Two datasets from public databases were enrolled in this study. GSE108097 from Gene Expression Omnibus (https://www.ncbi.nlm.nih.gov/geo/, RRID: SCR_005012), containing scRNA-seq data of 3,756 cells from two 6-to-10-week-old female C57BL/6J mice uteri, was selected to build the digital mouse cell atlas of mouse uterus (Han et al., 2018). Mouse 1 contained 2041 cells, and mouse 2 contained 1715 cells. Based on the expression levels of proliferating cell nuclear antigen, differentially expressed genes of two samples, and gene ontology terms, we concluded that mouse 1 was in the maturational stage and mouse 2 was in the regeneration stage. Another scRNA-seq data of endometrial tissue was obtained from a previous study to compare the cultured endometrial stromal cells (human endometrial cells in the secretory stage were collected and cultured for 8 days in the cell culture medium consisting of a 1:1 mixture of DMEM (Cat No E15-892, GE Healthcare, United States)/Ham’s F12 (Cat No E15-890, GE Healthcare), supplemented with 10% FBS and Antibiotic-Antimycotic solution (GE Healthcare) and uncultured ones (endometrial biopsies collected from women in the secretory stage) (Krjutškov et al., 2016). We used R software (version 3.6.3; http://www.Rproject.org, RRID: SCR_001905) and Seurat package of R (version 2.3, https://satijalab.org/seurat/, RRID: SCR_016341) (Satija et al., 2015) to process the data. Seurat is a computational strategy to process and explore scRNA-seq data. The filter criteria of cells were determined as default. It is not necessary to obtain ethical approval because we used published online datasets.

Principal component analysis (PCA), uniform manifold approximation and projection (UMAP), t-distributed stochastic neighbor embedding (tSNE), and visualization were performed by the Seurat package of R (Satija et al., 2015). Seurat is a computational strategy to process and explore scRNA-seq data. The fate decisions and pseudotime trajectories of endometrium stromal cells and the mononuclear phagocyte system were reconstructed by the Monocle package of R (version 2.10.1, http://cole-trapnell-lab.github.io/monocle-release/docs/, RRID: SCR_016339) (Cole et al., 2014; Qiu et al., 2017; Wu et al., 2017). In short, Monocle performs differential expression and time-series analysis for single-cell expression experiments, which orders individual cells according to progress through a biological process, without knowing ahead of time which genes define progress through that process. Possible stem cells identification was performed by RaceID3 (https://github.com/dgrun/RaceID, RRID: SCR_017045) and StemID2 (https://github.com/dgrun/StemID, RRID: SCR_017242) (Grün et al., 2015). RaceID3 is an algorithm for rare cell type identification in complex populations of single cells, while StemID2 is an algorithm based on RaceID3 for the inference of differentiation trajectories and the prediction of the stem cell identity.

Gene Ontology (GO) analyses were conducted through clusterProfiler (clusterProfiler, RRID: SCR_016884) (Yu et al., 2012; Walter et al., 2015) and online GO resource website (Ashburner et al., 2000; 2019). Statistically significant GO terms (p < 0.05) were identified.

Histology and immunofluorescence were conducted according to our previous procedures (Wu et al., 2019). Briefly, vagina smears were firstly used to determine the mouse estrous cycle (Supplementary Figure S1) (Bertolin and Murphy, 2014). Then mouse uterine tissues were collected and fixed in 4% (w/v) paraformaldehyde, and then dehydrated in an ethanol gradient. Then the paraffin sections of 10 μm thickness were stained with hematoxylin and eosin. Immunostaining was carried out as follows: The series of 10 μm-thick sections were rehydrated, fixed with 4% (w/v) paraformaldehyde for 30 min, antigen retrieval was conducted by incubating in citrate antigen retrieval solution at 65°C overnight, rinsed three times with PBS, and treated with blocking solution (1% BSA) for 30 min, prior to incubation with primary antibodies at 4°C overnight. The primary antibodies rabbit anti-mouse antibodies against KRT7 (Abcam, ab181598), the primary antibodies mouse anti-mouse antibodies against PCNA (Abcam, ab29), the primary antibodies rat anti-mouse antibodies against CD34 (Biolegend, 119307) were used to detect the expression of selected proteins within the uterine tissues. The goat anti-rat-cy3 secondary antibody (Beyotime Biotechnology, A0507), goat anti-rabbit-488 secondary antibody (Invitrogen, A11008), donkey anti-mouse 546 secondary antibody (Invitrogen, A10036), and DAPI (Beyotime Biotechnology, C1002) were used to visualize the respective primary antibodies and the cell nuclei. All procedures were carried out according to the manufacturer’s instructions.

