Single-cell RNA sequencing of endometrium uncovers dynamic characteristics and dysregulation of perivascular CD9⁺SUSD2⁺ cells in thin endometrium

Abstract

Background

Thin endometrium (TE) has been widely recognized as a critical cause of various obstetric and gynecological conditions. Although the role of stem cells in endometrial functions and their pathologies has been suggested, the identity and molecular mechanisms of such stem cells remain unclear.

Methods

Data analysis of a publicly available single-cell RNA sequencing (scRNA-seq) was performed to unravel the cellular and molecular characteristics of endometrial CD9⁺ SUSD2⁺ cells in normal endometria (n = 3 ) across the menstrual cycle and proliferative phase of thin endometrium (n = 3 ). Then, CD9⁺ SUSD2⁺ cells were isolated for analysis. Flow cytometry and colony-forming assays were utilized to assess changes in CD9⁺ SUSD2⁺ cell proliferation and differentiation. Additionally, CellChat, Western blotting, and multiplex immunofluorescent analysis were performed to elucidate the tissue distribution of the CD9⁺ SUSD2⁺ cells and their molecular regulatory effects on the pathogenesis of TE.

Results

A total of 59,770 cells were grouped into 13 distinct clusters in normal proliferative, secretory endometrium, and thin endometrium. Our findings revealed that perivascular CD9⁺ SUSD2⁺ cells as putative progenitor stem cells based on pseudotime trajectory and enriched functions in ossification, stem cell development, and wound healing. Histological analysis unveiled a significant perivascular expression pattern of CD9⁺ SUSD2⁺ cells in different menstrual cycle phases. The scRNA-seq of endometrial samples during the proliferative phase from patients with TE and controls revealed TE-associated shifts in cell function, manifesting as increased fibrosis and attenuated cell cycle and adipogenic differentiation. Cell-cell communication network mapping underscored aberrant crosstalk among specific cell types, implicating crucial pathways such as collagen over-deposition around perivascular CD9⁺ SUSD2⁺ cells, indicating a disrupted response to endometrial repair in TE, particularly remodeling of the extracellular matrix.

Conclusions

Our study provides potential molecular mechanisms underpinning perivascular CD9⁺ SUSD2⁺ cells in the context of thin endometrium. The mechanistic insights could establish new therapeutic strategies for endometrial regeneration and repair.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13287-025-04658-y.

Introduction

Thin endometrium (TE) is a common disease in reproductive medicine that can lead to infertility and low live birth rates. It is crucial for the endometrial lining to have a thickness of at least 7 mm on the day of embryo transplantation, as this provides the necessary environment for successful implantation and pregnancy [1]. However, there is no consensus on the exact definition of thin endometrium. The most widely accepted measurement is 7 mm, as a thickness below this has been linked to lower chances of implantation and pregnancy [2, 3]. Various treatments have been used to address TE, including extended estrogen administration, low-dose aspirin, pentoxifylline, vaginal sildenafil, tocopherol, and intrauterine perfusions with granulocyte colony-stimulating factor and platelet-rich plasma (PRP) [4]. Despite these options, some refractory patients with TE do not respond to treatment, leading to repeated cancellations of frozen-thawed embryo transfer (FET) cycles or failed recurrent embryo implantation. Therefore, restoring a TE remains a challenging task.

The most common causes of TE arise from inappropriate endometrial repair after curettage or surgical separation of intrauterine adhesions, which can lead to disrupted blood vessel distribution and sparse glands [5, 6]. Physiologically, the human endometrium undergoes dynamic remodeling, including massive physical shedding, rather than individual cell apoptosis, cyclically throughout a woman’s reproductive life [7]. During each menstrual cycle, which typically lasts 28 days, the uterus sheds its endometrial lining and then regrows new tissue to a thickness of approximately 4–10 mm within one week, under the influence of ovarian hormones [8]. This huge regenerative ability suggests that the endometrium has a stem cell basis supporting tissue regrowth, which was proposed decades ago [9], with the initial evidence of adult stem cell populations in the endometrium being published in 2004 [10]. Since then, the study of endometrial stem/progenitor cells has been highly developed. To date, based on cell types and identification techniques, there is some evidence that CD140b⁺CD146⁺ or sushi domain containing-2 (SUSD2⁺) endometrium-derived mesenchymal stem cells, endometrial epithelial stem/progenitor cells, and side population cells [11–13], which exhibit somatic stem cell properties, including extensive proliferative potential, self-renewal capacity, and tissue reconstitution. These endometrial stem/progenitor cells are known to be involved in the pathogenesis of endometrium-related disorders (such as endometriosis) and have been proposed to be implicated in the pathogenesis of TE [6, 14]. But it still cannot explain why only 2.4% of reproductive-age women with infertility develop into TE [15]. The discovery of specific markers for endometrial stem/progenitor cells has enabled the examination of their role in TE. Although the role of stem cells in endometrial functions and their pathological implications has been suggested, the specific cell population that plays a key role in TE pathogenesis remains unclear.

