Introduction

Myocardial infarction results in the permanent loss of cardiomyocytes (CM) in the ischemic myocardium significantly affecting the functional performance of the surviving heart and ultimately contributes to heart failure (HF). Every year, the myocardial ischemia results in the death of more than seven million sufferers globally and remarkably impacts the quality of life of the survivors [1]. Current therapeutic interventions including revascularization, drugs and medications, and cardiac resynchronization have been widely practiced which benefitted millions of sufferers across the globe. However, current management strategies are far from myocardial regeneration owing to the inherently poor regenerative capacity of the native heart tissue [1, 2]. Hence, the cell-based therapies have gained global interest for myocardial regeneration where diverse cell types including mesenchymal stem cells (MSCs) and fibroblasts have been attempted [3]. Among them, adipose-derived mesenchymal stem cells (ADMSCs) have been hailed as a practical source owing to the ease and minimal invasiveness of harvesting, autologous resources, excellent plasticity, and superior cardiac differentiation potential [4].

Interestingly, early reports unveiled the existence of resident stem cell population in the myocardium which represent less than 1% of the total cell population in the heart. These resident stem cell population possesses immense translational potential as they accelerate cardiac regeneration following an injury and maintain the cardiac lineage specificity; however, their limited numbers, difficulty in isolation and unavailability of myocardial tissue source offer roadblocks [5]. Importantly, the possibility of cardiac regenerative responses facilitated by adjacent tissues especially epicardial fat (EF) cannot be ignored as EF shares same microvasculature and exist in proximity with myocardium supporting normal cardiac function. In contrast, EF-derived proinflammatory milieu is intimately associated with coronary artery diseases and myocardial ischemia [6, 7]. However, EF remains to be an abundant repository for MSCs eliciting cardiac protective responses via paracrine and vasocrine signaling [8]. Unfortunately, the cardiac regeneration potential of epicardial adipose tissue derived stem cells (EATDS) warrants careful investigation with identification of specific stem cell phenotype.

MSCs are heterogeneous population and the identification of ideal phenotype is crucial in regenerative medicine to obtain reproducible outcomes; however, is challenging [9]. Moreover, information regarding the subpopulations and cardiac regenerative potential of EATDS are currently unavailable despite the recent advancements in adipobiology. In addition, the single cell genomics data regarding the heterogeneity of EATDS have not been reported yet. As EF closely reflects the cardiac pathology during ischemia, the present study aims to determine the EATDS subpopulations under in vitro simulated ischemic and reperfused conditions employing single cell RNA sequencing (scRNAseq) with a focus on the upregulated genes in distinct clusters.

Methodology

Isolation and Maintenance of EATDS

EATDS were isolated employing collagenase digestion method from three hyperlipidemic Yucatan microswine (Sus scorfa, Sinclair bioresources) post-sacrifice. Hyperlipidemia was induced to the animals (8–10 months of age, weighing 60–80 pound) by feeding with high cholesterol high fat diet, and the weight gain and blood chemistry were monitored regularly as reported in our previous study [10]. The harvested stem cells were maintained in DMEM with 10% fetal bovine serum (FBS) (Cat# 30–2020, ATCC) and antibiotics. The stem cells in passage 0–2 were used for the study and were characterized by the expression status for the biomarkers including CD90, CD105, αSMA, vimentin, Sox2, Oct3/4 CD34, CD44 and CD31 following our previously published reports [11].

Ischemia and Reperfusion

The isolated EATDS were grouped into Group I (Control), Group II (Ischemia only) (ISC), and Group III (Ischemia and reperfusion) (ISC/R). Ischemia was simulated by treating the cells in ischemic buffer (118 mM NaCl, 24 mM Na2HCO3, 1 mM Na2HPO4, 2.5 mM CaCl2, 1.2 mM, MgCl2, 20 mM sodium lactate, 16 mM KCl, 10 mM 2-deoxyglucose and pH 6.2) for 2 h and reperfusion was attained by replacing the ischemic buffer with complete culture media overnight. The treatment with ischemic buffer results in hypoxia and the reoxygenation was achieved using complete media and the experimental groups were defined accordingly. ISC group tolerated hypoxia/anoxia due to treatment with ischemic buffer, ISC/R group received reoxygenation following ischemia due to the treatment with complete media, and the cells grown in complete media served as control.

Single Cell RNA Sequencing (scRNA-seq)

EATDS were pooled from three swine and used for scRNAseq analysis. Single-cell library preparation was performed in the commercially available 10 × Genomics Chromium System (Children's Hospital Los Angeles SC2 Core, CA) exploiting the droplet method. The library generation was performed using the Chromium Single Cell 5′ v2 chemistry to capture ~ 10,000 cells per sample. For the analysis, the cells from three pigs were pooled in each experimental group. Reagents and cells were combined to generate GEMs (Gel Beads-in-Emulsion) in such a way that single cells were partitioned with unique, cell-linked molecular barcodes following the cell sorting and counting which were loaded on the microfluidics Chip K. The samples were sequenced on the Illumina HiSeq 2500 and the Illumina NextSeq 500 high output at ~ 20,000 reads/cell. The generated raw scRNA-seq data was processed using the 10 × Genomics CellRanger pipeline and the data obtained from the Illumina NextSeq platform was processed to Fastq files using the CellRanger mkfastq program. Then, the Fastq files were mapped to the sus_scrofa_11 database (https://uswest.ensembl.org/Sus_scrofa/Info/Index). The CellRanger count program was run on individual Fastq datasets from the different treatment conditions and CellRanger aggr was employed to generate aggregated unique molecular identifier (UMI) count matrices for the experimental datasets generated from the study.

Statistical analysis for the expression status of the scRNA-seq data was performed by the statistical program associated with the 10 × platform [12]. The processed data revealing the cell clusters was examined in the Loupe Browser 5.0.1. using graph-based analysis mode. The clusters of interests were analyzed in LibraryId mode to assess the gene expression and cell count in each experimental group. The comparison was performed based on the number of cells mapped in each cluster and the locally upregulated genes were listed compared to the control. The genes were sorted based on the fold-change of upregulation where FC < 2 was omitted from the listing. Similarly, the cell clusters/populations < 2% of the total population was omitted to obtain reliable data. Also, the biomarker expression was assessed from the scRNAseq data based on the gene/feature expression mode.

Results

Single Cell Genomics

Web-summary of the scRNAseq data profile for the spatial gene expression using barcoding-based transcriptomics technology of 10 Genomics platform revealed > 91% Fraction Reads in Cells (FRC) and > 98% valid barcodes in control, ISC and ISC/R groups suggesting the effective sample preparation providing healthy cells for the analysis. Estimated number of cells, mean reads per cell, and median genes per cell obtained were 6922, 49,027 and 4066, 9791, 28,912 and 3151 and 8808, 30,135 and 3483, respectively for control, ISC and ISC/R groups with a total estimate of 25,521 cells, 711,958,422 total sequencing reads (Post-normalization), 27,897 mean reads per cell (Post-normalization), and 3,299 mean genes per cells throughout all three samples. Overall, the Loupe analysis revealed heterogeneous cell clusters in ISC, and ISC/R groups compared to the control cells.

The EATDS were positive for the mesenchymal markers, vimentin and αSMA, and the stem cell markers, SOX2, CD90 and Oct3/4 and the adipocyte marker adiponectin, as evident from the immunostaining. The cells were negative for CD105, CD14, CD44, and CD34 (Fig. 1A). Similar trend was obtained from scRNAseq data where the transcriptomics of vimentin and αSMA was evident from the mean fold change > 100 in the violin plots (Fig. 1A). Similarly, the negative biomarkers displayed the fold change ≤ 1 as displayed in the violin plots (Fig. 1A). However, the details for SOX2 and Oct3/4 were unavailable in the database. Based on the characterization, the EATDS were defined to be Vim + /αSMA + /SOX2 + /Oct3/4/CD90-/CD105-/CD14-/CD44-/CD34/ADIPOQ-.

