Abstract

   Intrauterine adhesions (IUAs) are a complex condition that frequently
   results in menstrual disturbances, infertility, and obstetric
   complications. Unfortunately, the underlying pathophysiology of IUAs
   remains poorly understood, and current treatments often exhibit limited
   efficacy. We performed the single-cell RNA-sequencing (scRNA-Seq)
   comparison of 5 endometrial tissues, including patients with confirmed
   intrauterine adhesions and healthy controls(HCs). We profiled the
   transcriptomes of 55,308 primary human endometrial cells isolated from
   healthy controls and intrauterine adhesions patients at single-cell
   resolution. Compared with those in HCs, the number of fibroblasts
   derived from IUAs significantly decreased. Further analysis found that
   fibroblasts subcluster 3 were enriched in the IUAs, whereas opposite in
   HCs. GO enrichment analysis revealed that specific genes of fibroblasts
   subcluster 3 were markedly involved in the positive regulation of
   embryonic placenta development, the response to reactive oxygen species
   and female pregnancy, and the positive regulation of the mitotic DNA
   damage checkpoint and DNA damage response. In addition, the proportion
   of proliferating endothelial cells was significantly lower in IUAs. GO
   enrichment analysis revealed that the specific genes were markedly
   involved in the positive regulation of cell cycle arrest, the cellular
   response to interferon − gamma and the negative regulation of the
   mitotic cell cycle. According to the number of intercellular
   receptor–ligand pairs, we identified endothelial cells as the core cell
   population. Our study provides deeper insights into the endometrial
   microenvironment disorders that are potentially applicable to improving
   therapeutics for IUAs.

   Keywords: Single cell RNA-sequencing, Intrauterine adhesions,
   Endometrial cells

   Subject terms: Reproductive disorders, Genetics research

Introduction

   Intrauterine adhesions (IUAs) may develop following various endometrial
   injuries, including postpartum curettage, hysteroscopic myomectomy,
   intrauterine infections, and pregnancy termination procedures. While
   the incidence of IUAs after first-trimester abortion is estimated at
   15–20%, this risk escalates to 30–40% following repeated procedures or
   those complicated by infection^[34]1–[35]3. Hence, IUAs have become
   among the largest medical burdens in the field of reproductive health.
   Unfortunately, the pathophysiology of intrauterine adhesions remains
   unclear, and the clinical efficacy of treatments is poor.

   The full-thickness endometria in the uterus are essential for
   fertilization and embryonic development. Human endometria, which are
   composed mainly of endometrial epithelial and stromal cells, exhibit
   remarkable plasticity and undergo repeated shedding and regeneration.
   Full-thickness injury or dysfunction of the human endometria causes
   intrauterine adhesions, miscarriage, and uterine factor
   infertility^[36]4. The development of new regenerative technologies
   against intrauterine adhesions is hindered by our incomplete
   understanding of the molecular characteristics of the cell populations
   responsible for endometrial healing without scarring and fibrosis. The
   tissue microenvironment is indispensable during tissue development,
   homeostasis, regeneration, and disease progression^[37]5–[38]7.
   Single-cell analysis has been increasingly utilized to dissect cell
   heterogeneity and study dynamic cell population architectures and their
   regulation during biological processes such as development, tissue
   homeostasis, and pathology. Here, we used single-cell RNA sequencing
   (scRNA-seq) to identify cell populations in endometrial tissues. Our
   study provides novel insights into the cellular architecture of
   intrauterine adhesions and potential therapeutic strategies.

Results

Single-cell transcriptional landscape of endometrial tissue in healthy women
and women with intrauterine adhesions

   To reveal the cellular diversity and gene signatures, we performed
   single-cell sequencing from five patients, including three normal
   endometrial samples derived from three healthy control women (HCs;
   i.e., HC1, HC2 and HC3) and two adhesion endometrial samples derived
   from two women with intrauterine adhesions (IUAs; i.e., IUA1 and IUA2),
   via 10X Genomics Sequencing (Fig. [39]1A). The enrolled patients were
   diagnosed recently and had not received therapy in the last six months.
   After quality filtering, 55,308 cells were detected, among which 29,813
   cells were collected from HCs and 25,495 from IUAs. After
   dimensionality reduction and unsupervised cell clustering, we
   identified B cells (634; MS4 A1, CD79 A, and CD79B), ciliated cells
   (988; FOXJ1, PIFO, TPPP3, CCDC17, and SNTN), endothelial cells (4257;
   CDH5, CLDN5, PECAM1, VWF, and KDR), fibroblasts (20,348; LUM, DCN, COL1
   A2, COL1 A1, and PDGFRA), macrophages (2153; LYZ, CD14, C1QC, MRC1,
   CD68, XCR1, and CD1 C), mast cells (492, TPSAB1, TPSB2, and CPA3),
   mural cells (4992, RGS5, ACTA2, TAGLN, MYL9, and MYLK), proliferating
   epithelial cells (2076; Mki67, and Top2a), secretory epithelial cells
   (12,418, PAX8, MUC1, WFDC2, GABRP, TFF3, and KRT18), mural cells (434;
   CD79 A, JCHAIN and MZB1) and T cells (6516; CD2, CD3D, TRAC, and TRBC2)
   as eleven distinct lineages based on marker gene expression.

