{"paper_id":"ae160c76-aa2a-437c-9f44-b7ae7e47d55a","body_text":"ARTICLE | Article in Press\nScientific Reports https://doi.org/10.1038/s41598-026-70076-6\nScientific Reports\nThis Article in Press is shared early to give you faster access to new research. It is citable and carries a permanent DOI. \nThe final edited version will replace it automatically.\nXiulan Weng and Shunhe Lin indicates co-first authors.\nRotenone targeting the oxidative phosphorylation \npathway improves invasive lesions in ovarian \nendometriosis\nXiulan Weng · Shunhe Lin · Zhenna Wang · Chaobin Liu · Guan Lin · Pengming Sun · \nJingsong Yi\nReceived: 8 April 2026 / Accepted: 31 August 2026\n© The Author(s) 2026\nAbstract\nEndometriosis (EM) is a prevalent gynecological disorder characterized by diagnostic difficulty and high recur -\nrence rates. In this study, single-cell RNA sequencing (scRNA-seq) was performed to characterize endome -\ntrial stromal cell (ESC) subpopulations in EM, investigate the involvement of the oxidative phosphoryla -\ntion (OXPHOS) pathway in EM pathogenesis and fibrosis, and assess the therapeutic effects of Rotenone, a \nmitochondrial electron transport chain complex I inhibitor. Twelve endometrial single-cell samples from the \nGSE179640 dataset were analyzed to compare cellular composition between EM and control groups, and to \nidentify EM-associated pathways using Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis. The \nrole of OXPHOS was further validated using clinical samples, in vitro assays, and in vivo animal models, with \nevaluation of histopathology, fibrosis, and mitochondrial function. scRNA-seq analysis demonstrated increased \nproportions of stromal, lymphoid, and myeloid cells in EM. Subcluster analysis of ESC subclusters revealed \nenrichment of the eStromal_cycling population, with differentially expressed genes significantly associated \nwith the OXPHOS pathway. Ectopic EM lesions exhibited disrupted tissue architecture, increased collagen \ndeposition, elevated mitochondrial complex I activity and adenosinetriphosphate (ATP) levels. OXPHOS-\nrelated proteins were co-localized with Vimentin in ectopic tissues. In vitro, Rotenone suppressed proliferation, \nmigration, and invasion of ectopic endometrial stromal cells (ecESCs), and reduced complex I activity, ATP \nproduction, and OXPHOS protein expression. In vivo, Rotenone treatment reduced ovarian cystic lesions and \nadhesions, decreased fibrosis area, restored estradiol and progesterone levels, and reduced the co-localization \nof OXPHOS and stromal markers. These findings indicate that activation of the OXPHOS pathway in stromal \ncells is associated with disease progression and fibrosis. Inhibition of mitochondrial complex I attenuated \nlesion development and fibrosis in experimental models, suggesting that metabolic modulation may represent \na potential therapeutic approach for EM.\nKeywords Ovarian endometriosis · Single-cell analysis · Oxidative phosphorylation · ESC · Rotenone\n\nArticle in Press\nScientific Reports\nhttps://doi.org/10.1038/s41598-026-70076-6\nIntroduction\nEndometriosis (EM) is an estrogen-dependent chronic inflammatory disorder characterized by the presence of \nectopic endometrial-like tissue. It affects approximately 10–15% of women of reproductive age and is associated \nwith 30–50% of infertility cases 1–3. The widely accepted retrograde menstruation theory proposes that refluxed \nendometrial stromal cells (ESC) adhere to ectopic sites, invade surrounding tissues, and promote angiogen -\nesis, thereby contributing to lesion formation and fibrosis4. Several pathogenic mechanisms have been proposed, \nincluding retrograde menstruation, coelomic metaplasia, and immune dysfunction, reflecting the multifacto -\nrial nature of EM5–9. Increasing evidence indicates that EM pathogenesis involves complex interactions among \ngenetic, hormonal, immune, inflammatory, angiogenic, and metabolic factors. A better understanding of these \nmechanisms may facilitate the identification of diagnostic biomarkers and therapeutic targets.\nRecent advances in single-cell transcriptomics (scRNA-seq) have enabled the dissection of cellular heteroge -\nneity and functional states within EM lesions at single-cell resolution. ESC are not a homogeneous population; \nrather, they comprise quiescent, activated, proliferative, and secretory subpopulations, each contributing differ -\nently to disease progression10. However, largely large-scale scRNA-seq studies of EM have primarily focused on \nthe immune microenvironment or epithelial-mesenchymal transition11, with limited systematic exploration of the \nmetabolic characteristics of stromal subpopulations and their direct contribution to fibrosis.\nWhile metabolic reprogramming in endometriosis has attracted increasing attention, most existing studies have \nrelied on bulk tissue metabolomics or have described generalized mitochondrial dysfunction without resolving \ncell-type specificity or establishing functional causality. For example, Hawkins et al., identified reduced levels of \nmalic acid and flavin adenine dinucleotide (FAD) in endometriotic tissue from a nonhuman primate model using \nbulk metabolomics; however, these alterations could not be attributed to specific cell populations 12. Similarly, \nChen et al., observed swollen mitochondria with disrupted cristae in adenomyotic stromal cells by transmission \nelectron microscopy (TEM), but did not determine whether these structural abnormalities contributed to invasive \nbehavior or were secondary to inflammatory processes 13. Consequently, it remains unclear whether OXPHOS \nhyperactivation in a specific cell type functionally drives EM progression and whether this pathway represents a \nviable therapeutic target.\nOxidative phosphorylation (OXPHOS) coordinates energy supply with biological processes processes such as \ncollagen deposition and cell migration through the regulation of oxygen species (ROS) production and metabolic \nintermediates14,15. In EM, ESC exhibit tumor-like invasive and migratory properties, suggesting that OXPHOS \nremodeling may drive aberrant ESC proliferation, migration, and fibrosis activity. Rotenone, a reversible inhibi-\ntor of mitochondrial complex I, suppresses OXPHOS and reduces ATP and ROS production, and has been widely \nused in studies of neurodegenerative disorders and tumor metabolism 16,17. Therefore, If ESC within EM lesion \ndepend on enhanced OXPHOS activity to maintain their invasive phenotype, targeting mitochondrial complex I \nmay disrupt the metabolic support required for lesion establishment and fibrogenesis, potentially complementing \nhormonal therapy, surgical resection, and anti-inflammatory strategies to prevent recurrence.\nHere, we integrated publicly available scRNA-seq data (GSE179640) with clinical samples to characterize \nstromal subpopulations in EM and to identifiy associations between OXPHOS activation and disease phenotypes. \nWe further employed primary ESC isolates from eutopic and ectopic tissues to evaluate the effects of Rotenone \non cell proliferation, migration, and mitochondrial function. In addition, we established a mouse ovarian EM \nmodel to assess the impact of Rotenone impact on lesion burden, fibrosis, and circulating sex hormone levels. \nThe work aimed to elucidate the role of OXPHOS in EM progression and to explore the potential of metabolism-\ntargeted therapeutic strategies.