3.1 Preprocessing results of ScRNA seq dataset
After rigorous quality control, a total of 33,300 cells and 20,402 genes were selected for subsequent analysis. From these, we chose 5,000 genes with the most significant expression changes (Fig. 2A) for principal component analysis, and subsequently annotated them using the SingleR package.After processing through the standard procedure (Fig. 2B), a total of 10 cell types were identified (Fig. 2C), and the composition proportion of each cell type was further explored (Fig. 2D), including cardiomyocytes, endothelial cells, fibroblasts, myofibroblasts, macrophages, granulocytes, dendritic cells, monocytes, T cells, and B cells. By comparing the cellular composition among samples, we found that the proportion of macrophages and fibroblasts is higher in the MIRI group(Fig. 2E and F). Simultaneously, we focus on the transformation of macrophages under pathological conditions and use volcano plots to display the differential genes between macrophages in MIRI and those in the sham group (Fig. 3A).
A) Volcano plot showing the differential genes of macrophages in MIRI.B) Analysis of the trajectory from monocytes to Ccr2 + macrophages using Monocle.C) Expression changes of M1 and M2 macrophage-related genes during differentiation.D) Enrichment scoring results for different gene sets in each macrophage subset by ssGSEA/singscore/UCell/AUCell.
3.2 Differentiation trajectory of Ccr2 + macrophages
Following acute injury, Ccr2 + macrophages are almost entirely derived from monocytes. Thus, we used Monocle2 to analyze the differentiation process from monocytes to Ccr2 + macrophages (Fig. 3B) to investigate whether there was a phenotypic transition in Ccr2 + macrophages. The results indicated a gradual decrease in the expression of M1 macrophage-related genes(obtained from the official CellMarker website[38, 39]) such as Nos2 and Il1b during the differentiation of monocytes to Ccr2 + macrophages, while genes related to M2 macrophages[38, 39] began to be expressed (Fig. 3C). Furthermore, GOBP enrichment analysis of differential genes in pseudotime (Supplementary Fig. 1) revealed that during differentiation, genes related to lipid metabolism, signal transduction, and clearance of apoptotic cells were upregulated, whereas genes related to the negative regulation of T-cell proliferation and the complement receptor pathway were downregulated.
3.3 Functional Analysis of Macrophage Subgroups
AUCell, UCell, singscore, and ssgsea were used to score each cell, resulting in different enrichment scoring matrices. Subsequently, irGSEA calculated the differentially expressed gene sets in each cell subgroup within the various enrichment scoring matrices through the Wilcoxon test. An "up" or "down" indicates that the enrichment level of the differential gene set within the cell cluster is higher or lower than that in other clusters (Fig. 3D). To reduce bias from different statistical scoring methods, we focused only on the functional gene sets that were significantly enriched in Ccr2 + macrophages across different scores(The score result of AUcell is consistent with UCell). Ccr2 + macrophages showed higher enrichment in the areas of cytokine interaction, matrix remodeling, and angiogenesis (Fig. 3D) compared to other clusters (p<0.001).
Through pseudotime analysis involving changes in markers and the enrichment of functional gene sets, it was discovered that Ccr2 + macrophages undergo polarization towards the M2 phenotype. Interestingly, the significant enrichment of cytokine interaction gene sets within this subgroup indirectly suggests that they may play a crucial role in intercellular communication. Following this, the study delves into analyzing the intercellular communication associated with Ccr2 + macrophages.
A) Changes in the number and intensity of intercellular communications in MIRI. B) Changes in the number and intensity of interactions between cells in MIRI, with blue indicating a decrease and red indicating an increase in the number/intensity of communications. C)Differences in the output/input intensity of intercellular signaling pathways of Ccr2 + macrophages, fibroblasts, and myofibroblasts after ischemia-reperfusion compared to normal (sham surgery) conditions. D) Comparison of the intensity of signal pathways output by Ccr2 + macrophages to fibroblasts and myofibroblasts under MIRI conditions. E) Comparison of the interaction strength between Ccr2 + macrophages and various types of cells under MIRI conditions.
