Current advances in understanding endometrial epithelial cell biology and therapeutic applications for intrauterine adhesion.

The source of stem cells responsible for endometrial epithelial regeneration remains a subject of debate. Endometrial epithelial cells are considered to have originated from three types of cells: endogenous endometrial epithelial stem cells, endometrial mesenchymal stem/stromal cells, and bone marrow-derived cells. Endometrial glands, lined with columnar epithelium, extend from the luminal epithelium to the endometrial-myometrial junction [ 13 ]. It has long been assumed that the human endometrial gland was a single-blind tube-like structure and epithelial stem cells were located at their terminal ends [ 14 , 15 ]. With the in-depth study of endometrial 3D morphologies, people revealed that the endometrial glands branch and form a network in the stratum basalis, horizontally expanding along the muscular layer, and some endometrial glands share a plexus with other glands. This complex network of the stratum basalis could protect the epithelial stem cells in the human endometrium [ 16 , 17 ]. During menstruation, the basal component of the glands remains in the basalis layer [ 13 ]. Scanning electron microscopy analysis revealed that the residual stumps of endometrial glands proliferate, producing a continuous layer of fusiform cuboidal epithelial cells that cover the denuded surface of the uterine cavity [ 6 , 18 – 20 ]. Putative endometrial epithelial stem cells were initially identified as clonogenic cells in humans [ 4 ]. Their clonogenicity remains consistent across menstrual cycle stages and menopausal statuses [ 21 ]. These cells exhibit adult stem cell properties, such as self-renewal, differentiation, and high proliferative potential in vitro [ 22 ]. Label-retaining cells (LRCs) are stem-like cells that retain a DNA label over a longer period than frequently dividing cells. The incorporation of BrdU into the DNA of dividing cells results in the loss of the label after four cell divisions. Adult stem cells are slow-cycling and relatively quiescent; hence, they often retain DNA labels over a longer period compared to rapidly dividing cells [ 23 ]. LRCs have been detected in the luminal or glandular epithelium in mice [ 24 – 26 ]. Specifically, luminal LRCs have been implicated in mediating estrogen-induced endometrial regeneration in mice [ 25 , 26 ]. In a mouse model simulating menstruation, glandular LRCs proliferate less frequently compared to luminal epithelial cells. Glandular LRCs proliferate selectively and contribute to the re-epithelialization of the endometrial surface after menses in the absence of estrogen [ 27 ]. Side population (SP) cells, initially identified in the bone marrow, constitute a mixed cell population distinguished by their ability to actively efflux Hoechst 33,342, a DNA-binding dye [ 28 ]. This capability is linked to the expression of ATP-binding cassette transporters encoded by BCRP1/ABCG2 [ 29 ]. The SP phenotype, regarded as a universal marker for somatic stem cells, has been utilized to isolate these cells from adult tissues when specific stem cell markers remained unidentified. Freshly isolated human endometrial SP cells could give rise to endometrial-like tissues under the kidney capsule in NOD-SCID mice [ 30 , 31 ]. Using lentiviral marking techniques, these transplanted endometrial SP cell populations were shown to differentiate into various endometrial cell types, including stromal, epithelial, smooth muscle, and endothelial cells under the kidney capsule in NOD-SCID mice [ 32 ]. In contrast to differentiated cells, whose telomeres shorten with each division, the intermediate telomerase activity observed in endometrial SP cells suggests their potential as putative stem cells [ 31 , 33 ]. Researchers have successfully isolated a group of SP cells within the human endometrial epithelium. These cells are located in the basal layer of the endometrial epithelium [ 34 ] and there is no significant difference in the proportion of SP cells in the epithelium throughout reproductive life [ 33 ]. Epithelial SP cells exhibit long-term repopulation properties (> 24 weeks) and can differentiate into CD9( +)gland-like structures in vitro [ 35 ]. These findings conclusively revealed the existence of stem cells in the endometrial epithelium. These cells are capable of long-term self-renewal, proliferation, and multilineage differentiation, vital for both the maintenance and regeneration of the endometrial lining. However, the specific markers of endometrial epithelial stem cells remain elusive. The identification of epithelial stem cells facilitates their isolation and application in infertility treatment resulting from inadequate endometrial regeneration. We summarize existing candidate markers of endometrial epithelial stem/progenitor cells and relevant evidence in experiments involving both human and murine endometrium (Table  1 ). Table 1 Summary of candidate markers of endometrial epithelial stem/progenitor cells Maker Species Main Location Characterization Refs. SSEA-1 Human GE in basalis (P/S) GE in funtionalis (P) LE (P/S) Compared with SSEA-1(−) epithelial cells, SSEA-1( +) epithelial cells showed higher telomerase activity, longer mean telomere lengths, less proliferative, lower expression of steroid receptors (PR and ERα) and produced a significantly higher number of endometrial gland-like spheroids in 3D culture [ 36 – 39 ] SOX9 Human GE in basalis (P/S) GE in funtionalis (P) LE (P/S) SOX9 is a transcription factor located inside the cell and cannot be sorted by conventional methods. Given nuclear SOX9( +) cells were largely confined to the SSEA-1(a cell-surface marker) expressing epithelial cells, people study them by isolating and characterizing SSEA-1( +) cells [ 37 , 38 , 40 , 41 ] N-cadherin Human GE adjacent to the endometrial-myometrial junction in basalis (P/S) N-cadherin( +) endometrial epithelial cells could form larger clones and show higher cloning efficiency relative to N-cadherin( −) cells in vitro. N-cadherin( +) cells-derived colony-forming units exhibited higher self-renewal and proliferative capacity compared to those derived from N-cadherin( −) cells. Freshly isolated N-cadherin( +) epithelial cells were capable of differentiating into cytokeratin( +) gland-like organoids in 3D culture. N-cadherin( +) epithelial cells were quiescent, rarely co-localised with Ki-67 [ 37 ] LGR5 Human GE in funtionalis (P) LE (P/S) LGR5 expression was hormonally regulated in endometrium. Oral and local progestogens have been shown to obviously reduce LGR5 mRNA levels. Further in vitro functional experiments are still needed to validate the stem cell