Fig. 1
figure 1

(A) Immunofluorescence analysis for the protein expression of biomarkers in cultured EATDS. Images in the top panel show overlayed images of biomarkers with the nuclear stain DAPI. Images were acquired at 20 × magnification using CCD camera attached to the Leica Thunder microscope. The images right to the overlayed images reveal the UMAP analysis of the scRNAseq data revealing the expression status of each gene in the total cell population. Images in the lower panel show the violin plots showing the distribution of gene expression in each cluster.Distribution of cell populations and altered genes in ISC-EATDS: (B) UMAP analysis revealing three distinct population of cells as evident by green (ISC), orange (control) and blue (ISC/R) spots. (C) Heat map showing the spatial expression of 763 genes identified from EATDS based on the upregulation in ISC group. (D) Scatter plot showing the FC expression indicating the upregulation of genes in ISC than control and ISC/R. (E) Violin plot showing the distribution of upregulated genes in which the ISC group showing trend towards > 1 FC. (F) Heat map of highly altered 16 genes (FC > 2) revealing the upregulation in ISC group. (G) Scatterplot of highly expressed genes showing the minimal level of expression in control and ISC/R group and increased expression in ISC group. (H) Violin plot for highly expressed genes revealing the distribution towards upregulation in ISC compared to control and ISC/R. Volcano plots displaying the differential expression (both upregulated and downregulated) of key genes in (I) control, (J) ISC and (K) ISC/R groups

Full size image

UMAP analysis revealed three unique population of cells each representing control, ISC and ISC/R (Fig. 1B). The cells from control and ISC/R were clustered together whereas ISC cells were distinct from the other two groups. Also, 763 genes were highly altered in the global population of EATDS cells where the sorting was performed based on the extent of upregulation in the ISC group (Supplementary Table 1, Fig. 1C and D). The violin plot revealed that the distribution of cells, based on the expression of these genes, were trending to be FC > 1 whereas the cells in the control and ISC groups were clustered around FC = 1 displaying uniform distribution. However, a few genes in control cells tend towards FC > 1 (Fig. 1D and E). Sixteen genes (FC > 2) were significantly increased in ISC group on comparison with the control and ISC/R where PRXL2A (Peroxiredoxin like 2A) exhibited maximum FC of 5.21 (P < 0.0001) (Table 1, Fig. 1F-H). The FC values of these 16 genes in the control and ISC/R groups were distributed mostly around 1 whereas FC > 2 was evident in ISC group (Fig. 1F-H). Apart from the 16 genes in ISC, DHRS3 gene was only one to be upregulated (FC = 2.30, P < 0.0001) in the other two groups. Overall, the data suggest a unique population of ISC cells which tend to revert towards control lineage upon reperfusion. The volcano plot revealed the of downregulation IL-11, TNC, PRXL2A, and RNF121 (Ring Finger Protein 121) in the control cells with a concomitant upregulation of ISG15 (Interferon-stimulated gene 15), MX1 (MX dynamin like GTPase 1), DHRS3, and LMCD1 (LIM and Cysteine Rich Domains 1) (Fig. 1I). The ISC cells revealed the downregulation of ENSG00000171848 (RRM2, ribonucleotide reductase regulatory subunit M2), PCNA, MCM7 (mini chromosome maintenance function 7), TK1 (Thymidine Kinase 1) and PCLAF (PCNA Clamp Associated Factor) and the upregulation of PRXL2A and ALDH1A1 (Aldehyde dehydrogenase Family Member A1) (Fig. 1J). Reperfusion resulted in the downregulation of PRXL2A, DHRS3 and ALDH1A1 with an upregulation of IL-11, RRM2, TNC, PRXL2A, PCNA, MCM7, PCLAF, TK1 and RNF121 (Fig. 1K). The t-SNE plot revealed 18 unique clusters of cells overall representing control, ISC and ISC/R (Fig. 2A and B). The number of cells in each cluster and sub-clusters are displayed in Table 2.

Table 1 The list of genes globally upregulated (FC > 2) in ISC group
Full size table
Fig. 2
figure 2

Distribution of cell populations and altered genes in Cluster 1: (A) t-SNE plot revealing 18 distinct clusters of cells in the global population of EATDS. (B) Split view of Control, ISC and ISC/R groups and combined view of t-SNE plot showing the distribution of cells within the Cluster 1 based on the local expression of 959 genes. (C) Violin plot showing the distribution of upregulated genes in which the ISC group showing trend towards > 1 FC. (D) Scatter plot showing the FC expression indicating the upregulation of genes in ISC than control and ISC/R. (E) Heat map showing the local expression of 959 genes identified from EATDS based on the upregulation in ISC group. (F) Scatterplot of highly expressed 25 genes in ISC group showing the minimal level of expression in control and ISC/R group where the dotted lines indicate mean FC of 25 genes. (G) Violin plot for highly expressed genes revealing the distribution towards upregulation in ISC compared to control and ISC/R. (H) Heat map of highly altered 25 genes (FC > 2) revealing the upregulation in ISC group. (I) Scatterplot of highly expressed 5 genes (FC > 2) in the Control group showing the minimal level of expression in ISC/R and similar level of expression in ISC. (J) Heat map of highly altered 5 genes (FC > 2) in Control group compared to ISC and ISC/R groups. (K) Violin plot for highly expressed 5 genes in Control group revealing the trend towards upregulation in ISC and downregulation in ISC/R

Full size image
Table 2 Cell density in each cluster
Full size table

Cluster 1 Cells Favored ISC/R

Cluster 1 exhibited a total number of 3,300 cells where ~ 95% cells were mapped in ISC/R group with ~ 4% in control groups and ~ 0.7% cells in ISC group (Table 2). This suggests the existence of a predominant population of cells in ISC/R which is distinct from control and ISC group. Also, 959 genes were highly altered in the local population of EATDS cells in the cluster 1 where the sorting was based on the extent of gene expression in ISC group (Supplementary Table 2, Fig. 2B–E). The violin and scatter plots revealed that the distribution of cells, based on the expression of these genes, were trending to be FC > 1 in ISC group whereas the cells in the control and ISC groups were clustered around FC = 1 and displayed uniform distribution. However, a few genes in control cells tend towards FC > 1 (Fig. 2C–E). Twenty-five genes (FC > 2) were increased in ISC group on comparison with the control and ISC/R. The DNA binding protein H1-0 (Histone H1.0) exhibited maximum FC of 6.12 followed by the ECM component protein MGP (Matrix Gla Protein) with a FC of 5.58 whereas the FC of other three genes were less than 5 (Table 3, Fig. 2F-H). The FC values of these 25 genes in the control groups were distributed mostly around 1 whereas that of ISC/R were tend mostly towards FC < 1 (Fig. 2F-H). Five genes (FC > 2) were increased in Control group on comparison with the ISC and ISC/R where DHRS3 (Short-chain dehydrogenase/reductase 3) and TGFβ3 (Transforming growth factor beta 3) were highly upregulated (Table 3, Fig. 2I-K). The violin and scatter plots revealed that the distribution of cells was trending to be FC > 2 in control group and FC > 1 in ISC group whereas the cells in the ISC/R groups were clustered around FC < 1; however, displayed uniform distribution (Fig. 2I-K). The genes with FC > 2 was completely absent in ISC/R group suggesting the unique population of cells with basal level of expression.