All statistical analyses were performed by Graphpad Prism 8.0 software (https://www.graphpad.com/, RRID: SCR_002798) and R software (version 3.6.3; http://www.Rproject.org, RRID: SCR_001905). A two-sided probability value of p < 0.05 was considered being statistically significant.

The scRNA-seq data of two 6-to-10-week-old female C57BL/6J mice were processed using published Seurat pipelines (Satija et al., 2015; Han et al., 2018). In total, we analyzed 3,756 single cells and identified 16 cell clusters, which were grouped into eight major cell types (Figures 1A,B, Supplementary Table S1). Then we clarified the identities of each cluster (Figure 1C, Supplementary Table S2). Cluster 0, cluster 1, cluster 2, cluster 3, cluster 10 all highly expressed Col3a1 and Fn1, so we clarified them into stromal cells. Cluster 4, in the meantime, specifically expressed Acta2, which made us clarify them into myofibroblasts. Cluster 11, however, expressed Acta2 and Actg2 without Col3a1, so cells in cluster 11 belonged to muscle cells. Cluster 5, cluster 9, and cluster 12 all highly expressed Cd68 and Adgre1, so they were macrophages/monocytes. Cd7 and Nkg7 could be found highly expressed in cluster 8, which made cluster 8 natural killer (NK) cells. Ly6d and Cd79a were highly expressed in cluster 15, so we identified cluster 15 as B cells. Cluster 13 and cluster 7 both had high expression levels of Krt8, Krt18, and Epcam, making them epithelial cells. Cluster 6 and cluster 14 had high expression levels of Cldn5 and Pecam, making them endothelial cells. In this way, we clarified the identities of 16 cell clusters, building a solid foundation for further analysis.

We identified 5 sub-clusters of stromal cells (cluster 0, cluster 1, cluster 2, cluster 3, cluster 10) and 2 sub-clusters of epithelial cells (cluster 13 and cluster 7). According to their distances in the t-SNE, we divided the 5 sub-clusters of stromal cells into three groups: stromal cells 1, stromal cells 2, and stromal cells 3. The proportions of each cell type in two mice were shown (Figure 1D). Surprisingly, we found that stromal cells 1 were exclusively in mouse 2, and stromal cells 3 were exclusively in mouse 1. Stromal cells 2 could be found both in mouse 1 and mouse 2. This phenomenon was also seen in epithelial cells. Epithelial cells 1 were exclusively in mouse 2, while epithelial cells 2 were exclusively in mouse 1.

We believed the different cell contents of two mice had further biological significance. We found that proliferating cell nuclear antigen (Pcna) was highly expressed in epithelial cells 1 (mouse 2) (Figure 2A). The protein encoded by this gene is a cofactor of DNA polymerase delta, which is regarded as the marker of proliferation (Celis and Celis, 1985). Previous studies have reported that for humans experiencing estrous cycles, the divergence of the expression level of Pcna was mainly due to the states of the estrous cycle rather than age. It has been reported that the expression of Pcna showed a high peak in the proliferative phase, and then decreased sharply in the secretory phase (Li et al., 1993; Noci et al., 1995; Hamid et al., 2002). In mice, the treatment with estrogen to ovariectomized mice upregulates PCNA expression, while the co-treatment with estrogen and progesterone downregulates PCNA expression (Annie et al., 2019). In addition, we found that the expression levels of metalloproteinase (MMPs) were significantly suppressed in epithelial cells 1 (mouse 2) (Supplementary Table S2). So we preliminarily speculated that the uterus of mouse 2 was in the regenerative stage, and the uterus of mouse1 was in the maturational stage.