properties of LGR5( +) cells, when a reliable anti-human LGR5 antibody is available [ 38 , 41 ] Mouse Müllerian duct during embryogenesis LE at birth* The tips of developing glandular buds LE in adult LGR5 expression in the mouse endometrium dynamically changed with age and ovarian hormones. In-vivo lineage tracing/ablation/organoid culture assays indicated that endometrial Lgr5( +) cells were Wnt-dependent stem cells responsible for uterine gland development during the prepuberty period. However, adult endometrial Lgr5( +) cells became short-lived differentiated cells [ 42 – 44 ] CD44 Human GE in basalis (P/S) CD44( +) epithelial cells were ERα and PR negative and exhibited superior sphere-forming efficiency relative to CD44(-) epithelial cells [ 45 ] Mouse GE CD44( +) epithelial cells were capable of self-renewal, proliferation, and multilineage differentiation in vivo. They had a higher in vivo regenerative potential than CD44(−) epithelial cells. CD44( +) epithelial cells were actively cycling with a significantly elevated telomerase reverse transcriptase enzyme expression compared to CD44(−) epithelial cells. Selective blocking CD44( +) cells would decrease the sphere forming capacity of endometrial epithelium. CD44( +) epithelial cells did not express ERα and PR but were responsive to steroid hormones, which was likely mediated by paracrine signals from neighboring cells  [ 45 ] Axin2 Human Cytoplasm of GE in functionalis (P/S) Nuclei of GE in basalis (P/S) Since Axin2 expression occurs intracellularly and the absence of cell surface markers for live isolation Axin2( +) cells, the exact role of Axin2( +) cells in the human endometrial lining remains an area of ongoing study and discovery [ 46 ] Mouse LE at birth* GE adjacent to the endometrial-myometrial junction in adult Axin2( +) epithelial cells were long-lived and able to self-renew during normal estrus cycles as well as following multiple rounds of pregnancy. These cells were capable of generating both the ciliated and non-ciliated cells. The selective ablation of Axin2( +) cells led to the significant collapse of endometrial glands and markedly compromised the proliferation capacity of GE and LE cells. Axin2( +) cells but not Axin2(−) cells formed fully functional endometrial organoids, which could be passaged at least 10 times without an apparent loss of replating efficiency [ 47 , 48 ] ALDH1A1 Human GE in basalis (P/S) In the glands of the basal layer, ALDH1A1 frequently co-localized with N-cadherin, which is a marker of endometrial epithelial progenitor cells [ 49 ] * At birth, the murine endometrium consists of a tube lined with luminal epithelium; At postnatal day 7, the luminal epithelium invades the stroma and initiates the development of glandular epithelium Abbreviation: P Proliferating phase, S Secretory phase, LE Luminal epithelium, GE Glandular epithelium, ERα Estrogen receptor α, PR Progesterone receptor Summary of candidate markers of endometrial epithelial stem/progenitor cells GE in basalis (P/S) GE in funtionalis (P) LE (P/S) GE in basalis (P/S) GE in funtionalis (P) LE (P/S) GE in funtionalis (P) LE (P/S) Müllerian duct during embryogenesis LE at birth* The tips of developing glandular buds LE in adult Cytoplasm of GE in functionalis (P/S) Nuclei of GE in basalis (P/S) LE at birth* GE adjacent to the endometrial-myometrial junction in adult * At birth, the murine endometrium consists of a tube lined with luminal epithelium; At postnatal day 7, the luminal epithelium invades the stroma and initiates the development of glandular epithelium Abbreviation: P Proliferating phase, S Secretory phase, LE Luminal epithelium, GE Glandular epithelium, ERα Estrogen receptor α, PR Progesterone receptor SSEA-1 is a carbohydrate-associated molecule that is expressed in various stem and progenitor cells, including those in the retina [ 50 ], lungs [ 51 ], and thyroid [ 52 ]. It can be used to identify glandular epithelial cells in the human endometrial basalis [ 36 ]. SOX9, a transcription factor from the SOX family, regulates homeostasis across various stem cell populations by interacting with the canonical Wnt signaling pathway [ 53 ]. Valentijn et al. observed that nuclear SOX9 is predominantly expressed in SSEA-1( +) epithelial cells [ 36 ]. N-cadherin, encoded by CDH2, is a calcium-dependent, single-chain transmembrane glycoprotein critical for synaptogenesis and stem cell differentiation [ 54 , 55 ]. Dual-color immunofluorescence staining of N-cadherin and SSEA-1 showed that N-cadherin( +) cells are located in the deeper layers of the basalis and do not fully co-localize with SSEA-1 [ 37 ].The different localization of markers within the human endometrial epithelium implies a potential positional hierarchy of epithelial stem cells (Fig.  1 ). Nguyen et al. suggested that N-cadherin( +)SSEA-1( −) epithelial stem cells located at the gland bases adjacent to the myometrium are the most primitive. N-cadherin( +)SSEA-1( +) cells represent a transitional status, followed by N-cadherin(-)SSEA-1( +) progenitor cells located in the upper basalis and lower functionalis [ 37 ]. Fig. 1 Positional hierarchy of endometrial epithelial stem cells in humans. Figure was drawn using Adobe illustrator software Positional hierarchy of endometrial epithelial stem cells in humans. Figure was drawn using Adobe illustrator software LGR5, a transmembrane receptor that binds to R-spondins and enhances Wnt/β-catenin signaling through complex formation [ 56 ], serves as a stem cell marker in multiple organs and tissues, including the stomach [ 57 ], esophagus [ 58 ], liver [ 59 ], ovaries and fallopian tubes [ 60 ]. Unlike other stem cell markers, LGR5 exhibits the highest expression in the endometrial luminal epithelium of humans [ 41 ]. Tempest et al. hypothesized that, in addition to the basalis, the luminal epithelium contains another epithelial stem cell pool (LGR5 +  + SSEA-1 + SOX9 +) that supports the embryo implantation and the daily cellular loss in the luminal epithelium. This stem cell pool probably originates from the rapid proliferation and re-epithelialization of SSEA-1(+ +)SOX9(+ +) LGR5( +) progenitor cells in the basalis during the initial post-menstrual repair phase [ 38 ]. Tanwar et al. initially observed a Wnt activity gradient in the endometrial epithelium, ranging from the basal to the luminal regions of the gland. Conditional deletion of Wnt signaling in transgenic mice led to a complete failure of endometrial gland development [ 48 ]. Axin2, a well-established Wnt reporter gene, accurately identifies Wnt-responsive stem cells [ 61 ]. Lineage tracing studies in mice have demonstrated that Axin2( +) cells is critical for the development, regeneration, and