Fig. 1.

   [40]Fig. 1
   [41]Open in a new tab

   Single-cell transcriptional landscape of endometrial tissue in healthy
   women and intrauterine adhesion women. (A) Schematic diagram of
   scRNA-seq analysis workflow. (B) tSNE plots for cell type
   identification of 55,308 high-quality cells. (C) tSNE plot colors by
   spatial distribution of cells in healthy women and intrauterine
   adhesion women’s endometrial tissue. (D) Proportion of cell types in
   endometrial tissues from healthy controls and IUA patients. The Y-axis
   represents the percentage of each cell type within the total cell
   population. (E) Dot plot showing the top marker genes revealed by
   single-cell RNA sequencing of cell types defined in(B). ECs:
   endothelial cell;MPs:macrophages;HC:healthy control;IUA:intrauterine
   adhesion.

   In this project, eleven main cellular components were identified and
   visualized in UMAP plots (Fig. [42]1B, [43]C). Cell proportion analysis
   revealed that the most abundant cells in the endometrial tissue were
   fibroblasts (Fig. [44]1D). In addition, the most plentiful cells in the
   endometrial tissue derived from IUAs were also fibroblasts. However,
   compared with those in HCs, the number of fibroblasts derived from IUA
   patients significantly decreased, whereas the number of T cells
   significantly increased. All the clusters were assigned according to
   the expression levels of specific marker genes (Fig. [45]1E).

Identification of fibroblasts types and their marker genes across endometrial
tissue in healthy women and women with intrauterine adhesions

   Owing to the different dissociation efficiencies of different cell
   types, fibroblasts cells accounted for the largest proportion of all
   samples. In total, 24,448 fibroblasts were subjected to unsupervised
   clustering to reveal subtypes. In accordance with previous studies,
   genes (COL1 A1, COL1 A2, DCN, LUM, and PDGFRA) were used to identify
   fibroblasts (Fig. [46]2A). Five subclusters were subsequently
   identified with unique signature genes, including fibroblasts
   subclusters 1–5 (Fig. [47]2B): subcluster 1 (IGFBP5, IL17RB, and SLC26
   A7), subcluster 2 (WIF1, ZCCHC12, and LRRTM1), subcluster 3 (FOSB,
   ZFP36, and ATF3), subcluster 4 (CLSPN, FAM111B, and MCM10) and
   subcluster 5 (ADAMDEC1, PLA2G2 A, and RPS4Y1) (Fig. [48]2D). As
   depicted in Fig. [49]2C, fibroblasts subclusters 3 and 5 were
   predominantly present in IUAs. In contrast, the proportion of
   subcluster 1, 2, and 4 cells was significantly reduced in IUAs compared
   to HCs.

Fig. 2.

   [50]Fig. 2
   [51]Open in a new tab

   fibroblasts Cell Types and Their Marker Genes. (A) UMAP plots of
   expression of the marker genes (COL1 A1, COL1 A2, DCN, LUM, PDGFRA) for
   fibroblasts. (B) tSNE plots showing cell types identified by marker
   genes, including fibroblasts cells subclusters 1–5. (C) the proportion
   of five fibroblasts cell types in healthy women and intrauterine
   adhesion women.(D) Dot plot showing the top marker genes revealed by
   single-cell RNA sequencing of cell types defined in (B). (E) the
   heatmap of difference genes of fibroblasts subclusters 3 cells between
   two groups.(F-G) Gene ontology (GO) enrichment analysis showing the
   enriched pathways in fibroblasts subclusters 3 cells.

