Abstract Background The efficiency of mesenchymal stem cells (MSCs) in treating myocardial infarction (MI) remains inconsistent, which limits their therapeutic applications. Therefore, exploring the mechanism for the inconsistent efficacy of MSCs and identification the criteria for screening MSCs are important for improving the efficiency of MSCs. Methods Mouse model after MI was utilized to test the role of MSCs from different donors and the functional subpopulation in improving cardiac function. Heterogeneity of MSCs was identified using single-cell RNA sequencing (scRNA-seq) of MSC-GY. GSEA and Scissor analyses were used to find the functional subpopulations of MSCs that promote angiogenesis. The role of functional subpopulations in promoting angiogenesis was verified by detecting the secretory proteins, the ratio of N-CADHERIN^+/CD168^− subpopulations in MSCs, and the tube formation, migration, and proliferation of HUVECs after treatment with conditional medium (CM) derived from different MSCs. Results We found that umbilical cord-derived MSCs (UC-MSCs) from different donors have varied therapeutic efficacy in MI mice and UC-MSCs with higher therapeutic effectiveness exhibited the most potent pro-angiogenic effects by secreting elevated levels of angiogenesis-related proteins, such as MYDGF, VEGFA, and FGF2. ScRNA-seq of 10,463 UC-MSCs revealed that the N-CADHERIN^+/CD168^− subpopulation was closely associated with pro-angiogenic effects, and the ratio of this cell subpopulation was positively correlated with the angiogenic potential of MSCs. We also found that the N-CADHERIN^+/CD168^− subpopulation was the functional subpopulation of MSCs in improving cardiac function of MI mice. Conclusions Our study identified that the N-CADHERIN^+/CD168^− subpopulation was the functional subpopulation of MSCs in treating MI, which was essential for the development and utilization of MSCs in MI treatment. Supplementary Information The online version contains supplementary material available at 10.1186/s13287-024-04032-4. Keywords: MSCs, Myocardial infarction, ScRNA-seq, N-CADHERIN^+/CD168^−, Angiogenesis Introduction Myocardial infarction (MI) involves extensive death of cardiomyocytes due to coronary artery occlusion, leading to impaired cardiac function and progressive heart failure [[40]1]. The transplantation of mesenchymal stem cells (MSCs) has emerged as a promising therapeutic strategy for treating MI, attributed to their low immunogenicity and paracrine effects, as demonstrated in animal studies and clinical trials [[41]2, [42]3]. Lee and Amado demonstrated that MSC transplantation significantly reduced myocardial injury and improved cardiac function in the infarcted hearts of mice and pigs, respectively [[43]4, [44]5]. A meta-analysis by Lalu et al. of 23 clinical trials (involving 1,148 patients) confirmed that MSC application was safe and effective for treating MI [[45]6]. Additionally, Chullikana et al. found that MSC transplantation improved ejection fraction and reduced adverse cardiac events in MI patients [[46]7]. However, Nowbar et al. reported the limited effects of MSCs on improving left ventricular ejection fraction (LVEF) based on an analysis of clinical trial data from 1252 patients [[47]8]. Thus, achieving consistent and effective MSC treatment for MI remains a critical challenge. Single-cell RNA sequencing (scRNA-seq) enables characterization of gene expression at the single-cell level and resolves cellular heterogeneity of MSCs. MSCs have been reported to be a heterogeneous population, with different cellular subpopulations have varying biological properties [[48]9–[49]12]. Sacchetti and Zhou et al. found that CD146^+ and LepR^+ MSCs, highly expressed bone-related genes, exhibited significant osteogenic differentiation ability [[50]13, [51]14]. Similarly, Arufe and Mifune et al. identified the CD271^+ MSCs possessed high expression of cartilage-related genes and significant chondrogenic differentiation ability [[52]15, [53]16]. Additionally, the CMKLR1^+ MSCs have been reported to possess strong immunomodulatory and osteogenic capabilities [[54]17]. Thus, understanding the heterogeneity of MSCs would be crucial for identifying MSC subpopulations suitable for treating specific diseases. However, it remains unclear whether specific MSC subpopulation could improve the treatment efficacy for MI. The paracrine effects have been identified as a major mechanism of MSCs for the treatment of MI [[55]2, [56]18, [57]19]. MSCs can secrete anti-inflammatory factors such as tumor necrosis factor stimulated gene 6 (TSG6), transforming growth factor beta 1 (TGF-β1), and interleukin 10 (IL10), which regulate macrophage polarization and inhibit T cell activation [[58]20]. Lee et al. also found that MSCs improved cardiac function by regulating the inflammatory response, which was blocked by TSG6 knockdown [[59]4]. Furthermore, the angiogenesis-related factors such as vascular endothelial growth factor (VEGF), fibroblast growth factor 2 (FGF2), and hepatocyte growth factor (HGF) were secreted by MSCs to promote angiogenesis and cell survival [[60]21, [61]22]. Lee et al. has also demonstrated that the pro-angiogenic effect is the main mechanism by which MSCs improve cardiac function and enhance therapeutic efficacy for treating MI [[62]23]. Despite these insights, the key mechanism influencing MSC efficacy for MI treatment remains unclear. In this study, we demonstrated that variations in the angiogenic potential of MSCs affected their therapeutic efficacy in MI mice. Furthermore, we identified the N-CADHERIN^+/CD168^− subpopulation as a functional MSC subpopulation in regulating angiogenesis and improving the heart function of MI mice, with the ratio of this subpopulation significantly correlating with the therapeutic effects of MSCs. Thus, our study highlights a specific MSC subpopulation that improving the heart function of MI mice, which is essential for the development and utilization of MSCs in MI treatment. Methods Experimental animals The C57BL/6 male mice (6–8 weeks old) used in this study were purchased from Shanghai Laboratory Animal Center (SLAC), Shanghai, China. Mice were housed in the Laboratory Animal Research Center, Tongji University, and maintained under controlled temperature (22 °C ± 1 °C) and humidity conditions with a 12:12 h light:darkness cycle. Mice were euthanized by cervical dislocation and the death was confirmed by cardiorespiratory arrest. All animal experiments were conducted following the institutional animal ethics guidelines and were approved by the Animal Use Committee of Tongji University (TJAB04523102). Cell culture UC-MSCs were cultured in MSC medium (ScienCell #7501) supplemented with 5% FBS, 1 × mesenchymal stem cell growth supplement (MSCGS), and 1 × penicillin/streptomycin (PS) solution. Construction of MI model and UC-MSC transplantation Animals were randomly assigned to different cage locations to prevent positional effects. To establish a mouse model of MI, adult male C57BL/6 mice were anesthetized with 2% isoflurane inhalation, followed by ligation of the left main descending coronary artery (LCA), as previously described [[63]24]. Immediately after MI, 2 × 10^5 UC-MSCs suspended in 30 μL saline were injected intramyocardially into the infarct border zone at three distinct sites. Control MI animals received saline injection only, while sham-operated animals underwent all surgical procedures except for LCA ligation and intramyocardial injection. In the experiment evaluating the therapeutic effects of MSCs from different donors on MI, 38 adult male C57BL/6 mice were randomly assigned to the following groups: the Sham group (n = 7), the MI + Saline group (n = 7, with 2 mice dying during surgery, 7/9), the MI + MSC-CX group (n = 7, injected with 2 × 10^5 MSC-CX, with 1 mouse dying during surgery, 7/8), the MI + MSC-CJ group (n = 7, injected with 2 × 10^5 MSC-CJ), and the MI + MSC-GY group (n = 7, injected with 2 × 10^5 MSC-GY). A total of 35 mice survived, resulting in a survival rate of 92.1%. After evaluation on day 3, all 28 surviving infarcted mice had an LVEF below 40%. In the experiment evaluating the therapeutic impact of N-CADHERIN^+/CD168^− subpopulation within MSCs on