Table 3 The list of genes locally upregulated (FC > 2) in each group based on Cluster-1 cells
Full size table

Cluster 2 and 3 Tend Towards Normal

Cluster 2 displayed 2,454 cells where ~ 94% cells were mapped in control group with ~ 1% in ISC group and ~ 5% cells in ISC/R group (Fig. 3A, Table 2). This suggests the existence of a predominant population of cells in the control which is distinct from ISC and ISC/R groups. Interestingly, 789 genes were highly altered in the local population of EATDS cells in the Cluster 2 where the sorting was based on the extent of gene expression in ISC group (Supplementary Table 3, Fig. 3B–D). The violin and scatter plots revealed that the distribution of cells, based on the expression of these genes, were trending to be FC > 1 in control group whereas the cells in the ISC and ISC/R groups were clustered around FC = 1. However, a few genes in ISC cells tend towards FC > 1 (Fig. 3B–D). Three genes (FC > 2) were increased in ISC group on comparison with the control and ISC/R. The cell cycle mediator CDC20 (Cell division cycle protein 20 homolog) exhibited maximum FC of 3.28 whereas the FC of other two genes were less than 3 (Table 4, Fig. 3E–G). The FC values of these 3 genes in the control groups were distributed mostly around 1 whereas that of ISC/R were tend mostly towards FC < 1 (Table 4, Fig. 3E–G). Ten genes (FC > 2) were increased in control group on comparison with the ISC and ISC/R. The highly expressed genes were DHRS3 (FC = 3.94), and MGP (FC = 3.65) whereas the remining 8 genes displayed FC ≤ 3 (Table 4, Fig. 3H–J). The violin and scatter plots revealed that the distribution of cells was trending to be FC > 2 in control group and FC < 1 in ISC and ISC/R groups which displayed distribution around the median FC. The genes with FC > 2 was completely absent in ISC/R group suggesting the unique population of cells with basal level of expression.

Fig. 3
figure 3

Distribution of cell populations and altered genes in Cluster 2: (A) Split view of Control, ISC and ISC/R groups and combined view of t-SNE plot showing the distribution of cells within the Cluster 2 based on the local expression of 789 genes. (B) Violin plot showing the distribution of upregulated genes in which the ISC group showing uniform distribution around FC = 1 whereas the control cells tend towards FC > 1. (C) Scatter plot showing the FC expression indicating the upregulation of genes in ISC than control and ISC/R. (D) Heat map showing the local expression of 789 genes identified from EATDS based on the upregulation in ISC group. (E) Violin plot for highly expressed 3 genes revealing the distribution towards upregulation in ISC compared to control and ISC/R. (F) Scatterplot of highly expressed 3 genes in ISC group showing the minimal level of expression in control and ISC/R group where the dotted lines indicate mean FC. (G) Heat map of highly altered 3 genes (FC > 2) revealing the upregulation in ISC group. (H) Violin plot for highly expressed 10 genes in Control group revealing the trend towards upregulation whereas downregulation and uniform distribution in ISC and ISC/R. (I) Scatterplot of highly expressed 10 genes (FC > 2) in the Control group showing the minimal level of expression in ISC/R and similar level of expression in ISC. (J) Heat map of highly altered 10 genes (FC > 2) in Control group compared to ISC and ISC/R groups

Full size image
Table 4 The list of genes locally upregulated (FC > 2) in each group based on Cluster 2 cells
Full size table

Similarly, the cluster 3 displayed 2,612 cells where ~ 95% cells were mapped in control group with ~ 2% in the ISC group and ~ 3% cells in ISC/R group (Fig. 4A–D, Table 2) (Supplementary Table 4). This suggests the existence of a predominant population of cells in the control which is distinct from ISC and ISC/R group. Interestingly, 970 genes were detected to be highly altered in the local population of EATDS cells in the cluster 3 where the sorting was based on the extent of gene expression in ISC group. However, all genes except CCN5 (Cellular Communication Network Factor 5) in the control group exhibited FC < 2.

Fig. 4
figure 4

Distribution of cell populations and altered genes in Cluster 3 and 4: (A) Split view of Control, ISC and ISC/R groups of Cluster 3 cells and combined view of t-SNE plot showing the distribution of cells within the Cluster 3 tending towards the control based on the local expression of 970 genes. (B) Violin plot showing the distribution of upregulated genes in which the ISC and ISC/R groups showing the distribution around FC = 1 whereas the control cells tend towards FC > 1. (C) Scatter plot showing the FC expression indicating the upregulation of genes in control than ISC and ISC/R. (D) Heat map showing the local expression of 970 genes identified from EATDS based on the upregulation in ISC group. (E) Split view Control, ISC and ISC/R groups of Cluster 4 cells and combined view of t-SNE plot showing the distribution of cells within the Cluster 4 based on the local expression of 1431 genes. (F) Violin plot for highly expressed 1431 genes revealing the distribution towards upregulation in control and ISC group. (G) Scatterplot of highly expressed 1431 genes in Control group showing the minimal level of expression in ISC group where the dotted lines indicate mean FC. (H) Heat map of highly altered 1431 genes (FC > 2) revealing the upregulation in Control group. (I) Scatterplot of highly expressed 95 genes in Control group showing the minimal level of expression in ISC group where the dotted lines indicate mean FC. (J) Violin plot for highly expressed 95 genes revealing the distribution towards upregulation in control and ISC group. (K) Heat map of highly altered 95 genes (FC > 2) revealing the upregulation in Control group

Full size image

Cluster 4 Cells Tend Toward Ischemia

Cluster 4 displayed 2,585 cells where ~ 99.5% cells were mapped in the ISC group, and control group with ~ 0.5%; however, ISC/R group displayed negligible cells (Fig. 4E, Table 2). Hence, the comparison was between ISC and control groups. This suggests the existence of a predominant population of cells in the ISC group which is distinct from the control. Interestingly, 1,431 genes were detected in the locally in the Cluster 4 where the sorting was based on the extent of gene expression in control group (Supplementary Table 5, Fig. 4E–H). The violin and scatter plots revealed that the distribution of cells trending to be FC > 1 in control group whereas the cells in the ISC group were clustered around FC = 1 (Fig. 4F–H). Interestingly, 95 genes (FC > 2) were downregulated in ISC group on comparison with the control. The FC values of these 95 genes in the control groups were distributed mostly around 2 whereas that of ISC were tend mostly towards FC < 1 (Table 5). Similarly, the violin and scatter plots revealed that the distribution of cells was trending to be FC > 2 in control group and FC < 1 in ISC (Fig. 4I–K). The fibroblast differentiation protein NREP (Neuronal Regeneration-Related Protein) exhibited maximum FC of 6.13 whereas the FC RAMP1 (Receptor Activity Modifying Protein 1), PICK1 (Protein Interacting with PRKCA1) and ISG15 (Interferone Simulated Gene 15) were 4.59, 4.31, and 4.36 respectively (Table 5, Fig. 4I–K). FC for the remaining 91 genes were between 2 and 3 (Table 5). The genes with FC > 2 was completely absent in ISC group suggesting that this unique population of cells lineaging from the control cells by the downregulation of at least 95 genes in response to ischemia.

Table 5 The list of locally downregulated genes in ISC group with respect to control (FC > 2) group in Cluster 4 cells
Full size table

Cluster 5 Cells Favor Reperfusion

Cluster 5 displayed 1,998 cells where ~ 96% cells were mapped in ISC/R group, and control group contains ~ 4% cells; however, ISC group displayed negligible cells (Fig. 5A, Table 2). Hence, the comparison was between ISC/R and control groups. The data suggests the existence of a predominant population of cells in the ISC/R group which is distinct from the control. Interestingly, 2,768 genes were detected locally in the cluster 5 where the sorting was based on the extent of gene expression in ISC/R group (Supplementary Table 6, Fig. 5A–D). The violin and scatter plots revealed that the distribution of cells trending to be FC > 1 in both the groups (Fig. 5A–D). Interestingly, 28 genes (FC > 2) were upregulated in ISC/R group on comparison with the control (Table 6). The FC values of these 28 genes in the ISC/R group were distributed around 2 whereas that of control group were tend mostly towards FC < 1 (Table 6). Similarly, the violin and scatter plots revealed that the distribution of cells was trending to FC > 2 in ISC/R group and FC < 1 in control (Fig. 5E–G). Hematopoietic cytokine, IL-11, (FC = 8.10) and the DNA binding protein RNF121 (Ring Finger Protein 121) (FC = 8.03) were highly expressed in ISC/R cells and the FC of remaining 27 proteins were less than 5 (Table 6). Similarly, 21 genes (FC > 2) were upregulated in control group on comparison with the ISC/R (Table 6). The violin and scatter plots revealed that the distribution of cells was trending to be FC > 2 in control group and FC < 1 in ISC/R (Fig. 5I–K). MI associated protein PROCR (Protein C Receptor) exhibited maximum FC of 7.93 whereas the FC of remaining 27 proteins were less than 4 (Table 6). The data suggests that the unique population of cells in ISC/R group lineages from the control cells by the downregulation of at least 21 and upregulation of at least 28 genes in response to ischemia.