Specific Phenotypes of Subcluster 3 fibroblasts

   Since fibroblasts cluster 3 accounts for the highest proportion of
   intrauterine adhesions, we further explored it. Interestingly, we found
   that fibroblasts subcluster 3 with high expression of FOSB, ZFP36 and
   ATF3 were expanded in endometrial tissues derived from women with
   intrauterine adhesions (Fig. [52]2D). Furthermore, we also identified
   the specific genes associated with fibroblasts and their biological
   functions in HCs and IUAs. With the cut-offs of logFC = 0.25 and p <
   0.05, 105 specific genes were significantly upregulated, whereas 69
   specific genes were significantly downregulated in the IUAs. A heatmap
   of the differentially expressed genes in the fibroblasts subcluster 3
   revealed the top twenty upregulated and downregulated genes between the
   two groups (Fig. [53]2E). GO enrichment analysis revealed that specific
   genes were markedly involved in the positive regulation of embryonic
   placenta development, the response to reactive oxygen species and
   female pregnancy, the negative regulation of the mitotic DNA damage
   checkpoint and the DNA damage response (Fig. [54]2F, [55]G).

Reconstruction of the Temporal Dynamics of fibroblasts

   To investigate the underlying evolution among fibroblasts, the Monocle
   tool was adopted to reveal pseudotemporal ordering for the similarity
   of cell clusters with developmental lineages. For fibroblasts, the
   trends of pseudotime‐dependent genes along the pseudotime line were
   divided into five cell clusters of fibroblasts with diverse expression
   dynamics (Fig. [56]3A, [57]B). The results clearly demonstrated the
   uniform development of fibroblasts from cluster 1 to cluster 5
   (Fig. [58]3C). Furthermore, we observed that fibroblasts clusters 2 and
   3 from HCs were rich in the beginning of the differentiation process
   and were sequentially transformed into cluster 5 fibroblasts under
   regenerative conditions (Fig. [59]3D). Massive specific regulons of
   fibroblasts subclusters were identified via SCENIC analysis by
   calculating the regulon specificity score (RSS), and transcription
   factor expression was found to be heterogeneous across subclusters of
   fibroblasts (Fig. [60]3E). Some regulators were particularly enriched
   in certain specific cell subclusters. Retinoic acid receptor gamma
   (RARG), TEA domain family member 1 (TEAD1), nuclear factor I-A (NFIA),
   peroxisome proliferator activated receptor gamma (PPARG) and nuclear
   receptor subfamily 2 F member 1 (NR2 F1) were abundant in fibroblasts
   cluster 1, and dorsal root ganglion homeobox (DRGX), oestrogen receptor
   1 (ESR1), homeobox A10 (HOX1 A10), homeobox A11 (HOX1 A11) and Wilm
   tumour gene 1 (WT1) were highly enriched in fibroblasts cluster 2.
   Early growth response protein 2 (EGR2), Krüppel-like factor 4 (KLF4),
   Krüppel-like factor 10 (KLF10), basic helix–loop–helix family member
   e40 (BHLHE40) and MYC were highly enriched in fibroblasts cluster 3.
   MYB proto-oncogene like 2 (MYBL2), E2 F transcription factor 8 (E2 F8),
   E2 F transcription factor 2 (E2 F2), enhancer of zeste homologue 2
   (EZH2) and E2 F transcription factor 1 (E2 F1) were highly enriched in
   fibroblasts cluster 4. NK2 homeobox 3 (NKX2-3), Runt-related
   transcription factor 1 (RUNX1), avian reticuloendotheliosis viral
   (v-rel) oncogene related B gene (RELB), trefoil factor 3 (TFF3) and
   basic helix–loop–helix family member e22 (BHLHE22) were highly enriched
   in fibroblasts cluster 5.

Fig. 3.

   [61]Fig. 3
   [62]Open in a new tab

   Temporal Dynamics of fibroblasts. (A) The differential expressed genes
   along the trajectory were clustered into five gene sets. Different
   colors indicated the gene expression levels from low to high. (B)
   Expression patterns of representative genes along the reprogramming
   trajectory of fibroblasts subclusters. (C-D)Pseudotime ordering of
   fibroblasts The SCEINIC analysis of expression level of specific
   regulons in fibroblasts sub-clusters.

Identification of endothelial cell types and their marker genes across
endometrial tissue in healthy women and women with intrauterine adhesions

   Endothelial cells play a key role in controlling vascular function and
   regulating biological function. In total, 4257 endothelial cells were
   subjected to unsupervised clustering to reveal subtypes. In accordance
   with previous studies, genes (CDH5, CLDN5, KDR, PECAM1, and VWF) were
   used to identify endothelial cells (Fig. [63]4A). Four subclusters were
   subsequently identified with unique signature genes, including arterial
   endothelial cells(AECs) (FBLN5, [64]AC004947.1, and SSTR1)
   (Fig. [65]4B), capillary endothelial cells(CapECs) (ADASTML2, SPP1, and
   COL15 A1), proliferating epithelial cells (SHCBP1, DLGAP5, and HMMR),
   and venous endothelial cells(VECs) (ACKR1, CCL23, and CPE)
   (Fig. [66]4D). As shown in Fig. [67]4C, the proportion of proliferating
   endothelial cells was significantly lower in IUAs than in HCs.