Here, we utilized two publicly available single-cell RNA sequencing (scRNA-seq) data sets [16, 17] to demonstrate that endometrial perivascular CD9⁺SUSD2⁺ cells are functionally more responsive than CD9⁻ SUSD2⁺ cells during the proliferative phase of the endometrium. Additionally, we observed a higher abundance of CD9⁺SUSD2⁺ cells during the proliferative phase compared to the secretory phase through immunofluorescence analysis. By comparing endometrial tissue from normal controls and patients with TE, we identified a specific role of CD9⁺SUSD2⁺ cells in perivascular adaptation during endometrial regeneration and repair. Our analysis of cell interactions revealed that signaling pathways related to collagen deposition around blood vessels were markedly interrupted in TE, particularly in perivascular CD9⁺SUSD2⁺ cells. These findings suggest that CD9⁺SUSD2⁺ cells represent a subpopulation of endometrial perivascular cells that function as endometrial progenitors, potentially contributing to the pathogenesis of TE and providing insight into improving fertility.

Methods and materials

Data retrieval

Single-cell sequencing data of the proliferative phase (including thin and control samples) used in this study were downloaded from the Sequence Read Archive under accession number PRJNA730360. Single-cell sequencing data of secretory phase samples were downloaded from the Gene Expression Omnibus (GEO) database under accession number GSE183837.

Single-cell transcriptomic sequencing data analysis

The primary analyses in this study were conducted using the Seurat R package (version 5.0.1) [18]. Cells with fewer than 1,000 detected genes and less than 10,000 transcripts were excluded. Normalization of raw counts was performed using the “NormalizeData” function with default settings, applying the LogNormalize method and a scale factor of 10,000. Highly variable genes were identified using the “FindVariableGenes” function with the “vst” method, selecting the top 4,800 variable features, among which 3,800 highly variable genes were retained for principal component analysis (PCA). Based on the JackStrawPlot and ElbowPlot, 30 principal components were chosen for constructing a shared nearest neighbor graph. Cell clustering was conducted using the “FindClusters” function with a resolution parameter of 0.7, and dimensionality reduction was performed via t-SNE. Differential expression analysis (DEA) across clusters was carried out using the “FindAllMarkers” function with default settings. Cluster visualization was achieved through the “DimPlot” function, while gene expression patterns were displayed using “FeaturePlot,” “RidgePlot,” and “VlnPlot.” Additionally, heatmaps were generated with the “DoHeatmap” function. Gene ontology (GO) enrichment analysis was conducted using the clusterProfiler package (version 4.12.2) [19]. Statistical significance testing was conducted using the ggsignif R package (version 0.6.0). To identify genes specifically upregulated in CD9⁺ SUSD2⁺ eMSCs, we performed differential gene expression (DEG) analysis using the FindAllMarkers function with the default parameter. The reference population of all other cell clusters, except for the CD9⁺SUSD2⁺ cluster. Genes were considered significantly upregulated if they had an adjusted p-value < 0.05 and a log2 fold change > 0.25. The top 200 upregulated genes in the CD9⁺ SUSD2⁺ population were selected for Gene Ontology (GO) enrichment analysis.

RNA velocity analysis was performed using the scVelo package. First, the preprocessed single-cell RNA-seq data, including spliced and unspliced transcript counts, were obtained from the original dataset. Data preprocessing involved filtering and normalizing the gene expression matrix, followed by computing first- and second-order moments using the “pp.moments” function. RNA velocity was estimated using the “tl.velocity” function with the stochastic mode, and velocity vectors were projected onto the low-dimensional embedding using the “tl.velocity_graph” function. Finally, RNA velocity streamlines were visualized on the UMAP embedding with the “pl.velocity_embedding_stream” function, illustrating the direction of cellular state transitions.

Clinical sample collection

Human endometrial samples and procedures involved in this study were approved by the Ethics Committee of Shenzhen Zhongshan Urology Hospital (No. SZZSECHU-2022008). Informed written consent was obtained from each participant before the endometrial biopsy. Thirty-seven patients who needed hysteroscopic examination at the Fertility Center at Shenzhen Zhongshan Obstetrics & Gynecology Hospital were enrolled in this study, including 10 patients with thin endometria and 27 controls with normal endometria. The diagnosis of TE was based on their endometrial thickness < 7 mm at the mid-luteal phase or a history of embryo transfer cancellation in vitro fertilization procedures due to TE. Those patients also presented scanty menstruation, poor response to estrogen stimulations, and a normal uterine cavity by hysteroscopy. The normal controls had normal menstrual blood volume, and the endometrial thickness was between > 7 mm and < 12 mm at mid-luteal phase and normal ovarian function. All participants presented regular menstrual cycling, normal karyotype, and negative serological tests for human immunodeficiency virus, hepatitis B virus, hepatitis C virus, and syphilis. Patients with endometriosis, leiomyoma, adenomyosis, or polycystic ovary syndrome were excluded. The samples of endometrium were taken in the uterine body near the fundus under hysteroscopic guidance at the proliferating phase of the participant’s natural menstrual cycles. Of the 27 normal controls, 13 endometria underwent endometrial biopsies using an endometrial curette (Gynetics, Lommel, Belgium) during the proliferative phase, and 14 endometria were collected during the mid-luteal phase (LH days 7–9). Additionally, three normal controls with the proliferative phase, and four normal controls with the secretory phase were used for isolating endometrial CD9⁺SUSD2⁺ cells and CD9⁻ SUSD2⁺ cells for the following cell culture and treatment. Other endometrial samples were used for immunofluorescence. The clinical information of the other 30 patients is listed in Supplementary Table 1.