tumorigenesis of the endometrial epithelium [ 48 ]. As a robust stem cell validation tool, the mouse model allows us to observe cells physiological behavior and fate within their natural stem cell niche in intact tissues through lineage tracing [ 62 ], which does not alter the properties of the analyzed cells and is considered a gold standard for identifying epithelial stem cells [ 63 ]. In addition, the transgenic mouse model can isolate stem cells positive for markers expressed in the nucleus or cytoplasm for further functional validation. However, the gathered evidence shows that epithelial stem cells in mice and humans were not entirely congruent (Table  1 ). This discrepancy could be attributable to species-specific changes occurring in the endometrium during the estrous cycle. In rodents, the endometrial lining undergoes periodic apoptosis and regeneration, in contrast to the cyclic shedding and regeneration seen in humans [ 64 ]. Jin et al. revealed that mouse epithelial stem cells were located in the intersection zone between luminal and glandular epithelium. They proposed that mouse epithelial stem cells first generated luminal or glandular progenitors, which further differentiated into either luminal epithelium or glandular epithelium [ 65 ]. Tracing the glandular epithelium using Foxa2-CreERT2; Rosa26-tdTomato mice demonstrated that FOXA2( +) glandular epithelial cells did not contribute to luminal epithelium under physiological conditions and significantly reduced after 36 uterine cycles. This indicated that long-term self-renewing stem cells did not exist in mouse endometrial glands [ 65 ]. Similarly, Axin2( +) epithelial stem cells located at the basal ends of glands did not give rise to small numbers of luminal epithelial cells until after about 18 estrous cycles or after multiple pregnancies [ 48 ], so Axin2( +) cells seem to function as glandular progenitors in mice. Further exploration is needed to characterize the epithelial stem cells in the intersection area and clarify their relationship with the epithelial stem cells located in the glands. Some scholars argue that beyond epithelial stem cells, endometrial mesenchymal stem/stromal cells also play an important role in the regeneration of the endometrial epithelium. Mesenchymal-epithelial transition (MET), characterized by mesenchymal cells losing their typical mesenchymal traits and acquiring epithelial characteristics, contributes to the development of the embryonic uterus and the process of decidualization [ 66 ]. Garry et al. observed that during the initial stage of menstrual repair, new surface epithelial cells were distributed at a site far from the glandular stump, and their appearance was markedly distinct from the morphology of glandular epithelium observed in the late secretory phase. Garry et al. suggested that these epithelial cells likely originate from the underlying stromal cells, rather than from the residual gland stumps in humans [ 67 , 68 ]. In subsequent studies, scientists employed various transgenic Cre-loxP-activated animal models to track the behavior of stromal cells and their progenies across different physiological states, including postpartum, menstruation, and estrous cycle states, aiming to explore the contribution of endometrial mesenchymal stem/stromal cells to the epithelium (Table  2 ). Table 2 Summary of stroma-derived epithelial cells under various conditions in different mouse models Animal Model Context Detection Time Epithelial Compartment Comments Refs. Amhr2-Cre; Rosa26-lacZ Estrous cycle 2–6 mo – lacZ( +) cells were restricted to endometrial stromal and myometrial cells [ 69 ] Postpartum PPD1/PPD3/PPM2  +  The patches of lacZ( +) cells appeared in the LE and GE after delivery and the proportion of positive cells increased at PPM2 Amhr2-Cre; Rosa26-EYFP Development PND14/PND25 – EYFP expression was observed in both stroma and myometrium, but not in GE or LE [ 70 ] Postpartum After 2–3 pregnancies  +  EYFP expression was not only observed in uterine mesenchymal tissue, but was also observed in the GE and LE Sm22a-Cre;Rosa26- mT/mG Menses 36 h after progesterone withdrawal – The number of GFP( +) cells was drastically increased in the stromal compartment during the early stage of regeneration [ 72 ] 72 h/96 h after progesterone withdrawal  +  GFP( +) cells were detected in the epithelium at the late stage of regeneration Amhr2-Cre; Rosa26-lacZ Development PND35  +  LacZ( +) cells were found in LE, GE and stroma of the uterus [ 75 ] Cspg4-Cre; Rosa26-lacZ Development PND35  +  Patches of lacZ( +) cells were present in both LE and GE in some uterine sections Tamoxifen treatment a PND56  +  Tamoxifen treatment caused massive expansion of rarely labeled endometrial epithelial cells Sm22a-Cre; Rosa26-lacZ Development PND7  +  – Estrous cycle PND56  +  – Thy1-Cre; Rosa26-lacZ – –  +  – Acta1-Cre; Rosa26-lacZ – –  +  – Cspg4-CreER; Rosa26-lacZ Development PND14 b – The lacZ expression was restricted to the stromal perivascular cells Estrous cycle PND56 c – 13mo/14.5mo d – Pdgfrb-rtTA; tetO-Cre; Rosa26-lacZ Estrous cycle PND365 e  +  Organoids developed from freshly isolated endometrial epithelial cells were some lacZ-labeled and some unlabeled PND365 f – LacZ( +) cells were restricted to the stromal compartment and were completely absent in the epithelium Amhr2-Cre; Rosa26-EYFP Estrous cycle 2–5 mo  +  There was a high percentage of EpCAM( +)EYFP( +)cells (average 58%) identified in virgin uteri during proestrus that declined to undetectable levels in diestrus [ 74 ] Postpartum PPD4/PPD14/PPD21/PPM2/PPM3  +  The percentage of EpCAM( +)EYFP( +)cells had no significant differences from PPD4 to PPM3 Amhr2-Cre; Rosa26-tTA; H2B-GFP Development PND0.5/3/8/14/21  +  The percentage of EpCAM( +)GFP( +) cells gradually decreased from PND0.5 to PND8, increased slightly at PND14 and was maintained through PND21 Pdgfrb-BAC-eGFP Menses 24 h/48 h after progesterone withdrawal  +  The number of GFP( +)EpCAM( +) cells reduced at 48 h and almost disappeared at 72 h, which might be due to GFP protein turnover [ 73 ] 72 h after progesterone withdrawal – Pdgfra-CreERT2; Rosa26-tdTm Menses 24 h/48 h/72 h after progesterone withdrawal g  +  A transient population of tdTm( +)CD90( +)EpCAM( +)cells was detected at 24 h, the proportion of tdTm( +)EpCAM( +) in epithelial cells rose from 6% at 24 h to 17% at 48 h and 15% at 72 h Cspg4-CreER; Rosa26-tdTm Menses 24 h/48 h/72 h after progesterone withdrawal h – Neither tissue sections nor flow cytometry detected tdTm( +)EpCAM( +) cells at any timepoint a Mouse was administered with intraperitoneal tamoxifen injection (2 mg/d) for five consecutive days from PND42 b,c Pregnant