\n\n\nArticle in Press\nScientific Reports\nhttps://doi.org/10.1038/s41598-026-70076-6\nMaterials and methods\nDate mining\nscRNA-seq data were retrieved from the Gene Expression Omnibus (GEO) database  (   h t t p s : / / w w w . n c b i . n l m . n i \nh . g o v / g e o     ) . The dataset GSE179640, generated using the GPL24676 platform and submitted by Tan et al., was \nselected for analysis. GSE179640 contains 59 samples, including 31 single-cell sequencing samples, 24 bulk-seq \nsamples, and 4 organoid-type samples 10. Bulk-seq and organoid samples were excluded to maintain single-cell \nresolution and primary tissue relevance. For single-cell samples, those with < 500 cells, > 20% mitochondrial \ngene expression, or > 10% doublet rate were excluded. After applying exclusion criteria, 12 samples remained \nfor analysis: 3 controls and 9 endometriosis samples (Table S1).\nData preprocessing and principal component analysis\nSingle-cell transcriptomic analysis was performed using Seurat (v5.0.0), and batch effects were corrected using \nHarmony (v1.2.4). The workflow included quality control filtering, data normalization, and feature selection. \nGraph-based clustering was conducted to identify distinct cell populations, followed by cell type annotation \nand evaluation of cellular composition. Principal component analysis (PCA) was used for initial dimensionality \nreduction, and visualization was performed using t-distributed stochastic neighbor embedding (t-SNE) and uni -\nform manifold approximation and projection (UMAP).\nIdentification of differentially expressed genes\nDifferential gene expression analysis was conducted on key subtypes using the R package limma. The criterion \nfor differential expression was |log 2FoldChange| > 1. The Benjamini-Hochberg method was applied to adjust \nP-values (FDR), with a threshold of adjusted P < 0.05 to define differentially expressed genes (DEGs). \nKEGG pathway analysis\nKyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis of DEGs was performed using the \nclusterProfiler (v4.16.0) in R. Statistical significance was defined as an adjusted P value < 0.05. Enrichment \nresults were visualized using bar plots generated with the ggplot2 package (v4.0.0).\nClinical sample collection\nPaired eutopic and ectopic endometrial tissue samples were collected from Fujian Maternal and Child Health \nHospital, including 14 eutopic EM tissues (controls) and 14 ectopic EM tissues. All samples were obtained dur-\ning the proliferative phase of the menstrual cycle. The patients were aged 25–40 years, with an average age of \n35.1 ± 10.4 years. All procedures involving human participants were conducted in accordance with the Declara -\ntion of Helsinki and relevant institutional guidelines. All patients procedures approved by the Ethics Committee \nof Fujian Maternal and Child Health Hospital (approval number: 2025KY253). All participants provided written \ninformed consent prior to participation.\nCell isolation and culture\nPostoperative eutopic and ectopic endometrial tissues were processed to obtain a single-cell suspension. Tis -\nsues were washed with PBS, transferred on ice, and digested with a pre-chilled solution containing Collagenase \nType IV (1 mg/mL; Solarbio, Cat# C8160) and DNase I (0.2 mg/mL) in RPMI-1640 (digestion solution; Procell, \n\n\nArticle in Press\nScientific Reports\nhttps://doi.org/10.1038/s41598-026-70076-6\nCat# PM150110) for 60 min at 37 °C with gentle shaking (220 rpm). The digest was passed through a 70-µm \ncell strainer (Bkmam, Cat# 110426002), and the filtrate washed with PBS. The suspension was subjected to red \nblood cell lysis in RBC lysis buffer (Solarbio, Cat# R1011), then centrifuged at 400 ×g for 5 min at 4 °C. The cell \npellet was resuspended in PBS, counted, and plated in DMEM complete medium (10 mL per dish, Procell, Cat# \nPM150210) and incubated at 37 °C with 5% CO 2; medium was changed after 24 h. Stromal cells were identi -\nfied by immunofluorescence for Vimentin (Proteintech, Cat# 10366-1-AP) and Cytokeratin 7 (Proteintech, Cat# \n17513-1-AP). For passaging, cells were trypsinized with 0.25% Trypsin-EDTA (0.25% Trypsin, Beyotime, Cat# \nC0201) for 1–3 min, and subculture ratios were 1:1 or 1:2 for the first passage and 1:3 for the second. Experi -\nments used cells from the second passage onward.\nCell treatments\nCells in the logarithmic growth phase were harvested, prepared as single-cell suspensions and counted. For the \nCCK-8 assay, 2 × 10^3 cells per well were seeded into 96-well plates with three technical replicates per condition. \nAfter overnight attachment, cells were treated with increasing concentrations of Rotenone (0–10 µM; MCE, Cat# \nHY-B1756) for 48 h. Following treatment, remove medium and add 100 µL of CCK-8 solution per well; incu -\nbated at 37 °C for 3 h. Absorbance was measured at 450 nm using a microplate reader. Dose-response curves were \ngenerated to determine the half-maximal inhibitory concentration (IC50). For subsequent experiments, cells were \ndivided into four groups: inESC (Intact endometrial stromal cells), inESC + Rotenone, ecESC (Ectopic endome-\ntrial stromal cells), ecESC + Rotenone. After 48 h, collect cells for downstream analyses.\nCCK-8 assay\nFor the four treatment groups, 2 × 10^3 cells per well were seeded into 96-well plate with three replicates per \ngroup. Cell viability was assessed using a CCK-8 assay kit (Beyotime, Cat# C0038) according to the manufac -\nturer’s instructions. Specifically, after 48 h of treatment, the culture medium was removed, add 10 µL of CCK-8 \nsolution per well, incubated at 37 °C for 3 h, and absorbance was measured at 450 nm using a microplate reader.\nEdU cell proliferation assay\nFour groups (n = 3 per group) were seeded at a density of 3 × 10^5 cells per well in a 6-well plate. After 48 h of \ntreatment, proliferation was evaluated using the BeyoClick EdU Cell Proliferation Kit (Beyotime, Cat# C0071S) \naccording to the manufacturer’s instructions. Briefly, EdU was added to a final concentration of 10 µM, and \nincubated 2 h at 37 °C. Cells were then fixed, permeabilized, and subjected to click chemistry labeling, nuclear \ncounterstaining, and observed and imaged under a fluorescence microscope.\nTranswell assay\nAfter 48 h of Rotenone treatment, cells were detached using 1 mL Trypsin-EDTA solution for 1–2 min, collected \nand counted. The cell suspension was adjust to 2.5 × 10^5 cells/mL. For the Transwell assay, 600 µL of culture \nmedium containing 10% FBS was added to the lower chamber of a 24-well plate. Transwell inserts (with or \nwithout Matrigel; Mogengel, Cat# 0827245) were placed into the wells, and added 200 µL cell suspension to the \nupper chamber. Incubate at 37 °C, 5% CO2 for 24 h. Removed inserts, wiped the interior to remove cells from the \nupper surface. Cells on the lower surface were fixed with 600 µL formaldehyde for 20 min. Stained with 0.1% \ncrystal violet (Beyotime, Cat# C0121) for 15 min. The inserts were rinsed three times with PBS and air-dried. \nMigrated or invaded cells were observed under a light microscope, and images were captured.\n\n\nArticle in Press\nScientific Reports\nhttps://doi.org/10.1038/s41598-026-70076-6\nMitoTracker Red CMXRos assay\nFour groups (n = 3 per group) were seeded at a density of 3 × 10^5 cells per well in a 6-well plate. MitoTracker \nRed CMXRos (200 nM, Beyotime, Cat# C1035) was added and incubated at 37 °C for 15–30 min. After stain-\ning, the dye was removed and replaced with fresh pre-warmed medium. Cells were observed and imaged under a \nfluorescence microscopy. Subsequently, cells were fixed with PBS containing 3.7% formaldehyde for 15–30 min, \nwashed three times with PBS containing 3% BSA, permeabilized with PBS containing 0.3% Triton X-100 for \n15 min. Nuclei were counterstained with Hoechst 33,342 (1×) for 10–15 min. Fluorescence images were then \nacquired using a fluorescence microscope.