To further investigate this speculation, functional enriched terms of the differentially expressed genes in each epithelial cell cluster were shown (Figures 2B,C, Supplementary Table S3). Terms including “response to steroid hormone,” “skin development,” “cellular response to epidermal growth factor stimulus,” “placenta development,” and “response to wounding” were enriched in epithelial cells 2 (mouse 1). These terms showed the active rebuilt of the uterus epithelium to prepare for subsequent fertilization in the maturational status. In this period, we also found terms that indicated “intrinsic apoptotic signaling pathway,” “increased cell mobility,” and “suppressed cell adhesion,” which were all consistent with the changes that happened in the maturational status as previously reported (Yip et al., 2013).

The epithelial cells from two uteri were then clustered exclusively to further verify our hypothesis. We gained 3 clusters (Figure 2D, Supplementary Table S4). According to their Pcna expression levels, cluster 1 and cluster 2 showed proliferating characteristics (Figure 2E). We concluded that mouse 2 was in the regenerative stage, for it was entirely composed of proliferating epithelial cells. As for mouse 1, only part of the epithelial cells showed proliferating characteristics (Figure 2F). In addition, the expression level of Esr1 was suppressed in the epithelial cells of cluster 0 in the mouse 1 (Supplementary Table S4).

Endometrial stromal cells (ESCs) perform a multitude of functions and undertake different functions in different estrous stages (Cottrell et al., 2017). We further explored the characteristics of stromal cells of two mice to further confirm the stages of two mice. The expression levels of MMPs were significantly upregulated in epithelial cells 2 (mouse 1) (Supplementary Table S2), which possibly linked with the active matrix degradation in the maturational stage.

We performed functional enrichment analyses of the differentially expressed genes in each stromal cell cluster to see their functional traits. The term “artery morphogenesis” was enriched in stromal cells in mouse 2 (Figure 2G, Supplementary Table S3). One of the key features of the regeneration stage is angiogenesis. Besides, “protein maturation” and “collagen catabolic process” were activated in this period.

The stromal cells in mouse 1 presented more dynamic cell communications (Figure 2H, Supplementary Table S3). The term “response to progesterone and other steroid hormones” further indicated that mouse 1 was in the maturational stage. The term “response to transforming growth factor beta” was enriched too. Also in the maturational stage, the accelerated cell movement was again seen in this stage, as enrichment analysis suggested the term “positive regulation of cell migration” was significantly enriched. We believed the elevated levels of cell communications and material transportation found in mouse 1 constituted a prosperous metabolism network in stromal cells.

In order to validate the data analysis results, vagina smear and hematoxylin and eosin (H&E) staining were used to determine the mouse estrous cycle (Figure 2I, Supplementary Figure S1) (Bertolin and Murphy, 2014). The expression levels of PCNA in the regenerative mice regenerative endometrium and maturational endometrium were presented (Figure 2I). We found that, like humans, expression level of PCNA in the regenerative endometrium was significantly higher than in the maturational endometrium, indicating the practicability of using PCNA expression level to define the estrus stage.

Taking into account the sheer differences in proliferation and apoptosis, cell adhesion and movement, angiogenesis, and extracellular matrix remodeling, we concluded that mouse 1 was in the maturational stage and mouse 2 was in the regeneration stage.

ESCs have been reported to perform a multitude of functions including hormonal regulation, decidualization, maternal-fetal communications, and embryo receptivity (Cottrell et al., 2017). After we clarified the estrous stages of two mice, we further explored the characteristics of stromal cells in each estrus period given their significances in the uterus.