Fig. 4.

   [68]Fig. 4
   [69]Open in a new tab

   Endothelial cells Types and Their Marker Genes. (A)UMAP plots of
   expression of the marker genes (CDH5, CLDN5, KDR, PECAM1, VWF) for
   endothelial cells. (B) tSNE plots showing cell types identified by
   marker genes, including AECs, CapECs, Proliferating ECs and VECs. (C)
   the proportion of four endothelial cells in healthy women and
   intrauterine adhesion women. (D) Dot plot showing the top marker genes
   revealed by single-cell RNA sequencing of cell types defined in (B).
   (E) the heatmap of difference genes of proliferating endothelial cells
   between two groups. (F-G) Gene ontology (GO) enrichment analysis
   showing the enriched pathways in proliferating endothelial cells.

Specific phenotypes of proliferating endothelial cells

   The data revealed that the proportion of proliferating endothelial
   cells was significantly decreased in IUAs, so we further explored this
   cluster. We identified the specific genes associated with proliferating
   endothelial cells and their biological functions in normal and IUA
   endometrial tissue. With the cut-offs of |FC|> 1.2 and p < 0.05, we
   found that 633 specific genes were significantly upregulated, whereas
   525 specific genes were significantly downregulated in IUAs compared
   with HCs. A heatmap of differentially expressed genes in proliferating
   endothelial cells revealed the top twenty upregulated and downregulated
   genes between the two groups (Fig. [70]4E). GO enrichment analysis
   revealed that the specific genes were markedly involved in the positive
   regulation of cell cycle arrest, the cellular response to
   interferon-gamma, and the negative regulation of the mitotic cell cycle
   (Fig. [71]4F, G).

Reconstruction of the Temporal Dynamics of Endothelial Cells

   To investigate the underlying evolution among endothelial cells, the
   Monocle tool was adopted to reveal pseudotemporal ordering for the
   similarity of cell clusters with developmental lineages. For
   endothelial cells, the trends of pseudotime‐dependent genes along the
   pseudotime line were divided into four cell clusters of endothelial
   cells with diverse expression dynamics (Fig. [72]5A, B). The results
   clearly demonstrated the uniform development of endothelial cells
   (Fig. [73]5C). Furthermore, we observed that proliferating endothelial
   cells in normal endometrial tissues were enriched at the beginning of
   the differentiation process (Fig. [74]5D). Massive specific regulons of
   endothelial cell subclusters were identified via SCENIC analysis by
   calculating the regulon specificity score (RSS). Some regulators were
   particularly enriched in certain specific cell subclusters
   (Fig. [75]5E). Retinoic acid receptor gamma (RARG),TEA domain family
   member 1 (TEAD1),Nuclear factor I-A (NFIA),Peroxisome proliferator
   activated receptor gamma (PPARG) and Nuclear receptor subfamily 2 group
   F member 1 (NR2 F1) were abundant in AECs, while dorsal root ganglia
   homeobox (HEYL), Oestrogen receptor 1 (ESR1), Homeobox A10 (HOXA10),
   Homeobox A11 (HOXA11) and Wilm tumour gene 1 (WT1) were highly enriched
   in CapECs. Early growth response protein 2 (EGR2),Krüppel-like factor 4
   (KLF4),Krüppel-like factor 10 (KLF10),Basic helix-loop-helix family
   member e40 (BHLHE40) and tumor protein 73(TP73) were highly enriched in
   proliferating endothelial cells. MYB proto-oncogene like 2 (MYBL2),E2 F
   transcription factor 8 (E2 F8),E2 F transcription factor 2 (E2
   F2),Enhancer of zeste homologue 2 (EZH2) and E2 F transcription factor
   1 (E2 F1) were highly enriched in VECs.

Fig. 5.

   [76]Fig. 5
   [77]Open in a new tab

   Temporal Dynamics of Endothelial Cells. (A) The differential expressed
   genes along the trajectory were clustered into four gene sets.
   Different colors indicated the gene expression levels from low to high.
   (B) expression patterns of representative genes along the reprogramming
   trajectory of endothelial cells subclusters.(C-D) Pseudotime ordering
   of endothelial cells. (E) The SCEINIC analysis of expression level of
   specific regulons in endothelial cells sub-clusters.

