Abstract Background Intrauterine adhesions (IUA) is one of the most common gynecological diseases and main causes of uterine infertility. Among proposed hypotheses on IUA development, the reduced endometrial regeneration resulting from loss of functional stem cells has been proposed as the key factor affecting the IUA prognosis. However, the underlying mechanisms mostly remain unclear. Because the eMSCs (endometrial mesenchymal stem/stromal cells) play a critical role in both supporting the gland development and also preparing the environment for embryo implantation through decidualization, the characteristics and functions were compared between the eMSCs derived from IUA and non-IUA patients, to uncover the important roles of eMSCs in IUA and also the underlying mechanisms. Methods Endometrium biopsies were collected from IUA patients and controls. The fibrosis features and eMSC distributions were investigated with IHC (immunohistochemistry). Then the eMSCs were isolated and their functions and characteristics were analyzed in vitro. Results Our results indicate that the scar tissues in IUA are characterized with hyper-activation of pro-fibrotic fibroblast and myo-differentiation, along with reduced number of eMSCs. The isolated eMSCs from IUA and controls show similar functions from the perspectives of cell morphology, proliferation, colony formation, exosome secretion, positive ratio of eMSC markers and conventional MSC markers, tri-differentiation efficiency, the ability of suppressing lymphocyte proliferation, cell aging, and promoting vascular tube formation. However, the eMSCs from IUA have reduced levels of decidualization and higher levels of cell migration, invasion, and also myofibroblast differentiation. Further investigations indicate that the TGF-β pathway, which is the major inducer of myofibroblast differentiation, is up-regulated and responsible for the enhanced myofibroblast differentiation potential of eMSCs from IUA. Conclusions In conclusion, we have demonstrated here that the scar tissues in IUA biopsy are characterized with enhanced differentiation of pro-fibrotic fibroblast and myofibroblast. The number of eMSCs is reduced in IUA tissues, and their myofibroblast differentiation capability is increased. Supplementary Information The online version contains supplementary material available at 10.1186/s13287-025-04183-y. Keywords: Intrauterine adhesions, Endometrial mesenchymal stem/stromal cells (eMSCs), Myofibroblast differentiation, IUA Background Intrauterine adhesions (IUA, also known as, Asherman's Syndrome) is one of the most common gynecological diseases and main causes of uterine infertility, in which the walls of the uterine cavity adhere and fuse together, leading to hypo-menorrhea, dysmenorrhea, amenorrhea, and even infertility [[40]1, [41]2]. IUA severely affects the health, life quality, and also the pregnancy outcomes in women [[42]3]. Patients with IUA have higher risks of pregnancy complications, such as recurrent miscarriage, premature delivery, abnormal placentation and placenta implantation, intrauterine growth restriction, and preterm premature rupture of membranes [[43]1, [44]2, [45]4–[46]8]. Although the incidence of IUA is only around 1.5% in all women, the ratio rises significantly up to 40% in women with the history of postpartum uterine curettage. And the induced abortion or curettage is the dominant risk factor of IUA. Because of the increased number of induced abortion and curettage, the prevalence of IUA has increased widely during the past decades [[47]9]. Currently, the trans-cervical resection of adhesion (TCRA) is the standard treatment for IUA. However, effective therapies are still missing, especially for patients with moderate to severe IUA [[48]10]. Their recurrence ratio is up to 60%, and pregnancy rate is down to around 20% [[49]4–[50]9]. Although different approaches have been developed to treat IUA after the surgery, such as placing an intrauterine device (IUD)/Foley’s catheter balloon/hyaluronic acid or estrogen/progesterone treatment, minimal progress has