female was administered with a single intraperitoneal tamoxifen injection (40 mg/kg) at gestation day 15.5 d Tamoxifen was administered by intraperitoneal injection (2 mg/d) for five consecutive days to 5–7 weeks old e Pregnant female was administered with doxycycline (5 mg/ml) in drinking water from embryonic day 8.5 to PND1 f Mouse was administered with doxycycline (5 mg/ml) in drinking water for two months from PND42 to PND102 g,h Mouse was administered with tamoxifen via oral gavage (100 μl/day) for three consecutive days and underwent a 4w washout period prior to the mouse model of menstruation Abbreviation: LE Luminal epithelium, GE Glandular epithelium, mo month, PPD Postpartum day, PPM Postpartum month, PND Postnatal day Summary of stroma-derived epithelial cells under various conditions in different mouse models a Mouse was administered with intraperitoneal tamoxifen injection (2 mg/d) for five consecutive days from PND42 b,c Pregnant female was administered with a single intraperitoneal tamoxifen injection (40 mg/kg) at gestation day 15.5 d Tamoxifen was administered by intraperitoneal injection (2 mg/d) for five consecutive days to 5–7 weeks old e Pregnant female was administered with doxycycline (5 mg/ml) in drinking water from embryonic day 8.5 to PND1 f Mouse was administered with doxycycline (5 mg/ml) in drinking water for two months from PND42 to PND102 g,h Mouse was administered with tamoxifen via oral gavage (100 μl/day) for three consecutive days and underwent a 4w washout period prior to the mouse model of menstruation Abbreviation: LE Luminal epithelium, GE Glandular epithelium, mo month, PPD Postpartum day, PPM Postpartum month, PND Postnatal day The endometrial epithelium of mice during the postpartum period was reported to be partially derived from mesenchymal cells, and these cells persisted during the two subsequent months of uterine homeostasis [ 69 ]. In the menses-like mouse model, a group of Pan-cytokeratin( +) vimentin( +) transitional cells could be observed in the stroma adjacent to the myometrium. These cells were gradually identified in the stromal compartment adjacent to the lumen after endometrial breakdown [ 70 , 71 ]. Another laboratory independently confirmed the presence of SM22α( +)CD34( +) transitional cells in the menses-like mouse model. These cells were significantly increased in the stroma and subsequently integrated into the epithelium [ 72 ]. A recent study involving single-cell RNA sequencing identified a distinct cluster of cells in a mouse model of menstruation. These cells were characterized by the expression of both mesenchymal markers (Pdgfrb, Vimentin) and epithelial markers (Epcam, Krt18), along with key MET-associated transcription factors, consistent with previous conclusions. They further identified this cell population as PDGFRα( +) fibroblasts, which underwent MET and finally incorporated into the luminal surface of the repaired endometrium [ 73 ]. Initially, MET was undetectable in the endometrial epithelium of mice in the absence of menses-like injury and pregnancy. This suggested that the extent of occurrence of MET in this context was subtle, potentially falling or below the detection threshold of earlier methods. Subsequent studies using more sensitive techniques, such as serial sections of the entire uterus and image-based flow cytometry analysis, confirmed the presence of mesenchymal-derived epithelial cells in the endometrium of mice during the estrous cycle [ 74 , 75 ]. These studies revealed that the proportion of EpCAM( +) stroma-derived epithelial cells varied throughout the estrous cycle in both virgin and postpartum uteri. The abundance of these epithelial cells declined to indiscernible levels during diestrus, possibly explaining why previous studies failed to detect positive epithelial cells [ 74 ]. Additionally, beyond the previously mentioned findings, MET was found to be involved in gland morphogenesis and differentiation after birth [ 74 ]. However, the validity of conclusions drawn from embryonic cell lineage tracing using mesenchymal cell marker gene-driven reporters is debatable. This can be attributable to the fact that the Müllerian duct epithelium, the origin of endometrial epithelium, is of the mesoepithelial type [ 76 ]. Therefore, Ghosh et al. proposed that the mesenchyme-specific, Cre-driven labeling observed in the endometrial epithelium might result from recombination occurring in the Müllerian duct epithelium, owing to its mesoepithelial nature. In Pdgfrβ-rtTA;tetO-Cre; Rosa26-lacZ mice, Cre was activated in adult mice, but no stroma-derived cells were observed in the epithelial compartment, even after a year-long observation period [ 75 ]. In addition to labeling stromal cells, Syed et al. generated Pax8-rtTA;tetO-Cre;Rosa26-EYFP and Pax8-rtTA;tetO-Cre;Rosa26-lacZ mice to trace epithelial cells [ 48 ]. Pax8 is ubiquitously expressed in all endometrial epithelial cells of pre-pubertal, pubertal, and adult stages [ 48 , 77 ]. They found that there was no flux of labeled cells between epithelial and mesenchymal compartments. The epithelial compartment was completely derived from self-renewed epithelial cells, but not from stromal cells or bone marrow-derived cells during development, the estrus cycle (over the 9-month-long reproductive lifespan of a mouse), multiple pregnancies, and post-injury repair [ 48 ]. Taken together, the reporting on the contribution of endometrial mesenchymal stem/stromal cells to epithelial regeneration varies significantly (Table  2 ), the discrepancies arise from differences in animal models, experimental designs, or other factors. In most experiments, the presence of epithelial cells derived from mesenchymal stem/stromal cells is very low or undetectable, contradicting the high turnover rate of epithelium. This indicates that endometrial mesenchymal stem/stromal cells play a supplementary rather than a primary role in epithelial regeneration. Bone marrow-derived cells (BMDCs) were once believed to possess the capacity to overcome lineage barriers and transdifferentiate into tissue-specific cells in a variety of organs [ 78 ]. In 2004, Taylor et al. first reported that bone marrow-derived cells obtained from female donors contributed to endometrial tissue regeneration in HLA-mismatched bone marrow transplant recipients [ 79 ]. A similar engraftment was observed in the endometrium of patients receiving sex-mismatched bone marrow transplants [ 80 , 81 ]. The bone marrow is comprised of various stem and progenitor cells, and opinions differ on which types of cells can repopulate the endometrial epithelium. CD45( +) hematopoietic progenitor cells [ 82 ], mesenchymal stem cells [ 83 ], and endothelial