Redefining the contribution of retrograde menstruation to endometriosis: single-cell analysis of endometriotic lesions suggests a process more complex than simple autografting

Female mice (c57BL/6 strain), six-to-eight-weeks-old, were purchased from Charles River Laboratories (Wilmington, MA, USA) while c57BL/6 female transgenic ubiquitin-GFP mice expressing GFP (green fluorescent protein) were obtained from Jackson Laboratory (stock #004353, Bar Harbor, ME, USA). These mice were maintained in the Yale animal facility center at the Yale School of Medicine (New Haven, CT, USA). Each mouse cage containing five mice were housed and kept in a room (21 °C ± 1 °C) with a 12-h light/dark cycle (7:00  am  to 7:00  pm ) with ad libitum access to food (Purina Chow; Purina Mills, Richmond, Indiana) and water. All experiments were carried out according to a protocol (#2023–07113) approved by the Institutional Animal Care and Use Committee (IACUC) at Yale University. The female transgenic mice expressing GFP were used as donors for uterus transplantation (allograft) to induce endometriosis in recipient wild-type (c57BL/6) mice. Endometriosis was induced surgically in c57BL/6 female mice of 6–8-week age as previously described with minor changes (Lee et al. 2009 ). The uterus was removed from GFP expressing female transgenic mice as donors under anesthesia by hysterectomy and immersed in Hank's Balanced Salt Solution (HBSS). Uterine horns were separated from the extracted uterus, and each horn was opened longitudinally and divided into four fragments for a total of four sections per horn to the total of eight fragments per uterus. Each fragment is approximately 3–5 mm in length and kept in HBSS on ice for the immediate implantation onto the peritoneal wall of the recipient mice. Twenty-two normal c57BL/6 mice underwent laparotomy under inhalation anesthesia with isoflurane for either endometriosis induction ( n  = 12) or sham surgery as a control ( n  = 10). Here, we briefly described the implantation procedure. For the transplantation of uterine fragments from the donor GFP mice into the recipient mice, the recipient mice were put under anesthesia by inhalation of isoflurane (2.5 L/min) in conjunction with oxygen (1.5 L/min) and its pelvic cavity was opened by a median longitudinal incision of approximately 2 cm long. Two uterine fragments were sutured to each side (right and left) of the parietal peritoneum using 5–0 polyglactin suture (Vicryl), approximately 1 cm apart. The peritoneum and skin were closed with 4–0 polyglactin suture. Sham surgeries were performed for the control group ( n  = 10) using the same surgical procedure giving V5.0 sutures without the introduction of extraneous uterine tissue. No hormonal supplementation or antibiotics were administered before or after surgery. Lesions were collected from the mice 18 weeks after induction of endometriosis and subjected to single cell suspension preparations. In-vivo imaging was carried out for GFP expression in donor mice as well as recipient mice in which GFP-uterine tissue fragments were transplanted. IVIS lumina X5 machine was used for the imaging. The excitation and emission wavelengths were 480 nm and 520 nm, respectively, for GFP expression determination. Endometriotic lesions collected from mice (18-weeks post-transplantation) were minced on ice into pieces of 1-mm 3 or smaller, followed by tissue digestion in enzymatic digestion buffer containing dispase at 37 °C for 30 min with intermittent shaking for every 5 min. Subsequently, the digested tissue was centrifuged at 1600 rpm for 5 min at room temperature, and the supernatant was collected, to which 0.04% of BSA (400 µg/ml) in phosphate buffered solution (PBS, calcium and magnesium-free) was added, followed by centrifugation at 1800 rpm at room temperature for 7 min. The collected cell pellet was then resuspended in 1.5 ml of red blood cell (RBC) lysis buffer and incubated at 4 °C for 15 min. After confirmation of RBC lysis, the samples were suspended in 1 ml of 0.04% BSA and filtered through a 40-µm cell strainer. The filtrate containing single cells was collected. The cell viability and concentration were determined using hemocytometer while cells were stained with Trypan Blue stain. The single cell preparations from each lesion were pooled and the final concentration of the cells was made to 1000 cells per μl and sent to 10X Genomic Center at Yale University, CT, USA for single cell sequence analysis. Single-cell RNA sequencing was carried out at the Yale Center for Genome Analysis (YCGA), West Haven, Connecticut, USA. A target count of 10,000 cells per sample was carried out by 10 × Chromium system (10 × Genomics; Pleasanton, CA) through single-cell droplet formation technique. Chromium Single Cell 3' Reagent v3 Kits were used for the preparation of scRNA-seq libraries. To generate single-cell gel beads in emulsions (GEM), single-cell suspensions were loaded on to a Chromium Single Cell Controller Instrument (10X Genomics). Briefly, about 10,000 cells were tagged for single cell sequence analysis. According to the manufacturer's protocol, single-cell RNA-seq libraries were prepared using the Chromium Single Cell 3’Library & Cell Bead Kit (10X Genomics) after generation of GEMs, and the generated scRNA-seq libraries were sequenced on an Illumina NovaSeq 6000. FASTQ files were generated with Cell Ranger v6.0.2 command  mkfastq , through demultiplexed base call files (BCL).The files were then aligned to the mouse reference genome GRCm39, followed by filtering and barcode processing, and unique molecular