Single-nucleus transcriptomics reveal cardiac cell type-specific diversification in metabolic disease transgenic pigs

Abstract

Cardiac damage is widely present in patients with metabolic diseases, but the exact pathophysiological mechanisms involved remain unclear. The porcine heart is an ideal material for cardiovascular research due to its similarities to the human heart. This study evaluated pathological features and performed single-nucleus RNA sequencing (snRNA-seq) on myocardial samples from both wild-type and metabolic disease-susceptible transgenic pigs (previously established). We found that transgenic pigs exhibited lipid metabolism disturbances and myocardial injury after a high-fat high-sucrose diet intervention. snRNA-seq reveals the cellular landscape of healthy and metabolically disturbed pig hearts and identifies the major cardiac cell populations affected by metabolic diseases. Within metabolic disorder hearts, metabolically active cardiomyocytes exhibited impaired function and reduced abundance. Moreover, massive numbers of reparative LYVE1⁺ macrophages were lost. Additionally, proinflammatory endothelial cells were activated with high expression of multiple proinflammatory cytokines. Our findings provide insights into the cellular mechanisms of metabolic disease-induced myocardial injury.

Keywords: Animals; Integrative aspects of cell biology; Model organism; Porcine cardiology; Transcriptomics.

Graphical abstract

Figure 1

Clinical characteristics of TG pigs (A) Average body weight and serum biochemical parameters. WT, wild-type group; TG, triple-transgenic group. HDL-C, high-density lipoprotein cholesterol; GLU, glucose. (B) Representative images of echocardiography of left ventricle in WT and TG groups (left). Left ventricular ejection fraction (EF) of WT and TG groups (n = 3 per group) (right). (C) Representative H&E staining images of left ventricle (top), interventricular septum (middle), and aortic vessel (bottom) of WT and TG pigs. The areas indicated by boxes are magnified on the right. Edema (black arrows) and myofiber rupture (green arrows) were indicated. Scale bars, 200 μm (left), 50 μm (right). (D) Representative images of cardiac mitochondrial morphology obtained from electron microscope. Normal (yellow arrow) and damaged (green arrow) myocardial ultrastructure were indicated. M, mitochondria; Mf, myofibril; SR, sarcoplasmic reticulum. Scale bars, 1 μm.

Figure 2

Cellular composition of WT and TG pig hearts (A) Schematic of snRNA-seq experimental design and workflow. Transmural samples were obtained from left ventricle of wild-type (WT) and triple-transgenic (TG) pigs (n = 1 WT, n = 1 TG), single nuclei were processed using Chromium 10× 3′DEG chemistry. (B) t-distributed stochastic neighbor embedding (t-SNE) clustering of 28,745 nuclei isolated from both WT and TG, split by cell type (left) or by source (right). Each dot represents a single cell. (C) Known marker genes for each cell type. Dot color and size indicate the relative mean expression level and proportion of cells expressing the gene in each cell type, respectively. (D) The proportion of cells within WT and TG hearts identified from nuclei. CM, (ventricular) cardiomyocytes; EC, endothelial cells; FB, fibroblasts; MD, myeloid cells; T, T cells; PC, pericytes; SMC, smooth muscle cells; NC, neuronal cells; ERY, erythroid cells; AD, adipocytes. (E) Correlation matrix between pig and human cardiac cells. Normalized average UMI values for each cell type were used in the calculation of correlation coefficient values. (F) Dot plots showing the number of up- and downregulated genes in TG cardiac cells relative to WT group (p < 0.05, fold change > 1.28).

Figure 3

Cardiomyocyte subpopulations and their alterations in TG pigs (A) t-SNE projection of five cardiomyocyte (CM) subpopulations. (B) Heatmap showing correlation of five CM subpopulations by Spearman correlation analysis. (C) Dot plot representing the signature gene expression in each CM subpopulation. Dot color and size indicate the relative mean expression level and proportion of cells expressing the gene in each subpopulation, respectively. (D) Violin plots showing the expression of top distinct genes of CM3 subpopulation. (E) Selected top GO enrichment terms for subcluster-specific genes of CM subpopulations. (F) The number of CMs of each CM subpopulation in WT and TG samples. (G) Co-staining of CM general marker TNNT2 and CM3 special marker TPT1 in left ventricular sections of WT sample, the white arrows indicate co-localized CMs (CM3). Scale bars, 100 μm. (H) Heatmap showing the top differentially expressed genes (DEGs) in CM3 subpopulations comparing WT to TG. (I) Significant downregulated KEGG pathways of CM3 subpopulation in TG group.

Figure 4

Endothelial cell subpopulations and their alterations in TG pigs (A) t-SNE projection of seven endothelial cell (EC) subpopulations. (B) Dot plot representing the signature gene expression in each EC subpopulation. Dot color and size indicate the relative mean expression level and proportion of cells expressing the gene in each subpopulation, respectively. (C) Selected top GO enrichment terms for subcluster-specific genes of BEC subpopulations. (D) Expression of key differentially expressed genes (DEGs) in all BECs subpopulations and bulk RNA-seq data. (E) Heatmap showing the top DEGs (left) and their enriched GO terms (right) in proinflammatory-EC2 subpopulations comparing WT to TG. (F) Co-staining of VCAM1 and general EC marker PECAM1 in WT versus TG sections. Scale bar, 100 μm (top), 50 μm (bottom).

Figure 5

Macrophage subpopulations and their alterations in TG pigs (A) t-SNE projection of four macrophage (MAC) subpopulations (left) and their selected signature genes (right). (B) Dot plot representing the selected marker gene expression in each MAC subpopulation. Dot color and size indicate the relative mean expression level and proportion of cells expressing the gene in each subpopulation. (C) Selected top GO enrichment terms for subcluster-specific genes of major MAC subpopulations. (D) Heatmap showing the average expression of marker genes of LYVE1⁺ MAC1 and MHC-II^hi MAC2 obtained from WT or TG samples, respectively. (E) The number of macrophages of each MAC subpopulation in WT and TG samples. (F) Co-staining of MAC1-specific marker LYVE1 and MAC pan-marker CD163, the white arrows indicate co-localized MACs. Scale bar, 50 μm. (G) Radar plot showing GO enrichment terms of highly expressed genes of major MAC subpopulations in WT and TG groups. Left, LYVE1⁺ MAC1, right, MHC-II^hi MAC2.

Figure 6

Fibroblast subpopulations and their alterations between WT and TG pigs (A) t-SNE projection of six fibroblast (FB) subpopulations. (B) The proportion of FB subpopulations in WT and TG pigs identified from nuclei. (C) Dot plot representing the signature gene expression in each FB subpopulation. (D) Selected top GO enrichment terms for subcluster-specific genes of FB subpopulations. (E) UMAP (Uniform Manifold Approximation and Projection) projection of FB populations split by source. (F) Entropy and pseudotime overlayed on UMAP projection split by source.

Figure 7

Intercellular connection network of cardiac cells (A) Network of cell-cell interactions in cardiac cells, dot color indicates the cell type, dot size indicates the total number of significant ligand-receptor pairs of each cell type, line thickness indicates the number of significant interactions between two cell types. (B) Dot plot representing the selected essential growth factors expression in major cardiac cell (sub)types. (C) Dot plot shows significant ligand-receptor pairs between changed cardiac cell subtypes. (D) Line plot showing ligands secreted by changed noncardiomyocyte subsets and their matching receptors. Dot color and size indicate the fold change and p value of the ligand in the subpopulation of TG group, respectively.