Cell-to-Cell interactions based on Ligand–Receptor interactions

   Intrauterine adhesion development is a complex process that
   necessitates the collaborative efforts of diverse cell lineages.
   Cell-to-cell communication across diverse cell types thoroughly governs
   the appropriate functions of metazoans and widely relies on
   interactions between secreted ligands and cell-surface receptors.
   Ligand–receptor interactions were matched on the basis of the marker
   genes. The numbers of ligand–receptor pairs for fibroblasts,
   endothelial cells and macrophages were 426, 595 and 526, respectively
   (Fig. [78]6A). According to the number of intercellular receptor–ligand
   pairs (Fig. [79]6B), we screened endothelial cells as the core cell
   population. We evaluated the biological functions of ligand and
   receptor genes in endothelial cells and fibroblasts in the two
   groups.Our results indicated that in the IUA group, ligand and receptor
   genes were primarily involved in peptidyl-tyrosine phosphorylation,
   rpeptidyl-tyrosine modification, and wound healing processes between
   endothelial cells and fibroblasts (Fig. [80]6C). In contrast, in the
   control group, these genes were mainly associated with
   peptidyl-tyrosine phosphorylation, rpeptidyl-tyrosine modification, and
   the regulation of chemotaxis between endothelial cells and fibroblasts
   (Fig. [81]6C). Additionally, we observed that cell migration was a
   significant biological function of the ligand and receptor genes
   interacting between endothelial cells and macrophages (Fig. [82]6D).

Fig. 6.

   [83]Fig. 6
   [84]Open in a new tab

   Cell–Cell Interactions Based on Ligand– Receptor Interactions. (A)
   Potential interactions between fibroblasts, endothelial cells and
   macrophages based on receptor-ligand pairs in IUAs. The width of the
   line represented the number of receptor-ligand pairs. (B) The number of
   ligands/receptors for fibroblasts, endothelial cells and macrophages in
   IUA and control group. (C) Gene ontology (GO) enrichment analysis
   showing the enriched biological function by these ligand and receptor
   genes between endothelial cells and fibroblasts in IUA. (D) Gene
   ontology (GO) enrichment analysis showing the enriched biological
   function by these ligand and receptor genes between endothelial cells
   and macrophages in IUA.

Discussion

   Endometrial fibrosis is considered the main histopathological
   lesion^[85]8,[86]9 in intrauterine adhesions. Unfortunately, few
   interventions have improved live-birth rates in patients with
   endometrial fibrosis until now, possibly because the underlying
   mechanisms are incompletely understood.

   In the present study, we aimed to further clarify the underlying
   mechanisms of intrauterine adhesions from the perspective of
   single-cell changes. We observed notable alterations in the cellular
   composition following intrauterine adhesions, with significant changes
   in the proportion of different cell types. Compared with HCs, the
   proportion of fibroblasts decreased from 45 to 26% after intrauterine
   adhesions. Consistent with previous studies, we found that fibroblasts
   are among the cell types with the most pronounced changes after
   intrauterine adhesions. Furthermore, by identifying unique signature
   genes, the fibroblasts were divided into five subclusters. Compared
   with the normal group, clusters 3 and 5 cells were significantly
   increased after intrauterine adhesions. Moreover, the significance of
   clusters 1, 2 and 4 cells decreased. Clusters 3 cells were the type
   with the most significant change in number. Hence, we believe that
   cluster 3 cells may play a key role in the pathological mechanism of
   fibroblasts regulation in IUAs. By GO and KEGG analyses, we found that
   fibroblasts cluster 3 cells can regulate the biological processes of
   pregnancy in females. We propose that fibroblasts cluster 3 cells are a
   pivotal cell type in intrauterine adhesions that lead to infertility,
   and future therapeutic strategies could focus on specifically
   modulating these cells.

   An adequate supply of nutrients is necessary for tissue damage repair.
   Suitable angiogenesis is a key process in ensuring the nutrient supply
   to the body^[87]10. The biological behaviour of vascular endothelial
   cells is considered to play a key role in regulating
   angiogenesis^[88]11,[89]12. Furthermore, current evidence has shown
   that endometrial tissues present with vascular closure in a model of
   intrauterine adhesion and that angiogenesis in endometrial tissues
   affects endometrial repair^[90]13,[91]14. In our study, the proportion
   of vascular endothelial cells increased from 6.5% to 9% after
   intrauterine adhesions and proliferating endothelial cells were the
   most significantly decreased subcluster. Many previous studies have
   indicated that endothelial cells are important players in regenerating
   tissue as well as in the vascularization of tumours and that
   endothelial cells are fundamental components of blood vessels, with
   their proliferation essential for new vessel
   formation^[92]12,[93]15–[94]17. Notably, our pseudotime analysis
   revealed dynamic transitions between EC subtypes. The observed
   reduction in proliferating endothelial cells suggest impaired
   angiogenic potential in IUAs consistent with prior reports of vascular
   rarefaction^[95]18. These findings align with clinical observations of
   diminished endometrial receptivity in IUAs. Hence, we suggest that the
   proliferation of endothelial cell subclusters may be a potential target
   for treating intrauterine adhesions.