\nProtein extraction and western blotting\nCells from Four groups ( n = 3 per group) were digested with 1 mL Trypsin-EDTA Solution (Beyotime, Cat# \nC0201) for 1–2 min, collected, and lysed in RIPA buffer (Beyotime, Cat# P0013B) to extract total protein. \nProtein extracts were mixed with 5 × SDS loading buffer (Sangon, Cat# C516031) and denatured at 100 °C for \n10 min. Samples were separated by SDS-PAGE and transferred onto NC membrane (Merck, Cat# HATF00010). \nMembranes were blocked with 5% BSA (Solarbio, Cat# SW3015) for 2 h, then incubated overnight at 4 °C with \nprimary antibodies against PPA1 (Proteintech, Cat# 14985-1-AP), ATP5PO (Proteintech, Cat# 10994-1-AP), \nNDUFB5 (Proteintech, Cat# 23855-1-AP), COX6C (Proteintech, Cat# 11429-2-AP), and GAPDH (1:5000; Pro-\nteintech, Cat# 10494-1-AP). After washing, membranes were incubated with HRP-conjugated secondary anti -\nbodies (1:10,000; Proteintech, Cat# SA00001-1, SA00001-2) for 2 h at 37 °C. Protein bands were visualized \nusing enhanced chemiluminescence (ECL) and detected with a chemiluminescence imaging system (Service -\nbio, Cat# SCG-W2000). Densitometric analysis of target proteins normalized to GAPDH was performed using \nImageJ software.\nAnimal experiments\nThirty-two female SPF C57BL/6J mice, 5–6 weeks old, were purchased from SPEF (Suzhou) Biotechnology \nCo., Ltd. All animals were housed under specific pathogen-free (SPF) conditions (25 ± 1 °C; 40–60% relative \nhumidity; 12 h light/dark cycle) with free access to standard chow and water. Mice were acclimated for 1 week. \nAll animal experiments were conducted in accordance with relevant guidelines and regulations. The proce -\ndures were approved by the Ethics Committee of Fujian Maternal and Child Health Hospital (approval number: \nAECSFY2025080).\nEight mice were used as donors, and the remaining 24 mice were randomly assigned to four groups (n = 8 per \ngroup): Sham group, ovarian endometriosis group (OE group), and OE with Rotenone group (OE + Rotenone \ngroup). The OE model was established by allogeneic uterine tissue transplantation as previously described 18. \nSpecifically, the donor-to-recipient ratio was 1:2. Prior to modeling, donor and recipient mice were gavaged with \n1 mg/kg estradiol valerate for two consecutive days and fasted for 12 h. Under deep anesthesia induced with 3% \nisoflurane (Ruiwode Life Science Technology Co., Ltd., Cat# 26675-46-7), donor mice were euthanized by cer-\nvical dislocation. Uteri were excised, trimmed of surrounding fat and connective tissue, rinsed in PBS, minced \ninto approximately ~ 0.5 mm fragments, and digested in Collagenase Type IV (1 mg/mL) at 37 °C for 30 min. \nThe digestion mixture was centrifuged at 8000 ×g for 5 min; the pellet was resuspended in 100 µL PBS and \ncentrifuged again to remove residual collagenase. Tissue fragments were used immediately for transplantation.\nFor OE modeling, recipient mice were anesthetized with 3% isoflurane, a 2–3 cm ventral midline laparotomy \nwas performed to expose the ovaries. The bilateral ovarian bursa were gently distended with 100 µL PBS to \nfacilitate identification of the membranes, and the ovarian capsule were opened. Half of a donor uterus tissue \nfragment was placed onto each ovary in a uniform manner. The ovaries were then returned to the peritoneal \ncavity, and the abdominal wall was closed in layers using 3 − 0 Vicryl sutures. The incision site was disinfected \n\n\nArticle in Press\nScientific Reports\nhttps://doi.org/10.1038/s41598-026-70076-6\nwith povidone-iodine. Mice were allowed to recover on a heating pad and returned to their cages after regain -\ning consciousness. Beginning on the day of surgery (day 0), OE mice were gavaged daily with estradiol valerate \n(1 mg/kg) for 5 consecutive days to promote ectopic lesion growth; On postoperative day 3, mice received an \nintraperitoneal injection of penicillin (8 IU in 0.2 mL) to prevent infection. The Sham group underwent identical \nanesthesia and laparotomy procedures, but no uterine tissue was transplanted; instead, PBS was injected into the \nperitoneal cavity. Starting on postoperative day 2, mice in the OE + Rotenone group were administered 0.5 mg/\nkg Rotenone by oral gavaged once daily for 6 weeks19.\nBlood and ovarian tissue collection\nAfter 6 weeks of treatment, mice were deeply anesthetized with 3% isoflurane. Blood was collected from the \nabdominal aorta and kept on ice for 1 h to allow clotting, followed by centrifuged at 4000 ×g for 20 min at 4 °C. \nSerum was separated and stored at -80 °C until further analysis. Mice were subsequently euthanized by cervical \ndislocation under deep anesthesia. The abdominal cavity was opened, and the reproductive organs were carefully \nexposed. Ovaries were examined for cyst formation and adhesions, then harvested. One ovary from each mouse \nwas fixed in 4% paraformaldehyde for histological analysis, while the contralateral ovary was stored at -80 °C \nfor subsequent experiments.\nH&E staining\nTissues (clinical endometrium tissues and animal ovaries) were processed for paraffin embedding, sectioned, \nand deparaffinized through graded xylene and graded ethanol. Sections were stained with hematoxylin and eosin \n(H&E) for 3–5 min, washed in running water, differentiated, blue, and washed again. Slides were counterstained \nwith eosin for 5 min, then dehydrated through graded ethanol and xylene, cleared, and mounted with neutral gum. \nImages were acquired and analyzed.\nMasson’s trichrome staining\nMasson’s trichrome stain kit (Servicebio, Cat# G1006) was used to analyze collagen fiber deposition in clinical \nendometrial tissues and mouse ovarian tissues according to the manufacturer’s instructions. In brief, paraffin-\nembedded sections were deparaffinized and rehydrated to distilled water, then immersed in Masson A solution \novernight and rinsed with running water. Subsequently, sections were immersed in a mixed solution of Masson \nB and Masson C (1:1 ratio) for 1 min, rinsed with running water, differentiated in 1% hydrochloric acid in etha-\nnol and rinsed again. Sections were then stained with Masson D solution for 6 min, rinsed, and counterstained \nwith Masson E solution for 1 min. After brief blotting, sections were directly stained with Masson F solution for \n2–30 s. Sections were differentiated in 1% acetic acid, dehydrated in ethanol, cleared in xylene, and mounted with \nneutral resin. Images were captured under a light microscope for subsequent analysis.\nImmunofluorescence\nParaffin-embedded tissues from different groups (clinical endometrium and animal ovaries) were sectioned at \n4–6 μm and mounted on poly-L-lysine-coated slides. After deparaffinization in xylene and rehydration through a \ngraded ethanol series, antigen retrieval was performed in 10 mM sodium citrate buffer (pH 6.0) at 95–100 °C for \n10 min. After cooling, sections were washed three times with PBS and blocked with 1% BSA for 30 min. Sections \nwere then incubated overnight at 4 °C with primary antibodies against PPA1 (1:200; Proteintech, Cat# 14985-\n1-AP), ATP5PO (1:200; Proteintech, Cat# 10994-1-AP), NDUFB5 (1:200; Proteintech, Cat# 23855-1-AP), \nCOX6C (1:200; Proteintech, Cat# 11429-2-AP ), and Vimentin (1:200; Proteintech, Cat# 10366-1-AP). After \nwashing with PBS, sections were incubated with HRP-conjugated secondary antibodies (1:10000; Proteintech, \n\n\nArticle in Press\nScientific Reports\nhttps://doi.org/10.1038/s41598-026-70076-6\nCat# SA00001-1, SA00001-2) for 1 h at 37 °C. Nuclei were counterstained with 1 µg/mL DAPI (Beyotime, Cat# \nC1006) for 5–10 min. Sections were mounted and imaged for analysis.