Fig. 5
figure 5

Distribution of cell populations and altered genes in Cluster 5: (A) Split view of Control, ISC and ISC/R groups of Cluster 5 cells and combined view of t-SNE plot showing the distribution of cells within the Cluster 5 tending towards the ISC/R group based on the local expression of 2,768 genes. (B) Violin plot for highly expressed 2,768 genes revealing the distribution towards upregulation in control and ISC/R group. (C) Scatterplot of highly expressed 2,768 genes in Control group showing the downregulation trend compared to ISC/R group. (D) Heat map of highly altered 2,768 genes (FC > 2) revealing the upregulation in ISC/R group. (E) Violin plot for highly expressed 28 genes revealing the distribution towards upregulation in ISC/R group compared to control. (F) Scatterplot of highly expressed 28 genes in ISC/R group showing the minimal level of expression in control group where the dotted lines indicate mean FC. (G) Heat map of highly altered 28 genes (FC > 2) revealing the upregulation in ISC/R group. (H) Violin plot for highly expressed 21 genes revealing the distribution towards upregulation in Control group compared to ISC/R group. (I) Scatterplot of highly expressed 21 genes in Control group showing the minimal level of expression in ISC/R group where the dotted lines indicate mean FC. (K) Heat map of highly altered 21 genes (FC > 2) revealing the upregulation in Control group

Full size image
Table 6 The list of genes locally upregulated genes (FC > 2) in ISC/R and control groups with respect to each other group in Cluster 5 cells
Full size table

Cluster 6 Cells Favors Ischemia

Cluster 6 displayed 1,947 cells where ~ 98% cells were mapped in ISC group, and control group contains ~ 2% cells; however, ISC/R group displayed negligible cells (Fig. 6A, Table 2). Hence, the comparison was between ISC and control groups. The data suggests the existence of a predominant population of cells in the ISC group which is distinct from the control. Interestingly, 2,908 genes were detected locally in cluster 6 where the sorting was based on the extent of gene expression in ISC group (Supplementary Table 7, Fig. 6A–D). The violin and scatter plots revealed that the overall distribution of cells was tending to be around FC = 1 in both the groups (Fig. 6A–D). Interestingly, 37 genes (FC > 2) were upregulated in ISC group on comparison with the control (Table 7). FC values of these 37 genes in the ISC group were distributed around FC > 2 whereas that of control group tend towards FC < 1 (Table 7). Similarly, the violin and scatter plots revealed that the distribution of cells was tending to FC > 2 in ISC group and FC < 1 in control (Fig. 6E–G). The contractile protein MYH11 (Myosin Heavy Chain 11) (FC = 4.9) and the calcium binding protein CRELD2 (Cysteine Rich with EGF Like Domains 2) (FC = 4.35) were highly expressed in ISC cells and the FC of remaining 35 proteins were less than 4 (Table 7). Similarly, 52 genes (FC > 2) were upregulated in the Control group on comparison with the ISC (Table 7). The violin and scatter plots revealed that the distribution of cells was towards FC > 2 in Control group and FC < 1 in ISC (Fig. 6I–K). The immunomodulatory protein OAS2 (2'-5'-Oligoadenylate Synthetase 2) exhibited maximum FC of 5.35 whereas the FC of remaining 51 proteins were less than 5 (Table 7). The data suggest that the unique population of cells in ISC group lineages from the control cells by the downregulation of at least 52 and upregulation of at least 37 genes in response to ischemia.

Fig. 6
figure 6

Distribution of cell populations and altered genes in Cluster 6: (A) Split view of Control, ISC and ISC/R groups of Cluster 6 cells and combined view of t-SNE plot showing the distribution of cells within the Cluster 6 tending towards the ISC group based on the local expression of 2,908 genes. (B) Violin plot for highly expressed 2,908 genes revealing the distribution towards upregulation in control and ISC groups. (C) Scatterplot of highly expressed 2,908 genes in Control group showing the downregulation trend compared to ISC group. (D) Heat map of highly altered 2,908 genes (FC > 2) revealing the upregulation in ISC group. (E) Violin plot for highly expressed 37 genes revealing the distribution towards upregulation in ISC group compared to control. (F) Scatterplot of highly expressed 37 genes in ISC group showing the minimal level of expression in control group where the dotted lines indicate mean FC. (G) Heat map of highly altered 37 genes (FC > 2) revealing the upregulation in ISC group. (H) Violin plot for highly expressed 52 genes revealing the distribution towards upregulation in Control group compared to ISC group. (I) Scatterplot of highly expressed 52 genes in Control group showing the minimal level of expression in ISC group where the dotted lines indicate mean FC. (K) Heat map of highly altered 52 genes (FC > 2) revealing the upregulation in Control group

Full size image
Table 7 The list of genes locally upregulated genes (FC > 2) in control and ISC groups with respect to each other group in Cluster 6 cells
Full size table

Cluster 7 Cells Tend Towards Reperfusion

Cluster 7 displayed 1,746 cells where ~ 92% cells were mapped in ISC/R group, ISC group contain ~ 1% cells and control group contain ~ 7% cells (Fig. 7A, Table 2). The comparison was between ISC/R, ISC and control groups and the data suggest the existence of a predominant population of cells in the ISC/R group which is distinct from the control and ISC. Interestingly, 2,037 genes were detected locally in cluster 7 where the sorting was based on the extent of gene expression in ISC/R group (Supplementary Table 8, Fig. 7A–D). The violin and scatter plots revealed that the overall distribution of cells was tending to be around FC = 1 in both the groups (Fig. 7A–D). Interestingly, 15 genes (FC > 2) were upregulated in ISC/R group on comparison with the control and ISC (Table 8). FC values of these 15 genes in the ISC/R group were distributed around FC > 2 whereas that of control and ISC groups tend towards FC < 1 (Table 8). Similarly, the violin and scatter plots revealed that the distribution of cells was tending to FC > 2 in ISC/R group, and FC = 1 in ISC and FC < 1 in control groups (Fig. 7E–G). Hematopoietic cytokine, IL-11, (FC = 10.27) and the DNA binding protein RNF121 (Ring Finger Protein 121) (FC = 7.27) were highly expressed in ISC/R cells and the FC of remaining 13 proteins were less than 4 (Table 8). Similarly, 49 genes (FC > 2) were upregulated in ISC group on comparison with the ISC/R and control (Table 8). The violin and scatter plots revealed that the distribution of cells was towards FC > 2 in ISC group and FC < 1 in ISC/R and control group (Fig. 7I–J). Histone protein H1.0 exhibited maximum expression (FC = 4.06) followed by the chaperon protein HSPA5 (Heat Shock Protein 70 Family Protein 5) (FC = 3.89) whereas the FC of other 47 genes were between 3.61 and 2 (Table 7). The membrane transporter protein, ABCB1 (ATP-binding cassette, sub-family B (MDR/TAP), member 1) (FC = 2.55) and the transcriptional regulator LMCD1 (LIM and Cysteine Rich Domains 1) (FC = 2.28) were the two genes upregulated in the control group compared with ISC/R and ISC group (Table 7). The data suggests that the unique population of cells in ISC/R group lineaging from the control cells by the downregulation of at least 49 genes in ischemia and 2 genes in control and the upregulation of at least 15 genes regarding ischemia and control.

