Shared and distinct molecular pathways in eutopic endometrium of endometriosis and adenomyosis patients: a transcriptomic study in Chinese women

Total RNA was extracted from the eutopic endometrial tissues using Qiagen RNeasy Mini Kit. RNA concentration, purity, and integrity were assessed using a NanoDrop spectrophotometer and an Agilent 2100 Bioanalyzer. Samples meeting quality criteria (RNA integrity number, RIN ≥ 7.0) were selected for further analysis. RNA sequencing libraries were constructed and sequenced on BGISEQ-2000(BGI, China) according to manufacture’s instruction. Raw sequencing reads were filtered using Cutadapt (v.2.5) [ 22 ] to remove low-quality sequences and adapter contaminants. Clean reads were aligned to the human reference genome (GRCh38) using HISAT2 (v.2.1.0) [ 23 ]. Gene expression levels were quantified using RSEM (v.1.3.3) [ 24 ] and normalized using DESeq2 (v.1.20) [ 25 ].

Comparative analysis identified 115 shared DEGs between EMs-vs-Control and AD-vs-Control (Fig.  4 A). Functional enrichment showed that these genes were involved in 43 BPs related to immune cell proliferation, chemotaxis, and epithelial cell adhesion (padj < 0.05) (Fig.  4 B, Table S13). KEGG annotation highlighted pathways such as IL-17 signaling and chemokine signaling (padj < 0.05) (Fig.  4 C, Table S14). PPI analysis identified hub genes including CD19, CXCR2, and MMP9 among these shared DEGs (Fig. S3). Fig. 4 Shared and distinct differences between EMs-vs-Control and AD-vs-Control. A  Shared and specific DEGs for EMs-vs-Control and AD-vs-Control. B  Functional enrichment of shared DEGs according to biological processes in gene ontology (GO) database. C  Functional enrichment of shared DEGs according to pathways in KEGG database. D  Functional enrichment of EMs-upregulated DEGs when comparing EMs with AD patients directly. The network was generated by default pipeline in STRING database Shared and distinct differences between EMs-vs-Control and AD-vs-Control. A  Shared and specific DEGs for EMs-vs-Control and AD-vs-Control. B  Functional enrichment of shared DEGs according to biological processes in gene ontology (GO) database. C  Functional enrichment of shared DEGs according to pathways in KEGG database. D  Functional enrichment of EMs-upregulated DEGs when comparing EMs with AD patients directly. The network was generated by default pipeline in STRING database Additionally, 255 DEGs were specific to EMs-vs-Control, enriched in immune-related BPs and KEGG pathways such as natural killer cell-mediated cytotoxicity (padj < 0.05) (Table S15-16). PPI analysis identified PTPRC, TNF, ITGAM, PECAM1, CD3E, ITGAL, ITGAX, ZAP70, CD38, and KLRK1 as the top ten EMs-specific DEGs with the highest degree nodes (Fig. S4). For AD-vs-Control, 194 specific DEGs were identified, linked to neutrophil and epithelial cell responses (padj < 0.05) (Table S17-18). PPI analysis identified CXCL8, IL1B, CSF3, FGF20, LTF, MUC5B, SDC1, BPIFA1, BPIFB1, and FOXP3 as the top ten AD-specific DEGs with the highest degree nodes (Fig. S5). Direct comparison between EMs and AD revealed 46 DEGs (23 upregulated in each group) (padj < 0.05, fold change ≥ 1.25) (Fig. S6, Table S19). Functional analysis of EMs-upregulated DEGs showed enrichment in processes such as antibacterial response and mucosal immunity (padj < 0.05), while no significant enrichment was found for AD-upregulated DEGs (Table S20). Specifically, EMs-upregulated DEGs were involved in 8 BPs, including antibacterial humoral response, adhesion of symbiont to host, myoblast migration, ureter development, positive regulation of branching involved in ureteric bud morphogenesis, regulation of branching involved in ureteric bud morphogenesis, cell aggregation, and innate immune response in mucosa (padj < 0.05) (Fig.  4 D).