\nBiochemical and ELISA assay\nMitochondrial respiratory chain complex I activity and ATP content were measured in clinical samples, four \ngroups of treated cells, and serum samples from three animal groups using the Mitochondrial Complex I/NADH-\nCoQ Reductase Activity Assay Kit (Solarbio, Cat# BC0515) and the ATP Content Assay Kit (Jonlnbio, Cat# \nJL-T0633), respectively, according to the manufacturer’s instructions.\nSerum concentrations of follicle-stimulating hormone (FSH), luteinizing hormone (LH), estradiol (E2), and \nprogesterone (PROG) were measured using commercial ELISA kits: Mouse FSH ELISA Kit (Jonlnbio, Cat# \nJL10239), Mouse LHR ELISA Kit (Jonlnbio, Cat# JL51835), Mouse Estradiol (E2) ELISA Kit (Jonlnbio, Cat# \nJL11790), and Mouse Progesterone (PROG) ELISA Kit (Jonlnbio, Cat# JL20678), following the manufacturer’s \ninstructions.\nStatistical analyses\nAll data are presented as mean ± standard deviation (SD). Normality of data distribution was assessed using the \nShapiro-Wilk test. Differences between two groups were analyzed by t-tests, and multiple-group comparisons \nwere conducted using one-way ANOV A followed by Tukey’s multiple comparisons test (GraphPad, La Jolla, CA, \nUSA). Statistical significance was set at P < 0.05. Denotations: * P < 0.05, ** P < 0.01, *** P < 0.001.\nResult\nSingle-cell transcriptomic analysis of endometrial cell subpopulations\nTo focus on stromal cell subpopulations, we rigorously filtered 31 samples from the GSE179640 dataset. Quality \ncontrol filtering retained high-quality cells with > 500 detected genes, 1,000-100,000 UMIs per cell, and mito -\nchondrial gene content < 25% (Figure S1A and S1B). The top 20 principal components (PCs) were selected for \ndownstream analyses based on the elbow plot (Figure S1C). Seurat clustering identified 21 distinct cell clusters \n(Figure S1D). Cell type annotation was performed using canonical markers as previously described10: epithelial \ncells (EPCAM, CDH1), stromal cells (COL1A1, PDGFRA), endothelial cells (VWF, PECAM1), myeloid cells \n(PTPRC, CD68, CD14), and lymphoid cells (PTPRC, CD2, CD3G). Violin plots illustrating marker expression \nacross Seurat clusters, identifying 9 epithelial, 5 stromal, 3 endothelial, 5 myeloid, and 1 lymphoid (Fig. 1A and \nB, Table S2). To eliminate potential doublets, DoubletFinder was applied (PCs = 1:10, pN = 0.25, pK = 0.09), \nyielding 5 major cell groups (Fig. 1C and F, Table S3). The distribution of subpopulations in control versus EM \ngroups shows stromal, lymphoid, myeloid, endothelial and epithelial cells comprising 76.49%, 76.75%, 74.01%, \n50.88% and 54.04% in EM, respectively. Notably, stromal cells, lymphoid cells, and myeloid cells are signifi -\ncantly enriched in EM compared with controls (Fig. 1E and F; Table S3).\nSingle-cell analysis reveals activation of oxidative phosphorylation in stromal cells of EM patients\nESC are recognized for their roles in immune modulation, tissue repair, and anti-inflammatory properties, making \nthem attractive candidates for cell therapy in endometrial diseases (e.g., Asherman syndrome)20. Therefore, ESC \nwere selected as the primary focus of this study. Following Marečková et al.,21, we performed ESC subpopulation \nanalyses of ESC in control and EM groups, identified three subpopulations: quiescent ESC specific to the pro -\nliferative phase (eStromal), activated eStromal with matrix metalloproteinases (eStromal_MMPs), and cycling \n\n\nArticle in Press\nScientific Reports\nhttps://doi.org/10.1038/s41598-026-70076-6\nFig. 1 Identification of marker genes and cluster analysis. (A) Expression of marker genes across Seurat clusters; (B) Expres-\nsion of markers in annotated clusters; (C) Doublet prediction in the dataset; (D) Clustering after removal of doublets; (E) \nCluster plots comparing control versus EM; (F) Relative abundances of cells in control versus EM.\n \n\n\nArticle in Press\nScientific Reports\nhttps://doi.org/10.1038/s41598-026-70076-6\neStromal (eStromal_cycling) exhibiting G2/M and S phase markers (Fig. 2A and B). Marker genes expression \nacross Seurat clusters was examined to annotate ESC subpopulations. High expression of MMP1, AR, ESR1, \nand PGR was observed in specific Seurat clusters. Within the annotated ESC subtypes, eStromal cells exhibited \nelevated expression of MMP11, MMP10, INHBA, COL8A1, MMP3, MMP1, AR, ESR1, and PGR, whereas \neStromal_MMPs were characterized by high expression of MMP11, AR, ESR1, and PGR. The eStromal_cycling \nsubpopulation displayed increased expression of proliferation-associated genes, including MKI67, PCNA and \nTGFB1 (Fig. 2C and D).\nComparison of subpopulation proportions between control and EM groups showed that all three subpopula -\ntions increased significantly in EM (Fig. 2E). The eStromal_cycling population is regarded as the differentiation \n“starting point” for stromal cells. Thus, KEGG analysis of DEGs in this subpopulation demonstrated significant \nenrichment of OXPHOS (Fig. 2F). The OXPHOS signature comprises 50 DEGs. To represent key components of \nthe mitochondrial electron transport chain and ATP production, we selected four representative genes: NDUFB5 \n(Complex I, NADH dehydrogenase), COX6C (Complex IV , cytochrome c oxidase), ATP5PO (Complex V , ATP \nsynthase), and PPA1 (a mitochondrial enzyme involved in energy metabolism). These genes collectively reflect \nelectron transport and ATP synthesis processes central to OXPHOS function. Their expression patterns across \nESC subtypes are shown in Fig. 2G. In addition, KEGG enrichment analysis identified significant enrichment of \nextracellular matrix-receptor interaction pathway including COL1A1, COL1A2, FN1, COL4A1, and COL4A2 \n(P < 0.05; data not shown, ranked 47th among enriched terms). Collectively, single-cell transcriptomic analysis \nindicates enhanced OXPHOS activity in stromal cells from EM patients.\nPathology and activation of the oxidative phosphorylation pathway in stromal cells from EM patients\nWe collected 14 eutopic endometrium tissue samples (control) and 14 ectopic endometrial tissue samples from \nEM patients. H&E and Masson’s trichrome staining (MTS) were performed on serial sections from the same \ntissue samples for histological evaluation. In control samples, the endometrium was intact, with glands lined by \na single layer of epithelial cells, orderly glandular arrangement, and no obvious hemorrhage or inflammatory \ninfiltration. In contrast, ectopic endometrial tissue from EM patients exhibited disrupted or partially lost glands, \nirregular glandular organization, the presence of foamy cells, and marked hemorrhage and inflammatory infiltra-\ntion (Fig. 3A). MTS revealed minimal collagen deposition in control endometrium, whereas EM tissue displayed \nconspicuous blue staining indicative of fibrosis (Fig. 3B). Quantitative analysis of MTS staining further revealed \nthat the collagen volume fraction (CVF) was significantly higher in EM tissue compared with controls (Fig. 3B). \nImmunofluorescence analysis confirmed the colocalization of PPA1, ATP5PO, NDUFB5, COX6C with Vimen-\ntin in both eutopic and EM endometrial tissues. Notably, the expression levels of PPA1, ATP5PO, NDUFB5, \nand COX6C were significantly elevated in EM tissue relative to eutopic endometrium, whereas Vimentin levels \nremained comparable between groups (Fig. 3C). Consistently, biochemical assays showed that mitochondrial \ncomplex I activity and ATP content were significantly increased in EM endometrium compared with controls \n(Fig. 3D and E). These findings indicate that EM endometrium exhibits pronounced histopathological alterations \nand fibrotic deposition, accompanied by enhanced OXPHOS activity.