in all three databases were retained. DEmRNA-containing mRNAs were then incorporated into the ceRNA network (Fig. 4E). Next, we constructed a circRNA-related ceRNA network and identified functional pathways whose targets were comparable to those of the DEcirRNAs (Fig. 4F). Construction of WGCNA network based on GEO databases The mRNAs with similar expression patterns were grouped into modules using average linkage clustering (Fig. 5). WGCNA was performed on DEGs derived from the analysis of GSE120103 and GSE141549 datasets to identify modules associated with clinical traits. Clinical traits, including non-pregnancy status, were retrieved from the GSE120103 (Fig. 5A). The heatmap of the module–trait correlations showed that the magenta module (module–trait correlation = 0.45, P = 0.006; Fig. 5B) and black module (module–trait correlation = 0.6, P = 1 × 10− 4 < 0.01; Fig. 5B) were significantly associated with non-pregnancy status. Clinical traits, including disease stage, were retrieved from the GSE141549 (Fig. 5C). The module-trait correlation heatmap identified the turquoise module as highly correlated with disease stage (correlation coefficient = 0.89, P = 9 × 10− 25 < 0.01; Fig. 5D). Finally, the intersection of gene sets associated with these clinical traits was identified. A Venn diagram was constructed, and five overlapping genes were identified (Fig. 5E). Validation of five clinical-trait-related genes in Ti-EVs Consistent with sequencing results, the expression levels of C7, ACTG2, DLK1, HOXC6, and PDLIM3 in Ti-EVs were validated using qRT-PCR. The expression patterns of all selected genes showed consistent trends between sequencing and qRT-PCR analyses. These findings indicate that these five associated genes may play important roles in the progression of OEM (Fig. 6). Expression of C7, DLK1 and HOXC6 proteins in ectopic, eutopic, and normal endometrium Immunohistochemical analysis showed that C7, DLK1, and HOXC6 expression levels were elevated in ectopic endometrial tissues(EM) compared to both eutopic endometrium (EU) and control endometrium (CTRL) (Fig. 7A–F). No statistically significant differences in DLK1 or HOXC6 protein expression were observed between the CTRL and EU groups (Fig. 7B, C, E, and F). However, C7 protein expression was significantly higher in the EU group than in the CTRL group (Fig. 7A and D). Correlation between C7, DLK1, and HOXC6 protein expression in ectopic endometrium and clinical characteristics (Table 1) Spearman and Kendall correlation analyses revealed that HOXC6 expression in women with OEM was significantly correlated with DLK1 expression (r = 0.438, P < 0.01), serum CA125 levels (r = 0.294, P < 0.05), and disease stage (r = 0.631, P < 0.01). DLK1 expression was also correlated with CA125 (r = 0.392, P < 0.05) and disease stage (r = 0.578, P < 0.01). Additionally, C7 expression was strongly correlated with disease stage (r = 0.762, P < 0.05), and CA125 levels were significantly correlated with disease stage (r = 0.515, P < 0.01).

