Identification of macrophage driver genes in fibrosis caused by different heart diseases based on omics integration

Abstract

Background: Myocardial fibrosis, a hallmark of heart disease, is closely associated with macrophages, yet the genetic pathophysiology remains incompletely understood. In this study, we utilized integrated single-cell transcriptomics and bulk RNA-seq analysis to investigate the relationship between macrophages and myocardial fibrosis across omics integration.

Methods: We examined and curated existing single-cell data from dilated cardiomyopathy (DCM), ischemic cardiomyopathy (ICM), myocardial infarction (MI), and heart failure (HF), and analyzed the integrated data using cell communication, transcription factor identification, high dimensional weighted gene co-expression network analysis (hdWGCNA), and functional enrichment to elucidate the drivers of macrophage polarization and the macrophage-to-myofibroblast transition (MMT). Additionally, we assessed the accuracy of single-cell data from the perspective of driving factors, cell typing, anti-fibrosis performance of left ventricular assist device (LVAD). Candidate drugs were screened using L1000FWD.

Results: All four heart diseases exhibit myocardial fibrosis, with only MI showing an increase in macrophage proportions. Macrophages participate in myocardial fibrosis through various fibrogenic molecules, especially evident in DCM and MI. Abnormal RNA metabolism and dysregulated transcription are significant drivers of macrophage-mediated fibrosis. Furthermore, profibrotic macrophages exhibit M1 polarization and increased MMT. In HF patients, those responding to LVAD therapy showed a significant decrease in driver gene expression, M1 polarization, and MMT. Drug repurposing identified cinobufagin as a potential therapeutic agent.

Conclusion: Using integrated single-cell transcriptomics, we identified the drivers of macrophage-mediated myocardial fibrosis in four heart diseases and confirmed the therapeutic effect of LVAD on improving HF with single-cell accuracy, providing novel insights into the diagnosis and treatment of myocardial fibrosis.

Keywords: Cardiomyopathy; Heart failure; Macrophage; Myocardial fibrosis; Myocardial infarction.

Fig. 1

Cell proportion of four single cell atlas. The cell ratio of DCM, ICM, HF and MI in normal and disease (A-D)

Fig. 2

Enrichment analysis of DEGs and signal transduction promoting fibrosis. Functional changes of fibroblasts in four diseases (A), functional analysis of macrophages in four diseases (B), and FN1 signal release network of macrophages in four diseases (C-F)

Fig. 3

Driver gene identification of macrophages in DCM. Functional analysis of fibroblasts receiving profibrotic signals (A) and macrophages sending profibrotic signals (B); TFs in different macrophage clusters (C); Key module gene analysis (D), hub gene identification of module gene (E); The expression of hub gene in profibrotic macrophages (F); Expression of MMT markers (G) and polarization markers (H)

Fig. 4

Driver gene identification of macrophages in ICM. The legend is the same as Fig. 3

Fig. 5

Driver gene identification of macrophages in HF. The legend is the same as Fig. 3

Fig. 6

Driver gene identification in MI macrophages. The legend is the same as Fig. 3

Fig. 7

Analysis of gene intersection and expression of profibrotic macrophages. The intersection analysis of hub gene (A) and TFs (B). The expression of hub gene in profibrotic macrophages of DCM, ICM and HF (C-E)

Fig. 8

Analysis of the expression of hub genes and transcription factor. Analysis of the expression of intersection hub genes in three Bulk RNAseq data sets (A, C, E), the expression of TFs in DCM patients (B) and the expression of common TFs in three data sets (B, D, F)

Fig. 9

LVAD regulates the gene expression of HF macrophages. cell grouping (A), cell annotation (B) and cell type ratio (C). Expression of hub genes (D), transcription factor (E), macrophage polarization marker (F) and MMT marker (G) in HF treated by LVAD

Fig. 10

Analysis of drug-protein binding stability. RMSD values of five drugs binding with EEF1A1, RPS2, RPS9 and RPL7 (A-D)