\nRotenone inhibits proliferation and migration of endometrial stromal cells\nEndometrial stromal cells in situ (inESC) and ectopic endometrial stromal cells (ecESC) were isolated by \nmechanical shear and collagenase digestion. Immunofluorescence showed high expression of the stromal marker \nVimentin in both inESC and ecESC, whereas the epithelial marker cytokeratin 7 was negative, indicating high \ncellular purity suitable for downstream experiments (Figure S2A). Dose-response analysis of Rotenone in ecESC \nyielded an IC50 of 0.9289 µM, which was used for subsequent treatments (Figure S2B). Further analyses assessed \nthe effects of Rotenone on proliferation, migration, and invasion of inESC and ecESC. The EdU assay revealed \na significantly more EdU positive cells in ecESC than in inESC, whereas Rotenone treatment reduced EdU \n\n\nArticle in Press\nScientific Reports\nhttps://doi.org/10.1038/s41598-026-70076-6\nFig. 2 Analysis of stromal cell subpopulations. (A) Stromal subpopulation Seurat clusters; (B) Annotated stromal subpopula-\ntions; (C) Marker gene expression profiles across stromal Seurat clusters (dot plot); (D) Marker gene expression profiles for \nannotated stromal subpopulations (dot plot); (E) Proportions of stromal subpopulations across treatments; (F) KEGG enrich-\nment for cycling stromal cells (bar graph); (G) Expression of oxidative phosphorylation marker genes (dot plot). NDUFB5 \n(Complex I), PPA1 (a mitochondrial enzyme involved in energy metabolism), COX6C (Complex IV), and ATP5PO (Com-\nplex V) were selected to span major electron transport chain complexes. Bubble size, percent expressed; color, average \nexpression.\n \n\n\nArticle in Press\nScientific Reports\nhttps://doi.org/10.1038/s41598-026-70076-6\nFig. 3 Analysis of clinical samples. (A) H&E staining to compare general morphology between control and EM (4×, bar \n= 500 μm; 10×, bar = 200 μm). (B) Masson’s trichrome staining to compare fibrosis between control and EM (4×, bar \n= 500 μm; 10×, bar = 200 μm). Quantitative analysis of collagen volume fraction (%), determined as the percentage of blue-\nstained area per total tissue area using ImageJ. (C) Colocalization analysis of OXPHOS-related pritein (PPA1, ATP5PO, \nNDUFB5, COX6C) with ESC marker. (D) Mitochondrial complex I activity in control versus EM. (E) ATP content in control \nversus EM. (H&E, Masson: n = 8; IF: n = 5; Mitochondrial complex I activity, ATP: n = 14); data are presented as mean ± SD. \nNormality of data distribution was assessed using the Shapiro-Wilk test. Comparisons between the two groups were per -\nformed using Student’s t-test. Significance is indicated by asterisks, *** P < 0.001.\n \n\n\nArticle in Press\nScientific Reports\nhttps://doi.org/10.1038/s41598-026-70076-6\nincorporation in both cell types, with a greater reduction in Rotenone-treated ecESC than in untreated ecESC \n(Fig. 4A and B). CCK-8 assay indicated higher viability of ecESC compared with inESC, and Rotenone mark -\nedly reduced viability in both groups (Fig. 4C). Transwell assays demonstrated increased migration and invasion \ncapacities in ecESC relative to inESC, these abilities were markedly suppressed following Rotenone treatment in \nboth cell types (Fig. 4D and E). Collectively, these results indicate that Rotenone suppressed proliferation, migra-\ntion, and invasion of ESC.\nAdditionally, we examined the impact of Rotenone on mitochondrial function in inESC and ecESC. Mitochon-\ndrial respiratory chain complex I activity and ATP content were significantly higher in ecESC than in inESC; \nRotenone treatment reduced these parameters in both cell types, the magnitude of suppression was greater in \necESC (Fig. 4F and G). MitoTracker Red CMXRos staining showed stronger mitochondrial fluorescence in \necESC compared with inESC. Rotenone markedly diminished mitochondrial fluorescence in both groups, with \nthe lowest intensity observed in ecESC + Rotenone group (Fig. 4H). Western blotting revealed that the OXPHOS-\nrelated proteins PPA1, ATP5PO, NDUFB5 and COX6C were upregulated in ecESC relative to inESC, and were \nsignificantly downregulated following Rotenone treatment in both groups (Fig. 4I).\nRotenone ameliorates lesions and fibrosis in a uterine endometriosis model by suppressing the \noxidative phosphorylation pathway\nAt 6 weeks post-surgery, mice in the sham, OE, and OE + Rotenone groups were euthanized, and ovarian tissues \nwere collected to assess surface cyst formation and surrounding adhesions. Ovaries from sham mice appeared \nmorphologically normal. In contrast, OE mice developed evident cystic lesions on one or both ovaries, whereas \nOE + Rotenone mice exhibited markedly reduced cystic lesions (Fig. 5A). H&E staining revealed intact ovar -\nian architecture and smooth surfaces without adhesions in sham mice. OE mice showed cystic dilation of the \novary accompanied by partial parenchymal atrophy, adhesions between the uterus and ovary, chronic inflamma-\ntory infiltration, and hemosiderin deposition at adhesion sites and within ovarian endometriosis. In comparison, \nOE + Rotenone mice demonstrated largely preserved ovarian and uterine structure, without ovarian-uterine adhe-\nsions, although focal ovarian endometriosis remained (Fig. 5B). Fibrosis is a hallmark of EM. MTS showed \nminimal fibrotic blue staining in sham mice, whereas OE mice exhibited prominent collagen deposition within \nthe cyst wall. The fibrotic blue staining was markedly attenuated in OE + Rotenone mice compared with OE mice, \nand quantitative analysis further demonstrated that the CVF was markedly elevated in OE mice (44.15 ± 18.60%) \ncompared with the sham group (3.60 ± 2.07%). Notably, CVF in OE + Rotenone mice (5.19 ± 1.43%) was sig -\nnificantly lower than in OE mice (Fig. 5C). Together, these results indicate that Rotenone can improve disease \nprogression in an OE model.\nBiochemical analyses were subsequently performed to assess mitochondrial function and sex hormone-related \nparameters. Compared with sham mice, OE mice showed significantly increased mitochondrial respiratory chain \ncomplex I activity and ATP content (Fig. 6A and B), along with elevated serum LH and FSH levels (Fig. 6C and \nD), while serum E2 and PROG levels were significantly reduced (Fig. 6E and F). In OE + Rotenone mice, com-\nplex I activity and ATP content were 402.85 ± 29.77 nmol/min/g and 657.77 ± 41.52 µmol/g, respectively. Serum \nE2, PROG, LH, and FSH levels were 1001.70 ± 16.59 pg/mL, 3037.50 ± 403.92 pg/mL, 88.43 ± 15.77 ng/mL, and \n298.13 ± 88.58 ng/mL, respectively, showing significant restoration compared with the OE group (Fig. 6A-F). \nFurthermore, immunofluorescence analysis of ovarian tissues demonstrated colocalization of OXPHOS-related \nproteins PPA1, ATP5PO, NDUFB5, COX6C with the stromal cell marker vimentin; OE ovaries showed mark -\nedly increased colocalization of these markers compared with sham mice, while OE+Rotenone mice showed \nreduced colocalization relative to OE mice (Fig. 7). Overall, these results suggest that Rotenone alleviated lesions \nand fibrosis in ovarian endometriosis, potentially through inhibition of OXPHOS.