been achieved [[51]11, [52]12]. Therefore, novel effective therapy strategies for IUA are urgently needed [[53]10]. Understanding the basic biological mechanisms of uterus under both physiological and pathological conditions, such as menstruation and IUA, is essential to develop novel therapeutic treatments [[54]13, [55]14]. Trauma to the endometrial basal layer or infection has been proposed as the main cause of IUA [[56]3, [57]11, [58]15]. Endometrial fibrosis, along with excessive myofibroblast differentiation, is one of those major pathological features in IUA [[59]3, [60]11, [61]15–[62]17]. In IUA scar tissues, the healthy stroma cells are invaded by myofibroblasts and filled with collagen enriched fibers; the functional glands become less structured with reduced proliferation abilities and less responding to the hormone stimulation. The precisely mechanisms of IUA development remain unclear. However, several proposed hypotheses have been demonstrated, including hypoxia, inflammation, dysregulated EMT (epithelial-mesenchymal transition) and MET (mesenchymal-epithelial transition), enhanced differentiation and hyper-activation of myofibroblast, loss of functional stem cells, and so on. Among them, the reduced endometrial regeneration resulting from loss of functional stem cells has been proposed as the key factor affecting the IUA prognosis [[63]14, [64]18, [65]19]. There are several types of stem/progenitor cells reside in endometrium [[66]20–[67]22]. Among them, eMSCs (endometrial mesenchymal stem/stromal cells) play a critical role in both supporting the gland development and also preparing the environment for embryo implantation through decidualization [[68]23–[69]26]. And serval types of eMSCs have been demonstrated, such as SUSD2^+, CD146^+, CD140β^+, and side populations [[70]20, [71]22]. Furthermore, inflammation is one of those key contributors in scar formation and IUA development [[72]27–[73]29], while MSCs have both immune suppressive and regenerative functions [[74]30]. Therefore, in the current study, the characteristics and functions were compared between the eMSCs derived from IUA and non-IUA patients, to uncover the important roles of eMSCs in IUA and the underlying mechanisms. Methods Participants and tissue sampling This study was approved by the ethics committee of Shenzhen Zhongshan Obstetrics & Gynecology Hospital (formerly Shenzhen Zhongshan Urology Hospital) and followed the tenants of the Declaration of Helsinki. Informed consent was obtained from each participant. Participants were recruited from September 2023 to June 2024 in the fertility center. Patients with moderate to severe IUA (the classification criteria of The American Fertility Society in 1988), confirmed by hysteroscopy examinations and pathological results, were assigned as IUA group. Patients undergoing embryo transplantation without any endometrium and myometrium lesions, confirmed by hysteroscopy examinations and pathological results, were assigned as control group. There were totally 10 IUA patients and 10 controls were recruited in the current study. Histologic examinations Scar or control endometrium biopsies were collected during TCRA surgery for IUA patients or regular endometrium function assessment before embryo transplantation. Tissues were washed 3 times with 0.9% saline solution in the operating room, and then fixed in 10% neutral buffered formalin (Sigma) for HE (hematoxylin–eosin staining) and IHC (immunohistochemistry) examinations, or incubated in DMEM/F12 (Gibco) plus 20% FBS (fetal bovine serum, Gibco) and 2% penicillin–streptomycin (Gibco) on ice for cell isolation. HE and IHC were performed with automatic Tissue Processor Thermo Scientific Revos (Thermo Scientific), BOND-RX Multiplex IHC Stainer (Leica Biosystems), SLIDEVIEW VS200 slide scanner (Olympus), and quantitative image analysis platform HALO (Indica Labs) according to the instructions. Antibodies were listed in Supplementary Table 1 (Table S1). Masson staining was conducted with the Masson`s Trichrome Stain Kit (Solarbio). Cell isolation, expansion