Fig. 7
figure 7

Distribution of cell populations and altered genes in Cluster 7: (A) Split view of Control, ISC and ISC/R groups of Cluster 7 cells and combined view of t-SNE plot showing the distribution of cells within the Cluster 7 tending towards the ISC/R group based on the local expression of 2,037 genes. (B) Violin plot for highly expressed 2,037 genes revealing the distribution towards upregulation in ISC and ISC/R groups. (C) Scatterplot of highly expressed 2, 037 genes in ISC/R and ISC group showing the upregulation trend compared to Control group. (D) Heat map of highly altered 2, 037 genes (FC > 2) revealing the upregulation in ISC/R and ISC group. (E) Violin plot for highly expressed 15 genes revealing the distribution towards upregulation in ISC/R group compared to control. (F) Scatterplot of highly expressed 15 genes in ISC group showing the minimal level of expression in control group and FC > 1 in ISC where the dotted lines indicate mean FC. (G) Heat map of highly altered 15 genes (FC > 2) revealing the upregulation in ISC/R group. (H) Violin plot for highly expressed 49 genes revealing the distribution towards upregulation in ISC group compared to Control and ISC/R groups. (I) Scatterplot of highly expressed 49 genes in ISC group showing the minimal level of expression in control and ISC/R group where the dotted lines indicate mean FC. (K) Heat map of highly altered 49 genes (FC > 2) revealing the upregulation in ISC group

Full size image
Table 8 The list of genes locally upregulated genes (FC > 2) in control and ISC groups with respect to each other group in Cluster 7 cells
Full size table

Cluster 8 Cells Tend Towards Ischemia

Cluster 8 displayed 1,721 cells where ~ 94% cells were mapped in ISC group, control group contain ~ 7% cells; however, ISC/R displayed negligible cells (~ 0.2%) (Fig. 8A, Table 2). The comparison was between ISC, and control groups and the data suggest the existence of a predominant population of cells in the ISC group which is distinct from the control. Interestingly, 1,895 genes were detected locally in cluster 8 where the sorting was based on the extent of gene expression in ISC group (Supplementary Table 9, Fig. 8A–D). The violin and scatter plots revealed that the overall distribution of cells was tending to be around FC = 1 in both the groups (Fig. 8A–D). Interestingly, 37 genes (FC > 2) were upregulated in ISC group on comparison with the control (Table 9). FC values of these 37 genes in the ISC group were distributed around FC > 2 whereas that of control tend towards FC < 1 (Table 8). Similarly, the violin and scatter plots revealed that the distribution of cells was tending to FC > 2 in ISC and FC < 1 in control groups (Fig. 8E–G). The chaperone protein, HSP90B1 (Heat Shock Protein 90 Beta Family Member 1) (FC = 4.29) and CRELD2 (Cysteine Rich with EGF Like Domains 2) (FC = 4.25) were highly expressed in ISC cells and the FC of remaining 35 proteins were less than 4 (Table 9). Similarly, 13 genes (FC > 2) were upregulated in Control group on comparison with the ISC (Table 9). The violin and scatter plots revealed that the distribution of these 13 genes was towards FC > 2 in control group and FC < 1 in ISC group (Fig. 8H–J). The extracellular cytokine, ISG15 (Interferon-stimulated gene 15) exhibited maximum expression (FC = 4.60) whereas the FC of other 14 genes were below 3 (Table 9). The data suggests that the unique population of cells in ISC group lineaging from the control cells by the downregulation of at least 13 and the upregulation of at least 37 genes in response to ischemia.

Fig. 8
figure 8

Distribution of cell populations and altered genes in Cluster 8: (A) Split view of Control, and ISC groups of Cluster 8 cells and combined view of t-SNE plot showing the distribution of cells within the Cluster 8 tending towards the ISC group based on the local expression of 1,895 genes. (B) Violin plot for highly expressed 1,895 genes revealing the distribution towards upregulation in ISC. (C) Scatterplot of highly expressed 1,895 genes in ISC group showing the upregulation trend compared to Control group. (D) Heat map of highly altered 1,895 genes (FC > 2) revealing the upregulation in ISC group. (E) Violin plot for highly expressed 37 genes revealing the distribution towards upregulation in ISC group compared to control. (F) Scatterplot of highly expressed 37 genes in ISC group showing the minimal level of expression in control group and FC > 2 in ISC where the dotted lines indicate mean FC. (G) Heat map of highly altered 37 genes (FC > 2) revealing the upregulation in ISC/R group. (H) Violin plot for highly expressed 13 genes revealing the distribution towards upregulation in Control compared to ISC group. (I) Scatterplot of highly expressed 13 genes in Control group showing the minimal level of expression in ISC group where the dotted lines indicate mean FC. (K) Heat map of highly altered 13 genes (FC > 2) revealing the upregulation in ISC group

Full size image
Table 9 The list of genes locally upregulated genes (FC > 2) in control and ISC groups with respect to each other group in Cluster 8 cells
Full size table

Cluster 9 Cells Tend Towards Ischemia

Cluster 9 displayed 1,286 cells where ~ 96% cells were mapped in ISC group, control and ISC/R groups contain ~ 2% cells each (Fig. 9A, Table 2). The comparison was between ISC, ISC/R and control groups and the data suggests the existence of a predominant population of cells in the ISC group which is distinct from the control. Interestingly, 3,093 genes were detected locally in cluster 9 where the sorting was based on the extent of gene expression in ISC group (Supplementary Table 10, Fig. 9A–D). The violin and scatter plots revealed that the overall distribution of cells was tending to be around FC = 1 in all the three groups (Fig. 9A–D). Interestingly, 37 genes (FC > 2) were upregulated in ISC group on comparison with the control (Table 10). FC values of these 25 genes in the ISC group were distributed around FC > 2 whereas that of control tend towards FC < 1 and ISC/R tend towards FC > 1 (Table 10). Similarly, the violin and scatter plots revealed that the distribution of cells was tending to FC > 2 in ISC and FC < 1 in the control and FC > 1 in ISC/R groups (Fig. 9E–G). Similar to the Cluster 8, CRELD2 (FC = 4.12) was highly expressed in ISC cells and the FC of remaining 24 proteins were less than 4 (Table 10). Similarly, 32 genes (FC > 2) were upregulated in the control group on comparison with the ISC (Table 10). The violin and scatter plots revealed that the distribution of these 32 genes was towards FC > 2 in control and ISC/R groups and FC < 1 in ISC group (Fig. 9H–J). The cytoskeletal protein ACTA1 (Actin Alpha 1, Skeletal Muscle) (FC = 9.84) exhibited maximum expression whereas the FC of other 21 genes were below 4 (Table 10). Additionally, 375 genes (FC > 2) were upregulated in ISC/R group on comparison with the ISC and control (Table 10). The violin and scatter plots revealed that the distribution of these 375 genes was towards FC > 2 in control and ISC/R groups and FC < 1 in ISC group (Fig. 9H–M). The DNA replication machinery component MCM3 (Minichromosome Maintenance Complex Component 3) (FC = 31.05) and the motor protein (Myosin XIX) (FC = 10.38) exhibited maximum expression whereas the FC of other 27 genes were between 10 and 4 and the remaining 346 genes were below 4 (Table 10). The data suggests that the unique population of cells in ISC group lineaging from the control cells by the downregulation of at least 32 and the upregulation of at least 25 genes in response to ischemia and the ischemic response was elicited by the upregulation of 375 genes as evident in ISC/R group.