3.4 Ccr2 + Macrophage-Mediated Intercellular Communication
Through the analysis of CellChat V2, we found that the intensity of intercellular communication significantly increased in the MIRI group compared to the sham group (Fig. 4A), with the change in communication intensity of myofibroblasts being the most significant (Fig. 4B). By analyzing the input and output signal strengths of various pathways for Ccr2 + macrophages, fibroblasts, and myofibroblasts under different conditions of MIRI and sham, it is evident that the Spp1 signaling pathway is the pathway with the highest output intensity for Ccr2 + macrophages, and also the pathway with the highest input intensity for fibroblasts. After the activation of fibroblasts into myofibroblasts, the Spp1 pathway continues to play a significant role (Fig. 4C). We then compared the various signals output by Ccr2 + macrophages to fibroblasts and myofibroblasts (Fig. 4D), and the results indicated that Spp1 remains the signal with the largest output weight. Ccr2 + macrophages not only interact closely
with fibroblasts but also have strong interactions with various types of immune cells (Fig. 4E). We categorized the immune cells into myeloid (Fig. 5A) and lymphoid series (Fig. 5B) and displayed the various ligands originating from Ccr2 + macrophages for each category. The results showed that apart from B cells primarily receiving APP signals from Ccr2 + macrophages, among the various signals received by T cells and other myeloid immune cells from Ccr2 + macrophages, the Spp1 signaling pathway has the highest weight.
A) Comparative analysis of signaling pathway intensity from Ccr2 + macrophages to myeloid immune cells under MIRI conditions.B) Comparative analysis of signaling pathway intensity from Ccr2 + macrophages to lymphoid immune cells under MIRI conditions.C) Comparison of output and input intensity of the Spp1 signaling pathway in different cell types under MIRI conditions.D) Identification of emitters, receivers, mediators, and affectees of the Spp1 signaling pathway under MIRI conditions.E) Comparison of autocrine and paracrine intensity of Spp1 in non-immune and immune cells: left figure shows the autocrine intensity in non-immune cells and paracrine intensity in immune cells, right figure shows the autocrine intensity in immune cells and paracrine intensity in non-immune cells.
3.5 Spp1 Signaling Pathway
By comparing the cell communication between sham and MIRI, it was found that the SPP1 pathway is the pathway with the greatest statistical difference between the two groups (Supplementary Fig. 2). In addition to Ccr2 + macrophages producing Spp1, T cells and dendritic cells are also important sources of Spp1 (Fig. 5C). Through the identification of the Spp1 signaling pathway network, we learned that Ccr2 + macrophages are the most important senders of Spp1. Besides autocrine signaling, myofibroblasts are the most important receptors of Spp1. T cells and dendritic cells, like Ccr2 + macrophages, also participate in both the output and reception of Spp1. Almost all cell types, except B cells, are influenced by this signaling pathway (Fig. 5D). By comparing the autocrine and paracrine analysis of Spp1 in immune cells and non-immune cells (Fig. 5E), we can observe that the interactions in the Spp1 signaling pathway are more significant in immune cells than in non-immune cells. Spp1 not only serves as a signal to recruit fibroblasts but also plays an important role in the communication among various immune cells, acting as a mediator for the communication of immune cells.
A) Ccr2 + macrophages exert biological effects on different target cells by outputting Spp1 to their receptors. B) The contribution of the expression strength of each receptor to the Spp1 signaling pathway. C) When comparing samples under MIRI conditions with the control group (sham group) through GSEA (gene set enrichment analysis), it was found that the ECM (extracellular matrix) interaction pathway in fibroblasts was significantly upregulated (P < 0.05).D) When comparing samples under MIRI conditions with the control group through GSEA, it was found that the PI3K Akt pathway in fibroblasts was significantly upregulated (P < 0.05).
