Activated fibroblasts drive cellular interactions in end-stage pediatric hypertrophic cardiomyopathy

Figures

• 

  Figure 1:. Single-nuclei analysis in end-stage, pediatric hypertrophic cardiomyopathy A) Study cohort description exhibiting HCM and control subjects with their respective ages. B) Study schematic and steps in analysis in both controls and those with HCM from time of procurement (1), tissue handling and RNA sequencing (2), identification of major cell types of interest (3), and downstream analyses (4). C) UMAP embedding of major cell types and subsequent clustering of major cell types of interest. D) Proportion of major cell types of interest within control subjects (C1 and C2) and HCM subjects (H1, H2, and H3) and proportions in pooled cohorts (Control vs HCM). E) PCA plot for pseudo-bulk RNA-seq analysis for all major cell types of interest by diagnosis (red: Control; blue: HCM). Ellipses represent the 95% confidence intervals.

• 

  Figure 2:. HCM cardiomyocytes are enriched in insulin receptor and actin binding pathway genes A) UMAP embedding of reiterative clustering of cardiomyocytes. B) Bar plot depicting proportion of major cell types of interest within control subjects (C1 and C2) and HCM subjects (H1, H2, and H3) and proportions in pooled cohorts (Control vs HCM). C) Left: Embedding derived from Milo k-nearest neighborhood differential abundance testing and layout derived from UMAP embedding. Nodes with false discovery rate <10% depict neighborhoods and are colored by log fold changes for HCM (blue) versus control (red) samples. Right: Beeswarm plot depicting significant log fold changes for HCM (blue) and control (red) neighborhoods derived from differential abundance testing stratified by cardiomyocyte clusters (y-axis). D) Heatmap of top 5 differently expressed genes derived from FindAllMarkers function in CM1, CM2, CM3, CM4, CM5, and CM6. E) Volcano plot derived from pseudo-bulk RNA differential gene expression. F) Enrichment map stratified by CM cluster depicting the proportional GO pathway analysis. Node size is determined by the number of genes within each GO pathway. G) Feature plots illustrating…

• 

  Figure 3:. Activated cardiac fibroblast clusters are enriched with cells derived from HCM myocardium A) UMAP embedding of reiterative clustering of fibroblasts. B) Bar plot depicting proportion of major cell types of interest within control subjects (C1 and C2) and HCM subjects (H1, H2, and H3) and proportions in pooled cohorts (Control vs HCM). C) Left: Embedding derived from Milo k-nearest neighborhood differential abundance testing and layout derived from UMAP embedding. Nodes with false discovery rate <10% depict neighborhoods and are colored by log fold changes for HCM (blue) versus control (red) samples. Right: Beeswarm plot depicting significant log fold changes for HCM (blue) and control (red) neighborhoods derived from differential abundance testing stratified by fibroblast clusters (y-axis). D) Heatmap of top 5 differently expressed genes derived from FindAllMarkers function in FB1, FB2, FB3, and FB4. E) Volcano plot derived from pseudo-bulk RNA differential gene expression. F) Enrichment map stratified by FB cluster depicting the proportional GO pathway analysis. Node size is determined by the number of genes within each GO pathway. G) Top left: Feature plots illustrating…

• 

  Figure 4:. Genes associated with angiogenesis and cell fate are enriched in HCM endothelial cell clusters A) UMAP embedding of reiterative clustering of endothelial cells. B) Bar plot depicting proportion of major cell types of interest within control subjects (C1 and C2) and HCM subjects (H1, H2, and H3) and proportions in pooled cohorts (Control vs HCM). C) Left: Embedding derived from Milo k-nearest neighborhood differential abundance testing and layout derived from UMAP embedding. Nodes with false discovery rate <10% depict neighborhoods and are colored by log fold changes for HCM (blue) versus control (red) samples. Right: Beeswarm plot depicting significant log fold changes for HCM (blue) and control (red) neighborhoods derived from differential abundance testing stratified by cardiomyocyte clusters (y-axis). D) Heatmap exhibiting differentially expressed genes for each of the EC clusters. Expression levels indicate the average expression for each cluster. Clusters are colored according to Figure 4A. E) Violin plot illustrating relative expression for RGCC and NOTCH4 in HCM versus control cells. F) Enrichment map stratified by EC cluster depicting the proportional GO pathway analysis.…

• 

  Figure 5:. Tissue resident macrophages are predominated by control rather than HCM myocardium A) UMAP embedding of reiterative clustering of myeloid cells. B) Left: Embedding derived from Milo k-nearest neighborhood differential abundance testing and layout derived from UMAP embedding. Nodes with false discovery rate <10% depict neighborhoods and are colored by log fold changes for HCM (blue) versus control (red) samples. Right: Beeswarm plot depicting significant log fold changes for HCM (blue) and control (red) neighborhoods derived from differential abundance testing stratified by myeloid cell clusters (y-axis). D) Bar plot depicting proportion of major cell types of interest within control subjects (C1 and C2) and HCM subjects (H1, H2, and H3) and proportions in pooled cohorts (Control vs HCM). E) Heatmap exhibiting differentially expressed genes for each of the ML clusters. Expression levels indicate the average expression for each cluster. G) Dot plots illustrating average expression (depicted on color scale) and percent of cells expressing each gene (depicted by size of dot), stratified by ML cells for CCR2+ markers (left) and CCR2− markers (right) derived from Bajpai et al.

• 

  Figure 6:. Cell interactions are increased in HCM myocardium and predominantly fibroblast driven A) Circle plot illustrating the differential number (top) and weight (bottom) of ligand-receptor interactions in HCM versus control cells. Red indicates increased number or weight of interactions in HCM and blue indicates decrease in number or weight of interactions in HCM. B) Circle plots exhibiting the number of ligand-receptor interactions in HCM (top) versus control (bottom) cells. C) )Circle plots exhibiting the weight of ligand-receptor interactions in HCM (top) versus control (bottom) cells. D) Scatter plots for control (top) and HCM (bottom) cells illustrating the strength of incoming (y-axis) and outgoing (x-axis) interactions. The size of each circle is dependent on the number of interactions. E) Scatter plots illustrating differential incoming and outgoing interactions between HCM and control cells in fibroblasts, cardiomyocytes, endothelial cells, and endocardium. Each point represents a signaling pathway. The shapes define if the pathway is shared (circle), incoming specific (square), outgoing specific (triangle), or both incoming and outgoing specific (diamond). The color defines…

• 

  Figure 7:. Pediatric HCM are enriched in ECM-related genes compared to adult HCM A) Lollipop charts for GO over-representation analysis showing top 10 GO pathways (Y-axis) ranked by fold enrichment (X-axis; calculated by mutate(x, FoldEnrichment = parse_ratio(GeneRatio) / parse_ratio(BgRatio))) in biological processes (left), molecular functions (center), and cellular components (right). Colors of circles correspond to adjusted-P values and size of circle corresponds to the number of genes implicated. B) Enrichment map stratified by pediatric vs adult HCM depicting the proportional GO pathway analysis (left) and KEGG pathway analysis (right; only pediatric specific pathways met significance). Node size is determined by the number of genes within each pathway. C) Boxplots illustrating relative expression level derived from activated LAMININ module score (left) and COLLAGEN module score (right) stratified by adult versus pediatric HCM. The center line in the box represents the median, the top edge of the box shows the upper quartile, the lower edge of the box shows the lower quartile, and the whiskers extend x1.5 of the interquartile range. P value <0.001 indicated by ***.