Fig. 9
figure 9

Distribution of cell populations and altered genes in Cluster 9: (A) Split view of Control, ISC and ISC/R groups of Cluster 9 cells and combined view of t-SNE plot showing the distribution of cells within the Cluster 9 tending towards the ISC group based on the local expression of 3,093 genes. (B) Violin plot for highly expressed 3,093 genes revealing the distribution towards upregulation in ISC/R group. (C) Scatterplot of highly expressed 3,093 genes in ISC group showing the upregulation trend compared to ISC and ISC/R groups. (D) Heat map of highly altered 3,093 genes (FC > 2) revealing the upregulation in ISC and ISC/R groups. (E) Violin plot for highly expressed 25 genes revealing the distribution towards upregulation in ISC group compared to control and ISC/R. (F) Scatterplot of highly expressed 25 genes in ISC group (FC > 2) showing the minimal level of expression in control group where the dotted lines indicate mean FC. (G) Heat map of highly altered 25 genes (FC > 2) revealing the upregulation in ISC group. (H) Violin plot for highly expressed 32 genes revealing the distribution towards upregulation in Control and ISC/R compared to ISC group. (I) Scatterplot of highly expressed 32 genes in Control group showing the minimal level of expression in ISC group where the dotted lines indicate mean FC. (J) Heat map of highly altered 32 genes (FC > 2) revealing the upregulation in Control and ISC/R group. (K) Violin plot for highly expressed 375 genes revealing the distribution towards upregulation in ISC/R compared to Control and ISC groups. (L) Scatterplot of highly expressed 375 genes in ISC/R group showing the minimal level of expression in ISC group where the dotted lines indicate mean FC. (M) Heat map of highly altered 375 genes (FC > 2) revealing the upregulation in ISC/R group

Full size image
Table 10 The list of genes locally upregulated genes (FC > 2) in control, ISC and ISC/R groups with respect to each other group in Cluster 9 cells
Full size table

Cluster 10 Cells Tend Towards Control and Ischemia

Cluster 10 displayed 1,261 cells where ~ 40% cells were mapped in ISC group, ~ 57% in control group and ISC/R groups contain ~ 3% cells (Fig. 10A, Table 2). The comparison was between ISC, ISC/R and control groups and the data suggests the existence of a predominant population of cells in the ISC group and control group which is distinct from the ISC/R. Interestingly, 2,546 genes were detected locally in cluster 10 where the sorting was based on the extent of gene expression in ISC group (Supplementary Table 10, Fig. 10A–D). The violin and scatter plots revealed that the overall distribution of cells was tending to be around FC > 1 in ISC and ISC/R groups and FC = 1 in the control group (Fig. 10A–D). Interestingly, 25 genes (FC > 2) were upregulated in ISC group on comparison with the control and ISC/R groups (Table 11). FC values of these 25 genes in the ISC group were distributed around FC > 2 whereas that of control and ISC/R tend towards FC ≤ 1 (Table 11). Similarly, the violin and scatter plots revealed that the distribution of cells was tending to FC > 2 in ISC and FC ≤ 1 in the control and ISC/R groups (Fig. 10–G). SLC16A3 (FC = 8.49) and PRXL2A (FC = 7.02) were highly expressed in ISC cells and the FC of remaining 23 proteins were less than 4 (Table 11). Similarly, 7 genes (FC > 2) were upregulated in the control group on comparison with the ISC (Table 11). The violin and scatter plots revealed that the distribution of these 7 genes was towards FC > 2 in control and FC ≤ 1 in the ISC and ISC/R groups (Fig. 10H–J). The ISG15 (FC = 7.55) exhibited maximum expression whereas the FC of other 6 genes were below 3 (Table 11). Additionally, 30 genes (FC > 2) were upregulated in ISC/R group on comparison with the ISC and control (Table 11). The violin and scatter plots revealed that the distribution of these 30 genes was towards FC ≤ 1 in control and ISC groups (Fig. 10H–M). The Ring Finger protein (RNF121) (FC = 17.16) and the ECM protein Tenascin C (TNC) (FC = 5.22) exhibited maximum expression whereas the FC of other 28 genes were between were below 4 (Table 11). The data suggests that the unique population of cells in ISC group lineaging from the control cells by the downregulation of at least 25 and the upregulation of at least 7 genes in response to ischemia and the ischemic response was elicited by the upregulation of 30 genes as evident in ISC/R group.

Fig. 10
figure 10

Distribution of cell populations and altered genes in Cluster 10: (A) Split view of Control, ISC and ISC/R groups of Cluster 10 cells and combined view of t-SNE plot showing the distribution of cells within the Cluster 10 tending towards the ISC group based on the local expression of 2,546 genes. (B) Violin plot for highly expressed 2,546 genes revealing the distribution towards upregulation in ISC and ISC/R group. (C) Scatterplot of highly expressed 2,546 genes in ISC group showing the upregulation trend compared to control and ISC/R groups. (D) Heat map of highly altered 2,546 genes (FC > 2) revealing the upregulation in ISC and ISC/R groups. (E) Violin plot for highly expressed 25 genes revealing the distribution towards upregulation in ISC group compared to control and ISC/R. (F) Scatterplot of highly expressed 25 genes in ISC group (FC > 2) showing the decreased level of expression in control group where the dotted lines indicate mean FC. (G) Heat map of highly altered 25 genes (FC > 2) revealing the upregulation in ISC group. (H) Violin plot for highly expressed 7 genes revealing the distribution towards upregulation in Control and ISC/R compared to ISC group. (I) Scatterplot of highly expressed 7 genes in Control group showing the minimal level of expression in ISC group where the dotted lines indicate mean FC. (J) Heat map of highly altered 7 genes (FC > 2) revealing the upregulation in Control and ISC/R group. (K) Violin plot for highly expressed 30 genes revealing the distribution towards upregulation in ISC/R compared to Control and ISC groups. (L) Scatterplot of highly expressed 30 genes in ISC/R group showing the minimal level of expression in ISC group where the dotted lines indicate mean FC. (M) Heat map of highly altered 30 genes (FC > 2) revealing the upregulation in ISC/R group

Full size image
Table 11 The list of genes locally upregulated genes (FC > 2) in control, ISC and ISC/R groups with respect to each other group in Cluster 10 cells
Full size table

Cluster 11 Cells Tend Towards Reperfusion

Cluster 11 displayed 4,134 cells where ~ 0.16% cells were mapped in ISC group, ~ 1.26% in control group and ISC/R groups contain ~ 98.58% cells (Fig. 11A, Table 2). The comparison was between ISC/R and control groups and the data suggests the existence of a predominant population of cells in the ISC/R group which is distinct from the control. Interestingly, 4,134 genes were detected locally in cluster 11 where the sorting was based on the extent of gene expression in ISC/R group (Supplementary Table 11, Fig. 11A–D). The violin and scatter plots revealed that the overall distribution of cells was tending to be around FC > 1 ISC/R group and Control group (Fig. 11A–D). Interestingly, 51 genes (FC > 2) were upregulated in ISC/R group on comparison with the control (Table 12). FC values of these 51 genes in the ISC/R group were distributed around FC > 2 whereas that of control tend towards FC ≤ 1 (Table 12). Similarly, the violin and scatter plots revealed that the distribution of cells was tending to FC > 2 in ISC/R and FC ≤ 1 in the control (Fig. 11E–G). RCAN1 (Regulator of Calcineurin 1) (FC = 5.81) was the highly expressed gene in ISC/R cells and the FC of remaining 50 proteins were less than 4 (Table 12). Similarly, 126 genes (FC > 2) were upregulated in the control group on comparison with the ISC/R (Table 12). The violin and scatter plots revealed that the distribution of these 126 genes was towards FC > 2 in control and FC ≤ 1 in the ISC/R group (Fig. 11H–J). The Protein phosphatase 4 regulatory subunit 4 (PPP4R4) (FC = 7.17) exhibited maximum expression followed by the G-protein-coupled neuropeptide receptor, NPFFR2 (Neuropeptide FF receptor 2) (FC = 6.61) whereas the FC for the remaining 124 genes were below 6 (Table 12). The data suggest that the unique population of cells in ISC/R group lineaging from the control cells by the downregulation of at least 51 and the upregulation of at least 126 genes in response to ischemia.