This study included 65 participants, stratified into three groups based on laparoscopic findings: endometriosis (EMs, n  = 25), adenomyosis (AD, n  = 22), and a Control group ( n  = 18) (Fig.  1 , Table  1 ). Eutopic endometrium samples were collected from all participants for transcriptome analysis. Fig. 1 Study design. GSEA: gene set enrichment analysis; DEGs: differentially-expressed genes; PPI: protein–protein interaction; NES: normalized enrichment score Table 1 Information of 65 participants Control ( n  = 18) EMs ( n  = 25) AD ( n  = 22) P value Age mean(min–max) 30.7(26–45) 31.4(22–43) 38.5(27–47) EvC: P  = 0.7500 AvC: P  = 0.0002 EvA: P  = 0.0003 CA125(U/ml) mean(min–max) 20.2(7.7–41.5) 106.6(12.9–650.2) 171.5(16.5–654.1) EvC: P  = 0.0048 AvC: P  < 0.0001 EvA: P  = 0.0096 CA199(U/ml) mean(min–max) 35.0(2.2–219.1) 76.9(4.9–606.3) 37.9(4.1–152.1) EvC: P  = 0.2800 AvC: P  = 0.3700 EvA: P  = 0.9000 HE4(pmolL) mean(min–max) 32.8(24.7–45.6) 36.6(22.4–68.2) 45.0(31.2–83) EvC: P  = 0.0410 AvC: P  = 0.0002 EvA: P  = 0.0130 AMH(ng/ml) mean(min–max) 4.8(2.59–9.08) 4.0(0.94–9.67) 2.0(0.29–3.25) EvC: P  = 0.3900 AvC: P  = < 0.0001 EvA: P  = 0.0001 Dysmenorrhea  No 10 11 1 EvC: P  = 0.9308  Mild 4 8 3 AvC: P  < 0.0001  Moderate 2 3 0  Severe 2 3 18 EvA: P  < 0.0001 Menstrual flow  Low 1 0 1 EvC: P  = 0.1909  Normal 14 24 5 AvC: P  = 0.0005  High 3 1 16 EvA: P  < 0.0001 Gravidity mean(min–max) 1(0–4) 0.96(0–6) 3.09(0–11) EvC: P  = 0.6400 AvC: P  = 0.0020 EvA: P  = 0.0004 Parity mean(min–max) 0.61(0–3) 0.52(0–4) 1.27(0–3) EvC: P  = 0.6400 AvC: P  = 0.0220 EvA: P  = 0.0025 Endometrial cycle  Proliferative phase 4 9 9 EvC: P  = 0.7170  Secretory phase 14 14 12 AvC: P  = 0.6361  Unknown 0 2 1 EvA: P  = 1 EMs endometriosis, AD adenomyosis, EvC, AvC and EvA in the P value column represents EMs-vs-Control, AD-vs-Control and EMs-vs-AD, respectively. For Age, CA125, CA199, HE4, AMH, Gravidity and Parity, the statistical significance was assessed by Wilcoxon rank sum test with the p value was adjusted via BH correction. For other factors, the statistical significance was assessed via Fisher’s exact test Study design. GSEA: gene set enrichment analysis; DEGs: differentially-expressed genes; PPI: protein–protein interaction; NES: normalized enrichment score Information of 65 participants Age mean(min–max) CA125(U/ml) mean(min–max) CA199(U/ml) mean(min–max) HE4(pmolL) mean(min–max) AMH(ng/ml) mean(min–max) Gravidity mean(min–max) Parity mean(min–max) EMs endometriosis, AD adenomyosis, EvC, AvC and EvA in the P value column represents EMs-vs-Control, AD-vs-Control and EMs-vs-AD, respectively. For Age, CA125, CA199, HE4, AMH, Gravidity and Parity, the statistical significance was assessed by Wilcoxon rank sum test with the p value was adjusted via BH correction. For other factors, the statistical significance was assessed via Fisher’s exact test The age of patients in the AD group was significantly higher than that of the Control and EMs groups (padj < 0.05) (Table  1 ). Serum levels of CA125 and HE4 were markedly elevated in both the EMs and AD groups compared to the Control group, with the highest levels observed in AD patients (padj < 0.05). In contrast, serum AMH levels were significantly lower in the AD group compared to both the Control and EMs groups (padj < 0.05). Dysmenorrhea severity, irregular menstrual flow, gravidity, and parity also differed significantly in the AD group compared to the other two groups (padj < 0.05). However, no significant differences in these clinical features were observed between the Control and EMs groups. There was no statistical differences for the distribution of endometrial cycle phases between Control, EMs and AD groups.