MI, 30 adult male C57BL/6 mice were randomly assigned to the following groups: the MI + Saline group (n = 6, with 2 mice dying during surgery, 6/8), the MI + N-CADHERIN^+/CD168^− MSC-GY group (n = 6, injected with sorted N-CADHERIN^+/CD168^− cells from MSC-GY, with 2 mice dying during surgery, 6/8), the MI + others group (n = 6, injected with remaining MSC-GY cells excluding the N-CADHERIN^+/CD168^− subpopulation, with 1 mice dying during surgery, 6/7), and the MI + N-CADHERIN^+/CD168^− MSC-CX group (n = 6, injected with sorted N-CADHERIN^+/CD168^− cells from MSC-CX, with 1 mice dying during surgery, 6/7). A total of 24 mice survived, resulting in a survival rate of 80%. After evaluation on day 3, all 24 surviving infarcted mice had a LVEF below 40%. In both experiments, all surgeries and injections were performed by a single experienced operator in a blinded manner. Additionally, The LVEF measurements and data analyses were conducted with the investigator blinded to the group assignments, ensuring unbiased assessment of cardiac function across all groups. Assessment of cardiac function Echocardiography was performed to assess the cardiac function of MI mice 28 days after cell transplantation (Visual Sonics Vevo 2100 system equipped with a 40-MHz 550 s probe). LVEF, fraction shortening, end-diastolic diameter, end-systolic diameter were measured and calculated using cardiac echocardiography software [[64]25]. Masson trichrome staining Masson trichrome staining was performed as previously described [[65]24, [66]26]. Briefly, the mouse hearts (4 weeks after MI) were fixed in 4% paraformaldehyde for 24 h at 4 °C and then dehydrated using a sucrose gradient. Hearts were sectioned from below the ligation site to the apex at 8 μm thickness, with samples taken at 500 μm intervals for staining (Yeasen). Infarct size was quantified as the average of five sections using the formula: infarct size = [infarct perimeter (infarct epicardium + infarct endocardium) × 100]/left ventricle perimeter (left ventricle epicardium + left ventricle endocardium). Immunostaining Immunostaining was performed as previously described [[67]27]. Fixed slides were blocked for 1 h at room temperature in PBS containing 10% donkey serum and 0.1% Triton X-100. Slides were incubated overnight at 4 °C with primary antibodies, followed by incubation with fluorescent secondary antibodies and Hoechst33342 for nuclear staining at room temperature for 1 h. CD31 and α-SMA positive cells were quantified from three randomly selected sections per slide. Flow cytometry (FACS) UC-MSCs were dissociated with TrypLE (Gibco), washed with PBS, and 3 × 10^5 cells were incubated with 100 μL PBS containing 1 μL CD168 (RHAMM)-FITC antibody (bs-4736R-FITC, Bioss) and 10 μL CDH2 (N-CADHERIN)-PE antibody (11039-R001-PE, Sino Biological) for 30 min at 4 °C. Cells were analyzed using a BD FACSVerse, and data were processed with FlowJo software. For sorting of the N-CADHERIN^+/CD168^− MSC subpopulation. MSCs were stained with CD168 (RHAMM)-FITC antibody (4 μL/10^6 cells) and CDH2 (N-CADHERIN)-PE antibody (30 μL/10^6 cells) for 30 min at 4 °C. After washing with PBS, the N-CADHERIN^+/CD168^− subpopulation was sorted by BD FACSAria II. Tracing of MSCs MSCs were transfected with lipophilic tracker-DiR (DiIC18(7)) (APEx BIO, B8806; 10 μM in PBS) labeling for 20 min before implantation into the hearts. At day 1 and 7 after transplantation, mice were anesthetized with isoflurane and photographed with an IVIS Lumina XRMS Series III instrument. The region of interest (ROI) of DiR-labeled MSCs was measured using living image4.4 software (Radiant Efficiency). Production of UC-MSCs conditioned medium UC-MSCs (2 × 10^6 cells) from different donors were cultured in 10 cm dishes until 80–90% confluent. Cells were washed with PBS and switched to low glucose DMEM. After 5 days, supernatants were collected and stored at 4 °C for further use. Tube formation assay Tube formation assay was performed as previously described [[68]22]. Basement membrane matrix (Matrigel®, BD Biosciences) was solidified in 24-well plates at 37 °C for 30 min. HUVECs (8 × 10^4 cells/well) were plated on Matrigel in 1640 medium with 10% FBS and 30% conditioned medium from UC-MSCs. Tube formation was imaged after 8 h, and structures were quantified using Image J. Cell migration assay HUVECs (4 × 10^4 cells/well) were cultured in 24-well plates for 24 h. After creating a scratch, cells were cultured in RPMI 1640 medium with 30% UC-MSC conditioned medium for 16 h. Images of migrated cells were taken at 0, 8, and 16 h, and the migration rate was calculated using Image J. Cell proliferation assay HUVECs (8 × 10^4 cells/well) were cultured in 12-well plates in RPMI 1640 medium with 30% UC-MSC conditioned medium for 48 h. Cell numbers were counted at 0, 24, and 48 h using CounterStar software. scRNA-seq analysis Single-cell RNA sequencing was conducted using 10 × Genomics technology. Data were processed into a Seurat object, filtering out cells with fewer than 200 features or more than 5% mitochondrial gene expression. Data normalization and feature selection identified 2,000 highly variable features. Principal Component Analysis (PCA) reduced dimensionality to 30 components, which were used for clustering at a resolution of 0.25. UMAP and t-SNE techniques visualized cell populations. Statistical analysis All data were presented as mean ± S.E.M. of three independent experiments. Statistical significance was determined using one-way ANOVA with Tukey’s post hoc test for more than two groups (GraphPad Prism software). The significance levels were set at *, ^#P < 0.05, **, ^##P < 0.01, and ***, ^###P < 0.001. The work had been reported in line with the ARRIVE guidelines 2.0. Results MSCs from different donors exhibit distinct effects on improving cardiac function in mice post-MI We obtained UC-MSCs from three donors (MSC-GY, MSC-CJ, and MSC-CX), with flow cytometry results showed that these UC-MSCs were positive for CD44, CD90, CD105, and CD73, while negative for CD34 and HLA-DR (Fig. [69]1A). We then transplanted these three UC-MSCs into MI mice hearts via intramyocardial injection. Four weeks post-transplantation, echocardiography was performed to evaluate the cardiac function. Our results demonstrated that MSC-GY and MSC-CJ significantly improved LVEF and left ventricular fraction shortening (LVFS). Correspondingly, we observed a decrease in left ventricular internal diameter in diastole (LVIDd) and systole (LVIDs) in the mice transplanted with MSC-GY and MSC-CJ, whereas MSC-CX transplantation had no effect on cardiac function post-MI (Figs. [70]1B, 1C and S1). Additionally, the Masson's trichrome staining revealed that MSC-GY and MSC-CJ transplantation reduced the injury area in the MI mice (Fig. [71]1D). Quantification of DiR-labeled MSCs revealed that MSC-GY and MSC-CJ exhibited a higher retention ratio compared to MSC-CX (Fig. S2). These results indicated that UC-MSCs derived from different donor exhibit variations in therapeutic efficacy in mice post-MI. Fig. 1. [72]Fig. 1 [73]Open in a new tab MSCs from Different Donors Exhibit Distinct Effects on Improving Cardiac Function in Mice Post-MI. (A) Flow cytometric analysis of UC-MSCs from different donors, showing MSC-positive markers (CD44, CD90, CD105, and CD73) and MSC-negative markers (CD34 and HLA-DR). (B) Echocardiography analysis of ejection fraction (EF) and fraction shortening (FS) in Sham, MI, MI + MSC-CX, MI + MSC-CJ, and MI + MSC-GY at 4 weeks post-MI (n = 7). **P < 0.01 versus Ctrl; ##P < 0.01 versus MSC-CX. (C) Echocardiographic analysis of left ventricular end-diastolic diameter (LVIDd) and left ventricular end-systolic diameter (LVIDs) in Sham, MI, MI + MSC-CX, MI + MSC-CJ, and MI + MSC-GY at 4 weeks post-MI (n = 7). *P < 0.05, **P < 0.01 versus Ctrl. (D) Representative images of Masson trichrome staining of hearts sections (left) and quantification of infarct size (right) at 4 weeks post-MI. *P < 0.01 versus Ctrl; # P < 0.05 versus MSC-CX. Statistical significance was determined using one-way ANOVA followed by Tukey’s post hoc. Scale bar: 500 μm Variable angiogenic efficacies of MSCs from different donors MSCs could promote angiogenesis and reduce inflammation. Vagnozzi