and characterization The eMSCs were derived from endometrium biopsy as described previously [[75]31–[76]33, [77]33, [78]34, [79]34, [80]35, [81]35]. Briefly, the biopsy was minced, digested with 1 mg/mL collagenase B (STEMCELL Technologies) for 1 h, filtered with 70 μm cell strainer (Corning), and expanded with DMEM/F12 plus 10% FBS and 1% platelet lysate (Sigma), 2 U/ml heparin, and antibiotics (100 units/ml of penicillin and 100 μg/ml of streptomycin). The eMSCs were passaged with TrypLE (Thermo Scientific) and stimulated with 10 ng/ml IFN-γ plus 10 ng/ml TNF-α (PeproTech). The differentiation and characterization were performed with a StemPro® Adipogenesis Differentiation Kit (Gibco), StemPro® Osteogenesis Differentiation Kit (Gibco), and StemPro® Chondrogenesis Differentiation Kit (Gibco) as described previously [[82]31–[83]33, [84]33, [85]33, [86]34, [87]34, [88]34, [89]35, [90]35, [91]35]. The immune suppression assessment was performed with CellTrace™ CFSE (Thermo Fisher Scientific) and Dynabeads® Human T-Activator CD3/CD28 (Thermo Fisher Scientific) as described previously [[92]31]. Cell aging was performed with the Senescence β-Galactosidase Staining Kit (Beyotime) according to the instructions. Cell migration and invasion were conducted as described previously [[93]36]. Exosome purification and vascular tube formation Medium for exosome production was centrifuged at 120,000 g for 140 min to eliminate the exosomes pre-existed in the medium. Then the eMSCs were maintained in exosome-free medium for 48 h. Exosomes from eMSCs were extracted as follows: 10,000 g for 60 min to eliminate the cells and debris; 120,000 g for 140 min to precipitate the exosomes. The concentration and size of the exosome were characterized with the NanoSight NS300 (Malvern Panalytical). For vascular tube formation, 8 × 10^4 HUVEC (human umbilical vein endothelial cells, Thermo Fisher Scientific) were plated on Matrigel coated p24 plate with/without 5 × 10^9 exosomes. Tube formation was pictured 3 h later. Myofibroblast differentiation The eMSCs were expanded in DMEM/F12 plus 2% FBS for 3 days. Then they were seeded onto p24 plate (2 × 10^4 cells/well), and treated with/without 10 ng/mL TGF-β1 for 4 days. The myofibroblast differentiation was visualized with anti-α-SMA immunofluorescence and quantified with fluorescence intensity with ImageJ. Decidualization induction The eMSCs were plated onto p12 plates with 5 × 10^4 cells/well. 24 h later, the medium was refreshed with DMEM/F12 plus 2% FBS and 10 nM E2 for 48 h. Then the medium was replaced with decidualization induction medium (DMEM/F12 containing 2% FBS, 10 nM E2, 0.5 mM 8-Br-cAMP, 1 μM Medroxyprogesterone acetate) for 96 h. Statistics Data are shown as the mean ± SEM (standard error of the mean) and were analyzed with GraphPad Prism 8 for Windows. Student’s t test was applied to the two-group comparison. One-way ANOVA was applied to the multiple group comparison. P < 0.05 indicates statistical significance. Results Fibrosis features of IUA samples To uncover the critical roles of eMSCs in the pathology of IUA, 10 biopsies from IUA patients and 10 from controls were recruited (Fig. [94]1). Among 10 IUA samples, 7 samples have both the scar region and also functional endometrium, providing the opportunity to comparing the scar tissue with functional endometrium within the same sample (Fig. [95]1A). Based on the relative locations to the scars and the distribution pattern comparing with the healthy controls, the IUA sample could be divided into healthy region—location No. 1, fibrosis region—location No. 2, peri-scar region—location No. 3, and scar region—location No. 4 (Fig. [96]1A, B). Masson staining showed the classical pattern in which higher level of fibrosis around the scars (location No. 2—4 in the IUA sample) and less fibrosis in the healthy region (location No. 1 in the IUA sample) and healthy samples (Fig. [97]1A, B). Interestingly, the classical fibrosis marker COL1A1, which has been well demonstrated in other organs [[98]37, [99]38], didn`t show obvious differences among