A multi-omics approach to identify molecular alterations in a mouse model of heart failure

Figure 1. Clustering and annotation of single cells in control and heart failure (HF) mice. (A) Initial clustering of 23033 single cells from mouse heart tissue in both control and HF groups visualized in UMAP dimension reduction. (B) Overlapping of cells from HF and control groups visualized in UMAP dimension reduction. (C) Heatmap of prediction scores matching each cluster of the current dataset to published mouse single-cell profile. (D) Verifying the cell type annotation by visualization of known markers. (E) A UMAP dimension-reduced view of annotated major cell types. (F) The top differentially expressed genes for each cell type visualized in heatmap. EC: endothelial cell; FB: fibroblast; BC: B cell; TC: T cell; CM: cardiomyocyte.

Figure 2. The changes in the immune cells between control and heart failure (HF) mice. (A) The proportion of each cell type in control and HF groups. (B, C) Immunofluorescent staining and quantitative analysis for CD74 and Coro1a in control and HF groups. (n = 5, *p < 0.05). EC: endothelial cell; FB: fibroblast; BC: B cell; TC: T cell; CM: cardiomyocyte.

Figure 3. Differentially expressed miRNAs and corresponding downstream targets. (A) Volcano plot of differentially expressed miRNAs in heart failure (HF) and control mice (p < 0.05 and absolute log2FC ≥1). (B) The downstream targets of above miRNAs overlapped with the differentially expressed genes identified from scRNA-seq analysis.

Figure 4. The miR-3473a/Rora regulatory pathway in cardiac fibroblasts. (A) Differentially expressed genes in fibroblast cells overlapping with the potential targets from miRNA profiling between control and heart failure (HF) groups, visualized in violin plot. (B) Immunofluorescent staining and quantitative analysis of Rora in control and HF mice (n = 5, *p < 0.05). Col1a1 was used to label fibroblast cells. (C) FISH experiments were performed to confirm the interaction between miR-3473a and Rora in mouse cardiac fibroblasts treated with Ang II (1 μg/ml).

Figure 5. ATAC-seq analysis and the accessible chromatin regions. (A, B) TSS enrichment of the ATAC-seq for open-chromatin regions. Genes with accessible chromatin regions from control and heart failure (HF) mice were first determined by the consensus identification. The TSS regions were then merged and plotted among samples of each group. (C) Common and different accessible chromatin regions were identified by ATAC-seq in control and HF mice. (D, E) Annotated regions and distribution of the accessible chromatin regions in control and HF mice were clustered around the promoter regions.

Figure 6. ATAC-seq revealed gene expression changes in failing hearts. (A) Common and different accessible promoter regions were identified by ATAC-seq in control and heart failure (HF) mice. (B) Gene ontology analysis revealed enriched pathways for the genes with open promoters in HF group but not in control group (yellow), and vice versa (blue). (C) HF samples have enhanced ATAC-seq signals at the promoter regions of Tnni3, Btg2 and Nr4a1.

Figure 7. scATAC-seq profiling of heart tissue in control and heart failure (HF) mice. (A) Five major cell clusters were identified by scATAC-seq in myocardial tissue. (B) UMAP embedding showing the cells from control and HF groups. (C) UMAP embedding showing the gene scores of cell type-specific marker genes. (D) Heatmap plot of cluster specific marker genes. (E) Enriched motifs of the chromatin open regions for the major cell clusters.

Figure 8. Gene expression changes in endothelial cells (ECs) of failing hearts. (A) Bar plot showing the cell numbers of different cell populations in control and heart failure (HF) mice. (B) Volcano plot showing the differentially expressed genes in ECs. (C) Gene ontology analysis of the up-regulated genes in ECs and the gene scores of Nos3 in control and HF groups.