Isolation and culture of endometrial CD9⁺SUSD2⁺ cells

The functional capacity of CD9⁺SUSD2⁺ cells and CD9⁻ SUSD2⁺ cells isolated from the proliferative and secretory phase of the menstrual cycle samples was compared. Endometrium was scraped from the tissue, minced with a scalpel, and digested with collagenase type Ⅳ (Cat# LS004189, Worthington Biochemical, USA) and DNase I (Cat# 10104159001, Roche, Switzerland) for 1 h at 37 ℃, with regular vigorous shaking in a humidified incubator. The digested tissue was filtered through a 100 μm and 25 μm cell strainer (Cat# 352235, Falcon, USA), respectively. The filtrate was centrifuged at 300 × g for 5 min, and the cell pellet was resuspended in RBC Lysis buffer (Cat# 420301, BioLegend, USA) for 10 min. After re-centrifugation at 300 × g for 5 min, the cell pellet was blocked with 1% BSA (Cat# 9048468, SIGMA, USA) for 30 min, followed by a final centrifugation at 300 × g for 5 min. The endometrial cells were then re-suspended in PBS and stained with human SUSD2-PE (Cat# 327406, BioLegend, USA), CD9-APC (Cat# 312107, BioLegend, USA), and Fix viability dye 506 (Cat# 65086618, Thermo Fisher, USA) for 20 min at room temperature. Then, perivascular CD9⁺SUSD2⁺ cells and CD9⁻SUSD2⁺ cells were sorted using a BD FACSAria™ II with FACS Diva software (BD Bioscience, Germany). The acquired CD9⁺SUSD2⁺ cells and CD9⁻SUSD2⁺ cells were then cultured in DMEM/F12 medium containing 10% FBS in 5% CO₂ at 37 ℃ for 48 h.

Flow cytometry

CD9⁺SUSD2⁺ cells were isolated from the proliferative endometrium to analyze the mesenchymal stem cells (MSCs) markers by flow cytometry. In total, 1 × 10⁵ cells were incubated with 1% bovine serum albumin (BSA) (Cat# A8020, Solarbio, China)/DPBS (Cat# 14190250, Gibco, USA) for 30 min to block nonspecific antigens. Subsequently, the cells were incubated in the dark with fluorescein isothiocyanate (FITC)-labeled anti-CD73 (Cat# 10904-MM07-F, Sino Biological, China), CD90 (Cat# 16897-MM10-F, Sino Biological, China), CD105 (Cat# 561443, BD Pharmingen, USA), CD34 (Cat# 68035-XM01-F, Sino Biological, China), CD45 (Cat# 10086-MM05-F, Sino Biological, China), CD11b (Cat# 03211-50, BioGems, USA), CD19 (Cat# 11880-MM17-F, Sino Biological, China), HLA-DR (Cat# 68038-XM01-F, Sino Biological, China), and Isotype Control FITC (Cat# 44212-50, BioGems, USA) at room temperature for 30 min. The cells were washed twice with PBS and resuspended in 200 µL of PBS. The data were acquired using CytoFLEX (Beckman Coulter, USA). Analysis was performed using CytExpert (Beckman Coulter, USA).

Immunofluorescence

The paraffin-embedded tissue sections were deparaffinized and processed using a Leica BOND RX to automate staining with the PANO Multiplex IHC kit (Cat# 10234100100, PANOVUE, China), according to the manufacturer’s instructions. The automated processing included heat-induced epitope retrieval at 95 ℃ for 10 min in buffer ER1/ER2. The tissue sections were then incubated with primary antibodies for 30 min and secondary antibodies for 15 min at room temperature. The primary antibodies and dilution ratios were as follows: SUSD2 (1:600, Cat# MAB9056, R&D, USA), CD9 (1:150, Cat# MAB25292, R&D, USA), CD31 (1:200, Cat# MA5-13188, Invitrogen, USA), COL1A1 (1:1500, Cat# PA5-29569, Thermo Fisher, USA), ITGB1 (1:800, Cat# RA011880A0HU, Cusabio, China), α-SMA (1:300, Cat# 19245T, CST, USA), and Ki67 (1:200, Cat# ab16667, Abcam, USA). The secondary antibody working solution was added, followed by additional fluorescence staining using Tyramide Signal Amplification (TSA) kits with different fluorophores (TSA480, TSA520, TSA570, TSA650, and TSA780). All sections were counterstained with DAPI and visualized using fluorescent signals from different lasers. Images were captured using the VS200 whole-slide scanner (Olympus, Japan) and analyzed using the HALO-LINK analysis system.

Western blotting

Protein extracts were prepared from the endometrium using RIPA lysing buffer(Cat# P0013B, Beyotime, China). Proteins in the lysates were quantified using a bicinchoninic acid (BCA) assay kit (Cat# P0010, Beyotime, China). Proteins (20 µg in each sample) were separated by SDS-PAGE and transferred onto polyvinylidene difluoride membranes. Non-specific binding sites were blocked with Blotting Grade (Cat# P0216, Beyotime, China) for 1 h at room temperature. The membranes were incubated with primary antibodies against COL1A1 (1:1500, Cat# PA5-29569, Thermo Fisher, USA), ITGB1 (1:1500, Cat# RA011880A0HU, Cusabio, China), α-SMA (1:2000, Cat# 19245T, CST, USA), or GAPDH (1:4000, Cat# #97166, CST, USA) at 4 ℃ overnight. After washing with Tris-Buffered Saline with Tween-20, the blots were incubated with a peroxidase-conjugated anti-rabbit IgG secondary antibody (Cat# A0208, Beyotime, China) and anti-mouse IgG secondary antibody (Cat# A0216, Beyotime, China) for 1 h at room temperature. After washing, protein bands on the blots were visualized using iBright1500 (Thermo Fisher, USA). Analysis was performed using Image J (National Institutes of Health, USA).