EVs isolated from endometrial epithelial cells have been reported previously. However, studies on Ti-EVs derived from ectopic endometrium in OEM are limited [15]. In this study, we successfully isolated EVs from human endometrium and characterised their cargo using whole-transcriptome high-throughput sequencing. We identified aberrantly expressed lncRNAs, circRNAs, miRNAs, and mRNAs by analysing whole-transcriptome expression profiles of EM-EVs and CTRL-EVs. Functional enrichment analysis showed that the DEmRNAs were mainly involved in inflammatory response, immunoglobulin production involved in immunoglobulin-mediated immune response, negative regulation of interferon-gamma production, cell surface receptor signalling pathway, and sensory perception of pain. In addition, GSEA confirmed significant enrichment of cytokine-cytokine receptor interaction and chemokine signalling pathways. It remains unclear whether immune dysregulation contributes to the development of OEM or occurs as a consequence of the disease [16]. Nevertheless, immune components play a major role in the pro-inflammatory phenotype observed in lesions and the eutopic endometrium [17]. Evidence indicates that disruption of the immune environment is widely involved in EM pathology [18]. Furthermore, the chemokine signalling pathway was upregulated in EM-EVs compared with that in CTRL-EVs. Network analysis identified several hub genes, including IL6, IL10, and CD40. These findings suggest concurrent activation of pro-inflammatory cytokines and compensatory elevation of anti-inflammatory factors in endometriosis. Cytokine imbalance ensures both chronic inflammation damaging tissues and immunosuppression preventing elimination of ectopic implants [19]. EM severity is associated with elevated IL6, IL10, and CD40 levels [20, 21]. Notably, DEmRNAs were also enriched in sensory perception of pain, suggesting a potential regulatory mechanism underlying pelvic pain in EM. LncRNAs play important roles in human diseases, including endometriosis [22]. However, the contribution of lncRNAs within Ti-EVs to EM remains unclear. We identified differentially expressed lncRNAs associated with OEM and constructed lncRNA–miRNA interaction networks. GO and KEGG enrichment analyses were performed to evaluate the functions of DElncRNAs. Enriched terms included translational initiation, SRP-dependent co-translational protein targeting to membrane, and viral transcription. Within the constructed network, lncRNA-FAM30A exhibited the highest number of interactions. FAM30A has been reported to be associated with B lymphocyte biology [23]. Therefore, lncRNA-FAM30A may contribute to immune pathway regulation during endometriosis progression. In addition, miR1291 and miR1296-5p, which interact with lncRNA-FAM30A, may serve as potential markers of disease progression and prognosis [24]. However, the functional role of this regulatory axis in endometriosis has not been reported and warrants further investigation. CircRNAs are a recently identified class of closed-loop endogenous non-coding RNAs. Although their role in endometriosis is not fully understood, previous studies suggest that circRNAs may influence endometrial stromal cell proliferation, migration, and invasion and may serve as molecular markers of disease [25]. We investigated circRNAs in Ti-EVs derived from ectopic endometrium and comprehensively analysed circRNA expression profiles and circRNA-associated ceRNA networks in endometrium-derived EVs. Differential expression analysis was performed using DESeq2, and circRNAs were filtered using an adjusted P value threshold of 0.2 to retain potentially biologically relevant candidates. Functional analysis of parent genes of DEcircRNAs revealed enrichment of proliferation-related terms, including microRNAs in cancer. Histone ubiquitination-related terms were also enriched. Histone ubiquitination is a major epigenetic modification of histone tails and plays an important role in regulating gene transcription. A ceRNA network based on DEcircRNAs was constructed, and GO enrichment analysis was performed for genes within this network. These findings suggest that the circRNA-associated network may influence endometriosis pathogenesis through multiple mechanisms, including regulation of histone ubiquitination, presynaptic endocytic zone membrane processes, and ubiquitin ligase complex activity. These results are