location No. 1–4 and the healthy control (Fig. [100]1B). Furthermore, statistical analysis showed that the expression level of COL1A1 was significantly higher in healthy controls (Fig. [101]1C), and no differences between the scar regions and non-scar regions within the IUA samples (Fig. [102]1D). Thus, the fibrosis in IUA might differ from the fibrosis process in other organs, at least from the perspective of the type of collagen deposited in scars. Fig. 1. [103]Fig. 1 [104]Open in a new tab Fibrosis features of IUA samples. A Representative Masson staining of biopsies from IUA patient with healthy region-location No. 1, fibrosis region-location No. 2, peri-scar region-location No. 3, and scar region-location No. 4. B Representative Masson staining and IHC analysis of COL1A1, S100A4, α-SMA, and CD31 in biopsies from IUA patients and controls. C Expression levels of COL1A1, S100A4, α-SMA, and CD31 were analyzed between samples from IUA patients and controls with HALO platform (n = 10). D Expression levels of COL1A1 (n = 5 for scar region; n = 3 for non-scar region), S100A4 (n = 6 for scar region; n = 5 for non-scar region), α-SMA (n = 5 for scar region; n = 5 for non-scar region), and CD31 (n = 5 for scar region; n = 5 for non-scar region) were analyzed between scar region and non-scar region within biopsies from IUA patients with HALO platform. IUA, intrauterine adhesions; COL1A1, collagen type I alpha 1 chain; S100A4, S100 calcium binding protein A4; α-SMA, alpha-smooth muscle actin; CD31, platelet and endothelial cell adhesion molecule 1; N.S., no significant. * indicates P < 0.05. Scale bar, 200 μm Fibrotic fibroblast has been demonstrated as the major inducer of fibrosis and effector in collagen deposit and scar formation [[105]19, [106]37]. And the pro-fibrotic fibroblast marker S100A4 had similar distribution patterns with Masson staining (Fig. [107]1B, location No. 4). However, it also more clearly showed that the S100A4 was more enriched in the blood vessels in the scar region (Fig. [108]1B). Although the statistical analysis showed that the IUA samples had higher level of S100A4, there was no significant differences between the scar regions and non-scar regions (Fig. [109]1C, D). Interestingly, the smooth muscle marker α-SMA (alpha smooth muscle actin) was specifically expression in the scar region (Fig. [110]1B), and also significantly higher expressed in the IUA samples and scar regions within the IUA samples (Fig. [111]1C, D). Therefore, the α-SMA might be a more suitable marker for IUA rather than COL1A1. Hypoxia resulting from insufficient blood supply has been suggested as one contributor in IUA development, which also might regulate the functions of stem cells and MSCs [[112]3]. However, our data showed that the scar tissue also had sufficient blood vessels, identified with CD31 (Fig. [113]1B). And there was no significant difference between IUA samples and healthy controls, and also between the scar regions and non-scar regions (Fig. [114]1C, D; Figure S1). Thus, the scar tissues might also have sufficient blood and oxygen supply. In summary, scar in IUA is characterized as fibrosis. Pro-fibrotic fibroblast and myo-differentiation might be the major contributor in IUA development. COL1A1 deposit and lack of oxygen supply might not be the basic characteristics or inducer in IUA. Decreased eMSCs in IUA Then the markers of eMSCs (SUSD2, CD146, CD140β) and also the conventional markers of MSCs (CD73, CD90, CD105) were analyzed in IUA and control samples (Fig. [115]2). SUSD2 was barely detected in stroma, and it was mainly expressed around blood vessels in both IUA and control samples (Fig. [116]2A). Higher expression level of SUSD2 was detected in IUA samples while there was no significant difference between scar regions and non-scar regions (Fig. [117]2B, C). And the higher level of SUSD2 in IUA samples might result from more arteries existed in the scar regions (Fig. [118]2A–C). Indeed, the SUSD2 was used to label the eMSCs surrounding the spiral arteries in the endometrium [[119]20]. CD146 was enriched