identifier (UMI) counting to generate count matrices for each sample. The dataset was created after count matrices were merged and then subjected to the standard quality control and clustering pipeline in Seurat v4.0 (Satija et al. 2015 ). Briefly, cells with greater than 5% mitochondrial counts, as well as cells with unique feature counts over 2,500 or under 200, were excluded from further analyses. The dataset was analyzed for the presence of GFP sequence using the Seurat package in R studio. Data were normalized with log normalization using the default scale factor of 10,000. FindVariableFeatures  function was used for identifying genes with high variability in expression across cells. Data were then scaled, and linear dimension reduction was performed using principal component analysis (PCA). To identify cell populations, cells were clustered with  FindNeighbors  and  FindClusters  with a resolution of 0.75, then subjected to UMAP nonlinear dimensional reduction. Markers for each cluster were identified with FindAllMarkers  and used to distinguish cell types which were determined by literature search and PangaloDB (Franzén et al. 2019 ). Harmony v0.1was used to integrate cell populations across the time points (Korsunsky et al. 2019 ) while ShinyGO v0.76.3 ( http://bioinformatics.sdstate.edu/go/ ) was used to perform gene ontology analysis with the differentially expressed genes (DEGs) across pseudo time. Differentially expressed genes (DEGs) were identified using the FindMarkers function (test.use = MAST) in Seurat. The cell specific marker genes (top 7) used to identify different cell clusters are showed in Suppl. Table 1, while their expression validation is showed in Suppl. Figure 1. P value  0.79 were set as the threshold for significantly differential expression. We also specifically examined several cytokines and chemokines that have previously been shown to be increased in endometriosis: Ccl8 (Henlon et al. 2024 ), Ccl7 and Ccl4 (Laudański et al. 2006 ), Ccl12 , Ccl2 and Ccl5 (Zhou et al. 2025 ), Cxcl12 (Zhou et al. 2023 ), Cxcl16 (Manabe et al. 2011 ), and Il1β (Herrmann Lavoie et al. 2007 ). Further, we assessed the expression of Kirsten rat sarcoma virus oncogene homologue ( Kras ), insulin-like growth factor-1 ( Igf-1 ), matrix metalloproteinases-2 ( Mmp-2 ), and interleukin-1 beta ( Il-1β ) by the endogenous endometrial-derived cells in the lesions.

Figure  1 A shows the image of a GFP transgenic mouse (positive control). Figure  1 B represents a control mouse which was not expressing GFP (negative control) and Fig.  1 C shows GFP expressing uterine tissue fragments from donor mice that were transplanted into the peritoneal cavity of recipient non-GFP mice. Four fragments that were transplanted into each mouse developed into endometriosis lesions as shown in Fig.  1 D. All recipient mice showed 4 lesions per mouse. Figure  1 E shows the sham surgery in control mice demonstrating no lesion development, scar or cellular collection observed in the peritoneal cavity at any suture site. Fig. 1 In-vivo imaging of GFP expression in mice. A Mice expressing GFP (green color). B Mice not expressing GFP (no green color). C Mice in which uterine tissue was transplanted. Image showing GFP expression only in the transplanted area in the pelvic cavity of the recipient mice (green color). D Representative images showing ectopic endometriotic lesions developed after 18 weeks of uterine tissue transplantation in the recipient mice. E Absence of lesions, suture material or inflammatory reaction at site of sham surgery in controls. (Original magnification 5X) In-vivo imaging of GFP expression in mice. A Mice expressing GFP (green color). B Mice not expressing GFP (no green color). C Mice in which uterine tissue was transplanted. Image showing GFP expression only in the transplanted area in the pelvic cavity of the recipient mice (green color). D Representative images showing ectopic endometriotic lesions developed after 18 weeks of uterine tissue transplantation in the recipient mice. E Absence of lesions, suture material or inflammatory reaction at site of sham surgery in controls. (Original magnification 5X) Mice were sacrificed 18 weeks post-surgery, and lesions were collected. Average size of each lesion was 75.77 mm 3 calculated using the formula: V = L x W x H x π/6. Single cell suspensions were prepared from the lesions and subjected to Single Cell RNA Sequence (scRNA-seq) analysis by 10X Genomics. Total cells were clustered into distinct cell types based on the specific cell markers as showed in Fig.  2 A. We detected the presence of fibroblasts, smooth muscle cells, enterocytes, endothelial cells, natural killer cells, T cells, B cells, neutrophils and macrophages. The number of cells and percentage of each cell type are shown in Table  1 . Fibroblasts had the cell count (1233, 31%), followed by B cells (684, 17%), macrophages (570, 15%), and T cells (431, 11%). The marker genes used for each cell type and their expression validation is shown in Suppl. Table 1 and Suppl. Figure 1 respectively. Fig. 2 UMAP projections of scRNA-seq data for ectopic endometriotic lesion. A Mapping of total cells derived both exogenously and endogenously into clusters of different cell types present in the endometriotic lesion. B Heat map showing the expression of different genes by various cell types Table 1 Number and Percentage of Cell Types in Total Cells Cell type Number of Cells Percentage (%) Fibroblasts 1233 31.37 B Cells 684 17.40 Macrophages 570 14.50 T Cells 431 10.96 Natural Killer Cells 363 9.23 Neutrophils 278 7.07 Endothelial