\n\n\nArticle in Press\nScientific Reports\nhttps://doi.org/10.1038/s41598-026-70076-6\nFig. 4 Rotenone inhibits proliferation and migration of ESC. (A) EdU assay showing proliferation of Rotenone-treated \ninESC and ecESC; bar = 50 μm. (B) EdU/DAPI cell numbers ratio. (C) CCK-8 assay for cell viability after Rotenone treat-\nment. (D) Transwell migration after Rotenone treatment. (E) Transwell invasion after Rotenone treatment. (F) Biochemical \nmeasurement of mitochondrial complex I activity after Rotenone treatment in inESC and ecESC. (G) ATP content after \nRotenone treatment. (H) Mito-Tracker Red CMXRos staining indicating changes in mitochondrial membrane potential \nafter Rotenone treatment; bar = 50 μm. (I) Western blot analysis of oxidative phosphorylation markers PPA1, ATP5PO, \nNDUFB5, COX6C after Rotenone treatment. Original blots are presented in Supplementary info file. n = 3; data are pre -\nsented as mean ± SD. Normality of data distribution was assessed using the Shapiro-Wilk test. Comparisons among multiple \ngroups were performed using one-way ANOV A followed by Tukey’s multiple comparisons test. differences versus inESC \nare marked with *, ** P < 0.01, *** P < 0.001; differences versus ecESC are marked with #, ### P < 0.001.\n \n\n\nArticle in Press\nScientific Reports\nhttps://doi.org/10.1038/s41598-026-70076-6\nFig. 5 Rotenone improves pathology and fibrosis in the ovarian endometriosis model. (A) At 6 weeks post-surgery, mice \nwere euthanized and ovarian lesions photographed to assess disease status. In the OE group, cystic lesions were clearly \nobservable, whereas in the OE+Rotenone group the lesions were improved. (B) HE staining confirms the severity of ovar -\nian endometriosis; upper panels show 4× magnification (bar = 500 μm) and lower panels show the corresponding magnified \nregion (4×, bar = 100 μm). (C) Masson staining confirms the extent of fibrosis in ovarian endometriosis; upper panels show \n4× magnification (bar = 500 μm) and lower panels show the magnified region (4×, bar = 100 μm). Quantitative analysis of \ncollagen volume fraction (%), determined as the percentage of blue-stained area per total tissue area using ImageJ. n = 3; data \nare presented as mean ± SD. Normality of data distribution was assessed using the Shapiro-Wilk test. Comparisons among \nmultiple groups were performed using one-way ANOV A followed by Tukey’s multiple comparisons test. Differences versus \nsham are marked with *, ** P < 0.01; differences versus OE are marked with #, # P < 0.05. Sham, sham operation; OE, ovar-\nian endometriosis.\n \n\n\nArticle in Press\nScientific Reports\nhttps://doi.org/10.1038/s41598-026-70076-6\nDiscussion\nIn this study, we employed single-cell transcriptomics to systematically characterize the heterogeneity and \nmolecular features of stromal cells in EM and, for the first time, identified a pivotal role of OXPHOS in disease \ninitiation and progression. Through in vitro experiments and animal model, we further validated the therapeutic \npotential of targeting mitochondrial electron transport chain complex I (Mitochondrial ETC Complex I) using \nRotenone, thereby providing a new rationale and target for precision treatment of EM.\nIn the initial phase, we analyzed the GSE179640 dataset to define cellular composition and subtypes in the \nendometrium. Five major cell types and 21 subtypes were identified, consistent with previously reports10. Nota-\nbly, the proportions of ESC, lymphocytes, and myeloid cells were increased in EM samples, reaching 76.49%, \n76.75%, and 74.01%, respectively. The expansion of ESC likely reflects enhanced extracellular matrix remodel-\ning and fibrosis within ectopic tissues. ESCs play essential roles in normal endometrial physiology, inflamma -\ntion, angiogenesis, and tissue repair 22–24. The increased proportions of lymphocytes and myeloid cells supports \nthe presence of a chronic inflammatory microenvironment in EM, in which infiltrating immune cells may \npromote survival, proliferation, and invasion of ectopic endometrium tissue 25,26. Subsequent subclustering of \nESCs revealed three functionally distinct subpopulations: quiescent eStromal, activated eStromal_MMPs, and \nFig. 6 Rotenone treatment alters mitochondrial function and sex hormone levels in the mouse ovarian endometriosis model. \nAt 6 weeks post-surgery, ovarian tissues were collected after euthanasia for biochemical analyses of different treatment \ngroups. (A) Mitochondrial respiratory chain complex I activity in the ovary; (B) ovary ATP levels; and serum concentrations \nof (C) E2, (D) PROG, (E) LH, (F) FSH in mice from different treatment groups. n = 8; data are shown as mean ± SD. Normal-\nity of data distribution was assessed using the Shapiro-Wilk test. Comparisons among multiple groups were performed using \none-way ANOV A followed by Tukey’s multiple comparisons test. Differences versus sham are marked with *, *** P < 0.001; \ndifferences versus OE are marked with #, ### P < 0.001. Sham, Sham operation; OE, ovarian endometriosis; E2, Estradiol; \nPG, Progesterone; LH, Luteinizing Hormone; FSH, Follicle-Stimulating Hormone.\n \n\n\nArticle in Press\nScientific Reports\nhttps://doi.org/10.1038/s41598-026-70076-6\nproliferative eStromal_cycling. This classification is broadly consistent with Marečková et al., 21, although we \nobserved distinct expression patterns specific to EM. The eStromal_MMPs subgroup exhibited high expression \nof matrix metalloproteinases (MMPs, including MMP1, MMP3, MMP10, and MMP11), which are involved in \nextracellular matrix degradation, angiogenesis, and cell migration 27. In EM, MMP overexpression may facili -\ntate lesion invasiveness and neovascularization. This subgroup also shows high expression of ESR1, PGR, and \nAR, suggesting a prominent role for hormone signaling in stromal activation. The eStromal_cycling subset, \ncharacterized by high expression of cell cycle–related genes (MKI67, PCNA) and TGFB1, may represent a \nproliferative progenitor-like population. TGF-β signaling plays a central regulatory role in fibrosis, and may be \na key mechanism underlying EM-associated fibrosis28,29. Importantly, this subset showed significant enrichment \nFig. 7 Immunofluorescence analysis of \ncolocalization between oxidative phosphor-\nylation markers and stromal cell marker in \novaries after Rotenone treatment. Immuno-\nfluorescence detection of PPA1, ATP5PO, \nNDUFB5, COX6C with Vimentin in ovaries \nfrom the different groups, illustrating the \ndegree of colocalization.\n \n\n\nArticle in Press\nScientific Reports\nhttps://doi.org/10.1038/s41598-026-70076-6\nof OXPHOS-related pathways, suggesting disease-associated metabolic reprogramming in proliferative stromal \ncells.