Firstly, functional enrichment analysis of upregulated genes of myofibroblasts was performed (Figure 3A, Supplementary Table S3). Terms like “muscle structure development” and “regulation of smooth muscle cell proliferation” confirmed our previous classification of its identity.

The analyses of two stromal subsets in the last section showed that stromal cells in different physiological stages had huge differences in their functions. Therefore, we wanted to further explore how these two cell groups gained their unique traits by differentially gene expression. The pseudotime trajectory of stromal cells in two physiological stages was constructed (Figure 3B). The results showed that there were two differentiation branches: the regenerative stromal cells would transform into myofibroblasts or maturational stromal cells at the first decisional point, and a small part of the maturational cells still had the ability to differentiate into myofibroblasts at the second decisional point. In this way, an estrus cycle was completed. All the significantly expressed branch 1-dependent genes were clustered into three categories by unsupervised clustering (Figure 3C). Among these genes, the myofibroblasts-branch genes had high expression levels of Acta2 and Adamts1 (Figure 3D). In the meantime, the maturational_stromal-branch genes had high expression levels of Col6a4, Fbln2, Tcf4, and Wnt5a (Figure 3E). These divergent expression patterns of representative genes further confirmed our previous classification. We listed other significant branch-dependent genes, which may also play key roles in the stromal cell differentiation and need further validation (Table 1).

Embryo implantation and tumor progression are similar to some extent (Holtan et al., 2009). The maternal immune system needs to find the balance and provides an appropriate environment for the fetus to grow. The current findings of NK, monocytes, and dendritic cells (DCs) during the menstrual cycle are limited. For example, findings concerning NK cell number and cytotoxic activity are conflicting (Oertelt-Prigione, 2012). In order to gain reliable results with regard to uterine immune cells at the single-cell level, all immune cells were clustered exclusively into 4 clusters (Figure 4A, Supplementary Table S5). The heatmap of the top 50 markers for each cluster showed that we gained a convincing clustering result (Figure 4B).

Using known markers, the identities of each cluster were clarified (Figure 4C). Cluster 0 expressed Adgre1 and Mrc1, so we clarified them into macrophages. Cluster 1 expressed Nkg7 and Cd7, which made us clarify them into NK cells. Cluster 2 expressed Cd83 and Cd209a, so cells in cluster 2 belonged to DC cells. For their high expression levels of Ly6c2, we clarified cluster 3 into monocytes. We found a clear difference in immune cell composition in these two estrus stages. The proportions of macrophages and NK cells in the two stages varied considerably (Figure 4D). Macrophages dominated in the regenerative stage, while NK cells dominated in the maturational stage.

The mechanism of monocytes differentiating into DCs or macrophages is poorly understood. The developmental trajectory of the mononuclear phagocyte system was constructed (Figure 4E). The branch point led to two differentiation paths: macrophages or DCs. Branch-dependent genes were clustered into three categories according to their expression patterns (Figure 4F). The expression levels of marker genes of each branch over pseudotime verified the identities of two differentiation paths (Figure 4G). We listed other significant branch-dependent genes (Table 2). Among them, we detected three transcription factors: Mafb, Irf7, and Nr4a1. Mafb was detected in the macrophage branch, Irf7 expression was seen in the monocyte branch, and Nr4a1 expressed highly in the DC branch (Figure 4G).

As for NK cells (Table 3), NK cells in the regenerative stage showed intense inflammatory responses with high expression levels of Tnf, Il1b, and S100a8. It also expressed Cxcl2 and Ccrl2, which respectively promoted the recruitment of neutrophils and themselves (Regan-Komito et al., 2017). In contrast, upregulated genes of NK cells in the maturational stage mostly were immune-suppressive genes like Serpinb9, Stat3, Cd96, and Cd55. They also secreted substances that were advantageous for follow-up pregnancy like Ccl2.