After myocardial ischemia-reperfusion, Spp1 mediates changes in different pathways of target cells by acting on different receptors on the surface of target cells (Fig. 6A). Each receptor contributes differently to the pathways, with CD44, mainly expressed on the surface of myeloid immune cells, ranking at the top (Fig. 6B). Lymphoid immune cells and non-immune cells recognize Spp1 through the expression of integrin complex receptors. GSEA results indicate that the extracellular matrix interaction pathways of fibroblasts are upregulated after myocardial ischemia-reperfusion (p < 0.05)(Fig. 6C). The activation of integrin receptors is a crucial upstream event in the activation of the intracellular PI3K-Akt signaling pathway[13]. By interacting with integrin receptors on the surface of fibroblasts, Spp1 may promote the activation of the PI3K-Akt pathway within fibroblasts(Fig. 6D), thereby triggering key biological responses such as adhesion and migration of fibroblasts.
A) Parameter selection in hdWGCNA.B) Co-expression modules of Ccr2 + macrophage-like cells.C) Analysis of disease-related modules in Ccr2 + macrophage-like cells.D) Intersection of hub genes in disease-related modules, Ccr2 + macrophage marker genes, and differential genes in bulk RNA-seq to obtain key genes.E) Further screening of molecular markers through Lasso regression, Support Vector Machine, and Random Forest.F) Importance ranking of four key genes in the Support Vector Machine
3.6 Identification of Molecular Markers
Through hdWGCNA analysis (Fig. 7A and B), we obtained MIRI-related modules (Fig. 7C) and 764 hub genes within the modules. We intersected the 20 marker genes of Ccr2 + macrophages calculated through Findallmarker with the 764 hub genes to identify markers highly related to MIRI, resulting in 10 genes (Fig. 7D). To further filter out genes that are not only significant at the single-cell level but also at the tissue level, we intersected the marker genes with differential genes from bulk RNAseq processed using the limma package (log2FC > 1, adj.p < 0.05), resulting in four genes: Spp1, Ctss, Fn1, and Col1a1. Subsequently, we employed Lasso, SVM, and RF machine learning methods on the bulk RNAseq expression matrix of these four genes to identify the most critical markers (Fig. 7E), selecting spp1 as the marker distinguishing between Ccr2 + macrophages in steady-state and post-myocardial ischemia-reperfusion. The analysis of feature importance using SVM reveals the ranking of importance for Spp1, Ctss, Fn1, and Col1a1 (Fig. 7F),.
In addition, we performed GOBP enrichment analysis on the hub genes of the three MIRI-related modules in Ccr2 + macrophages. The enrichment results indicated that the turquoise module is related to the regulation of ERK1 and ERK2 cascades, leukocyte proliferation, and catabolic metabolism. The blue module is associated with leukocyte migration, lipid metabolism, and Ras protein signal transduction. The yellow module is related to wound healing. (Supplementary Table 2)
A) T-test results based on gene sets GSE168160 (n = 4), GSE130217 (n = 6), and GSE210494 (n = 3) demonstrate that Spp1 is significantly upregulated in MIRI (P < 0.05).B) Analysis of variance (ANOVA) results from the GSE194000 dataset (n = 3) indicate a gradually increasing trend in the expression of Spp1 during the early stages of reperfusion injury.C, D) Pseudotime analysis confirms that the expression of Spp1 gradually increases during the differentiation process of Ccr2 + macrophages.E) Immune infiltration analysis on GSE168160 confirms that the expression of the M2 macrophage gene set is significantly upregulated under MIRI conditions (P < 0.05).
3.7 External Validation through Other Datasets
To validate the representativeness of the molecular markers calculated through the combination of ScRNAseq GSE227088 and bulk RNAseq GSE130217, we processed three bulk RNAseq datasets, GSE168610, GSE210494, and GSE194000, using the limma package and conducted external validation. The results confirmed that, compared to the sham group, the expression of spp1 significantly increased in the MIRI group of the other three datasets (P < 0.05) (Fig. 8A and B). Subsequently, we also verified in the pseudotime analysis of Ccr2 + macrophages (Fig. 8C) that the expression of Spp1 significantly increases during the differentiation process from monocytes to this subgroup (P < 0.05) (Fig. 8D). Finally, through immune infiltration analysis of GSE168610, the results show that the expression of markers related to M2-type macrophages significantly increases 24 hours after myocardial ischemia-reperfusion in the mouse model (Fig. 8E).