Colony-forming unit assay

A colony-forming unit (CFU) assay was performed to determine the colony-forming efficiency and proliferation capacity of CD9⁺SUSD2⁺ cells. Perivascular CD9⁺SUSD2⁺ cells from proliferative and secretory endometrium were seeded in 6-well plates at 1000 cells per well and cultured in 5% CO₂ at 37 ℃. After culturing for 48 h, non-adherent cells were discarded, and adherent cells were washed twice with PBS. The growth medium was replaced with a fresh medium every 3 days. After 2 weeks, colonies were stained with 1% crystal violet solution and counted under the inverted microscope.

Tri-lineage differentiation

CD9⁺SUSD2⁺ cells and CD9⁻SUSD2⁺ cells were used for the tri-lineage differentiation (adipogenesis, osteogenesis and chondrogenesis).

For adipogenic differentiation, cells were seeded in a 12-well plate at a density of 2 × 10⁴ cells/cm² with 0.5 mL of media per well. Once the cells reached 90% confluence, they were transferred to Adipogenic Induction Media (Cat# UCHX-D102R, HyCyteTM, China) and cultured for three days. Following this, they were cultured for one day using an Adipogenic Maintenance Medium and then for an additional three days using an induction medium. This cycle was repeated for a total of 21 days for complete induction. For staining, the medium was removed and the cells were washed with PBS. They were then fixed with 4% neutral formaldehyde for 30 min at room temperature and washed again with PBS. Next, the cells were treated with Oil Red O working solution for 30 min and washed twice with PBS. Finally, the cells were stained with Adipored and observed under fluorescent microscopy.

For osteogenic differentiation, cells were seeded in a 12-well plate at a density of 2 × 10⁴ cells/cm² with 0.5 mL of media per well. When the cells reached 60–70% confluence, the medium was changed to the Osteogenic Induction Medium (Cat# UCHX-D101R, HyCyteTM, China) and cultured for 21 days with media changes every 2–3 days. Cells were subsequently fixed with 4% neutral formalin for 30 min at room temperature. Cells were stained with Alizarin Red Solution for 5 min, then washed twice with PBS. Osteogenic differentiation was observed and assessed under a microscope. Images were captured for evaluation.

For chondrogenic differentiation, 2.0 × 10⁵ cells were centrifuged at 500 g for 10 min to form a small pellet in a 24-well plate. The cell pellets were cultured in Chondrogenic Induction Medium (Cat# UCHX-D203, HyCyteTM, China) and incubated at 37℃, 5% CO₂ for 3 h to allow adherence. Subsequently, 1 mL of complete chondrogenic differentiation medium was gently added to each well. The medium was changed every 2–3 days. After 3 weeks, the medium was aspirated and wells were rinsed once with PBS. Cells were fixed with 4% neutral buffered formalin for 30 min at room temperature, washed twice with PBS, and stained with alcian blue solution for 30 min in the dark. Chondrogenic differentiation was evaluated and documented by light microscopy.

Statistical analysis

Data were analyzed in GraphPad Prism v.8.2.1. If three or more groups were compared, analysis of Friedman’s test followed by Dunn’s multiple comparison for non-normally distributed paired groups, and Kruskal-Wallis test followed by Dunn’s multiple comparison for non-normally distributed unpaired groups. The bioinformatics data were statistically analyzed using an unpaired, two-tailed Mann-Whitney U test with the R language. P < 0.05 was considered to indicate statistical significance.

Results

Single-cell RNA-sequencing analysis reveals a unique population of perivascular CD9⁺SUSD2⁺ cells and dynamic expression of endometrial tissues

To generate a cellular map of the human endometrium across the menstrual cycle, we integrated two publicly available single-cell RNA-seq data sets [16, 17]. After qualifying and normalizing the single-cell RNA-seq dataset, a total of 59,770 cells were grouped into 13 distinct clusters in normal proliferative, secretory endometrium, and thin endometrium (Fig. 1A). Based on the expression of known markers, 13 cell types were identified with typical cell markers and visualized by uniform manifold approximation and projection (UMAP) (Fig. 1B). To gain insight into endometrial cell differentiation, we generated RNA velocity maps for the endometrial subpopulation (Fig. 1C). Notably, perivascular CD9⁺SUSD2⁺ cells were at the start site in the pseudotime trajectory, suggesting that these cells are the progenitor endometrial stem cells. To validate the spatial distribution of CD9⁺SUSD2⁺ cells in endometrial tissue, our immunohistochemical experiments confirmed that CD9⁺SUSD2⁺ cells were indeed localized around the spiral artery (Fig.S1). To further identify the dynamic cellular activities of CD9⁺SUSD2⁺ cells, we performed gene ontology (GO) enrichment analysis of differentially expressed genes (DEGs) between the two groups. Genes were enriched in “ossification”, “stem cell development”, “somatic stem cell division”, and “wound healing” in perivascular CD9⁺SUSD2⁺ cells (Fig. 1D). To determine the stem cell properties of perivascular CD9⁺SUSD2⁺ cells, we used flow cytometry to analyze their immunophenotype using standard mesenchymal stem cell (MSCs) markers. As shown in Fig. 1E, perivascular CD9⁺SUSD2⁺ cells exhibited a typical stem cell profile, with high expression of CD73⁺ (91.1%), CD90⁺ (92.7%), and CD105⁺ (90.9%), while showing negative results for HLA-DR⁻, CD34⁻, CD45⁻, CD11b⁻, and CD19⁻. These results suggested that this perivascular cell type contains progenitor stem cell characteristics.