in stroma, blood vessels, and also epithelium (Fig. [120]2A). And its expression levels were not significantly different between IUA samples and controls, and also between scar regions and non-scar regions (Fig. [121]2B, C). CD140β was mainly expressed in stroma, and its expression level was significantly down-regulated in scar regions when comparing with the non-scar regions in IUA samples (Fig. [122]2B, C). However, the total expression level of CD140β in IUA samples was not significantly different from controls (Fig. [123]2B, C). Therefore, the eMSC marker CD140β might be involved in IUA development. Fig. 2. [124]Fig. 2 [125]Open in a new tab Distribution of eMSCs in IUA. A Representative IHC analysis of SUSD2, CD146, CD140β, CD73, CD90, and CD105 in biopsies from IUA patients and controls. B Expression levels of SUSD2, CD146, CD140β, CD73, CD90, and CD105 were analyzed between biopsies from IUA patients and controls with HALO platform (n = 10). C Expression levels of SUSD2 (n = 7 for scar region; n = 3 for non-scar region), CD146 (n = 5 for scar region; n = 5 for non-scar region), CD140β (n = 5 for scar region; n = 5 for non-scar region), CD73 (n = 6 for scar region; n = 5 for non-scar region), CD90 (n = 8 for scar region; n = 5 for non-scar region), and CD105 (n = 8 for scar region; n = 7 for non-scar region) were analyzed between scar region and non-scar region within biopsies from IUA patients with HALO platform. IUA, intrauterine adhesions; SUSD2, sushi domain containing 2; CD146, melanoma cell adhesion molecule; CD140β, platelet derived growth factor receptor beta; CD73, ecto-5'-nucleotidase; CD90, Thy-1 cell surface antigen; CD105, endoglin; N.S., no significant. * indicates P < 0.05. Scale bar, 200 μm CD73 was enriched in stroma and epithelium, while the CD90 was mainly expressed in stroma and blood vessels (Fig. [126]2A). The expression levels of both CD73 and CD90 were not significantly different between the IUA samples and controls, and also between scar regions and non-scar regions (Fig. [127]2B, C). Similar to the expression pattern of CD90, CD105 also was mainly expressed in stroma and blood vessels (Fig. [128]2A). In contrast, the expression level of CD105 was significantly down-regulated in IUA samples when comparing to controls (Fig. [129]2B, C). However, no significant difference was detected between scar regions and non-scar regions within IUA samples (Fig. [130]2B, C). Thus, the conventional MSC marker CD105 might be involved in IUA development. In summary, the eMSC marker CD140β and conventional MSC marker CD105 are reduced in IUA, indicating that the reduced number of eMSCs might contribute to the IUA development. Increased myofibroblast differentiation of eMSCs from IUA To further characterize the functions of eMSCs, they were isolated, expanded and characterized in vitro. Data showed that there were no significant differences between eMSCs from IUA and controls, from the perspectives of cell morphology (Fig. [131]3A), proliferation (Fig. [132]3B), colony formation (Fig. [133]3C, D), exosome secretion (Fig. [134]3E), positive ratio of eMSC markers (Fig. [135]3F) and conventional MSC markers (Fig. [136]3G), tri-differentiation efficiency (Fig. [137]3H, [138]I), and also the abilities of suppressing lymphocyte proliferation (Fig. [139]3J, K). Thus, it seems that there is no significant difference between the eMSCs from IUA and controls, from the perspectives of conventional MSC functions. Fig. 3. [140]Fig. 3 [141]Open in a new tab Characterization of eMSCs—MSC related functions. A Representative cell morphology of eMSCs from IUA patients and controls. B Cell proliferation of eMSCs was determined by cell number counting (n = 5). C Representative figures of colony formation. D Colony number of eMSCs was determined by counting after stained with crystal violet (n = 5). E Exosome number secreted from eMSCs was determined by NanoSight NS300 (n = 5). F Ratio of SUSD2, CD146, and CD140β positive cells within eMSCs was determined by flow cytometry (n = 5). G Ratio of CD73, CD90, CD105, CD45, CD34, CD19, CD11b, and