Cells 179 4.55 Smooth Muscle Cells 113 2.87 Enterocytes 80 2.04 UMAP projections of scRNA-seq data for ectopic endometriotic lesion. A Mapping of total cells derived both exogenously and endogenously into clusters of different cell types present in the endometriotic lesion. B Heat map showing the expression of different genes by various cell types Number and Percentage of Cell Types in Total Cells The corresponding heatmap of these cell clusters is shown in Fig.  2 B. Each lesion contained two types of cells: allograft cells that express GFP (GFP-cells) derived from donor mice of allografted uterine tissue, and host cells that are not expressing GFP (Non-GFP), derived endogenously from the circulation of the recipient or host mice. Among the 10,000 tagged cells analyzed from each lesion, 35% are allograft cells (green) while 65% are host cells derived from the circulation (grey) as showed in Fig.  3 A. Figure  3 B shows the violin plots of allograft cell types: fibroblasts, smooth muscle cells, endothelial cells, and enterocytes. Cells from allograft and host were clustered into different cell types based on the specific cell markers assigned to the cells. Fig. 3 UMAP projections and violin plots of scRNA-seq data for GFP cells in ectopic endometriotic lesion. A Mapping of GFP (green fluorescent protein) expressing cells that are derived exogenously or from uterus of the donor mice. B Violin plots showing the expression levels of GFP in various cell types. Fibroblasts (orange), smooth muscle cells (blue), endothelial cells (purple) and enterocytes (pink). Different cell clusters colored by cell lineage and types of cells present in various cell clusters UMAP projections and violin plots of scRNA-seq data for GFP cells in ectopic endometriotic lesion. A Mapping of GFP (green fluorescent protein) expressing cells that are derived exogenously or from uterus of the donor mice. B Violin plots showing the expression levels of GFP in various cell types. Fibroblasts (orange), smooth muscle cells (blue), endothelial cells (purple) and enterocytes (pink). Different cell clusters colored by cell lineage and types of cells present in various cell clusters Figure  2 A showed different cell populations present in the total cells of a lesion while Fig.  4 A, C showed different cell types present in GFP expressing cells and non-GFP cells of a lesion respectively. Figure  4 A shows the cell clusters of allografts: fibroblasts, stromal fibroblasts, endothelial cells, stromal cells, macrophages, natural killer cells, enterocytes, smooth muscle cells, and B cells. The number of cells and the percentage of each cell type present in allografts were showed in Table  2 . The host contributes the bulk of the immune cells to the lesion. Most of the cells in the allograft are stromal cells (1016, 68.13%) followed by endothelial cells (138, 9%) and stromal cells (101, 7%); very few immune cells come with the allograft. The corresponding heatmap of these cell clusters is showed in Fig.  4 B. Host cell clusters were shown in Fig.  4 C: B cells, macrophages, T cells, natural killer cells, neutrophils, and endothelial cells. The number of cells and the percentage of each cell type originated from the host were shown in Table  3 . The highest number of cells was found in B cells (631, 26%), followed by macrophages (487, 20%), T cells (401, 16%), and natural killer cells (329, 13%). The corresponding heatmap of these cell clusters is showed in Fig.  4 D. Fig. 4 UMAP projections and heat maps of scRNA-seq data for GFP and non-GFP cells in ectopic endometriotic lesion. A Mapping of GFP cells derived exogenously into different cell types present in the endometriotic lesion. B Heat map of the GFP cells showing the differential gene expression. C Mapping of non-GFP cells derived endogenously into different cell types present in the endometriotic lesion. D Heat map of the non-GFP cells showing the differential gene expression Table 2 Number of Cells and Percentage of Cell Types in GFP Cells Cell type Number of Cells Percentage (%) Fibroblasts 1016 68.13 Endothelial Cells 138 9.26 Macrophages 78 5.23 Natural Killer Cells 68 4.56 Enterocytes 67 4.49 Smooth Muscle Cells 62 4.16 B Cells 62 4.16 Table 3 Number and Percentage of Cell Types in Non-GFP Cells Cell type Number of Cells Percentage (%) B Cells 631 25.86 Macrophages 487 19.96 T Cells 401 16.43 Natural Killer Cells 329 13.48 Fibroblasts 282 11.56 Neutrophils 271 11.11 Endothelial 39 1.60 UMAP projections and heat maps of scRNA-seq data for GFP and non-GFP cells in ectopic endometriotic lesion. A Mapping of GFP cells derived exogenously into different cell types present in the endometriotic lesion. B Heat map of the GFP cells showing the differential gene expression. C Mapping of non-GFP cells derived endogenously into different cell types present in the endometriotic lesion. D Heat map of the non-GFP cells showing the differential gene expression Number of Cells and Percentage of Cell Types in GFP Cells Number and Percentage of Cell Types in Non-GFP Cells Here, syngeneic transplant allows the introduction of GFP expressing tissue using an immunocompetent mouse model without rejection. To exclude the possibility of minor rejection reaction we examined the expression of MHC class II molecules and other rejection mediators. Suppl. Table 2 shows that there is no significant upregulation of MHC class II molecules such as H2-Aa, H2-Ab1 and H2-Eb1 between allograft lesions and normal uterus. There is no significant expression of T cell activation markers CD69 and CD25 (Suppl. Figure 2 A) and the rejection-related