consistent with enrichment findings for DEmRNAs and DElncRNAs, suggesting that EM-EVs may contribute to endometriosis pathogenesis through coordinated transcriptomic regulation. Because the sample size was limited, we couldn’t apply WGCNA to construct gene co-expression networks in our study cohort. To obtain additional insights, we retrieved two GEO datasets (GSE120103 and GSE141549) and performed an integrated analysis combining WGCNA and differential gene expression analysis. Based on these results, the magenta and black modules were significantly correlated with non-pregnancy status, whereas the turquoise module was significantly associated with disease stage. Five key genes were identified by intersecting hub genes from the relevant modules, including ACTG2, C7, DLK1, HOXC6, and PDLIM3. These five genes were selected for further validation by qRT-PCR, and the results were consistent with the RNA sequencing findings. All five hub genes were highly expressed in EM-EVs. Immunohistochemical analysis demonstrated that C7, DLK1, and HOXC6 protein expression levels were higher in ectopic endometrium than in control endometrium. Correlation analysis showed that DLK1 and HOXC6 expression levels were associated with endometriosis stage and serum CA125 levels, whereas C7 expression was correlated only with disease stage. Therefore, these three hub genes may serve as candidate molecular markers for the diagnosis and treatment of OEM. The complement system possesses the ability to recognise foreign pathogens and plays a key role in regulating adaptive immune responses [26]. As a central component of the terminal complement cascade, C7 plays a pivotal role in immune regulation. Previous studies have reported significantly increased C7 protein expression and a higher number of CD3⁺FOXP3⁺ regulatory T cells in endometriotic lesions than in normal endometrium. Importantly, C7 protein expression in endometriotic lesions was positively correlated with the number of CD3⁺FOXP3⁺ regulatory T cells. C7 has also been reported to be upregulated in endometriosis and was identified as a commonly upregulated gene in the present study [27]. Various uterine pathologies, including endometrial cancer, adenomyosis, and endometriosis, have been linked to dysfunction of tissue-resident, very small embryonic-like stem cells (VSELs). These stem cells exhibit epigenetic alterations and loss of imprinting at loci, such as DLK1-Meg3 [28]. Abnormal overexpression of DLK1 markedly increases expression of cell cycle-related proteins, including E2F1, CDK2, and Cyclin D1, suggesting a potential molecular mechanism by which DLK1 promotes cell proliferation. Moreover, DLK1 has been reported to influence cellular processes, such as proliferation, apoptosis, and inflammatory responses, through involvement in the Notch and MAPK signalling pathways. HOXC6 is an oncogenic transcription factor involved in multiple regulatory pathways. HOXC6 was identified as one of the most aberrantly expressed genes in both the infertile women and women with endometriosis [29]. Furthermore, HOXC6 was reported to be overexpressed in endometriosis and may promote proliferation, adhesion, invasion, and migration of endometrial stromal cells through the TGF-β1/Smad signalling pathway [30]. These findings are consistent with our results. Our study has several limitations. The sample sizes used for qRT-PCR and immunohistochemical analyses were relatively small, which may limit statistical power. Larger studies are required to validate these findings. In addition, because of the relatively subtle transcriptional differences observed between groups for circRNAs, a less stringent significance threshold (P < 0.2) was applied during initial screening to retain potentially biologically relevant candidates that might be excluded using conventional cut-offs. We attribute these subtle transcriptional differences to the small sequencing sample size and relatively low abundance of circRNAs in Ti-EVs compared with other long RNAs. Although we identified associations between candidate genes and clinical features and explored potential pathways using bioinformatic analyses, further experimental studies are required to clarify underlying biological mechanisms and assess their diagnostic value. Nevertheless, we characterised transcriptomic features of endometrial Ti-EVs in EM and CTRL groups. These findings expand current understanding of EV-mediated regulatory mechanisms in the pathogenesis of OEM.