   IUAs is characterized by fibrosis, microvascular rarefaction and
   chronic inflammation^[96]19,[97]20. The uterine cavity microenvironment
   is organized into different cell types, which strongly interact with
   each other by a number of paracrine or autocrine factors. Cell-to-cell
   communication is indispensable for tissue development but also plays a
   vital role in regulating the response of the uterus to damage^[98]21.
   The proportions of endothelial cells and fibroblasts underwent
   significant changes in IUAs, making them key contributors to the
   condition following uterine injury or infection. Historically, research
   on intrauterine adhesions has largely focused on individual cell types,
   such as endothelial cells or fibroblasts^[99]22,[100]23. The role of
   cell-to-cell communication between endothelial cells and fibroblasts in
   intrauterine adhesions remains unclear. Our study revealed that the
   proportion of intercellular communication with endothelial cells as the
   core increased from 15.3% to 16.9% after intrauterine adhesions, while
   that with fibroblasts as the core decreased from 13.3% to 12.1%. These
   findings suggest that endothelial cells may be among the most active
   cell types following intrauterine adhesions. Given that endothelial
   cells form the basic structure of blood vessels and are widely
   distributed around other cells in the uterine cavity microenvironment,
   they have a significant advantage in intercellular communication.
   Moreover, paracrine signals mediated by endothelial cells have multiple
   positive or negative effects on tissue repair^[101]24. Therefore,
   identifying the factors secreted by endothelial cells that can
   significantly improve the severity of intrauterine adhesions by
   regulating cell-to-cell communication will be an important direction
   for future research. The findings of this study may also provide a new
   and effective target for improving outcomes after intrauterine
   adhesions through the regulation of intercellular communication.

   The primary limitation of our study is the small sample size,
   consisting of only five subjects, which restricts the generalizability
   of our findings and the power of the statistical analyses. This
   constraint was due to the specificity of the patient population and the
   invasive nature of endometrial tissue collection. Although we
   implemented rigorous quality control and advanced bioinformatics to
   maximize insights from the available data, we acknowledge the need for
   larger, more diverse cohorts in future studies to validate our results
   and explore stratified analyses. Increasing the sample size will also
   allow for multivariable analyses and control of potential confounders,
   enhancing the robustness of the conclusions drawn.

Methods

Human endometrial tissue acquisition

   All tissue samples were collected from Xiangya Hospital Central South
   University.All participants provided written informed consent. The
   study protocol adhered to the Declaration of Helsinki and was approved
   by Xiangya Hospital Ethics Committee (No.2022111228) The samples in the
   experimental group were collected from the discarded pathological
   tissue after hysteroscopic adhesion separation, whereas the control
   group were collected from the endometrial tissue after hysterectomy due
   to uterine myoma. All the clinical information were showed in S table
   as supplementary material.

Tissue dissociation and preparation

   The fresh tissues were stored in the sCelLive™ Tissue Preservation
   Solution (Singleron) on ice within 30 min after surgery. The samples
   were washed with Hanks’ balanced salt solution (HBSS) three times,
   minced into small pieces, and then digested with 3 mL of sCelLive™
   Tissue Dissociation Solution (Singleron) with a Singleron PythoN™
   Tissue Dissociation System at 37 °C for 15 min. The cell suspension was
   collected and filtered through a 40-μm sterile strainer. Afterwards,
   GEXSCOPE® red blood cell lysis buffer (RCLB, Singleron) was added, and
   the mixture [cell:RCLB = 1:2 (volume ratio)] was incubated at room
   temperature for 5–8 min to remove red blood cells. The mixture was then
   centrifuged at 300 × g and 4 °C for 5 min to remove the supernatant,
   after which the mixture was gently suspended in PBS. Finally, the
   samples were stained with Trypan blue, and the cell viability was
   evaluated microscopically.

RT, amplification and library construction

   Single-cell suspensions (2 × 10^5cells/mL) in PBS (HyClone) were loaded
   onto a microwell chip using the Singleron Matrix® Single Cell
   Processing System. Barcoding beads are subsequently collected from the
   microwell chip, followed by reverse transcription of the mRNA captured
   by the barcoding beads to obtain cDNA and PCR amplification. The
   amplified cDNA was then fragmented and ligated with sequencing
   adapters. The scRNA-seq libraries were constructed according to the
   protocol of the GEXSCOPE® Single Cell RNA Library Kit
   (Singleron)^[102]25. Individual libraries were diluted to 4 nM, pooled,
   and sequenced on an Illumina NovaSeq 6000 with 150 bp paired-end reads.