HLA-DR positive cells within eMSCs was determined by flow cytometry (n = 5). H Representative images of adipocyte differentiation stained with Oil Red O, osteocyte differentiation stained with Alizarin Red, chondrocyte differentiation stained with Alcian blue before and after sectioning. I Chondrogenic sphere forming efficiency analysis (n = 10). J PBMC proliferation assay after coculture with eMSCs stimulated without/with 10 ng/ml TNF-α and 10 ng/ml IFN-γ for 48 h (n = 5). K Representative results of PBMC proliferation assay after coculture with eMSCs stimulated without/with 10 ng/ml TNF-α and 10 ng/ml IFN-γ for 48 h, determined by flow cytometry. IUA, intrauterine adhesions; SUSD2, sushi domain containing 2; CD146, melanoma cell adhesion molecule; CD140β, platelet derived growth factor receptor beta; CD73, ecto-5'-nucleotidase; CD90, Thy-1 cell surface antigen; CD105, endoglin; sti-, stimulated; N.S., no significant. * indicates P < 0.05 Differing to other types of MSCs, eMSCs also have their uterus-specific functions. Although there were no significant differences from the perspectives of cell aging (Fig. [142]4A, B) and promoting vascular tube formation (Fig. [143]4C, D), the eMSCs from IUA had significant impaired functions of decidualization in vitro (Fig. [144]4E, F). Furthermore, they also had enhanced cell migration (Fig. [145]4G, H), invasion (Fig. [146]4I, J), and also myofibroblast differentiation (Fig. [147]4K, L). Therefore, the eMSCs from IUA have enhanced myofibroblast differentiation capabilities and impaired decidualization potentials. Fig. 4. [148]Fig. 4 [149]Open in a new tab Characterization of eMSCs—endometrium related functions. A Representative figures of cell aging, analyzed through measuring the SA-β-gal activity. B The ratio of aging cells, analyzed through measuring the SA-β-gal activity (n = 42). C Representative figures of vascular tube formation mediated by exosome from eMSCs. D The relative number of branches and total length of vascular tube formation mediated by exosome from eMSCs (n = 28). E Representative figures of eMSC decidualization induced by 8-Br-cAMP plus MPA. F The relative mRNA levels of decidualization marker gene IGFBP1 and PRL, determined by qPCR (n = 15). G Representative figures of cell migration of eMSCs. H The relative area of cell migrated, analyzed by ImageJ (n = 18). I Representative figures of cell invasion of eMSCs stained with crystal violet. (J) The invaded cell number of eMSC, in the transwell-Matrigel assay and analyzed by ImageJ (n = 30). K Representative figures of myofibroblast differentiation of eMSCs induced by TGF-β1 and visualized through immunofluorescence anti-α-SMA. L The ability of myofibroblast differentiation of eMSCs, determined by measuring the fluorescence density of anti-α-SMA and analyzed by ImageJ (n = 50). IUA, intrauterine adhesions; SA-β-gal, senescence-associated beta-galactosidase; hVECs, human vascular endothelial cells; IGFBP1, insulin like growth factor binding protein 1; PRL, prolactin; TGF-β1, transforming growth factor-β1; N.S., no significant. * indicates P < 0.05 Hyper-activation of TGF-β pathway in eMSCs from IUA It has been demonstrated that the enhanced myofibroblast differentiation induces fibrosis development, and the TGF-β pathway is the major mediator [[150]37, [151]38]. However, the precise roles of TGF-β pathway in the uterus, especially during the IUA development, remain uncovered. Therefore, the bulk RNA-seq data of clinical IUA samples and healthy controls were downloaded from NCBI database (PRJNA870310, PRJNA916532). After GC-content normalization with EDASeq package in R, there were totally 1810 up-regulated and 1398 down-regulated genes in IUA samples when compared with the healthy controls (Fig. [152]5A, B; Table S2, S3). Although both GO (gene ontology) and KEGG (kyoto encyclopedia of genes and genomes) analyses didn`t show the enrichment of TGF-β pathway (Fig. [153]5C, D), TGF-β pathway had slightly up-regulated in IUA transcriptome (Fig. [154]5E). Therefore, the TGF-β pathway is hyper-activated in