cytokine IFN-γ (Suppl. Figure 2B). However, we observed that TNF-α is significantly expressed at low levels in macrophages (Suppl. Figure 2B), a phenomenon that is normally observed in endometriosis. These results suggest the absence of an alloimmune response and support immune infiltration linked to endometriosis pathology rather than rejection. To investigate the role of these host circulation-derived cells in the development and growth of endometriotic lesions, we performed the analysis of various genes that were expressed in host cell clusters. Figure  5 shows the violin plots revealing a significant increase in the density of endometriosis markers of Sf1 , Kras , Igf1 , and Mmp2 as showed in Fig.  5 A, B, C and D respectively. These genes have previously been implicated in the pathophysiology of endometriosis. Among these genes, Sf1 and Kras were widely expressed in most of the host immune cell types (B cells, macrophages, T cells, and neutrophils). Sf1 was also expressed in natural killer cells and endothelial cells. Igf1 was expressed only in two cell clusters (macrophages and fibroblasts) while Mmp2 was expressed only in the fibroblast cell cluster from the host. We further investigated the differential expression of cytokines that are responsible for inducing inflammation in endometriosis using violin density plots. Ccl8 , Ccl7 , Ccl4 , Ccl12 , Ccl2, Ccl5 , Cxcl12 , Cxcl16 , and Il1β have been shown to be upregulated in endometriosis as described above and were highly expressed in cell clusters that originated from the host in our study (Fig.  6 ). In addition, Cxcl21a is upregulated in our study as shown in Fig.  6 . Ccl8 , Ccl2 , Ccl7 , Ccl12, and Cxcl16 were exclusively expressed in macrophages and absent in other cell types. Ccl4 was expressed in natural killer cells and macrophages while Ccl5 was expressed in T cells and natural killer cells but absent in other cell types. The chemokine Cxcl12, a ligand for CXCR4 and CXCR7 chemokine receptors, was expressed only in fibroblasts and endothelial cells. Fig. 5 Violin Plots showing the expression levels of various endometriosis markers in different types of cells that express GFP. A Sf1, B Kras, C Igf1 D Mmp2 Fig. 6 Violin Plots showing the expression levels of various cytokines in different types of cells that express GFP. macrophages (green): Ccl8, Ccl7, Ccl12, Ccl2, Ccl4, Cxcl16 and Il1β. natural killer cells: Ccl4 & Ccl5. T cells: Ccl5. fibroblasts: Cxcl12. neutrophils: Ccl4 & Il1b. endothelial cells: Cxcl12 & Ccl21a Violin Plots showing the expression levels of various endometriosis markers in different types of cells that express GFP. A Sf1, B Kras, C Igf1 D Mmp2 Violin Plots showing the expression levels of various cytokines in different types of cells that express GFP. macrophages (green): Ccl8, Ccl7, Ccl12, Ccl2, Ccl4, Cxcl16 and Il1β. natural killer cells: Ccl4 & Ccl5. T cells: Ccl5. fibroblasts: Cxcl12. neutrophils: Ccl4 & Il1b. endothelial cells: Cxcl12 & Ccl21a The functions of cells in different cell clusters were analyzed by screening the top 10 genes that were expressed as shown in Fig.  7 . The violin plots in supplemental Fig. 3, 4, 5, 6, 7, 8 and 9 demonstrate the detailed distribution of expression levels of the top 10 genes that were expressed in NK cells, B cells, macrophages, T cells, fibroblasts, neutrophils, and endothelial cells, respectively. Klrb1c, Xcl1, IL2rb, and Gzmb expressed in NK cells (Suppl. Figure 3) have been associated with endometriosis while Cd19 has been implicated in the formation of tertiary structures and B cell recruitment (Suppl. Figure 4). Macrophages (Suppl. Figure 5) showed high expression of C1qb , C1qa , Aif1 , Mrc1 , C3ar1 , Cxcl16, and Lgmn genes, which have also been reported to play a role in the development of endometriosis. C1qb and C1qa are implicated in immune infiltration and correlated with tissue factor respectively in endometriosis. Icos and Cxcr6 expressed in the T cell cluster (Suppl. Figure 6) are known to play a role in endometriosis. Genes Dpt , Col1a2 , Aebp1 , Rarres2 , Dcn , Serping1 , and Col6a2 in fibroblasts (Suppl. Figure 7) as well as Hp , Trem1 , Slc7a11 , Tgm2 , and Fgr genes expressed in neutrophils (Suppl. Figure 8) are also involved in the pathogenesis of endometriosis. Hp is differentially expressed in advanced stages of endometriosis while Trem1 expression indicates impaired inflammatory reaction of the immune system. Slc7a11 is involved in the regulation of the miR-21-3p/p53/SLC7A11 signaling pathway. Tgm2 promotes cell proliferation, and Fgr may accelerate immune response to endometriosis. Cldn5 is abnormally expressed in patients with infertility and endometriosis, while Flt4 promotes cell proliferation (Suppl. Figure 9, endothelial cells). Fig. 7 Different cell types infiltered from host into lesions showing top 10 genes expression. Violin plots represent the expression levels of top 10 genes in each cell cluster. NK cells: Klrb1c, Nkg7, Cd7, Ctsw, Xcl1, Klre1, Il2rb, Klra8, Gzmb and, Klra4. B cells: Cd79a, Cd79b, Ly6d, H2-DMb2, Mzb1, Ms4a1, Cd19, Ebf1, H2-Ob and Fcmr. Macrophages: C1qb, C1qc, C1qa, Cxcl16, Aif1, Ms4a7, Mrc1, Lgmn, C3ar1 and Lyz2. T cells: Cd3e, Cd3d, Cd3g, Lat, Il7r, Icos, Thy1, Ramp3, Cxcr6 and Itk. Fibroblasts: Dpt, Serping1, Col1a2, Col6a2, Aebp1, C1s1, Rarres2, Serpinh1, Dcn and Col6a1. Neutrophils: Clec4e, Gsr, Hp, Msrb1, Trem1, F10, Slc7a11, Clec4d, Tgm2 and Fgr. Endothelial cells: Egfl7, Mmrn1, Cldn5, Tie1, Ccl21a, Cdh5, Flt4, Sdpr, Ptprb and Reln Different cell types infiltered from host into lesions showing top 10 genes expression. Violin plots represent the expression levels of top 10 genes in each cell cluster. NK cells: Klrb1c, Nkg7, Cd7, Ctsw, Xcl1, Klre1, Il2rb, Klra8, Gzmb and, Klra4. B cells: Cd79a, Cd79b, Ly6d, H2-DMb2, Mzb1, Ms4a1, Cd19, Ebf1, H2-Ob and Fcmr. Macrophages: C1qb, C1qc, C1qa, Cxcl16, Aif1, Ms4a7, Mrc1, Lgmn, C3ar1 and Lyz2. T cells: Cd3e, Cd3d, Cd3g, Lat, Il7r, Icos, Thy1, Ramp3, Cxcr6 and Itk. Fibroblasts: Dpt, Serping1, Col1a2, Col6a2, Aebp1, C1s1, Rarres2, Serpinh1, Dcn and Col6a1. Neutrophils: Clec4e, Gsr, Hp, Msrb1, Trem1, F10, Slc7a11, Clec4d, Tgm2 and Fgr. Endothelial cells: Egfl7, Mmrn1, Cldn5, Tie1, Ccl21a, Cdh5, Flt4, Sdpr, Ptprb and Reln