progenitor cells [ 84 ] have been proven capable of colonizing the uterine epithelium in mice. Researchers used transgenic or sex-mismatched donor mice to trace BMDCs in recipient mice. The recipients were subjected to myeloablation via irradiation or gonadotoxic chemotherapy. However, the contribution of BMDCs to the mouse endometrial epithelium was substantially limited [ 83 – 86 ]. Some studies aimed to rule out the possibility that decreased endometrium turnover, due to impaired ovarian function, compromises the migration of BMDCs to the uterus. To address this, researchers utilized ovarian transplantation or a non-gonadotoxic mouse bone marrow transplantation model to restore and protect ovarian function. Nonetheless, neither method raised the number of BMDCs in the endometrial epithelium [ 83 , 87 ]. In some experiments, the epithelial cells originating from BMDCs were scattered in the endometrium without showing significant clonal expansion. This suggests that these cells were terminally differentiated [ 80 , 81 , 83 , 85 , 86 ]. Side population cells are considered the adult stem cells responsible for endometrial regeneration [ 35 ], BMDCs also did not contribute to the abundance of side population cells [ 81 ]. In most experiments, BMDCs-derived epithelial cells are very limited in both human and mouse endometrium (Tables  3 and 4 ). Additionally, a study suggested that previously identified BMDCs-derived endometrial cells might be macrophages and highlighted the failure of BMDCs to transdifferentiate into mouse endometrial epithelial cells [ 88 ]. Overall, there is no reliable evidence that BMDCs can differentiate into endometrial epithelial cells. Table 3 Contributions of BMDCs to the endometrial epithelium in humans Age, year Time from transplantation to examination, month Maker Donor-derived epithelial cells, % Characterization Refs. 43 157 HLA antigen 48 Calcitonin( +) CD45(−) [ 79 ] 25 129 10 32 35 4 28 24 0.2 44 116 Y chromosome 2.6 CD10( +) ERα( +) CD45(−) [ 80 ] 41 33 8.4 27 54 0.6 38 132 Y chromosome 1.07 CD9( +) CD45(−) [ 81 ] 29 30 0.5 41 210 0.81 34 116 0.85 28 45 0.45 Table 4 Contributions of BMDCs to the endometrial epithelium in mice Recipient Donor Time from transplantation to examination Maker Donor-derived epithelial cells, % Characterization Refs. C57BL/6 J female mice C57BL/6-Tg [ACTB-EGFP]1Osb/J mice 8–12mo GFP Detectable L.tetragonolobus lectin( +) Pan-cytokeratin ( +) CD45(−) [ 82 ] C57BL/6 J-Tg [CD45/Cre-Z/EG] female mice C57BL/6J-Tg [CD45/Cre-Z/EG] female mice 1w GFP 0 6w 0 12w 0.5 20w 6 12w(pregnant) 82 C57BL/6 female mice C57BL/6 male mice 6mo Y chromosome 0.02 Cytokeratin( +) F4/80(−) CD45(−) [ 86 ] C57BL/6 female mice C57BL/6 male mice 8mo Y chromosome 0.014 ± 0.00148 (no ovary transplant) 0.013 ± 0.00089 (ovary transplant) Cytokeratin( +) F4/80(−) CD45(−) vWF(−) [ 83 ] C57BL/6 J female mice C57BL/6-Tg [UBC-GFP]30Scha/J mice C57BL/6-Tg [CAG-EGFP]131Osb/LeySopJ mice 3mo GFP Undetectable Located at epithelial compartment [ 85 ] 5mo Barely detectable 9mo Barely detectable 12mo 0–2.7 (in LE) 0–2.1 (in GE) C57BL/6 J female mice C57BL/6-Tg [CAG-EGFP] mice 12w GFP 0 (from UBCs) 0 (from HPCs) 4.78 (from EPCs) 0.36 (from MSCs) Located at epithelial compartment [ 84 ] C57BL/6 J female mice C57BL/6 J-Tg [Ubiquitin-GFP] male mice 1mo GFP 0 – [ 87 ] C57BL/6 J female mice C57BL/6 J-Tg [mTert-GFP] mice 17-24w GFP 0 – [ 88 ]  C57BL/6 J female mice C57BL/6-Tg [CAG-EGFP]1Osb/J mice 4-10mo GFP 0 Abbreviation: UBCs Unfractionated bone marrow cells, HPCs Hematopoietic progenitor cells, EPCs Endothelial progenitor cells, MSCs Mesenchymal stem cells, mo month, w week Contributions of BMDCs to the endometrial epithelium in humans Calcitonin( +) CD45(−) CD10( +) ERα( +) CD45(−) CD9( +) CD45(−) Contributions of BMDCs to the endometrial epithelium in mice L.tetragonolobus lectin( +) Pan-cytokeratin ( +) CD45(−) Cytokeratin( +) F4/80(−) CD45(−) 0.014 ± 0.00148 (no ovary transplant) 0.013 ± 0.00089 (ovary transplant) Cytokeratin( +) F4/80(−) CD45(−) vWF(−) C57BL/6-Tg [UBC-GFP]30Scha/J mice C57BL/6-Tg [CAG-EGFP]131Osb/LeySopJ mice 0–2.7 (in LE) 0–2.1 (in GE) 0 (from UBCs) 0 (from HPCs) 4.78 (from EPCs) 0.36 (from MSCs) Abbreviation: UBCs Unfractionated bone marrow cells, HPCs Hematopoietic progenitor cells, EPCs Endothelial progenitor cells, MSCs Mesenchymal stem cells, mo month, w week

Intrauterine adhesion (IUA), also known as Asherman’s syndrome, is characterized by the loss of functional endometrial tissue and fibrosis within the uterine cavity induced by infection, trauma, and other causes [ 89 ]. It manifests as periodic lower abdominal pain, oligomenorrhea, amenorrhea, recurrent abortion, and infertility [ 90 ]. The primary treatment for IUA is hysteroscopic transcervical resection of adhesion, with a high recurrence rate of 40–62.5% [ 91 ]. Postoperative adjuvant therapies typically involve pharmaceuticals such as hormones and vasoactive agents, or the insertion of a physical barrier in the uterine cavity; however, their effectiveness in repairing the endometrium and preventing adhesion is substantially limited [ 92 ]. Due to the inadequate efficacy of traditional treatments, there is a critical need to explore alternative therapeutic options. Researchers have suggested that local stem cell loss might contribute to endometrial repair disorders, as endometrial regeneration is closely linked to stem cells in the basal layer [ 93 ]. In cases of IUA, severe damage to the germinal compartment, where these stem/progenitor populations and their surrounding stem cell niches occur, leads to an inability to regenerate the functionalis. The damaged endometrium is subsequently replaced by fibrous tissue [ 94 ]. Thus, stem cells transplantation has emerged as a novel therapeutic option. They function by directly replacing damaged tissues or through their paracrine activity and immunomodulatory effects [ 95 ]. Moreover, platelet-rich plasma and various stem cell derivatives, such as conditioned media, growth factors, extracellular vesicles, and non-coding RNAs, have shown significant promise as innovative cell-free therapies [ 96 ]. In addition, drug-mediated treatments, including intrauterine injections of botulinum toxin A [ 97 ], curcumin [ 98 ], oral metformin [ 99 ], or traditional Chinese medicine Dingkun pills [ 100 ], are gaining popularity in the treatment of IUA. Considering that the endometrium in IUA exhibits a significant loss of endometrial epithelial cells and marked disruption in normal interactions between epithelial and stromal cells [ 12 ], directly transplanting epithelial stem cells is considered a favorable option. However, challenges are associated with the isolation and cultivation