\nMetabolic reprogramming is a hallmark of malignant cells, typically involving enhanced glycolysis and altered \nmitochondrial respiration to meet bioenergetic and biosynthetic demands. Although EM is a benign condition, \nectopic endometrial tissue exhibits invasive and migratory properties reminiscent of malignancy. Recent studies \nreport increased glycolytic activity in ectopic EM cells, likely reflecting elevated energy requirements 30. Gly-\ncolysis provides pyruvate for mitochondrial oxidation, fueling the TCA cycle and ATP generation via OXPHOS. \nElevated ATP levels may also contribute to EM-associated pain and infertility, as ATP acts as a purinergic signal-\ning molecule via receptors such as P2 × 3, P2 × 4, P2 × 7, and P2Y 31. Our in vitro findings demonstrated that \nOXPHOS was markedly upregulated in ESC derived from EM patients, accompanied by increased expression of \nPPA1, ATP5PO, NDUFB5 and COX6C. Enhanced mitochondrial respiratory chain complex I activity and ATP \nproduction were observed in ecESCs, together with augmented proliferative, migratory, and invasive capacities. \nThese data suggest that OXPHOS activation may provide the energetic support required for ectopic lesion growth \nand progression. Mitochondrial dysfunction is closely linked to inflammatory responses. Increased OXPHOS \nactivity may elevate ROS production, leading to oxidative stress and inflammatory signaling. In EM, ectopic \nlesions are frequently associated with chronic inflammatory infiltration, which may be exacerbated by mitochon-\ndrial ROS overproduction32,33. In turn, inflammation may further impair mitochondrial homeostasis, creating a \nself-perpetuating pathological cycle.\nFurthermore, Rotenone effectively inhibits proliferation, migration, and invasion of ESC, with more pro -\nnounced effects in ecESC. In vivo, Rotenone treatment significantly improved pathological features in the OE \nmodel, including reducted cystic lesions, fewer adhesions and ameliorated fibrosis. Importantly, serum sex hor -\nmone levels were also significantly restored in the OE + Rotenone group, suggesting that Rotenone not only \nmitigates local pathology but also helps restore ovarian endocrine function.\nNevertheless, several limitations should be acknowledged. First, the single-cell transcriptomic data were \nderived from publicly available databases with limited clinical annotation, which may restrict interpretability. \nSecond, the surgically induced murine model does not fully recapitulate human endometriosis. Mice neither \nmenstruate nor spontaneously develop EM; therefore, this model may not entirely reflect the chronic inflamma -\ntion, hormonal cyclicity, and progressive fibrosis observed in patients 34,35. These interspecies differences limit \nclinical translation. Third, the human validation cohort was relatively small and requires confirmation in larger \npopulations. Fourth, although Rotenone effectively inhibited complex I activity in our study, its documented neu-\nrotoxicity precludes clinical application36,37. Here, Rotenone was used solely as a pharmacological tool to estab-\nlish the functional relevance of OXPHOS/complex I in EM. For clinical translation, Metformin, Mdivi-1, SS-31, \nand mitochondrial-targeted antioxidants as potential therapeutics for clinical translation 38–40. Finally, while our \npharmacological findings implicate OXPHOS/complex I in EM pathogenesis, definitive causal inference will \nrequire genetic manipulation combined with rescue experiments, which remain to be addressed in future studies.\nConclusion\nOur analysis identifies aberrant activation of the OXPHOS in ESC, which is a key driver of disease progression \nand fibrosis. Targeted inhibition of mitochondrial complex I markedly suppresses ESC proliferation, migration, \nand invasion, improves lesion pathology and fibrosis in the OE model, and partially restores ovarian endocrine \nfunction. These findings provide a novel metabolism targeted strategy for EM intervention.\nSupplementary Information  The online version contains supplementary material available at https://doi.org/10.1038/\ns41598-026-70076-6.\nAuthor contributions Xiulan Weng and Shunhe Lin conducted all the experiments in this article, and wrote the manuscript. \nZhenna Wang and Chaobin Liu visualized the data and combined the results. Guan Lin provided methodological support. \n\n\nArticle in Press\nScientific Reports\nhttps://doi.org/10.1038/s41598-026-70076-6\nPengming Sun and Jingsong Yi conceived and designed the manuscript, and revised the manuscript. All the authors reviewed \nthe manuscript and approved it for publication.\nFunding We are very grateful for the funding support from Joint Funds for the innovation of science and Technology, \nFujian province (Grant number: 2025Y9611), and Fujian Provincial Traditional Chinese Medicine Science and Technology \nProgram (Category B) (Grant number: 2025YBB014).\nData availability The authors confirm that the data supporting the findings of this study are available within the article and/\nor its supplementary materials, and from the corresponding authors upon reasonable request.\nDeclarations\nCompeting interests The authors declare no competing interests.\nEthical statement The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy \nor integrity of any part of the work are appropriately investigated and resolved.\nEthical approval All patients procedures were approved by the Ethics Committee of Fujian Maternal and Child Health \nHospital (approval number: 2025KY253). All animal procedures were approved by the Ethics Committee of Fujian Maternal \nand Child Health Hospital (approval number: AECSFY2025080).\nOpen Access  This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, \nsharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the \noriginal author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The \nimages or other third party material in this article are included in the article’s Creative Commons licence, unless indicated \notherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your \nintended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly \nfrom the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.\nReferences\n 1. Begum, I. A. The connection between endometriosis and secondary dysmenorrhea. Journal Reproductive Immunology, \n104425. (2025).\n 2. Carranco, R. C. et al. Peritoneal retraction pocket defects and their important relationship with pelvic pain and \nendometriosis. J. Minim. Invasive. Gynecol. 28 (2), 168–169 (2021).\n 3. Smolarz, B., Szyllo, K. & Romanowicz, H. Endometriosis: Epidemiology, Classification, Pathogenesis, Treatment and \nGenetics (Review of Literature). Int. J. Mol. Sci. 22 (19). https://doi.org/10.3390/ijms221910554 (2021).\n 4. Ma, J. et al. Single-cell transcriptomic analysis of endometriosis provides insights into fibroblast fates and immune cell \nheterogeneity. Cell. bioscience. 11 (1), 125 (2021).\n 5. Bedrick, B. S. et al. A Systematic Review of Epigenetics of Endometriosis. F S Rev. 5 (1).  h t t p s : / / d o i . o r g / 1 0 . 1 0 1 6 / j . x f n r . \n2 0 2 4 . 0 1 . 0 0 3     (2024).\n 6. Jerman, L. F. & Hey-Cunningham, A. J. The role of the lymphatic system in endometriosis: a comprehensive review of \nthe literature. Biol. Reprod. 92 (3), 64. https://doi.org/10.1095/biolreprod.114.124313 (2015).\n 7. Longo, L. D. Classic pages in obstetrics and gynecology. Aberrant portions of the mullerian duct found in an ovary: \nWilliam Wood Russell Johns Hopkins Hospital Bulletin. 1899 Am. J. Obstet. Gynecol. 10 (2), 8–10 (1979).\n 8. Maruyama, T. & Yoshimura, Y . Stem cell theory for the pathogenesis of endometriosis. Front. Biosci. (Elite Ed). 4 (8), \n2754–2763. https://doi.org/10.2741/e589 (2012).\n 9. Ochoa Bernal, M. A. & Fazleabas, A. T. The Known, the Unknown and the Future of the Pathophysiology of Endometriosis. \nInt. J. Mol. Sci. 25 (11). https://doi.org/10.3390/ijms25115815 (2024).\n 10. Tan, Y . et al. Single-cell analysis of endometriosis reveals a coordinated transcriptional programme driving \nimmunotolerance and angiogenesis across eutopic and ectopic tissues. Nat. Cell. Biol. 24 (8), 1306–1318.  h t t p s : / / d o i . o r g \n/ 1 0 . 