Integrated whole-genome gene expression analysis reveals an atlas of dynamic immune landscapes after myocardial infarction

Figure 1. Activation of pro-inflammatory and glycolytic pathways in PBMCs from the MI patients. Two-dimensional PCA plots were used to visualize the distribution and variation of samples before (A) and after (B) batch normalization (n = 199). Each dot represented one sample and was colored by group. (C) Representative gene sets enriched in PBMCs from the MI patients at admission. Each vertical bar represented one gene, with the location of the bar indicating the occurrence of that gene in the gene list, and the height of the bar representing the relative fold change. Gene sets with a |NES| > 1, p value<0.05, p.adj < 0.25 were considered to be significantly enriched. (D) Heatmap depicted the expression of the leading-edge genes contributed most to the indicated pathways. (E) The relative abundance of leukocytes in PBMCs was computed based on xCell algorithm. Two-group comparisons were made using the Mann–Whitney non-parametric test with p value <0.05 signifying statistical significance. The Box-Whisker plots denoted median + quartiles (box) and range (whiskers). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. BP, biological process; CC, cell component; KEGG, Kyoto Encyclopedia of Genes…

Figure 2. Stratification of the MI PBMCs based on inflammatory response. (A) A two-cluster solution of PBMCs yielded by consensus clustering (n = 199). (B) CDF plot depicted the cumulative distribution functions of the consensus matrix for k = 2–7. (C) Delta area plot depicted the relative change in area under CDF curve for k = 2–7. (D) Volcano plots showed the DEGs when STEMI was compared to CAD. (E) Volcano plots showed the DEGs when STEMI subgroup 1 was compared to CAD. (F) Volcano plots showed the DEGs when STEMI subgroup 2 was compared to CAD. The DEGs with an adjusted p value<0.05 and |log2FC| > 1 were circled and labeled. (G) Heatmap visualized the pathway activities of PBMCs quantified by GSVA. The color scale reflected Z-scores with red indicating relatively high expression and blue relatively low expression. CDF, cumulative distribution functions; DEGs, differential expressed genes; GSVA, gene set variation analysis; HF, heart failure; SCAD, stable coronary artery disease; STEMI, ST-elevation myocardial infarction.

Figure 3. Increased expression of pro-inflammatory and glycolytic genes in PBMCs and peripheral whole blood from the MI patients. (A–D) GSEA plots showed a significant enrichment of gene sets involved in inflammatory response and glycolysis in the MI patients. Gene sets with p value<0.05, p.adj < 0.25 and |NES| > 1 were considered to be significantly enriched. (E) The Box-Whisker plots showed the relative abundance of monocytes and neutrophils in PBMCs. The central line represented the median value and the box bounds represented the interquartile range; whiskers represented min to max. Two-group comparisons were determined using the Mann–Whitney non-parametric test, with p value<0.05 signifying statistical significance. (F) Heatmap depicted the expression of representative leading-edge genes contributed most to the enriched gene sets. *p < 0.05, **p < 0.01. ACS, acute coronary syndrome; AMI, acute myocardial infarction; MI, myocardial infarction; NC, normal control; NES, normalized enrichment score; PBMCs, peripheral blood mononuclear cells; p.adj, adjusted p value; p-val, p value; SCAD, stable coronary artery disease; STEMI, ST-elevation myocardial infarction.

Figure 4. Increased expression of pro-inflammatory genes in CECs, platelets and exosomes from the MI patients. GSEA plots showed a significant enrichment of gene sets involved in inflammatory response in CECs (A) and platelets (B) from the MI patients. Gene sets with p value<0.05, p.adj < 0.25 and |NES| > 1 were considered to be significantly enriched. (C) GSEA plots of the exosomal gene sets. (D) Heatmap depicted the expression of the representative pro-inflammatory genes. (E) Scatter plots showed the correlations between the abundance of monocytes or neutrophils and the expression level of the representative pro-inflammatory genes in both SCAD patients and MI patients, using the combined and normalized dataset of GSE59867 and GSE62646. Statistical analysis was performed using Spearman’s correlation test. Correlation coefficient and p value were noted in the top left corner of each plot. ACS, acute coronary syndrome; AMI, acute myocardial infarction; CECs, circulating endothelial cells; LAD, left anterior descending; NC, normal control; NES, normalized enrichment score; p.adj, adjusted p value; p-val, p value; SCAD, stable coronary artery disease; STEMI, ST-elevation myocardial infarction.

Figure 5. Temporal evolution of the post-MI PBMCs gene expression profiles assessed by GSVA and xCell. Two-dimensional PCA plots visualized the distribution and variation of samples before (A) and after (B) batch normalization (n = 534). Each dot in the plots represented one sample. (C) Boxplots depicted the pathway activities of PBMCs quantified by GSVA. (D) Boxplots depicted the abundance of the indicated cell components quantified by xCell. The Box-Whisker plots denoted median + quartiles (box) and range (whiskers). Multiple-group comparisons were made using Kruskal–Wallis test with Dunn’s correction. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. OXPHOS, oxidative phosphorylation; SCAD, stable coronary artery disease.