In this study, we carried out scRNA-seq analysis on endometriotic lesions and showed the infiltration of multiple cell types into endometriotic lesions in a murine model of endometriosis. For the first time we describe the relative contribution of grafted endometrium vs circulating host cells to the endometriotic lesion. We also determined the differential expression of endometriosis related genes, including multiple cytokines that are responsible for endometriosis associated inflammation. We clearly distinguished the origin of cell types either exogenously or endogenously in endometriotic lesions by identifying the presence or absence of Green Fluorescent Protein (GFP). The majority of cells in endometriosis do not arise from endometrial fragments corresponding to retrograde menstruation in humans. Most of the lesion is derived from circulating cells infiltrating into the lesion after initiation. Specifically, 65% of the cells in the lesions were not from the endometrium; only 35% of cells in lesions were derived from the endometrium. Fibroblasts, natural killer cells, macrophages, B cells, and endothelial cells were commonly derived from both origins. However, neutrophils and T- cells were found exclusively in non-GFP cells that were non-endometrial in origin, while stromal cells, stromal fibroblasts, enterocytes, and smooth muscle cells were nearly exclusively GFP-expressing indicating endometrial origin. Our findings regarding the presence of immune cells such as natural killer cells, B cells, T cells, and macrophages that were embedded into the lesions from the recipient mice through the circulation suggest a major role for the immune system rather than the endometrium in composing endometriosis lesions (Li et al. 2024 ; Marečková et al. 2024 ; Quan et al. 2024 ; Shin et al. 2023 ; Wu et al. 2024 ). The infiltration of endogenous cells, especially immune cells, into the lesions may either inhibit the progression of endometriosis or enhance the development of lesions. Our laboratory previously reported the involvement of the immune system including macrophages' transition from M1 to M2 (Habata et al. 2023 ), lymphocyte infiltration (A. Tal et al. 2021a , b ), and immune checkpoints (Chen et al. 2021 ) in the inhibition or pathogenesis of endometriosis (Habata et al. 2023 ; Lv et al. 2024 ). This study confirms the preponderance of immune cells in endometriosis. Recently, single-cell RNA sequencing (scRNA-seq) has been employed by others to explore the involvement of specific cell types in the pathophysiology of endometriosis. Wu et al. identified the activation of pro-inflammatory macrophages, NK cell exhaustion, and aberrant proliferation of IQCG +  and KLF2 +  epithelium in ovarian endometriomas in humans (Wu et al. 2024 ). The absence of a unique subcluster of uterine natural killer (uNK) cells compromised decidualization of cultured stromal cells that were enriched in pro-inflammatory, senescent cells and B cells from the menstrual effluents of endometriosis patients (Shih et al. 2022 ). Zhu et al. reported the abundance of myofibroblasts, pericytes, endothelial cells, and macrophages in ectopic endometrium (Zhu et al. 2023 ). Fonseca et al. demonstrated the dysregulation of pro-inflammatory pathways and the upregulation of complement proteins, pro-angiogenic factors, and pro-lymphangiogenic factors (Fonseca et al. 2023 ). Proinflammatory cytokine levels in the endometrial immune cells were higher in women with endometriosis (Huang et al. 2023 ). Stromal–epithelial cell coordination via transforming growth factor beta (TGFβ) signaling in the functionalis, as well as signaling among fibroblasts, decidualized stromal cells, and macrophages in the basalis, are dysregulated in endometriosis (Marečková et al. 2024 ). Zou et al. showed that immune dysfunction occurs in the peritoneal fluid (Zou et al. 2021 ), while Ma et al. described the fate of fibroblasts and heterogeneity in immune cells in women with endometriosis (Ma et al. 2021 ). Patients with endometriosis treated with oral contraceptives showed a coordinated transcriptional program that could influence immunotolerance and angiogenesis (Tan et al. 2022 ). All the above studies were carried out on human ectopic endometrial tissue where the origin of different cell types was unknown. Our sc-RNA-seq data sets confirm that some of the endogenous endometrial-derived cells expressed well characterized endometriosis associated molecules such as Kras, Igf-1, Mmp-2, and Il-1β. Kras is an oncogene, and its somatic activating mutations play a critical role in the growth and development of endometriosis (Anglesio et al. 2017 ; Orr et al. 2023 ; Vu et al. 2024 ; Yachida et al. 2021 ). Our data suggests that a significant increase in Kras expression by macrophages, T cells, B cells, and neutrophils derived from the host could play a role in the growth and development of endometriosis. Similar to our study, upregulated IGF-1 has previously