These findings highlight transcriptomic alterations in EM-EVs and support their potential as minimally invasive biomarkers for ovarian endometriosis. Data availability The datasets analysed in the current study are available in the Gene Expression Omnibus under accession numbers GSE120103 and GSE141549.The repository links are provided below: [https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=gse120103](https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi). Abbreviations - OEM: - Ovarian endometriosis - ceRNA: - Competing endogenous RNA - CTRL: - Control - DEmRNAs: - Differentially expressed mRNAs - EM: - Endometriosis - EVs: - Extracellular vesicles - FDR: - False discovery rate - GEO: - Gene Expression Omnibus - GO: - Gene Ontology - GO-BP: - Gene Ontology biological process - MM: - Module membership - NTA: - Nanoparticle tracking analysis - PBS: - Phosphate-buffered saline - qRT-PCR: - Quantitative real-time PCR - SRP: - Signal recognition particle - TEM: - Transmission electron microscopy - Ti-EVs: - Tissue-derived extracellular vesicles - WGCNA: - Weighted Gene Co-expression Network Analysis

Kapoor R, Stratopoulou CA, Dolmans MM. Pathogenesis of endometriosis: new insights into prospective therapies. Int J Mol Sci. 2021;22(21):11700. Chen Y, Ye L, Zhu J, Chen L, Chen H, Sun Y, Rong Y, Zhang J. Disrupted Tuzzerella abundance and impaired L-glutamine levels induce Treg accumulation in ovarian endometriosis: a comprehensive multi-omics analysis. Metabolomics. 2024;20(2):32. Yin D, Xie D, De Vos S, Liu G, Miller CW, Black KL, Koeffler HP. Imprinting status of DLK1 gene in brain tumors and lymphomas. Int J Oncol. 2004;24(4):1011–5. Yurtkal A, Oncul M. Comparison of dienogest or combinations with ethinylestradiol/estradiol valerate on the pain score of women with endometriosis: A prospective cohort study. Med (Baltim). 2024;103(27):e38585. Della Corte L, Di Filippo C, Gabrielli O, Reppuccia S, La Rosa VL, Ragusa R, Fichera M, Commodari E, Bifulco G, Giampaolino P. The burden of endometriosis on women’s lifespan: a narrative overview on quality of life and psychosocial wellbeing. Int J Environ Res Public Health. 2020;17(13):4683. Saunders PTK, Horne AW. Endometriosis: Etiology, pathobiology, and therapeutic prospects. Cell. 2021;184(11):2807–24. Wang H, Gan Z, Wang Y, Hu D, Zhang L, Ye F, Duan P. A Noninvasive Menstrual Blood-Based Diagnostic Platform for Endometriosis Using Digital Droplet Enzyme-Linked Immunosorbent Assay and Single-Cell RNA Sequencing. Res (Wash D C). 2025;8:0652. Canday M, Yurtkal A, Makav M, Kuru M. Anti-inflammatory, antioxidant, antiangiogenic, and therapeutic efficacy of neroli oil in rats with endometriotic lesions. J Obstet Gynaecol Res. 2023;50(3):516–25. Wu Y, Toldo N, Fabbri M. The interplay between the extracellular matrix and extracellular vesicle-associated microRNAs. Cell Commun Signal. 2026;24(1):89. Patton JG, Franklin JL, Weaver AM, Vickers K, Zhang B, Coffey RJ, Ansel KM, Blelloch R, Goga A, Huang B, et al. Biogenesis, delivery, and function of extracellular RNA. J Extracell Vesicles. 2015;4:27494. Pang H, Fan W, Shi X, Li J, Wang Y, Luo S, Lin J, Huang G, Li X, Xie Z, et al. Characterization of lncRNA Profiles of Plasma-Derived Exosomes From Type 1 Diabetes Mellitus. Front Endocrinol (Lausanne). 2022;13:822221. Nazri HM, Greaves E, Quenby S, Dragovic R, Tapmeier TT, Becker CM. The role of small extracellular vesicle-miRNAs in endometriosis. Hum Reprod. 2023;38(12):2296–311. Lee SY, Koo YJ, Lee DH. Classification of endometriosis. Yeungnam Univ J Med. 2021;38(1):10–8. Théry C, Witwer KW, Aikawa E, Alcaraz MJ, Anderson JD, Andriantsitohaina R, Antoniou A, Arab T, Archer F, Atkin-Smith GK, et al. Minimal information for studies of extracellular vesicles 2018 (MISEV2018): a position statement of the International Society for Extracellular Vesicles and update of the MISEV2014 guidelines. J Extracell Vesicles. 2018;7(1):1535750. Abudula M, Fan X, Zhang J, Li J, Zhou X, Chen Y. Ectopic Endometrial Cell-Derived Exosomal Moesin Induces Eutopic Endometrial Cell Migration, Enhances Angiogenesis and Cytosolic Inflammation in Lesions Contributes to Endometriosis Progression. Front Cell Dev Biol. 2022;10:824075. Vallvé-Juanico J, Houshdaran S, Giudice LC. The endometrial immune environment of women with endometriosis. Hum Reprod Update. 