Fig. 1.

Single-cell transcriptomic data revealed that CD9⁺SUSD2⁺ cells were likely to be the progenitors in the endometrium. A UMAP of cells with the associated cell types in samples among proliferative (n = 3), secretory (n = 3), and thin endometrium (n = 3). Peri, perivascular cell; Str, stromal cell; pStr, proliferative stromal cell. B Expression of classical marker genes of each cell type in the endometrial samples. C Velocities derived from the dynamical model for endometrial samples are visualized as streamlines in a UMAP-based embedding. D Bar plot shows the representative GO terms of significant gene markers for perivascular CD9⁺SUSD2⁺ cells. E Flow cytometry revealed that perivascular CD9⁺SUSD2⁺ cells were positive for CD73, CD90, and CD105, but negative for HLA-DR, CD34, CD45, CD11b, and CD19. Negative staining results were also generated for corresponding isotype control antibodies. Grey peaks show the isotype control and red peaks display the indicated marker

Fig. 2.

CD9⁺SUSD2⁺ cells exhibit stronger stemness among perivascular cells. A Heatmap showing the scaled expression of differentially expressed genes between perivascular CD9⁺SUSD2⁺ cells and the remaining perivascular cells. B Violin plots displaying the expression levels of representative marker genes in perivascular CD9⁺SUSD2⁺ versus the remaining perivascular cells. C Gene Set Enrichment Analysis (GSEA) illustrating pathway enrichment between perivascular CD9⁺SUSD2⁺ and the remaining perivascular cells. D-E Representative in vitro colony formation assays (D) and quantification of cloning efficiency (E) between CD9⁺SUSD2⁺ and CD9⁻SUSD2⁺ cells isolated from the proliferative (Pro) and secretory (Sec) phases. Data are represented as the means ± SD. **P < 0.01; ***P < 0.001

Gene expression signatures and differentiation potential of perivascular CD9⁺SUSD2⁺ cells

To thoroughly depict the signature of CD9⁺SUSD2⁺ cells globally, genes of CD9⁺SUSD2⁺ cells and the remaining perivascular cells were clustered based on their expression patterns (Fig. 2A). We found that CD9⁺SUSD2⁺ cells are characterized by a distinct gene signature enriched for genes involved in cell growth and development, including MYH11, LRRC10B, STC2, AOC3, and LBH (Fig. 2B). In contrast, the remaining perivascular cells exhibited higher expression of DCN, EMC1, SDRP1, OLFML3, PDGFRA and APOD. We further identified through GSEA analysis and found that a significant enrichment of multiple stem cell self-renewal-associated genes of CD9⁺SUSD2⁺ cells was markedly enriched in the signaling pathways regulating the pluripotency of stem cells and the PI3K/AKT signaling pathway (Fig. 2C), suggesting that perivascular CD9⁺SUSD2⁺ cells possess enhanced stem cell-like properties compared to the remaining perivascular cells [20–22]. Given that SUSD2 has been widely recognized as a marker for endometrial mesenchymal stem cells, we further investigated whether CD9 could serve as a more precise stem cell marker. To this end, we performed a CFU assay to compare the proliferative potential between CD9⁺SUSD2⁺ cells and CD9⁻SUSD2⁺ cells isolated from the proliferative and secretory phases of the menstrual cycle, respectively. As expected, clonogenic CD9⁺SUSD2⁺ cells were found in the proliferative group (Pro) and the secretory group (Sec) groups, and the cloning efficiencies were significantly higher than those in CD9⁻SUSD2⁺ cells (Fig. 2D and E). Notably, the cloning efficiency of CD9⁺SUSD2⁺ and CD9⁻SUSD2⁺ cells was similar between the Pro and Sec groups (Fig. 2D and E). Meanwhile, to determine whether perivascular CD9⁺SUSD2⁺ cells possess stem cell differentiation potential, we also sorted the CD9⁺SUSD2⁺ cells and CD9⁻SUSD2⁺ cells from the endometrium during the proliferative and secretory phases, respectively. The CD9⁺SUSD2⁺ cells from the proliferative phase demonstrated a significantly higher differentiation potential into adipocytes compared to the CD9⁻SUSD2⁺ cells from the proliferative phase (Fig. S2B). However, both the CD9⁺SUSD2⁺ cells and CD9⁻SUSD2⁺ cells showed similar potential for osteogenic and chondrogenic differentiation throughout the proliferative and secretory phases (Fig. S2A and S2C). These findings suggested that the perivascular CD9⁺SUSD2⁺ cells were potential regulators in promoting cell proliferation rather than cell differentiation during endometrial growth and repair.