been reported to have a role in regulating ectopic endometrial cell growth and proliferation through an autocrine-paracrine signaling by its anti-apoptotic and mitogenic actions (Blontzos et al. 2023 ; Heidari et al. 2021 ; Laganà et al. 2019 ). Here we also show differential MMP expression. MMPs overexpression contributes to the initial development of endometriosis. MMPs participate in the reconstruction of the extracellular matrix in endometrium implantation sites by promoting the penetration and invasion of ectopic tissues (Nisolle & Donnez 2019 ). Upregulation of MMP-2 in endometriosis is one of the key factors that help in the progression and pathogenesis of this disease (Kleimenova et al. 2024 ; Laronha & Caldeira 2020 ; Pitsos & Kanakas 2009 ). The differential expression of cytokines and chemokines in endometriosis are also essential for lesion progression. Among pro-inflammatory cytokines, interleukin-1β ( IL-1β ) is one of the most increased in women with endometriosis (Akoum et al. 2008 ; Peng et al. 2020 ). Similar results were observed in our scRNA-seq data, where macrophages and neutrophils expressed significant levels of Il-1β in non-GFP cells from endometriotic lesions. Many of the alterations in gene expression in endometriotic lesions are in the immune infiltrate rather than in the allograft. Most of the cells in endometriotic lesions are immune cells and the majority are not derived from the endometrium through retrograde menstruation. Endometriosis is clearly a disease of the immune system and immunotherapies have been reported for this disease (Kotlyar et al. 2021 ; Pluchino et al. 2020 ). Could immune therapies be used to treat endometriosis? All currently approved therapies are hormonal, targeting the auto-graft stromal cells, which are not the majority of cells in the lesion. Increased immune function has been described as a critical contributing mechanism in the development of endometriosis. Among immune cells, macrophages are abundant and play a key role in the growth of endometriotic lesions while promoting vascularization and cell proliferation and inducing pain symptoms. Recently, we reported that TET3-overexpressing macrophages as key pathogenic contributors to and therapeutic targets for endometriosis (Lv et al. 2024 ), demonstrating the potential of therapies that target the immune system. Though endometriosis is not a malignant disease, it resembles cancers in the deregulation of immune system. Tumor-associated macrophages (TAMs) significantly influence tumor dynamics and progression (Hambardzumyan et al. 2016 ). TAMs secrete different biological substances/molecules such as growth factors, chemokines, and cytokines, which regulate the pathophysiological mechanisms including tumor initiation, angiogenesis, proliferation, cancer cell migration, and immunosuppression (Pyonteck et al. 2013 ). Further, we are considering TAMs as potential prognostic markers and therapeutic targets as TAM subgroups are linked to resistance to radiotherapy with reduced survival time (Komohara et al. 2012 ). Due to the above characteristics of TAMs, they have gained much importance in cancer research for evaluating the efficacy of therapies targeting TAMs. Here we provide a more comprehensive understanding of disease specific macrophages in endometriosis (Dean & Hooks 2022 ). Immunotherapies such as adoptive cell therapy, immune checkpoint inhibitors, antibody treatment and vaccinations allow the immune system to recognize and fight malignant cells. These types of therapies are often used in cancers and may also be applicable to treat endometriosis. NK cells are a subset of lymphocytes that are reduced in patients with endometriosis, resulting in a favorable conditions for the survival and growth of ectopic endometrial tissue (Symons et al. 2018 ) similar to what has been reported in cancer (Bernal et al. 2009 ; Morcillo-Martín-Romo et al. 2025 ). Current studies in NK cell-based therapies have significantly enhanced their potential in cancer immunotherapy; these may be helpful in endometriosis as well. The findings suggest potential new therapeutic approaches. Enhanced immune cell presence in the lesion contradicts the notion that a simple potentiation of the immune response would be a valuable therapeutic approach. Both the blocking as well as the enhancement of the immune reaction have potential to treat endometriosis. Cell or cytokine specific therapies hold promise for endometriosis treatment as they are currently used for other inflammatory diseases. New targeted therapies could replace current treatment paradigms that rely on globally disrupting sex steroid signaling. As with the use of any model system, this study also has limitations including the lack of menses in mice which may influence immune cell recruitment in endometriosis. Additionally, while the allogenic transplant from the same strain of mice should preclude immune rejection, we cannot rule out that the C57BL/6 syngeneic uterine grafts might still trigger a subtle immune recognition difference despite the identical genotype, leading to an alloimmune response. This is unlikely given the lack of rejection of transplants in this strain using other organs and tissues and by lack of expression of key rejection mediators in our model. Another limiting factor in our study is pooling of single cell preparations from the lesions. Finally, our murine model looks at endometriosis that is relatively newly established, while endometriosis is not diagnosed in humans until years after initial establishment; immune populations likely shift over time.