GSEA identified 38 GO BPs distinguishing the AD group from the Control group (padj < 0.05) (Table S6). Among these, five BPs were enriched in AD, including metallo-sulfur cluster assembly and nucleoside bisphosphate metabolism (Fig.  3 A). Only two AD-enriched MFs were identified: phospholipase A1 and olfactory receptor activity (padj < 0.05). Fig. 3 Distinct eutopic endometrium transcriptome files between AD and Control. A  Enriched biological processes assessed by GSEA. B  Enriched KEGG pathways assessed by GSEA. C  Protein–protein interactions of DEGs (r ≥ 0.3) based on aligning with STRING database. The network was generated by default pipeline in STRING database Distinct eutopic endometrium transcriptome files between AD and Control. A  Enriched biological processes assessed by GSEA. B  Enriched KEGG pathways assessed by GSEA. C  Protein–protein interactions of DEGs (r ≥ 0.3) based on aligning with STRING database. The network was generated by default pipeline in STRING database Conversely, 33 Control-enriched BPs were identified, predominantly related to innate and adaptive immunity. The top processes included neutrophil chemotaxis, humoral immune response, and granulocyte migration (Fig.  3 A, Table S6). KEGG pathway analysis revealed seven AD-enriched pathways, such as arachidonic acid metabolism and ovarian steroidogenesis, while the Control group showed enrichment in 23 pathways, including TNF signaling, IL-17 signaling, and cytokine-cytokine receptor interaction (padj < 0.05) (Fig.  3 B, Table S7). DESeq2 analysis identified 309 DEGs between AD and Control, with 73 upregulated in AD and 236 upregulated in Control (padj < 0.05, fold change ≥ 1.25) (Fig. S2, Table S8). Control-upregulated DEGs were associated with immune-related BPs, such as neutrophil migration and T cell responses, while AD-upregulated DEGs were linked to the Wnt signaling pathway and epithelial cell proliferation (padj < 0.05) (Table S9-10). KEGG annotation indicated significant enrichment of Control-upregulated DEGs in pathways such as cytokine-cytokine receptor interaction and IL-17 signaling (padj < 0.05) (Table S11). In contrast, no significant enrichment of AD-upregulated DEGs was found according to KEGG annotation. PPI analysis identified the top 15 hub genes including IL1B, CXCL8, and CD19 (Fig.  3 C). Functional clustering revealed key networks related to neutrophil chemotaxis (10 genes) and B cell receptor signaling (5 genes) (Table S12).

This study recruited 98 female patients who received laparoscopic surgery due to vaginal irregular bleeding at Department of Obstetrics and Gynecology, Peking University Shenzhen Hospital. Among those attenders, qualified eutopic endometrium of 43 were included following pathology diagnosis, being stratified into two groups: EMs (n = 25), and benign uterine fibroids or ovarian cysts (Control, n = 18). Another 22 AD patients were included and eutopic endometrium were collected during surgeries. Patient recruitment and qualified sample collection was performed from June 18, 2023 to March 3, 2025. Patients with previous hormonal therapy, malignancies, or severe systemic diseases were excluded. Clinical characteristics, including age, serum levels of CA125, HE4, anti-Müllerian hormone (AMH), dysmenorrhea severity, menstrual regularity, gravidity, parity, and menstrual cycle phases at the time of sampling were recorded. Eutopic endometrial tissue samples were collected during laparoscopic surgery (for EMs and Control group) or surgical therapies (for AD group). Immediately after collection, samples were snap-frozen in liquid nitrogen and stored at −80 °C until RNA extraction.