Fig. 3.

Characterization of perivascular CD9⁺SUSD2⁺ cells in patients with thin endometrium. A Proportions of perivascular CD9⁺SUSD2⁺ cells and remaining perivascular cells in proliferative, secretory, and thin endometrium samples. B Immunofluorescence co-staining of cell markers with CD9, SUSD2, and DAPI in proliferative, secretory, and thin endometrium tissues (n = 10 per group). The representative image was taken at a magnification of 150 × field of endometrial cells; Scale bar, 100 μm. C Quantitative analysis showing the percentage of perivascular CD9⁺SUSD2⁺ cells in proliferative, secretory, and thin endometrium tissues. D Scatter plot showing the number of upregulated (red) and downregulated (blue) genes in the control (proliferative phase) compared to the thin endometrium. E Gene Set Enrichment Analysis (GSEA) showing pathway enrichment between control (proliferative phase) and thin endometrium samples. F-G Immunofluorescence for SUSD2, CD9, Ki67, DAPI, and the merged images are shown in the proliferative endometrium from the normal control (F) and the thin endometrium (G). The representative image was taken at a magnification of 400 × field of endometrial cells; Scale bar, 25 μm. H Quantification of the percentage of Ki67⁺CD9⁺SUSD2⁺ cells (n = 10 per group). Ctrl-Pro, normal proliferative endometrium; Sec-Pro, normal secretory endometrium; Thin-Pro, thin proliferative endometrium. Data are represented as the means ± SD. *P < 0.05; ***P < 0.001; ns, not significant

Characterization of perivascular CD9⁺SUSD2⁺ cells in thin endometrium

We then explored the number of perivascular CD9⁺SUSD2⁺ cells in natural endometrium across the menstrual cycle and the proliferative endometrium of TE through scRNA-seq transcriptomics. We found that the percentage of perivascular CD9⁺SUSD2⁺ cells in the proliferative phase (Ctrl-Pro group) was slightly higher than that in the secretory phase (Ctrl-Sec group) (10.7% vs. 9.1%, P > 0.05) (Fig. 3A). Moreover, the percentage of perivascular CD9⁺SUSD2⁺ cells was similar in the proliferative endometrium between normal endometrium and TE (Thin-Pro group) (10.7% vs. 10.4%, P > 0.05). To validate the protein expression of perivascular CD9⁺SUSD2⁺ cells, we performed multiple immunofluorescent stainings for different tissues (Fig. 3B). Interestingly, the results illustrated that the percentage of perivascular CD9⁺SUSD2⁺ cells in the proliferative phase was significantly higher than that in the secretory phase in normal endometrium (11.39 ± 4.05% vs. 5.93 ± 2.25%, P < 0.01) (Fig. 3C), suggesting the perivascular CD9⁺SUSD2⁺ cells in the endometrium have been altered in response to endometrial envrionment remodeling. However, no significant difference in the percentage of perivascular CD9⁺SUSD2⁺ cells was observed between the Ctrl-Pro and Thin-Pro groups (Fig. 3C). This result implied that the number of perivascular CD9⁺SUSD2⁺ cells in the endometrium during the proliferative phase may not be a pathogenic factor for TE. Consequently, we next identified genes that were changing in perivascular CD9⁺SUSD2⁺ cells of the proliferative phase between the Ctrl-Pro and Thin-Pro groups, and characterized those associated with endometrial regeneration and repair. Compared to the Ctrl-Pro group, we observed that 368 genes were significantly upregulated and 534 genes were markedly downregulated in the Thin-Pro group (Fig. 3D). Interestingly, functional enrichment analysis also showed that the toll-like signaling pathway was highly enriched in the perivascular CD9⁺SUSD2⁺ cells of thin endometrium, along with decreased cell cycle and osteoclast differentiation (Fig. 3E). Moreover, immunofluorescence staining showed that the expression of Ki67 in perivascular CD9⁺SUSD2⁺ cells was significantly decreased in the Thin-Pro group (2.21 ± 1.14% vs. 4.34 ± 2.20%, P < 0.05) (Fig. 3F and H). These results suggest that perivascular CD9⁺SUSD2⁺ cells may be in a proinflammatory but limited proliferation state in TE, resulting in a decrease in tissue repair and regenerative capacity.