of endometrial epithelial stem cells in vitro due to their limited proliferation capacity, tendency to undergo senescence, and loss of polarity under traditional 2D culture conditions [ 101 ]. Consequently, stem cell-based therapies have predominantly focused on utilizing mesenchymal stem cells from various sources [ 89 , 93 , 102 ]. Recent advancements in optimized medium development, organoid technology, and tissue engineering have facilitated the in vitro expansion and in vivo orthotopic transplantation of epithelial stem cells. These innovations offer new avenues for the treatment of endometrial injuries (Fig.  2 ). Fig. 2 Summary of current research on epithelial cell-based treatments for the endometrial injury. Figure was drawn using Adobe illustrator software Summary of current research on epithelial cell-based treatments for the endometrial injury. Figure was drawn using Adobe illustrator software The transition and expansion medium (TEM), comprised of a mixture of growth factors and bioactive small molecules, such as hepatocyte growth factor, epidermal growth factor, lysophosphatidic acid, CHIR99021, Y27632, and A8301, facilitated the long-term culture and expansion of SSEA-1( +) endometrial epithelial stem cells. In comparison to SUSD2( +) endometrial mesenchymal stem cells, SSEA-1( +) cells demonstrate higher secretion of VEGFA and superior proangiogenic potential. Additionally, they possess the ability to enhance the migratory capacity of endometrial stromal cells by inducing the expression of several genes related to the cell adhesion molecules pathway [ 39 ]. Chitosan is a synthetic, high molecular weight polysaccharide with unique physicochemical and biological properties, including non-toxicity, biodegradability, and biocompatibility. It can serve as an effective carrier for delivering stem cells, bioactive molecules, or drugs for the repair of various tissues and organs [ 103 , 104 ]. SSEA-1( +) cells embedded in chitosan demonstrated in vivo survival for 48 h. This treatment process significantly enhanced endometrial thickness, increased glandular numbers, and reduced fibrotic area in the IUA rat model [ 39 ]. Liu et al. devised a protocol for conditional reprogramming that facilitated the long-term culture and propagation of primary epithelial cells from healthy human tissues. They co-cultured human epithelial cells with an irradiated J2 strain of Swiss-3T3 mouse fibroblasts in the presence of a rho kinase inhibitor (Y-27632) [ 105 ]. Building upon these findings, Xia et al. successfully achieved the expansion of conditional reprogrammed-mouse endometrial epithelial cells (CR-MEECs) in vitro. These CR-MEECs not only expressed epithelial stem cells makers, such as P63 and CD44, but also maintained the potential for tissue-specific differentiation. Furthermore, the allogeneic transplantation of CR-MEECs led to reduced fibrosis and enhanced the regeneration of luminal and glandular epithelial cells, subsequently enhancing pregnancy rates in mice with IUA [ 106 ]. Mechanistic experiments further revealed that CR-MEECs exerted therapeutic effects by inhibiting the activation of Indian hedgehog (Ihh)-krüppel-like factor 9 (KLF9) signaling in the injured endometrium [ 107 ]. Injury-induced abnormal activation of Ihh-KLF9 signaling inhibited the progesterone receptor function in the endometrium, potentially influencing the estrogen and progesterone receptor balance and disturbing the proliferation of endometrial epithelial cells [ 107 ]. Organoid is a 3D cellular structure derived from adult stem cells, embryonic stem cells, or induced pluripotent stem cells [ 108 ]. It exhibits remarkable abilities for self-renewal and self-organization, faithfully reproducing the functionality of its original tissue [ 109 ]. In contrast to the complexity of animal models and the simplicity of 2D cell culture models, organoids provide a more relevant and sophisticated platform for studying human-specific disease modeling, tissue developmental biology, and personalized medicine [ 110 – 112 ]. Additionally, organoid transplantation is gaining attention as a promising method for tissue or organ replacement and regeneration [ 108 ]. This technique has proven effective for restoring and regenerating damaged cells in various organs. Notable examples include successful organoid transplantation in the intestine, colon [ 113 – 115 ], liver [ 116 – 119 ], pancreatic islets [ 120 , 121 ], and the pituitary gland [ 122 ]. These instances underscore the versatility and potential of this technique in regenerative medicine. Zhang et al. demonstrated positive therapeutic effects and improved pregnancy outcomes following the transplantation of allogeneic mouse endometrial epithelium-derived organoids into a model IUA uterus [ 123 ]. Xu et al. developed multi-lineage endometrial organoids (MLEOs) by co-culturing endometrial mesenchymal stem cells (EMSCs) and endometrial epithelial cells (EECs) in Matrigel. EMSCs significantly boosted the expression of stem cell-related genes within endometrial epithelium organoids (EEOs) by providing a supportive stem cell niche. Both MLEOs and EEOs were effective in promoting endometrial regeneration and reducing fibrosis. Despite MLEOs showing slightly better outcomes, the difference was not statistically significant compared to EEOs [ 124 ]. This suggests that the therapeutic effect mainly originates from the epithelial component of the organoids, while EMSCs play a supportive role. Additionally, this study established an MLEO-HAAM patch by seeding MLEOs on human acellular amniotic membrane (HAAM), which could help improve their survival after transplantation and effectively enhance pregnancy outcomes [ 124 ]. Jiang et al. successfully generated endometrial epithelial progenitor cells (EEPCs) from human embryonic stem cells-9 (H9-ESCs) through the optimization of induction factors. Compared to primary endometrial epithelial cells, EPPCs derived from H9-ESCs were more easily obtained and could be expanded in large quantities. This provides a potentially unlimited source of endometrial epithelium. These EEPCs, along with human endometrial stromal cells, formed endometrial membrane organoids in 3D culture. In a rat model of endometrial injury, the EEPC-derived organoids/Matrigel group showed superior regenerative potential and vascular benefits compared to the EEPCs/Matrigel group [ 125 ]. Recent research has shown that endometrial organoid-originated mitochondria were essential for the anti-fibrotic and pro-repair effects of organoids. Endometrial organoids derived from mice and humans can be engrafted onto the endometrium and reverse mitochondrial dysfunction by transferring mitochondria into