2019;25(5):564–91. Abramiuk M, Grywalska E, Małkowska P, Sierawska O, Hrynkiewicz R, Niedźwiedzka-Rystwej P. The role of the immune system in the development of endometriosis. Cells. 2022;11(13):2028. Ahn SH, Monsanto SP, Miller C, Singh SS, Thomas R, Tayade C. Pathophysiology and Immune Dysfunction in Endometriosis. Biomed Res Int. 2015;2015:795976. Gorji-Bahri G, Moradtabrizi N, Vakhshiteh F, Hashemi A. Validation of common reference genes stability in exosomal mRNA-isolated from liver and breast cancer cell lines. Cell Biol Int. 2021;45(5):1098–110. Sun Z, Shi M, Xia J, Li X, Chen N, Wang H, Gao Z, Jia J, Yang P, Ji D, et al. HDAC and MEK inhibition synergistically suppresses HOXC6 and enhances PD-1 blockade efficacy in BRAF(V600E)-mutant microsatellite stable colorectal cancer. J Immunother Cancer. 2025;13(1):e010460. Bashir ST, Redden CR, Raj K, Arcanjo RB, Stasiak S, Li Q, Steelman AJ, Nowak RA. Endometriosis leads to central nervous system-wide glial activation in a mouse model of endometriosis. J Neuroinflammation. 2023;20(1):59. Wang M, Zheng L, Lin R, Ma S, Li J, Yang S. A comprehensive overview of exosome lncRNAs: emerging biomarkers and potential therapeutics in endometriosis. Front Endocrinol (Lausanne). 2023;14:1199569. de Lima DS, Cardozo LE, Maracaja-Coutinho V, Suhrbier A, Mane K, Jeffries D, Silveira ELV, Amaral PP, Rappuoli R, de Silva TI, et al. Long noncoding RNAs are involved in multiple immunological pathways in response to vaccination. Proc Natl Acad Sci U S A. 2019;116(34):17121–6. Maier IM, Maier AC. miRNAs and lncRNAs: potential non-invasive biomarkers for endometriosis. Biomedicines. 2021;9(11):1662. Feng Y, Zhang T, Wang Y, Xie M, Ji X, Luo X, Huang W, Xia L. Homeobox Genes in Cancers: From Carcinogenesis to Recent Therapeutic Intervention. Front Oncol. 2021;11:770428. Dunkelberger JR, Song W-C. Complement and its role in innate and adaptive immune responses. Cell Res. 2010;20(1):34–50. Bae S-J, Jo Y, Cho MK, Jin J-S, Kim J-Y, Shim J, Kim YH, Park J-K, Ryu D, Lee HJ, et al. Identification and analysis of novel endometriosis biomarkers via integrative bioinformatics. Front Endocrinol. 2022;13:942368. Singh P, Bhartiya D. Mouse uterine stem cells are affected by endocrine disruption and initiate uteropathies. Reproduction. 2023;165(3):249–68. Li Q, Xi M, Shen F, Fu F, Wang J, Chen Y, Zhou J. Identification of Candidate Gene Signatures and Regulatory Networks in Endometriosis and its Related Infertility by Integrated Analysis. Reprod Sci. 2022;29(2):411–26. Jiang J, Zhang L, Li L. Homeobox C6 is Up-Regulated and Affects the Pathogenesis of Endometriosis. Reprod Sci. 2025;32(8):2574–82.

The authors thank all research assistants of Tongji Hospital affiliated with Tongji University for their assistance in the collection of materials. Funding This study was supported by grants from the Shanghai Science and Technology Commission Project (Nos.20Z21900400 and NO.22410761100) and the Clinical Research Project of Tongji Hospital (ITJ(ZD)2208). Author information Authors and Affiliations Contributions Yanqiu Wang and Qizhen Chen designed the study; Qizhen Chen and Ao Li performed the experiments. Shana Guo reviewed the data and provided guidance on the study; Jing Chen and Yanqiu Wang supervised the study. Qizhen Chen and Ao Li wrote the manuscript. Wen Li has made significant intellectual contributions to the revision of this work, particularly in the data analysis, and also supervised the study. Corresponding authors Ethics declarations Ethics approval and consent to participate This study was approved by the ethics committee of Tongji Hospital affiliated with Tongji University (Ethics approval number: SBKT-2022-105). Clinical trial registration was not applicable as this study did not involve any interventional procedures. Formal written informed consent was obtained from all participants. Consent for publication Not applicable. No personal details that could compromise anonymity are included in the manuscript. 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/. About this article Cite this article Chen, Q., Li, A., Guo, S. et al. Long RNA profiles of endometrial extracellular vesicles provide new insights into the pathogenesis of ovarian endometriosis. BMC Women's Health 26, 246 (2026). https://doi.org/10.1186/s12905-026-04431-0 Received: Accepted: Published: Version of record: DOI: https://doi.org/10.1186/s12905-026-04431-0