Functional skewing toward increased collagen deposition is coupled to fibrosis

Finally, to understand the mechanism of the aberrant function of perivascular CD9⁺SUSD2⁺ cells in patients with TE, we used CellChat to investigate cellular cross-talk between different cell subpopulations and perivascular CD9⁺SUSD2⁺ cells of the proliferative phase from normal endometrium and TE. We found that the collagen-related signalling was aberrantly increased in patients with TE. Notably, the most intense crosstalk was between stromal and perivascular CD9⁺SUSD2⁺ cells (Fig. 4A). Perivascular CD9⁺SUSD2⁺ cells exhibited the highest expression of ITGA1 and ITGB1 genes (Fig. 4B). Expression of COL1A1, a gene encoding the type I collagen alpha 1 chain, was higher in stromal cells of thin endometrium than in those from normal endometrium, suggesting significant ECM remodeling changes in the endometrium of patients with TE. To begin with, we validated the presence of COL1A1, ITGB1, and α-SMA protein, which indicated a significant increase in COL1A1, ITGB1, and α-SMA protein levels in the Thin-Pro group (P < 0.05 each) (Fig. 4C and D). To further understand the changes in stromal cells and perivascular CD9⁺SUSD2⁺ cells, we conducted immunohistochemical staining for COL1A1, ITGB1, α-SMA, CD9, SUSD2, and vimentin (Fig. 4E). Our results showed a robust induction of COL1A1 expression in the stroma at sites surrounding the CD9⁺SUSD2⁺ cells in patients with TE (Fig. 4E and F). Additionally, we observed a significant increase in ITGB1 (Fig. 4G) and α-SMA (Fig. 4H) expression in perivascular CD9⁺SUSD2⁺ cells in TE patients, indicating an imbalance in the endometrial ECM remodeling. This was further confirmed by Masson staining (Fig. S3). Taken together, our unbiased investigation of the signaling pathways between perivascular CD9⁺SUSD2⁺ cells and other cell subpopulations highlighted that the collagen-related signaling pathway was an important regulator of perivascular CD9⁺SUSD2⁺ cells’ function and that their aberrant expression contributed to insufficient endometrial regeneration during the proliferative phase.

Fig. 4.

Collagen deposition surrounding perivascular CD9⁺SUSD2⁺ cells is closely associated with thin endometrium. A CellChat analysis showing the number of incoming signaling interactions from other cell types to perivascular CD9⁺SUSD2⁺ cells in normal versus thin endometrium samples. B Expression patterns of collagen signaling-related genes in normal and thin endometrium. C Western blotting analysis of the COL1A1, ITGB1, α-SMA, and GAPDH in normal and thin endometrium (n = 6 per group). DThe relative protein levels of CD9, SUSD2, and ITGB1 were analyzed between normal and thin endometrium. E Immunofluorescence co-staining of CD9, SUSD2, ITGB1, α-SMA, COL1A1, and DAPI in proliferative and thin endometrial tissues (n = 10 per group). F-H Quantitative analyses of the percentage of COL1A1 in stromal cells (F), ITGB1 (G), and α-SMA (H) in CD9⁺SUSD2⁺ cells in proliferative and thin endometrial tissues. Ctrl-Pro, normal proliferative endometrium; Thin-Pro, thin proliferative endometrium. Data are presented as mean ± SD. *P < 0.05; **P < 0.01

Disscussion

Thin endometrium represents an incompletely understood gynecological disorder. To address these gaps, in this study, we performed scRNA-seq analysis on thin and normal endometria from two publicly available databases. This allowed us to comprehensively characterize the mechanisms of endometrial regeneration and repair at the proliferative phase of the menstrual cycle, which is particularly critical for diagnosing and treating TE. We identified unique perivascular CD9⁺SUSD2⁺ cells, which have the ability to self-renew, have high proliferative potential, and can differentiate into adipogenic and osteogenic cells. We also found evidence of dysfunction in perivascular CD9⁺SUSD2⁺ cells in patients with TE, as shown by increased collagen deposition and changes in the ECM microenvironment. The identification of pervascular CD9⁺SUSD2⁺ cells, aberrant collagen-related signaling programs, and regional distinctions advances our understanding of the intricate molecular landscape in TE pathogenesis.

We focus on comparing the regeneration and repair characteristics of thin and normal endometrium and seek targeted treatment strategies for TE. An increasing body of evidence suggests the existence of endometrial stem/progenitor cells (eMSCs) in the human endometrium [23]. This study identified unique perivascular CD9⁺SUSD2⁺ cells as eMSCs through single-cell analysis. SUSD2 (also known as W5C5) has been reported as a marker of adult stem cells [24]. In humans, SUSD2⁺ eMSCs are essential for uterine gland development and angiogenesis, and are mainly expressed in the pericyte and perivascular cells of the endometrial basalis and functionalis [6, 25]. CD9 is a transmembrane four-family of molecules, and is expressed in the menstrual cycle and associated with integrins α3, α6, and β1 in the human endometrium [26, 27]. CD9 was also reported to be expressed on mouse and rat spermatogonial stem cells [28]. In the present study, immunofluorescent analysis revealed the presence of uterine CD9⁺SUSD2⁺ cells in the perivascular niche in both the proliferative and secretory endometrial tissues. The percentage of perivascular CD9⁺SUSD2⁺ cells significantly increased in the proliferative endometrium and decreased in the secretory endometrium, suggesting their role in the growth of the endometrial functionalis stroma. In patients with TE, there was only a somewhat decreased percentage of perivascular CD9⁺SUSD2⁺ cells during the proliferative phase of the menstrual cycle, hinting at underlying molecular feature changes of CD9⁺SUSD2⁺ cells. It is therefore essential to understand how perivascular CD9⁺SUSD2⁺ cells integrate niche signals to maintain homeostasis.

In addition to their clonal capacity and multilineage potential, scRNA-seq revealed significant shifts in cellular composition and differentiation patterns in perivascular CD9⁺SUSD2⁺ cells. According to our findings, the perivascular CD9⁺SUSD2⁺ cells show a distinct molecular profile that is directly related to proangiogenic and anti-fibrotic MSC actions. Specifically, increased MYH11, LRRC10B, STC2, AOC3, and LBH gene expression in CD9⁺SUSD2⁺ cells is a potent mechanism by which MSCs exert their immunosuppressive, proangiogenic, and anti-fibrotic effects [29–32]. Furthermore, notable alterations in signaling pathways, including the toll-like signaling pathway, cell cycle, and osteoclast differentiation, indicate disrupted cellular function, potentially affecting the balance of the endometrial environment essential for tissue regeneration and repair.