Fig. 11
figure 11

Distribution of cell populations and altered genes in Cluster 11: (A) Split view of Control, ISC and ISC/R groups of Cluster 11 cells and combined view of t-SNE plot showing the distribution of cells within the Cluster 11 tending towards the ISC/R group based on the local expression of 4,134 genes. (B) Violin plot for highly expressed 4,134 genes revealing the distribution towards upregulation in ISC/R group. (C) Scatterplot of highly expressed 4,134 genes in ISC/R group showing the upregulation trend compared to control. (D) Heat map of highly altered 4,134 genes (FC > 2) revealing the upregulation in ISC/R groups. (E) Violin plot for highly expressed 51 genes revealing the distribution towards upregulation in ISC/R group compared to control. (F) Scatterplot of highly expressed 51 genes in ISC/R group (FC > 2) showing the decreased level of expression in control group where the dotted lines indicate mean FC. (G) Heat map of highly altered 51 genes (FC > 2) revealing the upregulation in ISC/R group. (H) Violin plot for highly expressed 126 genes revealing the distribution towards upregulation in Control compared to ISC/R group. (I) Scatterplot of highly expressed 126 genes in Control group showing the minimal level of expression in ISC/R group where the dotted lines indicate mean FC. (J) Heat map of highly altered 126 genes (FC > 2) revealing the upregulation in Control compared to ISC/R group

Full size image
Table 12 The list of genes locally upregulated genes (FC > 2) in control, and ISC/R groups with respect to each other group in Cluster 10 cells
Full size table

Discussion

Despite the heterogeneous population of MSCs, the passaging of subpopulations results in clonal expansion and alterations in their stemness hurdling their regenerative function. Hence, the thorough understanding regarding the heterogeneous nature of MSCs is necessary for harvesting/expanding unique subpopulations to sustain their regenerative potential for ensuring reproducibility of the outcomes [13]. Interestingly, scRNAseq offers a powerful tool to dissect and quantify the heterogeneous subpopulation by assessing the gene expression patterns in individual cells [14]. The scarcity of relevant information regarding cardiac regeneration and immense translational potential of EATDS encouraged us to explore the existing subpopulations based on single cell genomics. Excitingly, 18 unique clusters of cells were unveiled in our study which were further screened by the upregulation status of signature genes in each cluster based on the treatment (Control, ISC and ISC/R). Ischemia being the primary trigger for myocardial infarction, our major focus was emphasized on the upregulated genes in the ISC group (considering the number of cells, > 2%) in each cluster despite the bulk of data.

Cardiac ischemia activates the asymmetric division of diverse stem cell population resulting in the progenitors/precursors cells to cardiac lineage. Hence, the adverse events underlying the pathology stimulates the stem cell activation as a protective mechanism [15]. EF being at the close proximity and sharing same micro niche as cardiac muscle, the ischemic insults possibly activate the EATDS; however, the underlying mechanism is scarce [16,17,18]. As expected, the UMAP analysis revealed a distinct cluster of EATDS in ISC group where the reperfusion tends the cells towards the control group suggesting the existence of a unique population of EATDS in response to ischemia. Overall, the genes associated with tissue regeneration were downregulated/expressed constitutively in the control cells which were significantly upregulated under ischemic insults and reperfusion. IL-11 is a key mediator in promoting cell proliferation and differentiation of ADMSCs and IL-11 signaling has been involved in antioxidant responses [19]. Apart from maintaining ECM integrity, TNC plays a significant role in tissue regeneration by facilitating the mobilization of stem cells from their respective niche [20]. Similarly, PRXL2A is intimately involved in redox signaling offering the protection from oxidative injury [21]. The downregulation of IL-11, TNC and PRXL2A in the control cells suggest the basal level of these genes reflecting the cellular homeostasis. Interestingly, immune responsive mediator ISG15, the antiviral gatekeeper gene MX1 [22], and the metabolic and redox sensor DHRS3 [23] were upregulated in the control contributing to the cellular homeostasis. Moreover, PRXL2A was significantly upregulated in ISC EATDS along with the antioxidant responsive enzyme ALDH1A1 with a concomitant downregulation of cell cycle mediators. This points that the EATDS under ischemia prefers non-proliferative phenotypes. Importantly, upon reperfusion the cellular homeostasis and healing mediators, (downregulated in control cells) and cell cycle mediators (downregulated in ISC cells) were significantly upregulated with a concomitant downregulation of redox and antioxidant signals vividly suggesting the switch of ISC cells to proliferative and healing phenotype. Evidently, the highly upregulated gene in ISC cells, PRXL2A, is a potent cardioprotective mediator which plays a key role in antioxidant response, angiogenesis, and post-ischemic inflammation via the activation of TLR4 signaling [24]. Importantly, PRXL2A is crucial for the activation, maintenance and proliferation of MSCs through the activation of the pathways including AKT/glycogen synthase kinase 3 beta/β-catenin axis and peroxiredoxin/JNK axis [24, 25] suggesting its pivotal role in the activation and proliferation of an unique subtype of EATDS following ischemia.

Even though the Cluster 1 cells favored reperfusion, two genes H1-0 and MGP were significantly upregulated in the ISC group. Interestingly, H1-0 is a key epigenetic regulator which is intimately associated with the activation, differentiation, and plasticity of stem cells [26, 27]. H1-0 drives the heterogeneity of cancer stem cells responding to the biochemical alterations in the micro niche [28]; however, information is unavailable regarding its role in EATDS. Therefore, it is logical that similar mode of operation is possible in EATDS as evident from the upregulation of H1-0 in the ISC group with a concomitant decrease following the reperfusion; however, warranting further examination. Similarly, MGP is an effective inhibitor for various BMPs (bone morphogenic proteins) and prevents the vascular calcification and angiogenesis [29]. Interestingly, MGP positive MSCs displayed protective effects on atherosclerosis and angiostenosis eliciting immunomodulatory responses [30]. Moreover, MGP is actively involved in the maintenance of myocardial homeostasis and cardiac performance where its role in EATDS is unknown [31]. The increased level of MGP in Cluster 1 cells suggests the protective responses elicited by EATDS on encountering ischemia which was declined following reperfusion. Taken together, the ischemic insults result in a subpopulation of cells by upregulating the key mediators including H1-0 and MGP and further understanding of this subpopulation are required for regenerative cardiology.

Cluster 2 and 3 cells predominated in control group where the genes CDC20 and CCN5 were significantly upregulated in the ISC groups of Cluster 2 and 3, respectively. CDC20 is the cell cycle regulator which is crucial in determining the viability and proliferation status of cells [32]. The increased CDC20 reflects the actively proliferating cells despite the ischemia suggesting the adaptations of Cluster 2 cells to withstand the deleterious effects by increasing the cell density. Similarly, the matricellular protein CCN5 regulates the cellular integrity especially in responding to alterations in microenvironment [33]. Hence, the ischemic cells in the clusters 2 and 3 were survived respectively by increasing the cell cycle and maintaining the cellular integrity. Hence, it is reasonable to speculate that these subpopulations of cells exist in lower density acting as progenitors for specific cell lineage aiding in biological responses in cardiac tissue. Additionally, the exact functions of these EATDS in cardiac regeneration warrant further investigation.

Ischemia favored cluster 4 cells in which the downregulation of the major fibroblast differentiation protein, NREP was evident. NREP was highly upregulated in the control cells suggesting that the cluster 4 cells represent a predominant population of EATDS which aids in the prevention of fibrosis. Evidently, NREP expression was significantly increased in hypertrophic scars and its role in scarring mediated through TGF-β-Smad axis has been unveiled [34, 35]. Importantly, NREP accelerates the transdifferentiation of stem cells to fibroblasts [34]. However, the information of NREP on EATDS differentiation and cardiac fibrosis is limited. Additionally, the decreased level of NREP in EATDS under ischemic environment suggests the existence of an antifibrotic subpopulation which maintains/preserves the stemness and possibly exhibiting cardioprotective functions. In addition, the cluster 6 cells favoring ischemia upregulated the cardiac muscle biomarker MYH11 suggesting a progenitor subpopulation differentiating towards myocardial lineage. Interestingly, MYH11 has been reported to be a crucial biomarker for adult cardiac precursor cells which are destined to be cardiomyocytes by the activation of NOTCH signaling [36].