damaged fibrotic cells in the murine model of IUA [ 126 ]. While organoid transplantation for endometrial injury represents a promising frontier of medical research, it is still a nascent field with some unresolved challenges. Organoid culture requires expansion of stem/progenitor cells in Matrigel [ 127 ], an ill-defined mix of extracellular matrix proteins and growth factors from Engelbreth-Holm-Swarm mouse tumor tissues [ 128 ]. This reliance on tumor-originated Matrigel poses limitations for the clinical application of organoid technology. Alternative approaches, such as hydrogels derived from decellularized human and bovine endometrium, offer more defined and possibly safer options [ 129 , 130 ]. Organoids grown in decellularized endometrium hydrogels show greater similarity to native tissues compared to those in Matrigel [ 129 ]. However, further validation is required to assess the efficacy of these alternatives in treating endometrial injury. Regeneration of vasculature is essential for the survival and function of the graft after transplantation. Various strategies for vascularizing organoids are discussed in another review [ 131 ]. The inosculation of the host vascular system with organoid endothelial cell tubes can expedite vascularization and enhance graft survival [ 132 ]. However, the specific study of vascularization in endometrial organoids remains an area for future research. In addition to the direct transplantation of epithelial stem cells and epithelial organoids, endometrial epithelial cells can be utilized to construct bioengineered endometrial tissues. This is accomplished through integration with cell sheet technology, biodegradable scaffolding, or advanced 3D printing techniques. These innovative approaches hold promise in restoring the full thickness of the endometrium in vivo. Cell sheet engineering involves culturing cells on temperature-responsive poly(N-isopropyl acrylamide)-grafted cell culture dishes. This technique facilitates the harvest of cells as a cohesive “cell sheet” while preserving inter-cellular connections and the native extracellular matrix [ 133 ]. Notably, the scaffold-free cell sheet also avoids the host inflammatory response triggered by the scaffolds after transplantation [ 134 ]. Kuramoto et al. developed a three-layer endometrial cell sheet by layering endometrial epithelial cell sheets onto stromal cell sheets [ 135 ]. In a nude rat model with resected full-thickness endometrium, this cell sheet remained viable for over 4 weeks and successfully reconstructed the endometrial stromal and epithelial compartments at the injured site. Importantly, the cell sheet–regenerated endometrium could decidualize and support pregnancy, similar to normal endometrium (Fig.  3 ) [ 135 ]. Fig. 3 The cell sheet technique can effectively reconstruct endometrial structure and function at the site of injury. A . GFP-labeled primary endometrial epithelial and stromal cells were cultured on temperature-responsive dishes and harvested as cell sheets, the three-layer cell sheets were fabricated by layering epithelial cell sheets onto stromal cell sheets. Scale bar, 100 μm; B , C . The cell sheets could colonize and survive at the transplanted site, restore the full-thickness structure of endometrium and the uterine cavity patency. Scale bar, 500 μm ( B iv–vi); 5 mm ( B vii, viii); 200 μm ( B ix); 500 μm ( C ); D , E . Pregnancy was successfully established on the cell sheet–regenerated endometrium. Scale bar, 1 mm ( E i, iii); 200 μm ( E ii, iv). [ 135 ] Reprinted from Fertility and Sterility, 110, Kuramoto G, Shimizu T, Takagi S, Ishitani K, Matsui H, Okano T. Endometrial regeneration using cell sheet transplantation techniques in rats facilitates successful fertilization and pregnancy, 172–181.e4, 2018, with permission from Elsevier The cell sheet technique can effectively reconstruct endometrial structure and function at the site of injury. A . GFP-labeled primary endometrial epithelial and stromal cells were cultured on temperature-responsive dishes and harvested as cell sheets, the three-layer cell sheets were fabricated by layering epithelial cell sheets onto stromal cell sheets. Scale bar, 100 μm; B , C . The cell sheets could colonize and survive at the transplanted site, restore the full-thickness structure of endometrium and the uterine cavity patency. Scale bar, 500 μm ( B iv–vi); 5 mm ( B vii, viii); 200 μm ( B ix); 500 μm ( C ); D , E . Pregnancy was successfully established on the cell sheet–regenerated endometrium. Scale bar, 1 mm ( E i, iii); 200 μm ( E ii, iv). [ 135 ] Reprinted from Fertility and Sterility, 110, Kuramoto G, Shimizu T, Takagi S, Ishitani K, Matsui H, Okano T. Endometrial regeneration using cell sheet transplantation techniques in rats facilitates successful fertilization and pregnancy, 172–181.e4, 2018, with permission from Elsevier Poly-dl-lactide-coglycolide (PLGA) and polyglycolic acid (PGA) are biodegradable synthetic polymers with good biocompatibility and adjustable mechanical properties. They have been approved by the Food and Drug Administration for human use in a variety of applications [ 136 ]. Magalhaes et al. used a combination of uterine-derived cells and PGA/PLGA scaffolds to construct tissue-engineered uteri. Myometrium-derived cells were seeded on the outer layer of the scaffold, and endometrium-derived cells (including both epithelial and stromal cells) were seeded on the inner layer. The tissue-engineered uteri facilitated the effective rebuilding of the uterine configuration in a subtotal excised rabbit uterus [ 137 ]. The regenerated endometrial epithelium expressed functional estrogen and progesterone receptors, along with uteroglobin, a rabbit reproduction-associated secretory protein. The engineered uteri could support pregnancy and normal fetal development to term, resulting in offspring with no congenital malformations and within normal weight ranges [ 137 ]. Three-dimensional extrusion-based bioprinting refers to creating cell-incorporated constructs or scaffolds based on the extrusion technique. This method can generate complex biomimetic tissues by precisely positioning multiple materials and cells in a defined spatial pattern [ 138 ]. Using endometrial cell-loaded Alg-HA hydrogel as the bioink, Nie et al. printed a bilayer endometrial construct [ 139 ]. The bilayer endometrial construct featured a dense upper layer of EECs and a lower layer with a grid-like microstructure loaded with endometrial stromal cells (ESCs), recapitulating the constitutive and structural characteristics of the endometrium. This construct not only promoted the regeneration of vascularised endometrium and inner muscle layer, but also improved reproductive outcomes in the rat model of partial full-thickness uterine excision. Moreover, the EECs and ESCs within the endometrial construct could survive and maintain the bilayer structure after transplantation (Fig.  4 ). Fig. 4 3D bio-printed bilayer endometrial construct can restore the structure of the endometrial wall in severe endometrial injury. A . 