Construction of a deep learning model and identification of the pivotal characteristics of FGF7- and MGST1- positive fibroblasts in heart failure post-myocardial infarction.
Journal:
International journal of biological macromolecules
Published Date:
Apr 19, 2025
Abstract
Dysregulation of fibroblast function is closely associated with the occurrence of heart failure after myocardial infarction (post-MI HF). Myocardial fibrosis is a detrimental consequence of aberrant fibroblast activation and extracellular matrix deposition following myocardial infarction (MI). However, the heterogeneity of fibroblasts in normal cardiac tissue and heart failure tissue remains to be further investigated. We discovered that the abundance of FGF7MGST1 fibroblasts were down-regulated in post-MI HF according to scRNA-seq analysis. Key gene characteristics of FGF7MGST1 fibroblasts were uncovered through both differential expression analysis and hdWGCNA pipeline. Pseudotime analysis revealed that FGF7MGST1 fibroblasts were gradually decreased with the occurrence of heart failure. Cell-cell communication analysis indicated an enhanced secretory ability in FGF7MGST1 fibroblasts compared to other fibroblasts. Utilizing machine learning algorithms, we identified 17 feature genes of this cell population. A deep learning model capable of predicting heart failure was successfully built based on these feature genes and immune infiltration levels of post-MI HF. FGF7 was highly related to cardioprotective pathway terms, including "PI3K/AKT pathway" and "protein secretion". Parallelly, mendelian randomization analysis was adopted to better understand the causal relationships between feature genes and post-MI HF. Results indicated that MGST1 was causally associated with heart failure, consistent with single cell data. And the post-MI HF mouse model was constructed and qRT-PCR assays supported that both FGF7 and MGST1 were largely down-regulated in myocardial infarction area than other cardiac tissues. These findings provide new insights into the roles of FGF7MGST1 fibroblasts in post MI HF.