Fig 1: SOX9 opens chromatin and increases enrichment of active histone modifications at a subset of binding sites. (A) Heatmap displaying SOX9, ATAC, H3K4me1, H3K27ac and H3K27me3 signal within a 4 kb window around the summit of SOX9 bound regions. K-means clustering revealed four modes of SOX9 binding (C1–C4). (B) Representative loci from each cluster. Regions with a SOX9 peak are highlighted with gray boxes, and the C2 region in the MMP2 locus is highlighted with a red box. (C) Distribution of SOX9 bound regions between TSS/promoter regions (5000 bp upstream to 500 bp downstream of TSS), within gene bodies, or distal. (D) Log2 FPKM ratios for differentially expressed genes with an associated SOX9 bound region in C1–C4. Select genes are highlighted. (E) Scatterplots of log2 FPKM values of expressed genes (% shown) with an associated SOX9 bound region in C1–C4. Significance was evaluated as in Figure 1E. Mean is indicated with black line. (F) Average SOX9 CUT&RUN signal at the summits of SOX9 bound regions in C1–C4. Significance was evaluated as in Figure 1E. Mean is indicated with black line.
Fig 2: scATAC-seq study suggests SOX9 opens chromatin in cells undergoing EndMT in atherosclerosis. (A) UMAP representations of co-embedded and separate scRNA-seq and scATAC-seq mouse atherosclerosis datasets from Andueza et al. (B) Pseudotime progression of EndMT based on trajectory from Andueza et al. and Sox9 expression in each cluster. (C) Top enriched motifs in regions that gained chromatin accessibility between EC2 to EC6, EC6 to EC8, and EC8 to Fibro. (D) Representative loci with increased chromatin accessibility from EC6 to EC8 (highlighted with gray boxes). Right panel shows the expression pattern of the given genes.
Fig 3: SOX9 alters expression of EndMT genes. (A) Scatterplot comparing normalized average log2 FPKM values between cells transduced with empty vector or SOX9. Significantly differentially expressed genes are coloured red with select genes highlighted. (B) Heatmap of differentially expressed genes between cells transduced with SOX9 or empty vector. Heatmap shows log2 FPKM values while log2 ratios are displayed on the right with select genes highlighted. (C) Gene set enrichment analysis (GSEA) of enriched Hallmark gene sets (based on P-values) for upregulated and downregulated genes upon SOX9 expression. (D) GSEA cell signatures of genes upregulated (left) and downregulated (right) in HUVECs. (E) Log2 ratios of SOX9 OE over empty vector for annotated endothelial- or mesenchymal-specific genes with select genes highlighted.
Fig 4: SOX9 is expressed at the onset of EndMT in mouse embryonic hearts and atherosclerotic lesions and ectopic SOX9 expression in human endothelial cells induces a mesenchymal phenotype. (A) SOX9 and ERG immunostaining of mouse hearts at E9.5 and 10.5. Scale bar whole hearts, 100 μm. Scale bar areas of interest, 25 μm. (B) SOX9 expression in each cluster from Andueza et al. (left) and Chen et al. (right). Significance was evaluated between clusters EC8, Fibro, or SMC and EC1-EC7 (Andueza et al.) and between EC6, Fibro or SMC and EC1-EC5 (Chen et al.) by Wilcoxon rank sum tests with P-values (*P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; ns P > 0.05). (C) Schematic overview of experimental setup. Created with Biorender. (D) Fluorescence microscopy of HUVECs 6 days after transduction with SOX9 or empty vector, demonstrating the change in morphology with SOX9 expression. Scale bar, 100 μm. (E) Transwell assay indicating number of migrated HUVECs after transduction with SOX9 or empty vector. Significance was evaluated by unpaired, two-tailed t-tests with P-values (*P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; ns P > 0.05). Mean is indicated by gray line. (F) SOX9, ERG, CD31, POSTN and VIM immunostaining of HUVECs 6 days after transduction with SOX9 or empty vector. Scale bar, 100 μm.
Fig 5: Enrichment of H2A.Z and H3S28P determines SOX9 binding and chromatin opening. (A) Heatmap displaying SOX9, ATAC, H2A.Z (ENCODE), H3K4me3 (ENCODE), H3K9me3 (ENCODE), H3K36me3 (ENCODE), H3K79me2 (ENCODE) and H4K20me1 (ENCODE) signal within a 4 kb window around the summit of SOX9 bound regions. (B) Average H2A.Z ChIP-seq signal (ENCODE) at the summits of SOX9 bound regions in C1–C4. Significance was evaluated as in Figure 1E. Mean is indicated with black line. (C) Histone peptide array containing 384 different histone tail modification combinations incubated with recombinant SOX9 protein and detected with anti-SOX9 primary antibody. Spots containing H3S28P (alone or in combination with other histone modifications) are highlighted with pink circles. (D) Relative average spot intensity (from 4 arrays) for all single modified histone peptides (blue), unmodified peptides (green), and background array spots (red). H3S28P is indicated with arrow. (E) SOX9, H3K4me1, H3K4me3, H3K27ac and H3S28P immunostaining of mouse atrioventricular canals at E12.5 and PLA signal (yellow) on serial sections showing interactions occurring between SOX9 and the indicated histone modifications. Scale bar, 50 μm. Scale bar areas of interest, 10 μm. (F) SOX9, H3K4me1, H3K4me3, H3K27ac and H3S28P immunostaining of SOX9-expressing HUVECs and PLA signal (yellow) showing interactions occurring between SOX9 and the indicated histone modifications. Right panel shows PLAs between H3 and indicated histone modifications as a positive control. Scale bar, 100 μm. Scale bar areas of interest, 25 μm. (G) Heatmap displaying H3S28P signal within a 4 kb window around the summit of ATAC peaks in regions with increased chromatin accessibility.
Supplier Page from Abcam for Recombinant Human SOX9 protein (GST tag N-Terminus)