et al. found that intramyocardial injection of both live and cryopreserved MSCs resulted in the accumulation of CCR2^+ and CX3CR1^+ macrophage, which alleviated inflammation in mice post-MI [[74]28]. Based on this, we hypothesized that angiogenesis might be the primary mechanism affecting MSC efficacy in treating MI. We examined MSC-induced angiogenesis via CD31 and α-smooth muscle actin (α-SMA) immunostaining and observed that MSC-GY and MSC-CJ significantly increased α-SMA and CD31-positive cells in peri-infarct heart regions, while MSC-CX had minimal impact (Fig. [75]2A and B). To further validate the angiogenic effects of different MSCs, we treated human umbilical vein endothelial cells (HUVECs) with conditional medium (CM) derived from different MSCs. Results showed that the MSC-GY and MSC-CJ CM significantly enhanced the tube formation, migration, and proliferation of HUVECs, rather than the MSC-CX CM (Fig. [76]2C–E), consistent with the results in the MI model. We then performed the secretome analysis of three UC-MSCs CM. Gene Ontology (GO) and Gene Set Enrichment Analysis (GSEA) results indicated that upregulated secretory factors in MSC-GY and MSC-CJ CM were associated with endothelial cell proliferation, vascular development, and endothelial cell migration (Fig. [77]2F, Fig. S3 and S4). Angiogenic factors such as VEGF and FGF2 were significantly higher in the CM of MSC-GY and MSC-CJ compared to MSC-CX (Fig. [78]2G). These results suggested that the angiogenic capacity of MSCs influences their efficacy in MI treatment. Fig. 2. [79]Fig. 2 [80]Open in a new tab Variable Angiogenic Efficacies of MSCs from Different Donors. (A and B) Representative immunostaining images for CD31 and α-SMA in the border zone of infarcted hearts at 4 weeks post-MI (A) and statistical analysis (B). ***P < 0.001 versus Ctrl; # < 0.05, ## P < 0.01 versus MSC-CX. (C) Representative images illustrating the tube formation of HUVECs on Matrigel and statistical analysis. ***P < 0.001 versus Ctrl; ### P < 0.001 versus MSC-CX. (D) Representative images depicting HUVEC migration and statistical analysis. **P < 0.01, ***P < 0.001 versus Ctrl; ## P < 0.0,1 ### P < 0.001 versus MSC-CX; ns versus MSC-CJ. (E) Statistical analysis of HUVEC proliferation treated with conditional medium (CM) from UC-MSCs. **P < 0.01 versus Ctrl; ns versus MSC-CJ. (F) GO term analysis highlighting the secreted proteins highly expressed in the CM of MSC-GY and MSC-CJ. (G) Heatmap illustrating the expression levels of angiogenesis-related factors in the CM of MSC-GY, MSC-CJ, and MSC-CX. Statistical significance assessed using one-way ANOVA followed by Tukey’s post hoc. Scale bar: 100 μm Single-cell transcriptomic analysis reveals four major subpopulations within MSCs To explore the heterogeneity of UC-MSCs, we generated single-cell transcriptome profiles of MSC-GY, which exhibited the most significant therapeutic effects in MI treatment. After quality control, single-cell transcriptomes of 10,463 cells were acquired for further analysis. By analyzing variably expressed genes across all cells, we identified four clusters of MSCs: Cluster 1 (35.10%), Cluster 2 (27.85%), Cluster 3 (23.06%), and Cluster 4 (13.99%) (Fig. [81]3A). We then performed differential gene expression analysis to find unique signatures for each cluster, presenting the top 10 significantly differentially expressed (SDE) genes of each cluster (Fig. [82]3B). Fig. 3. [83]Fig. 3 [84]Open in a new tab Single-cell Transcriptomic Analysis Reveals Four Major Subpopulations within MSCs. (A) tSNE plot illustrating the clustering of 10,463 cells from MSC-GY (above), along with the statistical analysis indicating the distribution of cells in each subpopulation (below). (B) Heatmap displaying the top 10 marker genes for each subpopulation. (C) GSEA-KEGG showing the enriched pathway for each subpopulation. (D) tSNE plot demonstrating the distribution of each subpopulation across the cell cycle. (E) Pseudotime analysis reflecting the cell state transition and differentiation trajectories of these four subpopulations To reveal the biological functions