Single-nucleus RNA sequencing in ischemic cardiomyopathy reveals common transcriptional profile underlying end-stage heart failure

Abstract

Ischemic cardiomyopathy (ICM) is the leading cause of heart failure worldwide, yet the cellular and molecular signature of this disease is largely unclear. Using single-nucleus RNA sequencing (snRNA-seq) and integrated computational analyses, we profile the transcriptomes of over 99,000 human cardiac nuclei from the non-infarct region of the left ventricle of 7 ICM transplant recipients and 8 non-failing (NF) controls. We find the cellular composition of the ischemic heart is significantly altered, with decreased cardiomyocytes and increased proportions of lymphatic, angiogenic, and arterial endothelial cells in patients with ICM. We show that there is increased LAMININ signaling from endothelial cells to other cell types in ICM compared with NF. Finally, we find that the transcriptional changes that occur in ICM are similar to those in hypertrophic and dilated cardiomyopathies and that the mining of these combined datasets can identify druggable genes that could be used to target end-stage heart failure.

Keywords: CP: Molecular biology; dilated cardiomyopathy; druggable genome; heart failure; hypertrophic cardiomyopathy; ischemic cardiomyopathy; single-nucleus RNA sequencing.

Figure 1.. Cellular composition of non-failing and failing ischemic cardiomyopathy

(A) UMAP representation of 99,684 nuclei from 8 NF samples and 7 ICM samples showing 16 cell clusters identified with unsupervised Leiden clustering. VSMC, vascular smooth muscle cell. (B) Cell-type composition of each individual sample, with colors representing cell types labeled in (A). (C) Changes in composition observed between NF and ICM samples. Statistically credible changes are noted with an * (see STAR Methods). Boxplots are represented as: center line, median; box limits, upper and lower quartiles; whiskers, 1.5× interquartile range; points, outliers. (D) Number of significantly differentially expressed genes in each cell type in ICM compared with NF. Upregulated genes are shown in red and downregulated genes are shown in blue. See also Figures S1–S3 and Tables S1, S2, S3, S4, and S5.

Figure 2.. Subclustering of endothelial cells identifies seven distinct subclusters

(A) UMAP of 18,636 endothelial cell identifies 7 subclusters (left) represented across both disease states (right). EC, endothelial cell. (B) Dot plot of the top marker genes of each endothelial cell subcluster. The size of the dot represents the percent of nuclei expressing the gene at non-zero levels (Pct Expr > 0) and the shading represents the average log-normalized expression (Avg logExpr) of the gene. (C) Breakdown of endothelial cell composition for each sample relative to the entire composition of all cell types. Subclusters are colored as in (A). (D) Compositional changes in subclusters between ischemic cardiomyopathy and non-failing samples. Statistically credible increases in EC-lymphatic, EC-angiogenic, and EC-arterial cells were observed in ischemic cardiomyopathy and denoted with an *. Boxplots are represented as: center line, median; box limits, upper and lower quartiles; whiskers, 1.5× interquartile range; points, outliers. (E) Representative in situ hybridization images of EC-lymphatic cells probed with CCL21 (blue) alongside trichrome staining of a consecutive section and a cropped close up of an area with high cell numbers. Nuclei are stained with hematoxylin. Scale bars, 1 mm and 250 μm (on close up images). n = 7 (NF) and 6 (ICM). (F) Representative in situ hybridization images of EC-angiogenic cells probed with PECAM1 (red) and MYO1B (blue). Boxes show where close up images were taken and arrows identify cells stained with both PECAM1 and MYOB1. Scale bars, 1 mm and 50 μm (on close up images). Nuclei are stained with hematoxylin (blue). n = 7 (NF) and 6 (ICM). See also Figures S4 and S5 and Tables S6 and S7.