Endometriotic lesions mostly contain cells that do not originate from the uterus but infiltrate endogenously from the circulation. The majority are immune cells, where we characterize the expression of genes that exemplify the endometriotic lesions. These results support the theory that endometriosis, while occurring in part due to retrograde menstruation, progresses as a complex lesion where the immune cells have a preponderant presence. The complex immune environment suggests alternative ways to treat the disease.

Endometriosis is an estrogen dependent chronic inflammatory gynecological disease (Taylor et al. 2021 ) that affects more than 10% of reproductive-age women (Parasar et al. 2017 ). These women often suffer from pelvic pain and infertility (Parasar et al. 2017 ; Zondervan et al. 2020 ). Endometriosis affects nearly 190 million women worldwide (Vannuccini et al. 2022 ; Zondervan et al. 2020 ) and costs nearly $120 billion annually in the United States alone (Ellis et al. 2022 ; Soliman et al. 2018 ). The disease is characterized by the presence of uterine endometrial tissue growing as discreet lesions outside the uterine cavity, most commonly on the ovaries, fallopian tubes, and the peritoneum (Taylor et al. 2021 ). The causes of endometriosis are not known yet; however, several theories have been postulated to explain its development. The most widely accepted is the retrograde menstruation theory, which hypothesizes that during menstruation some endometrial cells escape into the peritoneal cavity and attach to the pelvic peritoneum. There they induce inflammation, pain, and fibrosis (Smolarz et al. 2021 ). Inflammation induced by cytokines secreted from endometriotic lesions also regulate immune responses during the development of endometriosis (Zhou et al. 2019 ). Changes in cytokine levels and subsequent immunomodulation play an important role in the formation and development of endometriotic lesions (Crispim et al. 2021 ; Izumi et al. 2018 ; Taylor et al. 2021 ). Though many studies have reported the involvement of specific immune cells (Peng et al. 2024 ), the function and role played by the immune system in the pathophysiology of endometriosis is still not fully characterized. Factors secreted by immune cells, including macrophages and neutrophils, regulate cell proliferation, growth, and invasion while promoting vasculature remodeling and angiogenesis (Gao et al. 2023 ; Păvăleanu et al. 2023 ; Rocha et al. 2013 ). Recent studies using single-cell sequencing reported immune dysfunction in women with endometriosis (Zou et al. 2021 ), immune cell heterogeneity (Ma et al. 2021 ), immunotolerance, and angiogenesis. Current medical treatments are aimed at suppressing ovarian activity through hormonal treatments that stop menstruation. The most commonly used initial treatment is a progestin-containing oral contraceptives or progestin alone (Vercellini et al. 2016 ). Second line therapies, that include gonadotropin releasing hormone (GnRH) agonists, GnRH antagonists (Taylor et al. 2017 ), and aromatase inhibitors (Buggio et al. 2017 ) are aimed at suppressing ovarian activity to suppress menstruation. However, the symptoms will return following cessation of the treatment (Zondervan et al. 2020 ). Removing lesions by laparoscopic surgery relieves symptoms temporarily but the recurrence rates are very high. In an attempt to find more effective therapeutic options, it is important to dissect the molecular and cellular mechanisms, which play a key role in the growth and development of endometriosis, to facilitate the development of new therapeutic agents. Understanding the etiology and full cellular origin of endometriosis lesions is needed to optimize both preventive and therapeutic approaches. Previous reports from our laboratory (Fang et al. 2022 ; R. Tal et al. 2021a , b ) describing the engraftment of BMDCs to the uterus prompted us to investigate the role and contribution of the endogenous host cells that are recruited or engrafted into lesions in the growth and development of endometriosis. In previous studies we were able to track large numbers of GFP cells derived from bone marrow after bone marrow transplantation (Chen et al. 2021 ; A. Tal et al. 2021a , b ; Zhou et al. 2011 ). To test our hypothesis, we modeled endometriosis using a green fluorescent protein (GFP) expressing syngeneic allograft. Cells in the allograft that contribute to the endometriosis lesion are GFP +ve and we will refer to them as allografted endometrium. These endometriotic lesions are also infiltrated by cells from the circulation or surrounding tissues of the host, which contribute to the development of the endometriotic lesion; these cells are GFP-negative and will be referred to as host derived cells or circulating host cells. We demonstrate that ectopic endometriotic lesions contain both uterine tissue as well as a surprisingly large number of cells derived endogenously from the circulation.