The cluster 5 cells favoring reperfusion upregulated IL-11, which is a crucial anti-inflammatory cytokine which promotes the activation, proliferation, differentiation and commitment of various progenitor cells [37]. A recent study reported that IL-11 promoted the engraftment, maintenance, survival, and differentiation of ADMSCs in ischemic tissues improving their therapeutic efficiency [19]. Interestingly, another seminal study reported the critical role of IL-11 in tissue regeneration [38] and the attenuation of cardiac fibroblasts following MI through IL-11/glycoprotein 130/STAT3 axis [39]. Hence, the cluster 5 represents a subpopulation of EATDS which favors the prevention of inflammation and attenuation of cardiac fibrosis following the reperfusion; however, further validations are warranted. Similarly, cluster 7 cells favoring reperfusion revealed the upregulation of IL-11 suggesting a closely related subpopulation of EATDS. The unique functions of these two subpopulations require further research for their application in translational cardiology.

EATDS in the clusters 8 and 9 tending towards ischemia upregulated HSP90B1 and CRELD2, respectively. HSP90B1 has been identified as chaperone for several Toll-like receptors (TLRs) playing crucial role in inflammatory responses and TLR mediated apoptosis following renal ischemia [40, 41]. In addition, CRELD2 regulates activating transcription factor 6 (Atf6) to induce ER-stress resulting in decreased protein translation and increased degradation [42, 43]. However, the potential role of HSP90B1 and CRELD2 in EATDS and their association with ischemic cardiac tissues are yet to be discovered. Similarly, cluster 10 cells favored ischemia with an upregulation of the glycolytic component SLC16A3. Interestingly, SLC16A3 is intimately involved in stem-cell associated hypoxic/ischemic microenvironment especially activating HIF1α signaling [44]. Importantly, SLC16A3 plays a crucial role in lactate extrusion and facilitates AMPK phosphorylation eliciting protective responses during cardiac ischemia and myocardial injury [45]. Hence, the cluster 10 subpopulation of EATDS are equipped to withstand and to survive the extreme ischemic insults in the myocardium during/following an injury.

Cluster 11 cells favoring the reperfusion displayed the upregulation of RCAN1 which is a negative regulator for TLR-MyD88-NF-κB signaling mediates through IκBɑ [46]. Interestingly, RCAN1 elicits protective role against ischemic insults in multiple tissue types including myocardium and brain [47, 48]. The upregulation of RCAN1 in the reperfused cells suggests the protective mechanisms elicited by cluster 11 subpopulation by preventing TLR-driven aggravated inflammation. The regulatory role of RCAN1 in EATDS are largely unknown; however, the identification of RCAN1 + population unveils the possible trans-acting cardioprotective stem cell population.

It has been believed that the epicardium retains the embryonic program to generate mesenchymal progenitor cells secreting paracrine factors to stimulate the growth and survival of cardiomyocytes and promote angiogenesis for repair and regenerative responses. Hence, the epicardium has been considered as a signal generator for cardiac regeneration [49]. Despite a few recent reports [49, 50], minimal information is available regarding the cellular heterogeneity of epicardial tissue which points the necessity of mapping the subpopulations of epicardial cells. Unfortunately, the information regarding the heterogeneity of EF and EATDS is unavailable in the literature. EF being a repository of translationally worthwhile EATDS for cardiac regeneration, it is relevant to explore the cellular heterogeneity to identify the ideal stem cell subpopulations for improved cardiac healing. As hyperlipidemia is integral to atherosclerosis and IHD, we utilized atherosclerotic Yucatan micro swine model to harvest EATDS which was manipulated in vitro to simulate ischemia and reperfusion. To the best of our knowledge, this is the first report dealing with the heterogeneity of EATDS based on ischemia and reperfusion unveiling 18 unique clusters.

The overall findings regarding the characteristic subpopulations are summarized in Table 13. The outcomes of this study are novel, encouraging potential avenues for phenotyping EATDS for cardiac regeneration. Also, the easiness of harvesting and possibilities of expansion encourage the application of EATDS in regenerative cardiology. The major focus of this study was to map the unique cell clusters using scRNAseq representing the major EATDS subpopulations and future research is warranted to further characterize these populations for cardiac applications. In addition, the bulk of the obtained data drive us to focus on the upregulated genes based on the treatment status (ischemia or reperfusion) and further emphasis was given only to the highly upregulated genes for deriving the conclusions. However, the data regarding the downregulated genes warrant similar assessment for designing the panel of biomarkers for defining each unique cluster. Importantly, atherosclerotic swine (n = 3) was used for ensuring the clinical relevance and the cells were pooled for post-isolation culture, simulation of ischemia and reperfusion, and scRNAseq analysis.

Table 13 Overview of the EATDS phenotypes identified in the clusters based on their differential gene expression and possible mechanism of action (based on published literature)
Full size table

The untreated EATDS harvested from hyperlipidemic swine were used control for ensuring the alteration of gene expression was solely due to ischemia and reperfusion. The data from normal swine is warranted to investigate the effect of hyperlipidemia/atherosclerosis in EATDS heterogeneity. Additionally, the analysis was based on the number of cells in each cluster and 8 clusters displaying < 1000 cells were exempted from screening which need further analysis. Furthermore, the validation of the gene expression using qRT-PCR is limited as the major focus of this study was to identify EATDS heterogeneity by screening the cell clusters employing the sc-RNAseq analysis. However, further investigations and validations are warranted to confirm the gene expression profiles in the clusters/sub-populations of interest. Additionally, Loupe Browser was solely used for the analysis and batch correction was not performed. However, batch variation has minimal impact as all the samples utilized same 10 × chemistry for data generation and followed identical sample preparation protocol. Moreover, the present study focused on differential gene expression to screen the cellular phenotypes where the batch effect is negligible. However, future data analysis warrants batch correction and utilization of computational and statistical programs. Moreover, the gene interactions and prediction/assessment of regenerative pathways using bioinformatics tools using the findings from this study pose additional translational relevance.

To the best of our knowledge, this is the first study unveiling EATDS heterogeneity in hyperlipidemic microswine using single cell genomics. Overall, the study unveiled the existence of subpopulations of EATDS unveiling the regenerative phenotypes and the identification of similar regenerative phenotypes in human system may open novel avenues for MI management. A seminal study demonstrated the conserved cardiac fibroblast subpopulation between porcine and human heart following MI suggesting the possibilities of the co-existence of similar subpopulation of EATDS in human system [51]. Unfortunately, the information regarding such human counterpart of EATDS are unavailable in the literature; however, warrants further investigations for the potential therapeutic interventions. Nevertheless, our findings revealed unique phenotypes of heterogeneous EATDS population which open potent translational avenues for regenerative cardiology in the management of myocardial damage.

Conclusions

The scRNAseq of EATDS under simulated ischemia and reperfusion revealed 18 unique cell clusters suggesting the existence of heterogeneous phenotypes. Interestingly, the functions of the key upregulated genes screened in the treatment group revealed the cardioprotective phenotypes which elicits diverse mechanism of action including epigenetic regulation of differentiation pathways (Cluster 1), maintenance of myocardial homeostasis (Cluster 1), maintenance of cell integrity and cell cycle (Clusters 2 and 3), prevention of fibroblast differentiation (Cluster 4), differentiation to myocardial lineage (Cluster 6), anti-inflammatory responses (Clusters 5, 8, and 11), prevention of ER-stress (Cluster 9), and increasing the energy metabolism (Cluster 10). These unique phenotypes of heterogeneous EATDS population warrant further characterization for clinical applications; however, open significant translational opportunities for myocardial regeneration and cardiac management.