of each cluster, we performed the GSEA of cluster-specific differentially expressed genes (DEGs). The results showed that Cluster 1 was significantly associated with vascular development, smooth muscle cell migration and adhesion, indicating that Cluster 1 might regulate angiogenesis. Cluster 2 showed enrichment in metabolic processes such as lipid metabolism and fatty acid metabolism, while Cluster 3 was associated with cell morphogenesis and cell junction assembly, and Cluster 4 was linked to cell proliferation-related biological processes such as cell division and microtubule polymerization or depolymerization. Based on their gene expression profiles and pathway enrichment analysis, we named Cluster 1 as the angiogenesis subpopulation, Cluster 2 as the metabolic regulation subpopulation, Cluster 3 as the niche support subpopulation, and Cluster 4 as the proliferation subpopulation (Fig. [85]3C). Furthermore, we predicted the cell cycle phases of each cell and found that cells in the proliferation subpopulation were predominantly in the G2/M phase, confirming the strong proliferative capacity of this subpopulation (Fig. [86]3D). The proposed trajectory analysis can infer the evolution of cell subpopulation by their expression characteristics. We performed pseudo-time reconstruction of MSC-GY to stratify the connection of each sub-population. Pseudo-time reconstruction revealed seven cell states, starting from state 7 and ending at state 1. The angiogenesis subpopulation served as the starting point of the trajectory, while the metabolic regulation, niche support, and proliferation subpopulation were mainly located at the terminal end (Fig. [87]3E). This pattern showed that the angiogenesis subpopulation also represents the stem-like cell subpopulation within UC-MSCs, which gradually transits into other subpopulations. The N-CADHERIN^+/CD168^− subpopulation is associated with the angiogenesis capacity of MSCs To identify the cell subpopulations associated with angiogenesis, we integrated single-cell RNA-seq data of MSC-GY with RNA-seq data from MSC-M01 and M02 in [88]GSE13491 by Scissor analysis [[89]29], which can identify the most phenotypically relevant cells from single-cell data. MSC-M02 in Lee et al. significantly promoted the angiogenesis in the heart of rats after MI. In this integrated analysis [[90]23], Scissor identified 1890 Scissor^+ cells with gene expression profiles similar to MSC-M02 (Fig. [91]4A), suggesting that these 1890 Scissor^+ cells in the single-cell data were similar to MSC-M02, which could promote angiogenesis and improve heart function post- MI. Statistical analysis showed that 1824 (96.5%) of these Scissor^+ cells originated from the angiogenesis subpopulation (Fig. [92]4B). Therefore, we concluded that the angiogenesis subpopulation is the functional component of MSCs that can promote angiogenesis and improve heart function in the MI model. Fig. 4. [93]Fig. 4 [94]Open in a new tab The N-CADHERIN^+/CD168^− Subpopulation is Associated with the Angiogenesis Capacity of MSCs (A) tSNE plot illustrating the Scissor-selected cells associated with the therapeutic effect of MSCs in MI treatment, accompanied by statistical analysis. (B) Expression levels of N-CADHERIN and CD168 in Scissor^+ cells compared to other cells. (C) tSNE plot showing the expression pattern of N-CADHERIN and CD168 in MSC-GY. (D) Pearson correlation analysis of N-CADHERHN and CD168 expression levels with angiogenesis-related gene expression. (E) tSNE plot illustrating the clustering of 9,039 cells from [95]GSE182158. (F) tSNE plot displaying the expression pattern of N-CADHERIN and CD168 in [96]GSE182158. (G and H) GO term analysis of up-regulated genes (G) and down-regulated genes (H) in the N-CADHERIN^+/CD168^− subpopulation Next, we performed a comprehensive analysis between the cell surface proteins and DEGs in the angiogenesis subpopulation, finding that the angiogenesis subpopulation was positive for N-CADHERIN and negative for CD168 (Fig. [97]4C). Analysis of data from [98]GSE165811 [[99]30] revealed that CD168 was negatively correlated with the expression