eMSCs from IUA, which might contribute to the enhanced myofibroblast differentiation of eMSCs in IUA. Fig. 5. [155]Fig. 5 [156]Open in a new tab Bioinformatic analysis of IUA samples. A Biasplot before and after GC-content normalization; B Volcano plot showing the DEGs in eMSCs from IUA patients and controls. C GO enrichment analysis of DEGs in eMSCs from IUA patients and controls. D KEGG enrichment analysis of DEGs in eMSCs from IUA patients and controls. E TGFβ pathway enrichment analysis among DEGs in eMSCs from IUA patients and controls. DEGs, differentially expressed genes; GO, gene ontology; KEGG, kyoto encyclopedia of genes and genomes; TGF-β, transforming growth factor-β To further investigate the underlying mechanism of enhanced myofibroblast differentiation of eMSCs from IUA, the eMSCs isolated from IUA patients and controls, treated with/without TGF-β1, were subjected to RNA sequencing. Data showed that there were only 27 DEGs (differentially expressed genes) between eMSCs from IUA and controls (Fig. [157]6A, B); and 19 DEGs after myofibroblast differentiation induced by TGF-β1 (Fig. [158]6A, B). Differing to the RNA-seq data on clinical samples containing different types of cells (Fig. [159]5C, D), the GO analysis of eMSC transcriptome showed that the collagen-containing extracellular matrix modification was enriched (Fig. [160]6C). Furthermore, the TGF-β pathway was enriched in the KEGG analysis, especially after myofibroblast differentiation induced by TGF-β1 (Fig. [161]6D). Among the three genes enriched in TGF-β pathway, FMOD (Fibromodulin) was up-regulated in eMSCs from IUA (Fig. [162]6E). Furthermore, the expression level of FMOD was significantly down-regulated after TGF-β1 treatment while TGFB2 was up-regulated (Fig. [163]6E), indicating that FMOD is an important responder and regulator of TGF-β pathway. FMOD, which is preferentially expressed in scarless foetal wound healing, could reduce scar formation through up-regulating α-SMA and reducing type I collagen expression [[164]39–[165]41]. Therefore, it is possible that the up-regulated FMOD is responsible for the enhanced myofibroblast differentiation of eMSCs in IUA. Fig. 6. [166]Fig. 6 [167]Open in a new tab Transcriptome analysis of eMSCs. The eMSCs from IUA patients and controls were treated with/without 10 ng/mL TGF-β1 for 4 days. A Volcano plot showing the DEGs in eMSCs. B Heatmap showing the DEGs in eMSCs. C GO enrichment analysis of DEGs in eMSCs. D KEGG enrichment analysis of DEGs in eMSCs. E The TPM value of BMP6, TGFB2, and FMOD in eMSCs. DEGs, differentially expressed genes; GO, gene ontology; KEGG, kyoto encyclopedia of genes and genomes; TGF-β, transforming growth factor-β; BMP6, bone morphogenetic protein 6; TGFB2, transforming growth factor β2; FMOD, Fibromodulin; TPM, transcript per kilobase per million mapped reads; N.S., no significant. * indicates P < 0.05 In summary, we have demonstrated here that the scar tissues in IUA biopsy are characterized with enhanced differentiation of pro-fibrotic fibroblast and myofibroblast. The number of eMSCs is reduced in IUA tissues, and their myofibroblast differentiation capability is enhanced. Discussion IUA is clinically characterized as endometrial fibrosis, along with excessive myofibroblast differentiation [[168]3, [169]11, [170]15–[171]17]. Myofibroblast differentiation from stromal cells is essential and beneficial to repair the wound in non-uterine organs [[172]37, [173]38]. Smooth muscle actin (SMA) expression and contraction would reduce the size of the wound; type I collagen deposit could stabilize the wound region and form barriers, separating the wound area from environment; at the resolution stage, the myofibroblast would be cleared through apoptosis or de-differentiation/trans-differentiation, the fibrous collagen-rich ECM is eliminated by MMPs and the wound is closed by cell regeneration [[174]19, [175]37, [176]38]. However, persistent myofibroblast differentiation or resistance to apoptosis would result in myofibroblast over-activation and scar formation, also known as fibrosis, a