The ECM is broadly defined as the non-cellular component of tissues and organs. The critical role of ECM in cell biology, and particularly in endometrial physiology, has been increasingly appreciated with the realization that endometrial cells and their ECM interact strongly [33–36]. The interaction forms a bidirectional feedback loop [37]. On the one hand, the ECM instructs cell morphology, migration, and differentiation, as well as mediates cell-cell interactions. On the other hand, cells actively remodel the composition, geometry, and mechanics of the ECM [38]. ECM accumulation and remodeling are thought to be crucial for fibrotic diseases such as uterine adhesion [39]. ECM deposition is characterized by enhanced levels of protein markers, such as type I collagen (collagen I) and alpha-smooth muscle actin (α-SMA) [40]. Our analysis also uncovered perturbations in crosstalk between perivascular CD9⁺SUSD2⁺ cells and stromal cells in patients with TE, emphasizing aberrantly increased collagen-related signalling. Collagen cross-linking mediated by COL1A1/ITGB1 has been thought to contribute to increasing fibrosis [41]. Our studies identified that, in certain contexts, increased collagen deposition leads to α-SMA expression—a myofibroblast marker— in perivascular CD9⁺SUSD2⁺ cells. This up-regulation of α-SMA likely promotes the enhanced myofibroblast differentiation of CD9⁺SUSD2⁺ cells in TE. Given that enhanced myofibroblast differentiation induces fibrosis development [42, 43], our findings suggest that the observed changes collectively contribute to promoting a pro-fibrotic niche in TE. This fibrogenic microenvironment may further dysregulate ECM remodeling, impairing endometrial regeneration and repair. Such mechanisms may not only be relevant to thin endometrial pathologies but also extend to other endometrial disorders characterized by aberrant ECM remodeling, even if their functional role in disease progression remains unclear. Although the causal relationship between overdeposition of ECM and TE has not been fully elucidated, our findings highlight the need for further research on the therapeutic effects of antifibrotic treatments (e.g., pirfenidone) for endometrial repair, particularly in our previously established TE model [44].

The strength of our study lies in the use of the scRNA-seq strategy to unravel the cellular and molecular characteristics of CD9⁺SUSD2⁺ cells associated with the perivascular progenitor population, providing evidence for the involvement of stem cell/progenitor cells in TE pathogenesis. It may reshape our understanding of TE pathology that aberrant perivascular CD9⁺SUSD2⁺ cells impair ECM remodeling and endometrial regeneration. While our study is informative, it has limitations that must be addressed in future studies. The sample size was limited, and the functional role of the perivascular CD9⁺SUSD2⁺ cells and signaling pathways in TE is yet to be fully understood. Future studies should aim to validate these findings in larger cohorts and employ functional assays to confirm the roles of this subpopulation and pathways in the progression of TE. Additionally, the clinical implications of our findings, such as their relationship to symptoms like thin endometrium, require further exploration.

Conclusions

In summary, this study elucidates the characteristics of perivascular CD9⁺SUSD2⁺ cells during the proliferative phase at single-cell resolution in both normal and thin endometrium, and describes detailed gene signatures and communication in this cell type. The study provides a potential regulatory mechanism of perivascular CD9⁺SUSD2⁺ cells involved in human endometrial remodeling and repair. Our study uncovers that the potential regulatory network of the COL1A1/ITGB1/α-SMA signal axis may help to develop precise diagnosis and treatment of TE, and highlights the need for clinical validation of these promising findings.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

TE

thin endometrium

PRP

platelet-rich plasma

FET

frozen-thawed embryo transfer

SUSD2

sushi domain containing-2

scRNA-seq

single-cell RNA sequencing

GEO

Gene Expression Omnibus

PCA

principal component analysis

DEA

differential expression analysis

CFU

colony-forming unit

UMAP

uniform manifold approximation and projection

GO

gene ontology

DEGs

differentially expressed genes

ECM

extracellular matrix

Author contributions

LBF conducted the experiment; QLB performed all bioinformative analysis; CMY collected the data; MCR collected the endometrial samples. DLH, DLJ, and ZT revised the manuscript; CX and QLB designed the research and wrote the manuscript. All authors read and approved the final manuscript.

Funding

This work was supported by the National Natural Science Foundation of China (No. 82401975), the Basic and Applied Basic Research Foundation of Guangdong (No. 2024A1515010478), the Shenzhen Medical Research Fund (No. A2403029), the Shenzhen Science and Technology Program (No. JCYJ20220818103207016).

Data availability

The data supporting the findings of this study are available within the article and its supplementary materials.

Declarations

Ethics approval and consent to participate

Experiments involving human endometrial tissues were approved by the Ethics Committee of Shenzhen Zhongshan Urology Hospital (No. SZZSECHU-2022008; Project title: Study on the mechanism of exosomes regulating endometrial function and its relationship with placenta-related diseases; Date of approval: December 27, 2022). Informed written consent was obtained from each participant before the endometrial biopsy.

Consent for publication

Not applicable.

Competing interests

The authors declare that the research was conducted without any commercial or financial relationships that could potentially create a conflict of interest.