Figure 3.. Top differentially expressed protein coding genes in endothelial cell subclusters

Top differentially expressed protein coding genes for each endothelial subcluster. The log fold change (shading) and p value (size) for each gene comparing ICM with NF within each endothelial subcluster is shown on the left. The relative expression in each endothelial subcluster separated by ICM and NF is shown on the right with the size of the dot representing the percent of nuclei expressing the gene and the shade representing the average log-normalized expression. Genes were selected as those expressed in at least 30% of nuclei from either the non-failing or ischemic cardiomyopathy groups with logFC > 1, FDR-adjusted p value < 0.05, and a low probability of being derived from background contamination. NF, non-failing; ICM, ischemic cardiomyopathy; EC, endothelial cell; logFC, log fold change comparing ICM with NF; Pct Expr, percent of nuclei expressing at non-zero levels; Avg logExpr, average log-normalized expression. See also Table S8.

Figure 4.. CellChat analysis of cell-cell communication between endothelial cells and other cell types

(A) Inferred number of interactions of each subcluster of endothelial cells with all cell types in NF and ICM. (B) Differential number of interactions of each subcluster of endothelial cell with each cell type, distinguishing between sender and receiver groups. Prof. Mφ, proliferating macrophage. (C) Relative information flow of signaling from endothelial cells to all cell types within NF or ICM samples. (D) Chord diagram demonstrating the LAMININ signaling pathway within endothelial cell types and all other cell types, highlighting the increase in signaling in ICM compared with NF. The thickness of each connection represents the strength of communication from the origin cell type to the receiver cell type. (E) Expression of ligand and receptors involved in LAMININ signaling in endothelial cell subclusters and all other cell clusters. The size of each dot represents the percent of nuclei expression the gene and the shading represents the average normalized expression. Red boxes represent changes between NF and ICM that are significant (p < 0.01).

Figure 5.. snRNA-seq data on nearly 700,000 nuclei from end-stage cardiomyopathies

(A) Joint UMAP representation of nuclei from hypertrophic cardiomyopathy (HCM), (n_patients = 15), dilated cardiomyopathy (DCM) (n_patients = 11), ischemic cardiomyopathy (ICM) (n_patients = 7), and non-failing (NF) (n_patients = 24) samples (n = 692,373) colored by cell-type clusters. (B) Principal-component (PC) analysis of pseudo-bulk expression of all samples (n = 57) as calculated by summing expression across all nuclei in each patient. Colors represent the disease state and shape represents the patient sex. (C) Flow chart outlining filtering steps carried out to identify protein coding genes that were significantly differentially expressed between any cardiomyopathy and NF, as well as between pairs of cardiomyopathies. DEG, differentially expressed gene; FDR, false discovery rate. (D) Heatmap of effect size estimates for protein coding genes that were significantly differentially expressed between ICM and NF, as well as ICM and DCM or ICM and HCM (FDR-adjusted p value < 0.05, expressed in >1% of nuclei from either group, low background probability). Also included are genes that were significantly differentially expressed between DCM or HCM and NF, as well as between DCM or HCM and ICM. Shading represents the log fold change (logFC) estimates for each patient population comparison and significant changes are denoted with a dot. VSMC, vascular smooth muscle cell. (E) Reactome pathways with significant enrichment for genes that were up- or downregulated in all three cardiomyopathies. Multiple testing correction was applied across all cell types/direction combinations using a Benjamini-Hochberg correction (FDR < 0.05, dotted line). Bar colors indicate the cell type in which the significant enrichment was identified, colored as in (A). See also Figure S7.

Figure 6.. Druggable genes shared by ischemic, hypertrophic, and dilated cardiomyopathies

(A) Flow chart of filtering steps carried out to identify cell-specific druggable genes that are up- or downregulated in ischemic, hypertrophic, and dilated cardiomyopathies. DEG, differentially expressed gene; FDR, false discovery rate; ICM, ischemic cardiomyopathy; NF, non-failing; DCM, dilated cardiomyopathy; HCM, hypertrophic cardiomyopathy; logFC, log fold change; AUC, area under the receiver operating characteristic curve. (B) Upregulated cell-type-specific druggable genes, with druggable tier noted below. (C) Downregulated cell-type-specific druggable genes, with druggable tier noted below. Boxplots are represented as: center line, median; box limits, upper and lower quartiles; whiskers, 1.5× interquartile range; points, outliers. CPM, count per million; EC, endothelial cell; VSMC, vascular smooth muscle cell. See also Tables S10 and S11.