3D bioprinting technology deposited EECs and ESCs evenly on the Alg-HA hydrogel. Scale bar, 500 μm; B . Scanning electron microscope images showed that the EECs and ESCs were attached to and grew within the scaffold after 24 h of culturing. Scale bar, 5 μm; C . Confocal fluorescence microscopy images showed that most of ESCs were alive in the 3D constructs up to 48 h after bio-printing and in vitro culture. Blue scale bar, 500 μm; white scale bar, 200 μm.; D – G . 3D bio-printed endometrial construct can promote the recovery of endometrial structure and function, and the regenerated endometrium can support embryo implantation. Scale bar, 20 μm ( D ); 100 μm ( E ); 200 μm ( F top); 5 μm ( F bottom). [ 139 ] Reprinted from Acta Biomaterialia, 157, Nie N, Gong L, Jiang D, Liu Y, Zhang J, Xu J, et al. 3D bio-printed endometrial construct restores the full-thickness morphology and fertility of injured uterine endometrium, 187–99, 2023, with permission from Elsevier 3D bio-printed bilayer endometrial construct can restore the structure of the endometrial wall in severe endometrial injury. A . 3D bioprinting technology deposited EECs and ESCs evenly on the Alg-HA hydrogel. Scale bar, 500 μm; B . Scanning electron microscope images showed that the EECs and ESCs were attached to and grew within the scaffold after 24 h of culturing. Scale bar, 5 μm; C . Confocal fluorescence microscopy images showed that most of ESCs were alive in the 3D constructs up to 48 h after bio-printing and in vitro culture. Blue scale bar, 500 μm; white scale bar, 200 μm.; D – G . 3D bio-printed endometrial construct can promote the recovery of endometrial structure and function, and the regenerated endometrium can support embryo implantation. Scale bar, 20 μm ( D ); 100 μm ( E ); 200 μm ( F top); 5 μm ( F bottom). [ 139 ] Reprinted from Acta Biomaterialia, 157, Nie N, Gong L, Jiang D, Liu Y, Zhang J, Xu J, et al. 3D bio-printed endometrial construct restores the full-thickness morphology and fertility of injured uterine endometrium, 187–99, 2023, with permission from Elsevier The endometrium is comprised of various cell types, including epithelial, stromal, smooth muscle, vascular endothelial, and immune cells [ 140 ]. In comparison to scaffolds loaded with only a single cell type [ 141 – 144 ], those incorporated with epithelial cells can more effectively mimic the tissue-specific structure and replicate the intricate environment of the endometrium. This approach could potentially result in a more successful regeneration and restoration of the endometrial lining in severe cases [ 139 ].

During the repair of endometrial injury, rapid re-epithelialization of the wound surface is a critical step in tissue regeneration. Stem cells responsible for endometrial epithelial regeneration may be derived from endogenous endometrial epithelial stem cells, endometrial mesenchymal stem/stromal cells, and bone marrow-derived cells. This review summarizes published studies and recent research, suggesting that endogenous epithelial stem cells are likely the primary source. It also provides a detailed summary of their markers, characteristics, and spatial distribution within the endometrium of humans and mice. Intrauterine adhesion is characterized by impaired endometrial regeneration and fibrotic tissue overgrowth, typically accompanied by the loss and dysfunction of epithelial cells, which severely impacts female reproductive health. Several studies have confirmed that epithelial stem cells have the ability to inhibit fibrosis and promote endometrial repair, showing promise in the application of intrauterine adhesion. Additionally, the combination of epithelial cells and various biomaterials has shown significant efficacy in the treatment of severe endometrial damage. Therefore, epithelial cell-based therapies can be regarded as a desirable strategy for managing intrauterine adhesion.

The female uterine endometrium undergoes approximately 400 cycles of development, differentiation, and shedding during a female’s reproductive years. Endometrial regeneration occurs after menstruation, parturition, curettage surgery, and post-menopause in females following treatment with ovarian steroid hormones [ 1 , 2 ], which highlights its pronounced regenerative capacity. The involvement of adult stem/progenitor cells in this process cannot be overlooked. The hypothesis that endometrial regeneration is driven by stem cells located in the basal layer of the endometrium was proposed decades ago [ 3 ], with initial evidence of adult stem cell populations in the endometrium being published in 2004 [ 4 ]. Menstruation results in the concurrent shedding and repair of the endometrium. Re-epithelialization begins within 24 h of the onset of bleeding [ 5 ], and typically concludes within 5–6 days [ 6 ]. Rapid re-epithelialization safeguards the denuded surface, protecting the underlying endometrium from infection and mitigating bleeding [ 2 ]. Estrogen is not required for endometrial re-epithelialization, as demonstrated in animal models of endometrial repair [ 7 , 8 ]. Estrogen-dependent endometrial proliferation manifests only after surface epithelium restoration [ 7 , 9 ]. Moreover, the endometrial epithelium is crucial for uterine function. Endometrial glands and their secretions play a role in pregnancy success by regulating stromal cell decidualization, placental growth, and post-implantation differentiation [ 10 , 11 ]. Identifying epithelial stem cells could enhance our understanding of endometrium regeneration and the role of these cells in endometrial proliferative disorders, such as endometrial bleeding disorders, uterine cancer, endometriosis, adenomyosis, and intrauterine adhesion (IUA). In the context of IUA, the endometrial epithelium is notably reduced, and the epithelial-stromal cell communication shifts towards stromal cell self-stimulation, indicating a disruption in endometrial epithelial regeneration [ 12 ]. Thus, direct transplantation of epithelial stem cells becomes a promising candidate for stem cell-based therapies targeting IUA. This review presents three different views on the origin of endometrial epithelial cells, with emphasis on endogenous epithelial stem cells in the epithelial compartment being the primary source, and outlines the efficacy and deficiency of epithelial cell-based treatments in IUA.