Identification of biomarkers and immune infiltration in acute myocardial infarction and heart failure by integrated analysis

Abstract

The mortality of heart failure after acute myocardial infarction (AMI) remains high. The aim of the present study was to analyze hub genes and immune infiltration in patients with AMI and heart failure (HF). The study utilized five publicly available gene expression datasets from peripheral blood in patients with AMI who either developed or did not develop HF. The unbiased patterns of 24 immune cell were estimated by xCell algorithm. Single-cell RNA sequencing data were used to examine the immune cell infiltration in heart failure patients. Hub genes were validated by quantitative reverse transcription-PCR (RT-qPCR). In comparison with the coronary heart disease (CHD) group, immune infiltration analysis of AMI patients showed that macrophages M1, macrophages, monocytes, natural killer (NK) cells, and NKT cells were the five most highly activated cell types. Five common immune-related genes (S100A12, AQP9, CSF3R, S100A9, and CD14) were identified as hub genes associated with AMI. Using RT-qPCR, we confirmed FOS, DUSP1, CXCL8, and NFKBIA as the potential biomarkers to identify AMI patients at risk of HF. The study identified several transcripts that differentiate between AMI and CHD, and between HF and non-HF patients. These findings could improve our understanding of the immune response in AMI and HF, and allow for early identification of AMI patients at risk of HF.

Keywords: acute myocardial infarction; diagnosis; expression profile; heart failure; immune cell infiltration; single cell sequencing.

Figure 1. The step-by-step flowchart of this research

Figure 2. Immune cell landscape in AMI patients after merging GSE59867 and GSE62646 datasets

(A) The differences in immune cell enrichment scores between the acute myocardial infarction (AMI) and coronary heart disease (CHD) group by using xCell algorithm (* means P<0.05, ** means P<0.01, *** means P<0.001, **** means P<0.0001). (B) Correlation between 14 significantly changed immune cell was visualized with a heatmap (*means P<0.05, ** means P<0.01). (C) An area under the ROC curve (AUC) of top five immune cell abundance in merged cohort. (D) Pathway enrichment analysis by ssGSEA between acute myocardial infarction (AMI) and coronary heart disease (CHD) group in merged dataset (* means P<0.05, ** means P<0.01, *** means P<0.001, **** means P<0.0001).

Figure 3. Identification of common immune-related DEGs and predictive performance in merged GSE59867 and GSE62646 datasets

(A) Differentially expressed genes between acute myocardial infarction and CHD samples in merged dataset (GSE59867+GSE62646). (B) Gene Ontology (GO) enrichment analysis were visualized with bubble plots. (C) The 25 common immune-related DEGs in merged dataset were found by intersecting common DEGs with genes from ImmPort database. (D) PPI network of DEGs using STRING and visualized it using Cytoscape software. (E) MCC (Maximal Clique Centrality) score of the DEGs was visualized with histogram plot. (F) Expression of five hub immune-related DEGs in each type of immune cell by the Human Protein Atlas (HPA) database. We downloaded the expression levels of five genes from the ‘immune cell’ module in HPA database, and showed the gene expression levels at the protein level of each immune cell by percentage histogram. (G) The expression levels of five hub immune-related genes from day 1 to 6 months of PBMC of AMI patients (** means P <0.01, *** means P <0.001). (H) An area under the ROC curve (AUC) of five hub immune-related genes in merged dataset.

Figure 4. Correlation between the immune-related core DEGs and immune cell infiltration in AMI patients and immune cell landscape between non-HF and post-AMI HF patients after merging datasets (GSE11947 and GSE123342)

(A) Correlation between the immune-related hub DEGs and immune cell infiltration score in merged dataset was visualized with heatmap. immune cell infiltration score was gained by xCell algorithm above. (B–F) Correlation between five hub immune-related genes (S100A12, AQP9, CSF3R, S100A9, and CD14) and macrophage cell infiltration proportion gained by xCell algorithm above. (G) The differences in immune cell abundance between the non-HF and HF group in merged dataset (GSE11947 and GSE123342) by using xCell algorithm (* means P<0.05, ** means P<0.01, *** means P<0.001, **** means P<0.0001). (H) Correlation between five significantly changed immune cell was visualized with a heatmap (* means P<0.05, ** means P<0.01). (I) An area under the ROC curve (AUC) of five immune cell abundance in merged dataset.

Figure 5. Further demonstration of immune microenvironment in peripheral blood of ischemic heart failure

(A,B) 18 cell clusters in single cell sequencing dataset GSE145154 was calculated by ‘Seurat’ package (resolution = 0.5), and using ‘singleR’ package, we classified 18 cell clusters into 9 cell populations automatically. (C) The markers of immune cells markers in 9 cell population. (E) The markers of T-cell specific markers in 18 clusters. (D,F) The proportion of multiple immune cell subsets in health and patients with ischemic heart failure. (G,H) Immune cell intersection established by ‘Cellchat’ R package between control and heart failure group.

Figure 6. Identification and Validation of five hub DEGs in HF after AMI using RT-qPCR

(A) Gene Ontology enrichment analysis of DEGs between HF and non-HF group merged dataset (GSE11947 and GSE123342). (B) PPI network of DEGs using STRING and visualized it using Cytoscape software. (C) MCC (Maximal Clique Centrality) score of the DEGs was visualized with histogram plot. (D) An area under the ROC curve (AUC) of seven hub DEGs to distinguish HF with non-HF patients in merged datasets with the AUC value >0.7. (E–K) Gene expression changes (FOS, DUSP1, CXCL8, NFKBIA, CEBPD, BCL2A1, and SAMSN1) were investigated in HF and non-HF on day 1 after AMI using RT-qPCR (**** means P <0.0001). (L–O) ROC curves for FOS, DUSP1, CXCL8, and NFKBIA; AUC, area under the curve; ROC, receiver operating characteristic.