Primary analysis of raw read data (scRNA-seq)

   The raw reads were processed to generate gene expression profiles using
   CeleScope v1.5.2 (Singleron Biotechnologies) with default parameters.
   Briefly, barcodes and UMIs were extracted from R1 reads and corrected.
   Adapter sequences and poly A tails were trimmed from R2 reads, and the
   trimmed R2 reads were aligned against the GRCh38 (hg38) transcriptome
   using STAR (v2.6.1b). Uniquely mapped reads were then assigned to exons
   with FeatureCounts (v2.0.1). Successfully assigned reads with the same
   cell barcode, UMI and gene were grouped together to generate the gene
   expression matrix for further analysis.

Single-cell RNA-seq data processing and quality control

   We used 6 samples for single-cell RNA-seq data. During the data
   analysis, the heterogeneity of the data derived from sample 1 in the
   intrauterine adhesion group was too high to be excluded. Scanpy
   v1.8.2^[103]26was used for quality control, dimensionality reduction
   and clustering with Python 3.7. For each sample dataset, we filtered
   the expression matrix according to the following criteria: 1) cells
   with a gene count less than 200 or with a top 2% gene count were
   excluded; 2) cells with a top 2% UMI count were excluded; 3) cells with
   a mitochondrial content > 20% were excluded; and 4) genes expressed in
   fewer than 5 cells were excluded. After filtering, 61,959 cells were
   retained for the downstream analyses, with an average of 1739 genes and
   4996 UMIs per cell. The raw count matrix was normalized by total counts
   per cell and logarithmically transformed into a normalized data matrix.
   The top 2000 variable genes were selected by setting the flavour
   ‘seurat’. Principal component analysis (PCA) was performed on the
   scaled variable gene matrix, and the top 15 principal components were
   used for clustering and dimensionality reduction. The batch effect
   between samples was removed by Harmony^[104]27 using the top 15
   principal components from PCA. The cells were separated into 23
   unsupervised clusters via the Louvain algorithm, and the resolution
   parameter was set to 1.0. The cell clusters were visualized via uniform
   manifold approximation and projection (UMAP).

Differentially expressed gene (DEG) analysis (scanpy) and pathway enrichment
analysis

   To identify differentially expressed genes (DEGs), we used the
   scanpy.tl.rank_genes_groups function based on the Wilcoxon rank sum
   test with default parameters and selected the genes expressed in more
   than 10% of the cells in both of the compared groups of cells and with
   an average log(fold change) value greater than 1 as DEGs. The logFC
   = 0.25 was selected based on preliminary analysis showing significant
   biological relevance at this threshold, while maintaining adequate
   statistical power given the sample size. The adjusted p value was
   calculated via Benjamini‒Hochberg correction, and a value of 0.05 was
   used as the criterion to evaluate the statistical significance. To
   investigate the potential functions of fibroblasts subcluster 3 and
   proliferation VECs, Gene Ontology (GO) and Kyoto Encyclopedia of Genes
   and Genomes (KEGG) analyses^[105]28were performed via the
   clusterProfiler R package v 3.16.1^[106]29[.]Pathways with p_adj values
   less than 0.05 were considered significantly enriched. Selected
   significant pathways were plotted as bar plots. For GSVA pathway
   enrichment analysis, the average gene expression of each cell type was
   used as input data^[107]30. Gene Ontology gene sets, including
   molecular function (MF), biological process (BP), and cellular
   component (CC) categories, were used.

Cell-type recognition with Cell-ID

   Cell-ID^[108]31 is a multivariate approach that extracts gene
   signatures for each individual cell and performs cell identity
   recognition using hypergeometric tests (HGTs). Dimensionality reduction
   was performed on the normalized gene expression matrix through multiple
   correspondence analysis, where both cells and genes were projected in
   the same low-dimensional space. Then, a gene ranking was calculated for
   each cell to obtain the most featured gene sets of that cell. HGT was
   performed on these gene sets against the Uterine_endometrium reference
   from the SynEcoSys™ (Singleron Biotechnology) database, which contains
   all cell-type genes. The identity of each cell was determined as the
   cell type with the minimal HGT p value. For cluster annotation, the
   frequency of each cell type was calculated in each cluster, and the
   cell type with the highest frequency was chosen as the cluster
   identity. The cell type identification of each cluster was determined
   according to the expression of canonical markers from the reference
   database SynEcoSys™ (Singleron Biotechnology). SynEcoSys™ contains
   collections of canonical cell type markers for single-cell sequencing
   data from CellMakerDB, PanglaoDB and recently published literature.

