Novel insights into post-myocardial infarction cardiac remodeling through algorithmic detection of cell-type composition shifts

Interpreting bulk RNA sequencing from heterogeneous tissues like the post-myocardial infarction (MI) heart is confounded by dynamic changes in cell-type composition. To address this, we developed a computational approach using single-nucleus RNA sequencing (snRNA-seq) references to estimate and correct for cell-type abundance shifts in bulk transcriptomic data. We applied this method to analyze infarct border zone transcriptomes from wild-type (WT) and cardiomyocyte-specific α1A-adrenergic receptor knockout (cmAKO) mice subjected to MI via left coronary artery ligation or sham surgery. Our analysis revealed exaggerated cardiomyocyte loss and fibroblast gain in cmAKO mice post-MI compared to WT, implicating α1A-ARs in maintaining cellular homeostasis. We then demonstrate the confounding effect of composition changes though simulations: a modest 10% change in the major cell type's abundance caused over 20% of transcripts to appear as differentially expressed genes (DEGs) when composition was ignored. Applying our correction method refined the interpretation of MI-induced transcriptomic changes, attributing many apparent DEGs, particularly those related to metabolism and inflammation, to shifts in cell abundance rather than direct transcriptional regulation. Importantly, the correction also unveiled previously masked biological processes associated with the cmAKO-specific response to MI, including pathways related to cell adhesion, cell cycle regulation, and stress response, highlighting potential intrinsic mechanisms of α1A-AR cardioprotection. RNAscope validation supported the composition-aware findings for key genes. This work presents a robust method for dissecting bulk RNA-seq data from complex tissues and provides refined insights into the cellular and molecular roles of cardiomyocyte α1A-ARs during cardiac injury and remodeling.