lethal condition [[177]19, [178]37, [179]38]. Similar to the fibrosis occurred in other tissues (such as liver, heart, kidney, and skin), myofibroblast differentiation/hyper-activation has also been proposed as the major contributor to IUA development [[180]14, [181]18–[182]22]. Myofibroblast could be differentiated from mesenchymal stromal cells directly or MSCs trans-differentiated from epithelial cells (epithelial–mesenchymal transition, EMT), endothelial cells (endothelial–mesenchymal transition, EndMT), or circulating macrophages [[183]37]. In addition to the important roles of MSCs in fibrosis, the endometrial MSCs (eMSCs) also play an important role in both supporting the gland development and also preparing the environment for embryo implantation through decidualization [[184]23–[185]26]. And the reduced endometrial regeneration resulting from loss of functional stem cells has been proposed as the key factor affecting the IUA prognosis [[186]14, [187]18–[188]22]. Thus, in the current study, the characteristics and functions were compared between the eMSCs derived from IUA patient and healthy controls, to uncover the important roles of eMSCs in IUA and also the underlying mechanisms. Our results indicate that the scar tissues in IUA are characterized with hyper-activation of pro-fibrotic fibroblast and myo-differentiation rather than COL1A1 deposit and lack of oxygen supply, along with reduced number of eMSCs, which is in accordance with previously investigations [[189]16, [190]42, [191]43]. Interestingly, the S100A4 expression pattern seems like more specific than conventional COL1A1 in IUA, although the statistical analysis didn`t show significant differences. Further study with more clinical samples is necessary to validate whether the S100A4 is a more specific fibrosis marker in IUA. The isolated eMSCs from IUA and controls show similar functions from the perspectives of cell morphology, proliferation, colony formation, exosome secretion, positive ratio of eMSC markers and conventional MSC markers, tri-differentiation efficiency, the ability of suppressing lymphocyte proliferation, cell aging and promoting vascular tube formation. However, the eMSCs from IUA have reduced levels of decidualization and higher levels of cell migration, invasion, and also myofibroblast differentiation. Further investigations indicate that the TGF-β pathway, which is the major inducer of myofibroblast differentiation, is up-regulated and responsible for the enhanced myofibroblast differentiation of eMSCs from IUA. Myofibroblast could be differentiated from a variety of types of mesenchymal cells, including perivascular cells, mesenchymal stromal/stem cells (MSCs) and bona fide fibroblasts [[192]19, [193]37, [194]38]. Among various pathways, TGFβ pathway is the most and best described mechanism responsible for myofibroblast [[195]37]. Indeed, hyper-activation of the TGFβ pathway results in increased myofibroblast differentiation and activation, eventually resulting in myofibroblast enriched genes expression (α-SMA and collagen type I), fibrosis and scar formation [[196]18, [197]19, [198]37, [199]38, [200]44]. Although it has been well established that the TGFβ pathway is the major driver to fibrosis development, therapeutic strategies targeting the TGFβ pathway have failed to show promising therapeutically beneficial outcome. In uterus, the TGFβ pathway also contributes to immune regulation and embryo implantation [[201]45–[202]50]. The multiple roles of TGFβ pathway indicate that simply blocking the TGFβ pathway might not be the feasible strategy to suppress IUA development/recurrence. Conclusions In conclusion, we have demonstrated here that the scar tissues in IUA biopsy are characterized with enhanced differentiation of pro-fibrotic fibroblast and myofibroblast. The number of eMSCs is reduced in IUA tissues, and their myofibroblast differentiation capability is enhanced. Supplementary Information [203]Additional file 1^ (108.4KB, pdf) [204]Additional file 2^ (13.4KB, docx) [205]Additional file 3^ (1.2MB, pdf) Acknowledgements