The number of macrophages was higher in the regenerative stage. In this stage, macrophages were pro-inflammatory by secreting resistin, which was a systemic pro-inflammatory cytokine targeting both leukocytes and adipocytes (Table 3) (Nagaev et al., 2006). However, it also showed high expression levels of selenoprotein Msrb1 and Hes1 to inhibit inflammatory responses (Lee et al., 2017; Zhang et al., 2019). As for macrophages in the maturational stage, it mainly showed the M2 phenotype with high expression levels of Il10, Cd206, and Ccl7, which were consisted of the characteristics of decidual macrophages (Xuan et al., 2015).

In the regenerative stage, the functions of DC were mainly embodied in influencing other cells. It secreted Ccl22 to recruit regulatory T (Treg) cells and Ccl5 to recruit CCR5-positive cytokine-induced killer cells (Lee et al., 2016; Yashiro et al., 2019). Its pro-inflammatory role was also embodied in the high expression of Ccrl2, which played a significant role in inflammation (Regan-Komito et al., 2017). DCs in the maturational stage were immune-suppressive (Table 3). DCs expressed PD-L1, Jchain, and Anxa1, which were correlated with immune tolerance and refrained the secretion of pro-inflammatory cytokines (Källberg and Leanderson, 2008; Sena et al., 2016).

Monocytes in the maturational stage secreted multiple chemokines like Thbs1, Ccl2, Ccl7, Ccl8, Lyn, and Anxa1 to facilitate the migration of themselves (Table 3) (Asano et al., 2015; Liu et al., 2015). Monocytes also expressed M1 markers like Cd86 and Lgals8 to enhance inflammatory responses (Patel et al., 2017). By contrast, monocytes in the regenerative stage showed more intense cytolytic ability by producing S100a8, Tnf, and Il1b. It also secreted Cxcl16 and Ccl3 to facilitate the recruitment of monocytes and Mmp14, Trem2, and Mif to induce inflammatory responses (Buck et al., 2013; Ruiz-Rosado et al., 2016).

Taken together, we found that the immune responses were suppressed in the maturational stage compared to the regenerative stage by reducing inflammatory cytokines secretion and facilitating differentiation into immune-suppressive cells.

The mesenchymal-to-epithelial transition has been proposed as a possible reason to explain the periodically endometrial epithelial tissue regeneration (Huang et al., 2012). A previous study has reported a group of cells expressed both the epithelial cell marker, pan-cytokeratin, and the stromal cell marker, vimentin as well (Patterson et al., 2013). We tried to identify multipotent ESCs in the scRNA-seq data using published pipelines (Grün et al., 2016). We failed to find the cell groups that expressed all the proposed stem cell surface molecules like Cd73, Cd90, and Cd105. We hypothesized that expressions of Cd73, Cd90, and Cd105 may be induced during in-vitro culture. Another scRNA-seq dataset containing cultured ESCs and uncultured ones was used to show the diversity of two culture environments (Krjutškov et al., 2016). The cells from two different culture environments clustered separately (Figure 5A). The stem cell surface markers mentioned above were expressed differently in the two cell groups. We found that stem cell makers like Cd90 (Thy1) and Cd44 had significant higher expressions in the cultured group (Figure 5B). We found that cell group had a higher entropy score and more connections with other cell types (Figures 5C,D,F, Supplementary Table S6), had the potential to differentiate into stromal cells, epithelial cells, and immune cells (Figure 5E), which make it the potential multipotent ESCs in vivo. This cell subcluster expressed higher expression of Cd34, Pdgfrb, and Aldh1a2 (Figures 5F–I). The representative immunofluorescence staining of CD34 in mice endometrium was shown, indicating the existence of this endometrial mesenchymal stem cells (Figure 5J).

To conclude, we found the transcriptional signature of endometrial mesenchymal stem cells in vivo. And the cell group 5 that we presented here showed a great possibility to be the cells responsible for re-epithelialization.