1 0 3 8 / s 4 1 5 5 6 - 0 2 2 - 0 0 9 6 1 - 5     (2022).\n 11. Wu, J. et al. Dissecting the cell microenvironment of ovarian endometrioma through single-cell RNA sequencing. Sci. \nChina Life Sci. 68 (1), 116–129. https://doi.org/10.1007/s11427-024-2638-9 (2025).\n 12. Atkins, H. M. et al. Endometrium and endometriosis tissue mitochondrial energy metabolism in a nonhuman primate \nmodel. Reprod. Biol. Endocrinol. 17 (1), 70. https://doi.org/10.1186/s12958-019-0513-8 (2019).\n\n\nArticle in Press\nScientific Reports\nhttps://doi.org/10.1038/s41598-026-70076-6\n 13. Chen, M. et al. Role of Pink1-mediated mitophagy in adenomyosis. PeerJ 11, e16497. https://doi.org/10.7717/peerj.16497 \n(2023).\n 14. Sabini, E., Arboit, L., Khan, M. P., Lanzolla, G. & Schipani, E. Oxidative phosphorylation in bone cells. Bone Rep. 18, \n101688. https://doi.org/10.1016/j.bonr.2023.101688 (2023).\n 15. Tsogtbaatar, E., Landin, C., Minter-Dykhouse, K. & Folmes, C. D. L. Energy Metabolism Regulates Stem Cell \nPluripotency. Front. Cell. Dev. Biol. 8, 87. https://doi.org/10.3389/fcell.2020.00087 (2020).\n 16. Ibarra-Gutierrez, M. T., Serrano-Garcia, N. & Orozco-Ibarra, M. Rotenone-Induced Model of Parkinson’s Disease: \nBeyond Mitochondrial Complex I Inhibition. Mol. Neurobiol. 60 (4), 1929–1948.  h t t p s : / / d o i . o r g / 1 0 . 1 0 0 7 / s 1 2 0 3 5 - 0 2 2 - 0 \n3 1 9 3 - 8     (2023).\n 17. Xiao, W. et al. Rotenone restrains colon cancer cell viability, motility and epithelial–mesenchymal transition and \ntumorigenesis in nude mice via the PI3K/AKT pathway. Int. J. Mol. Med. 46 (2), 700–708.  h t t p s : / / d o i . o r g / 1 0 . 3 8 9 2 / i j m m \n. 2 0 2 0 . 4 6 3 7     (2020).\n 18. Hayashi, S. et al. Novel ovarian endometriosis model causes infertility via iron-mediated oxidative stress in mice. Redox \nBiol. 37, 101726. https://doi.org/10.1016/j.redox.2020.101726 (2020).\n 19. Abdo, K. M. et al. Toxicity and carcinogenicity of rotenone given in the feed to F344/N rats and B6C3F1 mice for up to \ntwo years. Drug Chem. Toxicol. 11 (3), 225–235. https://doi.org/10.3109/01480548809017879 (1988).\n 20. Queckborner, S., Syk Lundberg, E., Gemzell-Danielsson, K. & Davies, L. C. Endometrial stromal cells exhibit a distinct \nphenotypic and immunomodulatory profile. Stem Cell. Res. Ther. 11 (1), 15. https://doi.org/10.1186/s13287-019-1496-2 \n(2020).\n 21. Mareckova, M. et al. An integrated single-cell reference atlas of the human endometrium. Nat. Genet. 56 (9), 1925–1937. \nhttps://doi.org/10.1038/s41588-024-01873-w (2024).\n 22. Cousins, F. L., Filby, C. E. & Gargett, C. E. Endometrial Stem/Progenitor Cells-Their Role in Endometrial Repair and \nRegeneration. Front. Reprod. Health. 3, 811537. https://doi.org/10.3389/frph.2021.811537 (2021).\n 23. Hong, I. S. Endometrial Stem Cells: Orchestrating Dynamic Regeneration of Endometrium and Their Implications in \nDiverse Endometrial Disorders. Int. J. Biol. Sci. 20 (3), 864–879. https://doi.org/10.7150/ijbs.89795 (2024).\n 24. Okada, H. et al. Regulation of decidualization and angiogenesis in the human endometrium: mini review. J. Obstet. \nGynaecol. Res. 40 (5), 1180–1187. https://doi.org/10.1111/jog.12392 (2014).\n 25. Amidifar, S., Jafari, D., Mansourabadi, A. H., Sadaghian, S. & Esmaeilzadeh, A. Immunopathology of Endometriosis, \nMolecular Approaches. Am. J. Reprod. Immunol. 93 (3), e70056. https://doi.org/10.1111/aji.70056 (2025).\n 26. Suszczyk, D., Skiba, W., Jakubowicz-Gil, J., Kotarski, J. & Wertel, I. The Role of Myeloid-Derived Suppressor Cells \n(MDSCs) in the Development and/or Progression of Endometriosis-State of the Art. Cells 10(3).  h t t p s : / / d o i . o r g / 1 0 . 3 3 9 0 \n/ c e l l s 1 0 0 3 0 6 7 7     (2021).\n 27. Vacca, A. et al. Human lymphoblastoid cells produce extracellular matrix-degrading enzymes and induce endothelial cell \nproliferation, migration, morphogenesis, and angiogenesis. Int. J. Clin. Lab. Res. 28 (1), 55–68.  h t t p s : / / d o i . o r g / 1 0 . 1 0 0 7 / \ns 0 0 5 9 9 0 0 5 0 0 1 8     (1998).\n 28. Xia, Y . et al. Neferine mediated TGF-beta/ERK signaling to inhibit fibrosis in endometriosis. Am. J. Transl Res. 15 (5), \n3240–3253 (2023).\n 29. Yang, M. et al. Integrated Proteomics Reveals Spliceosome Dysregulation and Highlights DHX9 as a TGF-beta Signaling-\nDriven Fibrosis Hub in Endometriosis. J. Proteome Res. 24 (8), 4044–4059.  h t t p s : / / d o i . o r g / 1 0 . 1 0 2 1 / a c s . j p r o t e o m e . 5 c 0 0 1 \n2 2     (2025).\n 30. Lu, C. et al. Phosphorylation of PFKFB4 by PIM2 promotes anaerobic glycolysis and cell proliferation in endometriosis. \nCell. Death Dis. 13 (9), 790. https://doi.org/10.1038/s41419-022-05241-6 (2022).\n 31. Wang, Y ., Zhai, Q., Tian, Z., Chang, X. & Zhu, H. Relationship between the change of ATP level controlled by \nectonucleotidases and endometriosis-associated infertility and pain. Gynecology Obstet. Clin. Medicine 4(2). (2024).\n 32. Kapur, A. et al. Oxidative phosphorylation inhibitors inhibit proliferation of endometriosis cells. Reproduction 165 (6), \n617–628. https://doi.org/10.1530/REP-22-0265 (2023).\n 33. Shen, D., Qi, C., Hu, P., Li, J. & Shen, Y . Mitochondrial Involvement in the Pathogenesis of Endometriosis and its \nPotential as Therapeutic Targets. Reprod. Sci. 32 (4), 935–949. https://doi.org/10.1007/s43032-025-01827-5 (2025).\n 34. Burns, K. A. et al. Endometriosis in the Mouse: Challenges and Progress Toward a ‘Best Fit’ Murine Model. Front. \nPhysiol. 12, 806574. https://doi.org/10.3389/fphys.2021.806574 (2021).\n 35. Howe, C. A. et al. Induction of Endometriosis in a Menstruating Mouse Model (Mus musculus): A Translational Animal \nDisease Model. J. Am. Assoc. Lab. Anim. Sci. 64 (4), 1–15. https://doi.org/10.30802/AALAS-JAALAS-24-160 (2025).\n 36. Sherer, T. B. et al. Mechanism of toxicity in rotenone models of Parkinson’s disease. J. Neurosci. 23 (34), 10756–10764. \nhttps://doi.org/10.1523/JNEUROSCI.23-34-10756.2003 (2003).\n 37. Tat, J., Heskett, K. & Boss, G. R. Acute rotenone poisoning: A scoping review. Heliyon 10 (7), e28334.  h t t p s : / / d o i . o r g / 1 \n0 . 1 0 1 6 / j . h e l i y o n . 2 0 2 4 . e 2 8 3 3 4     (2024).\n 38. Bordt, E. A. et al. The Putative Drp1 Inhibitor mdivi-1 Is a Reversible Mitochondrial Complex I Inhibitor that Modulates \nReactive Oxygen Species. Dev. Cell. 40 (6), 583–594e586. https://doi.org/10.1016/j.devcel.2017.02.020 (2017).\n\n\nArticle in Press\nScientific Reports\nhttps://doi.org/10.1038/s41598-026-70076-6\n 39. Wu, M. et al. Metformin and Fibrosis: A Review of Existing Evidence and Mechanisms. J Diabetes Res 2021, 6673525. \nhttps://doi.org/10.1155/2021/6673525 (2021).\n 40. Zhu, Y . et al. SS-31, a Mitochondria-Targeting Peptide, Ameliorates Kidney Disease. Oxid Med Cell Longev  2022, \n1295509. https://doi.org/10.1155/2022/1295509 (2022).\nPublisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional \naffiliations.\nAuthors and Affiliations\nXiulan Weng1,2 · Shunhe Lin1 · Zhenna Wang1 · Chaobin Liu1 · Guan Lin1 · \nPengming Sun1,2 · Jingsong Yi2,3\n Pengming Sun\nfmsun1975@fjmu.edu.cn\n Jingsong Yi\n3129671082@qq.com\n1 Department of Gynecology, Fujian Maternity and Child Health Hospital (Fujian Obstetrics and Gynecology \nHospital), Fuzhou 350001, Fujian, China\n2 College of Clinical Medicine for Obstetrics & Gynecology and Pediatrics, Fujian Medical University,  \nFuzhou 350001, Fujian, China\n3 Fujian Maternity and Child Health Hospital, No. 18 Daoshan Road, Gulou District, Fuzhou, Fujian  \n350001, PR China","source_license":"CC0","license_restricted":false}