Overall, this study provides additional population-based evidence for the molecular mechanisms underlying EMs and AD by analyzing gene expression profiles in the eutopic endometrium. The findings reveal distinct immune, metabolic, and epithelial remodeling pathways in each condition, alongside shared inflammatory processes, reflecting their overlapping but unique pathophysiology. These results enhance our understanding of the molecular basis of these disorders and identify potential biomarkers and therapeutic targets for future research. However, further studies addressing the limitations of this study are required to validate and expand these findings. Ultimately, this research contributes to the understanding of aberrant eutopic endometrium gene expression in EMs and AD, especially in Chinese populations.

EMs and AD are chronic gynecological disorders that impose severe health and economic burdens on millions of women worldwide. EMs is characterized by the presence of endometrial-like tissue outside the uterine cavity, while AD involves the invasion of endometrial tissue into the myometrium. Despite overlapping clinical features, their pathophysiology and molecular mechanisms differ in key ways [ 10 , 11 ]. Recent studies have emphasized the importance of analyzing eutopic endometrial gene expression to uncover the molecular basis of these diseases, as the eutopic endometrium reflects both systemic and local disease-related changes [ 6 – 9 ]. By identifying shared and distinct dysregulated pathways between EMs and AD, this study provides noval insights into their molecular pathogenesis especially in Chinese populations. These findings also provide important implications for diagnosis and treatment of EMs and AD. Our analysis highlight profound immune dysregulation in both EMs and AD compared to controls, consistent with prior studies demonstrating the central role of immune system dysfunction in these diseases [ 6 , 7 , 9 , 12 – 14 , 28 – 30 ]. In the EMs group, we observed a reduction in control-enriched pathways related to NK cell-mediated cytotoxicity, B cell receptor signaling, and cytokine-cytokine receptor interactions. These findings align with previous studies indicating that EMs is associated with impaired immune surveillance and chronic inflammation, which allow ectopic endometrial tissue to evade immune clearance [ 15 , 16 , 29 – 33 ]. These findings partially support the potential use of immunomodulatory therapies, such as checkpoint inhibitors or cytokine-targeting biologics, to restore immune function and prevent disease progression. In contrast, AD was associated with reduced neutrophil chemotaxis and humoral immune responses, indicating a distinct immune microenvironment that may contribute to tissue invasion and fibrosis [ 19 – 21 , 28 , 34 ]. This could explain why AD patients often exhibit refractory pelvic pain and poor response to hormonal therapies. Targeting neutrophil recruitment pathways (e.g., CXCR2 inhibitors) or enhancing humoral immunity may offer new therapeutic avenues for AD. Another distinct feature of EMs identified in this study was the upregulation of metabolic pathways, including oxidative phosphorylation, ribosome biogenesis, and ATP synthesis-coupled electron transport. These results are consistent with previous reports that EMs exhibits heightened metabolic activity and mitochondrial dysfunction [ 18 , 33 , 35 – 37 ]. Oxidative stress has been implicated