levels of angiogenesis-related genes such as VEGFA and FGF2, while N-CADHERIN was positively correlated with these genes (Fig. [100]4D). Additionally, analysis of single-cell data ([101]GSE182158) [[102]31] showed that genes highly expressed in the N-CADHERIN^+/CD168^− cell subpopulation were associated with angiogenesis (Fig. [103]4E–H). These results indicated that N-CADHERIN^+/CD168^− subpopulation is significantly associated with promoting angiogenesis of MSCs. N-CADHERIN^+/CD168^− cell subpopulation have beneficial effects on cardiac function of MI mice To elucidate the relationship between the proportion of N-CADHERIN^+/CD168^− cell subpopulation and the function of UC-MSCs in MI, we detected the ratio of N-CADHERIN^+/CD168^− subpopulation in three UC-MSCs. Flow cytometry analysis revealed that the ratio of the N-CADHERIN^+/CD168^− cell subpopulation in MSC-CX was significantly lower than that in MSC-GY and MSC-CJ (Fig. [104]5A and B). Correlation analysis also showed a positive correlation between the proportion of N-CADHERIN^+/CD168^− cells and the improvement of cardiac function across MSC-GY, MSC-CJ, and MSC-CX (Fig. [105]5C). These results indicated that the ratio of the N-CADHERIN^+/CD168^− subpopulation affects the efficiency of MSCs in treating MI. To explore the role of N-CADHERIN^+/CD168^− cell subpopulation on the cardiac function in mice after MI, we sorted out the N-CADHERIN^+/CD168^− MSC subpopulation from MSC-GY (Fig. S5), and our results showed that the N-CADHERIN^+/CD168^− subpopulation significantly improved heart function and reduced fibrosis in MI mice. In contrast, the other MSCs depleted of N-CADHERIN^+/CD168^− subpopulation failed in improving heart function of MI mice (Figs. [106]5D–F and S6), suggesting that the N-CADHERIN^+/CD168^− MSC subpopulation were indeed the functional subpopulation responsible for the therapeutic effect in treating MI. In addition, we also found that the N-CADHERIN^+/CD168^− cells from MSC-CX could improve heart function and reduce fibrosis (Figs. [107]5D–F and S6), which further supporting the consistent therapeutic role of N-CADHERIN^+/CD168^− cells in treating MI. Fig. 5. [108]Fig. 5 [109]Open in a new tab N-CADHERIN^+/CD168^− Cell Subpopulation Have Beneficial Effects on Cardiac Function of MI Mice. (A and B) Representative images depicting the percentage of N-CADHERIN^+/CD168^− subpopulation in the UC-MSCs from different donors (A), with quantitative analysis of percentages of cells (B). **P < 0.01, ***P < 0.001 versus MSC-CX. (C) Pearson correlation analysis of EF and FS with the percentage of N-CADHERIN^+/CD168^− cells. (D) Echocardiography analysis of EF and FS in MI (Ctrl), MI + N-CADHERIN^+/CD168^− from MSC-GY (N-CAD^+/CD168^−-GY), MI + MSC-GY depleted of N-CADHERIN^+/CD168^− subpopulation (Others-GY), and MI + N-CADHERIN^+/CD168^− from MSC-CX (N-CAD^+/CD168^−-CX) groups at 4 weeks post-MI (n = 6). ns, **P < 0.01, versus Ctrl. (E) Representative images of masson trichrome staining of heart sections at 4 weeks post-MI. Scale bar: 500 μm. (F) Quantification of infarct size at 4 weeks post-MI. ns, **P < 0.01, versus Ctrl. (G) Ranked results of the DEGs in the N-CADHERIN ^+/CD168^− subpopulation. (H) Prediction of transcription factors (TFs) for the N-CADHERIN^+/CD168^− subpopulation. Statistical significance was determined using one-way ANOVA followed by Tukey’s post hoc To identify the key upstream regulators maintaining the N-CADHERIN^+/CD168^− subpopulation, we reconstructed gene regulatory networks using SCENIC [[110]32] and explored the unique transcriptional regulation of DEGs in distinct subpopulations. We found that the angiogenesis subpopulation might be regulated by the transcription factors CEBPB, TCF21, and TAL1 (Fig. [111]5G and H). These results suggested that CEBPB, TCF21, and TAL1 may maintain the fate of the N-CADHERIN^+/CD168^− subpopulation and increase the ratio of N-CADHERIN^+/CD168^− subpopulation in MSCs. Discussion MSCs hold significant promise for MI therapy, yet growing evidence reveals inconsistent therapeutic efficacy among MSCs derived from different donors [[112]8, [113]33–[114]35]. Our study identified critical differences in the therapeutic potential of UC-MSCs from various donors, highlighting that the pro-angiogenic capacity of MSCs was a determinant of their efficacy in MI treatment. Given the variability in donor-derived MSCs used in clinical settings, our findings suggested that selecting MSCs with higher pro-angiogenic potential is essential for achieving more effective clinical outcomes. Paracrine effects are the primary mechanism by which MSCs confer therapeutic benefits in MI. Lee et al. demonstrated that MSCs enhanced cardiac function in MI mice through the secretion of TSG6, which modulated inflammatory responses [[115]4]. Additional studies emphasized the secretion of angiogenesis-related factors as a crucial mechanism. Hu et al. showed that pre-conditioning MSCs under hypoxic conditions (0.5% oxygen for 24 h) significantly elevated the levels of VEGF and other angiogenic factors, thereby enhancing angiogenesis in the infarct border zone, reducing cardiac fibrosis, and improving cardiac function in MI mice [[116]36]. Overexpressing the factors like VEGF and FGF2 had been shown to significantly boost the angiogenic potential of MSCs, thus improving their reparative capacity in damaged hearts after MI [[117]37, [118]38]. Our findings corroborated that effective MSCs secrete higher levels of pro-angiogenic factors, supporting the notion that enhancing angiogenesis via paracrine effects of MSCs is the key to improve cardiac function post-MI. ScRNA-seq enables detailed characterization of cellular expression profiles, offering an invaluable tool for exploring MSC heterogeneity [[119]39]. Recent studies have identified MSC transcriptome profiles based on the DEGs and differential pathways, without functional analysis of subpopulations. Some studies failed to define the functional subpopulations because of insufficient cell numbers. Huang et al. analyzed only 361 UC-MSCs and found that the heterogeneity was primarily influenced by cell cycle differences [[120]40], Hou et al. [[121]41] and Wolock et al. [[122]42] respectively identified the subpopulations in UC-MSCs and bone marrow MSCs, highlighting the differences in differentiation potentials. In this study, we expanded the analysis to 10,463 UC-MSCs, identifying four main subpopulations. Notably, the N-CADHERIN^+/CD168^− subpopulation emerged as a key functional group in regulating angiogenesis. We found that N-CADHERIN^+/CD168^− subpopulation was the functional MSCs in improving heart function of MI mice, and the ratio of this subpopulation was positively correlated with the pro-angiogenic effects of MSCs. Together, our study demonstrated that variations in angiogenic potential influence the therapeutic efficacy of UC-MSCs in MI treatment. This study has several limitations. First, our study found that the N-CADHERIN^+/CD168^− subpopulation was significantly associated with the pro-angiogenic capacity of MSCs and their efficacy in MI mice, which needed to be confirmed in a large animal model of MI (eg, swine or nonhuman primates) and quantified to determine the relationship between the ratio of the N-CADHERIN^+/CD168^− cell subpopulation and the efficacy of MSCs, providing a basis for improving the clinical application of MSCs. Second, we found that CEBPB, TCF21, and TAL1 may maintain the fate of N-CADHERIN^+/CD168^− subpopulation, but have not yet examined whether CEBPB, TCF21, and TAL1 can increase the ratio of functional cell subpopulations in MSCs. We will evaluate the role of these factors in future studies. Conclusion In summary, by exploring UC-MSC heterogeneity, we identified the N-CADHERIN^+/CD168^− subpopulation as crucial for the regulation of angiogenesis and cardiac function recovery in MI mice. The ratio of this subpopulation determined the angiogenic efficacy of UC-MSCs, providing a basis for selecting MSCs suitable for MI therapy. This finding deepened the understanding of UC-MSC heterogeneity and established a new criteria for screening MSCs for effective MI treatment. Supplementary Information [123]Additional file 1.^ (1.3MB, pdf) Acknowledgements