Supplementary Material 1. Supplementary Fig. 1. Expression of marker genes used to identify each cell type. Supplementary Material 2. Supplementary Fig. 2. Violin plots showing expression of T cell activation markers: CD25 and CD69. Rejection-related cytokines: IFN-γ and TNF-α. Supplementary Material 3. Supplementary Fig. 3. Non-GFP Natural Killer cells. Violin plots showing the expression of top 10 genes: Klrb1c, Nkg7, Cd7, Ctsw, Xcl1, Klre1, Il2rb, Klra8, Gzmb and Klra4. Supplementary Material 4. Supplementary Fig. 4. Non-GFP B cells. Violin plots showing the expression of top 10 genes: Cd79a, Cd79b, Ly6d, H2-DMb2, Mzb1, Ms4a1, Cd19, Ebf1, H2-Ob and Fcmr. Supplementary Material 5. Supplementary Fig. 5. Non-GFP macrophages. Violin plots showing the expression of top 10 genes: C1qb, C1qc, C1qa, Cxcl16, Aif1, Ms4a7, Mrc1, Lgmn, C3ar1 and Lyz2. Supplementary Material 6. Supplementary Fig. 6. Non-GFP T cells. Violin plots showing the expression of top 10 genes: Cd3e, Cd3d, Cd3g, Lat, Il7r, Icos, Thy1, Ramp3, Cxcr6 and Itk. Supplementary Material 7. Supplementary Fig. 7. Non-GFP fibroblasts. Violin plots showing expression of top 10 genes: Dpt, Serping1, Col1a2, Col6a2, Aebp1, C1s1, Rarres2, Serpinh1, Dcn, and Col6a1. Supplementary Material 8. Supplementary Fig. 8. Non-GFP neutrophils cells. Violin plots showing expression of top 10 genes: Clec4e, Gsr, Hp, Msrb1, Trem1, F10, Slc7a11, Clec4d, Tgm2, and Fgr. Supplementary Material 9. Supplementary Fig. 9. Non-GFP endothelial cells. Violin plots showing expression of top 10 genes: Egfl7, Mmrn1, Cldn5, Tie1, Ccl21a, Cdh5, Flt4, Sdpr, Ptprb and Reln. Supplementary Material 10. Supplementary Table 1: Marker genes used for identifying cell types in total cells. Supplementary Material 11. Supplementary Table 2: MHC II expression was not elevated in the endometriosis lesions compared to normal uterine endometrium indicating lack of rejection. Supplementary Material 1. Supplementary Fig. 1. Expression of marker genes used to identify each cell type. Supplementary Material 2. Supplementary Fig. 2. Violin plots showing expression of T cell activation markers: CD25 and CD69. Rejection-related cytokines: IFN-γ and TNF-α. Supplementary Material 3. Supplementary Fig. 3. Non-GFP Natural Killer cells. Violin plots showing the expression of top 10 genes: Klrb1c, Nkg7, Cd7, Ctsw, Xcl1, Klre1, Il2rb, Klra8, Gzmb and Klra4. Supplementary Material 4. Supplementary Fig. 4. Non-GFP B cells. Violin plots showing the expression of top 10 genes: Cd79a, Cd79b, Ly6d, H2-DMb2, Mzb1, Ms4a1, Cd19, Ebf1, H2-Ob and Fcmr. Supplementary Material 5. Supplementary Fig. 5. Non-GFP macrophages. Violin plots showing the expression of top 10 genes: C1qb, C1qc, C1qa, Cxcl16, Aif1, Ms4a7, Mrc1, Lgmn, C3ar1 and Lyz2. Supplementary Material 6. Supplementary Fig. 6. Non-GFP T cells. Violin plots showing the expression of top 10 genes: Cd3e, Cd3d, Cd3g, Lat, Il7r, Icos, Thy1, Ramp3, Cxcr6 and Itk. Supplementary Material 7. Supplementary Fig. 7. Non-GFP fibroblasts. Violin plots showing expression of top 10 genes: Dpt, Serping1, Col1a2, Col6a2, Aebp1, C1s1, Rarres2, Serpinh1, Dcn, and Col6a1. Supplementary Material 8. Supplementary Fig. 8. Non-GFP neutrophils cells. Violin plots showing expression of top 10 genes: Clec4e, Gsr, Hp, Msrb1, Trem1, F10, Slc7a11, Clec4d, Tgm2, and Fgr. Supplementary Material 9. Supplementary Fig. 9. Non-GFP endothelial cells. Violin plots showing expression of top 10 genes: Egfl7, Mmrn1, Cldn5, Tie1, Ccl21a, Cdh5, Flt4, Sdpr, Ptprb and Reln. Supplementary Material 10. Supplementary Table 1: Marker genes used for identifying cell types in total cells. Supplementary Material 11. Supplementary Table 2: MHC II expression was not elevated in the endometriosis lesions compared to normal uterine endometrium indicating lack of rejection.