in the pathogenesis of EMs, where increased mitochondrial activity leads to the production of reactive oxygen species (ROS), promoting inflammation, angiogenesis, and tissue remodeling [ 17 , 32 , 35 , 36 , 38 ]. Additionally, the enrichment of arachidonic acid metabolism in the EMs group reinforces the role of prostaglandins in mediating pain and inflammation in EMs [ 18 ]. Collectively, these findings highlight metabolic reprogramming of eutopic endometrium as drivers for EMs, and suggest that targeting these pathways may provide novel therapeutic opportunities. In AD, our results revealed significant enrichment of pathways related to epithelial proliferation, stromal remodeling, and hormonal regulation. Specifically, Wnt signaling and metallo-sulfur cluster assembly were upregulated in the AD group, consistent with the invasive and fibrotic features of AD [ 28 ]. Wnt signaling has been implicated in the epithelial-mesenchymal transition (EMT), a process that facilitates the invasion of endometrial glands into the myometrium [ 39 , 40 ]. Furthermore, the enrichment of ovarian steroidogenesis pathways in adenomyosis highlights the role of estrogen-driven proliferation and stromal changes. These findings are partially consistent with previous studies demonstrating that AD is characterized by hormone-dependent tissue remodeling and fibrosis [ 9 , 28 , 41 ], which contributes to AD incidence. Further comparative analysis of DEGs between EMs and AD groups revealed 115 shared genes when compared to controls, highlighting common inflammatory processes such as immune cell proliferation, chemotaxis, and epithelial adhesion. These shared pathways suggest a common inflammatory basis for both conditions, driven by dysregulated cytokine and chemokine signaling [ 6 , 7 , 9 , 12 – 14 ]. Meanwhile, the disease-specific signatures—antibacterial response in EMs and neutrophil activation in AD—may enable non-invasive differential diagnosis (e.g., via endometrial biopsy or liquid biopsy), reducing reliance on invasive imaging or surgery [ 28 , 34 , 42 , 43 ]. These findings emphasize the importance of identifying both shared and specific molecular signatures to better understand the distinct clinical manifestations and molecular mechanisms of EMs and AD. Despite the significant insights gained from this study, several limitations should be acknowledged. First, the relatively small sample size (n = 65) may reduce the strength of the analysis and limit the generalizability of our findings. Nevertheless, this should be mitigated by the research strategies, including surgery and diagnosis by the same physicians as well as strict statistic standard to screen DEGs. Larger cohort studies across diverse populations are needed to validate these results. Second, this study exclusively utilized transcriptomic data, which does not capture post-transcriptional or proteomic changes. Integrating multi-omics approaches, including proteomics and metabolomics, could provide a more comprehensive understanding of disease mechanisms. Third, the cross-sectional design precludes an assessment of temporal changes in gene expression, which are critical for understanding disease progression. Finally, the lack of functional validation of the identified DEGs and pathways limits the translational potential of this study. Experimental validation is necessary to confirm the functional roles of these genes and pathways in disease pathogenesis.