Figure 6. Temporal evolution of the post-MI PBMCs gene expression profiles assessed by GSEA. The MI group from the indicated time point was compared to the SCAD group. GSEA plots showed the representative gene sets involved in inflammation, coagulation, metabolism and stress response. Gene sets with p value<0.05, p.adj < 0.25 and |NES| > 1 were considered to be significantly enriched. ES, enrichment score; NES, normalized enrichment score; OXPHOS, oxidative phosphorylation; p.adj, adjusted p value; p-val, p value; SCAD, stable coronary artery disease.

Figure 7. Temporal evolution of the post-MI PBMCs gene expression profiles assessed by Mfuzz. (A) Mfuzz soft clustering identified two clusters of DEGs. (B) Heatmap depicted the gene expression patterns of DEGs. Top 5 upregulated and downregulated DEGs at admission were specified. (C) The Box-Whisker plots showed the expression patterns of the representative leading-edge genes contributed most to the indicated enriched gene sets. The Box-Whisker plots denoted median + quartiles (box) and range (whiskers). Multiple-group comparisons were made using Kruskal–Wallis test with Dunn’s correction. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. FC, fold change; SCAD, stable coronary artery disease.

Figure 8. Temporal evolution of the post-MI PBMCs gene expression profiles in the patients with and without the long-term HF. The HF group and the non-HF group were compared separately to the SCAD group. The representative gene sets involved in acute inflammatory response, coagulation, canonical glycolysis and OXPHOS were presented. Gene sets with a p value<0.05, p.adj < 0.25 and |NES| > 1 were considered to be significantly enriched. ES, enrichment score; HF, heart failure; NES, normalized enrichment score; OXPHOS, oxidative phosphorylation; p.adj, adjusted p value; p-val, p value; SCAD, stable coronary artery disease; STEMI, ST-elevation myocardial infarction.

Figure 9. Elevated and prolonged expression of genes implicated in inflammation and coagulation in HF patients. (A) Volcano plots depicted the DEGs. DEGs with p < 0.05 and |log2FC| > 1 were circled and labeled. (B) Heatmaps depicted the expression of representative leading edge genes contributed most to the indicated gene sets. Statistical enrichment was highlighted with the solid-line boxes. (C) Line graphs showed the dynamic and heterogeneous expression patterns of hallmark inflammation-associated transcripts in the HF and non-HF groups. The dashed line represented the average gene expression level in the SCAD group. Data were expressed as mean ± SEM. Two-group comparisons at each time point were made using the Mann–Whitney non-parametric test with p value<0.05 defining statistical significance. *p < 0.05, **p < 0.01. HF, heart failure; OXPHOS, oxidative phosphorylation; SCAD, stable coronary artery disease.

Figure 10. Metabolic reprogramming of PBMCs from the MI and post-MI HF patients. Comparison of the glycolysis and OXPHOS signals was made by GSEA between the MI group derived from the patients at admission and the SCAD group (A), the MI group derived from the patients 6 months after MI and the SCAD group (B) and the HF group and the non-HF group from the patients 6 months after MI (C). Gene sets with p value<0.05, p.adj < 0.25 and |NES| > 1 were considered to be significantly enriched. The top five leading edge genes contributed most to the indicated gene sets were specified on the plots. Boxplots (left panel) and line plots (right panel) showed the dynamic and heterogeneous expression patterns of glycolysis (D) and OXPHOS genes (E). In the boxplots, multiple-group comparisons were made using Kruskal–Wallis test with Dunn’s correction. In the line plots, data were presented as mean ± SEM, two-group comparisons at each time point were made using the Mann–Whitney non-parametric test. The dashed line represented the average gene expression level in the SCAD group. For all tests applied, statistical significance was defined by p < 0.05. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. ES,…

Figure 11. Diversity and heterogeneity of PBMCs from the HF patients. (A) A three-cluster solution of PBMCs yielded by consensus clustering (n = 197). (B) A scatter plot showed the sample distribution in the four quadrants defined by the relative median expression level of the co-expressed glycolysis and OXPHOS genes. (C) Heatmap depicted the relative expression level of the co-expressed glycolysis and OXPHOS genes across the indicated subgroups. (D) Violin plots showed the differences of the indicated parameters revealed by multiple-group comparisons using Kruskal–Wallis test with Dunn’s correction. (E) GSEA plots depicted the phenotypic heterogeneity of PBMCs from the HF patients. Gene sets with p value<0.05, p.adj < 0.25 and |NES| > 1 were considered to be significantly enriched. ES, enrichment score; Lymp, lymphocyte; Mono, monocyte; Neut, neutrophil; NES, normalized enrichment score; OXPHOS, oxidative phosphorylation; p.adj, adjusted p value; p-val, p value; S, subgroup.