Subtyping of major cell types, filtering of cell doublets and cell cycle
analysis

   To obtain a high-resolution map of T cells, fibroblasts, MPs, mural
   cells, secretory cells, proliferating cells and ECs, cells from the
   specific cluster were extracted and reclustered for more detailed
   analysis following the same procedures described above and by setting
   the clustering resolution from 0.3– to 1.2. The number of cell doublets
   was estimated on the basis of the expression patterns of canonical cell
   markers. Any clusters enriched with multiple cell type-specific markers
   were excluded from downstream analysis. The cell cycle score of each
   cell was calculated using the sc.tl.score_genes_cell_cycle function
   implemented in scanpy v1.8.2^[109]26.

Cell-to-cell interaction analysis: CellPhoneDB

   The cell-to-cell interactions (CCIs) between cell types were predicted
   on the basis of known ligand‒receptor pairs via CellPhoneDB
   (v2.1.0)^[110]32. The permutation number for calculating the null
   distribution of average ligand‒receptor pair expression in randomized
   cell identities was set to 1000. Individual ligand or receptor
   expression was thresholded by a cut-off on the basis of the average log
   gene expression distribution for all genes across each cell type.
   Predicted interaction pairs with p values < 0.05 and average log
   expression > 0.1 were considered significant and visualized by
   heatmap_plot and dot_plot in CellPhoneDB.

Cell differentiation potential evaluation: CytoTRACE

   CytoTRACE v0.3.3^[111]33 (a computational method that predicts the
   differentiation state of cells from single-cell RNA-sequencing data
   using gene counts and expression) was used to predict the
   differentiation potential of cell subpopulations.

Pseudotime trajectory analysis: Monocle2

   The cell differentiation trajectory of monocyte subtypes was
   reconstructed via Monocle2 v 2.10.0^[112]34. To construct the
   trajectory, the top 2000 highly variable genes were selected by Seurat
   (v3.1.2). Variable features were identified, and dimension reduction
   was performed via the DDR tree. The trajectory was visualized via the
   plot_cell_trajectory function in Monocle2.

Transcription factor regulatory network analysis (pySCENIC)

   A transcription factor network was constructed by pySCENIC
   (v0.11.0)^[113]9 using a scRNA expression matrix and transcription
   factors in AnimalTFDB. First, GRNBoost2 predicted a regulatory network
   on the basis of the coexpression of regulators and targets. CisTarget
   was then applied to exclude indirect targets and to search for
   transcription factor binding motifs. After that, AUCell was used for
   regulon activity quantification for every cell. Cluster-specific TF
   regulons were identified according to regulon specificity scores
   (RSSs), and the activities of these TF regulons were visualized in
   heatmaps.

Supplementary Information

   [114]Supplementary Information 1.^ (13.5KB, docx)

Author contributions

   N.S., Y.J. and P.L. conceptualized and designed the experiments. N.S.,
   X.L. and P.L. performed the experiments and analyzed the data.
   P.L.,Z.T.and L.P. contributed some experiments, reagents, materials,
   and analytical tools. P.L. wrote the draft of the manuscript. Y.J. and
   N.S. revised and edited the manuscript. All the authors have read and
   agreed to the published version of the manuscript.

Funding

   National Natural Science Foundation of China,82101789,Changsha Natural
   Science Foundation,Kq2202369,Natural Science Foundation of Hunan
   Province,2022 JJ70079,Hunan Province Traditional Chinese medicine
   research project,D2022109

Data availability

   Sequence data that support the findings of this study have been
   deposited in the China National GeneBank (CNGB) with the primary
   accession code CNP0006637.

Declarations

Competing interests

   The authors declare no competing interests.

Ethics approval and consent to participate

   This study was conducted in accordance with the ethical standards of
   the institutional and national research committee and with the 1964
   Helsinki declaration and its later amendments or comparable ethical
   standards. It was approved by the Medical Ethics Committee of Xiangya
   Hospital, Central South University (2022111228). All the participants
   involved were informed of all experiment details and signed a written
   informed consent.For studies involving human tissue samples, consent
   was obtained from the donors for the use of their tissue in
   research.The details that might disclose the identity of the
   participants have been omitted.

Footnotes

   Publisher’s note

   Springer Nature remains neutral with regard to jurisdictional claims in
   published maps and institutional affiliations.

   Ping Li and Yu Jian These authors contributed equally to this work.

Supplementary Information

   The online version contains supplementary material available at
   10.1038/s41598-025-97433-1.

References