Statistical analysis were performed using R statistical software (version 4.0.5). Continuous clinical variables were compared among groups using wilcoxon rank-sum test. Categorical variables were compared using the Fisher's exact test. For both continuous clinical variables and DEGs, p value was corrected using the Benjamini–Hochberg procedure to obtain adjusted p-values (padj). The statistical significance was defined as padj < 0.05.

Endometriosis (EMs) and adenomyosis (AD) represent significant gynecological disorders affecting women's reproductive health, and had distinct epidemiology between developing and developed countries [ 1 , 2 ]. EMs is characterized by the presence of ectopic endometrium outside the uterus, affecting approximately 6–10% of reproductive-aged women worldwide. AD involves the invasion of endometrial tissue into the myometrium. EMs and AD share common symptoms including chronic pelvic pain, infertility, and dysmenorrhea. Current diagnosis relies heavily on clinical suspicion, imaging (ultrasound/MRI), and invasive confirmation (laparoscopy for endometriosis; hysterectomy for adenomyosis). Treatment strategies for EMs include hormonal suppression (e.g., progestins, GnRH analogs) for symptom control and laparoscopic excision for definitive management [ 3 ], though recurrence rates are high and surgical risks persist, particularly for deep infiltrating diseases. AD management similarly depends on hormonal therapy or hysterectomy, with uterine-sparing techniques limited by technical complexity and recurrence. Both conditions lack disease-modifying therapies, and fertility preservation remains a significant challenge. Advances in precision medicine and molecular profiling hold promise for future targeted interventions. However, the molecular pathogenesis remains poorly understood, limiting the development of effective diagnostic and therapeutic strategies. Emerging evidence suggests that aberrant gene expression in eutopic endometrium plays a critical role in disease initiation and progression [ 4 , 5 ]. Current fragmented knowledge necessitates further investigation into gene expression profiles and pathways involved in the eutopic endometrium of affected individuals, especially across different populations. Recent transcriptomic studies of eutopic endometrium have provided valuable insights into the molecular mechanisms underlying EMs and AD [ 6 – 9 ]. Comparative analyses have identified both shared and distinct molecular signatures between the two disorders, suggesting overlapping yet disease-specific molecular pathogenesis [ 10 , 11 ]. Several research teams have highlighted the dysregulation of immune-related pathways, including cytokine and chemokine signaling, in the eutopic endometrium of both EMs and AD patients [ 6 , 7 , 9 , 12 – 14 ]. Key pathways such as natural killer cell-mediated cytotoxicity, oxidative phosphorylation, and arachidonic acid metabolism have been emphasized in EMs [ 15 – 18 ], while epithelial remodeling and Wnt signaling are prominent in AD [ 19 – 21 ]. Despite those progress, knowledge gaps remain in the current literature, including limited evidence from diverse populations. This study aims to characterize the gene expression profiles of eutopic endometrium in Chinese patients with EMs and AD compared to controls. Through comprehensive transcriptomic analysis, we seek to identify shared and distinct aberrant gene expression patterns between EMs and AD patients. Our study identifies novel transcriptomic signatures in eutopic endometrium, elucidating shared (e.g., immune dysregulation) and distinct pathways between EMs and AD in Chinese patients, complementing prior findings in other cohorts.

Differentially expressed genes (DEGs) were identified using DESeq2 with correction of age due to its significant differences across groups. The standard of padj < 0.05 and |log 2 (foldchange)|≥ 1 was defined as statistical significance for DEGs. This represented ≥ twofold change of gene expression in EMs-vs-Control, AD-vs-Control and EMs-vs-AD. By applying R package ClusterProfiler [ 26 ], Gene Set Enrichment Analysis (GSEA) and functional enrichment analysis of DEGs were conducted to identify biological processes (BP), molecular functions (MF), and cellular components (CC) from Gene Ontology (GO) categories, as well as Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways. Protein–protein interaction networks were constructed using the STRING database [ 27 ].

Supplementary Material file 1. Supplementary Material file 2. Supplementary Material file 3. Supplementary Material file 4. Supplementary Material file 5. Supplementary Material file 6. Supplementary Material file 7. Supplementary Material file 8. Supplementary Material file 9. Supplementary Material file 10. Supplementary Material file 11. Supplementary Material file 12. Supplementary Material file 13. Supplementary Material file 14. Supplementary Material file 15. Supplementary Material file 16. Supplementary Material file 17. Supplementary Material file 18. Supplementary Material file 19. Supplementary Material file 20. Supplementary Material file 21. Supplementary Material file 22. Supplementary Material file 23. Supplementary Material file 24. Supplementary Material file 25. Supplementary Material file 26. Supplementary Material file 1. Supplementary Material file 2. Supplementary Material file 3. Supplementary Material file 4. Supplementary Material file 5. Supplementary Material file 6. Supplementary Material file 7. Supplementary Material file 8. Supplementary Material file 9. Supplementary Material file 10. Supplementary Material file 11. Supplementary Material file 12. Supplementary Material file 13. Supplementary Material file 14. Supplementary Material file 15. Supplementary Material file 16. Supplementary Material file 17. Supplementary Material file 18. Supplementary Material file 19. Supplementary Material file 20. Supplementary Material file 21. Supplementary Material file 22. Supplementary Material file 23. Supplementary Material file 24. Supplementary Material file 25. Supplementary Material file 26.