Background: Plasma extracellular RNAs have recently garnered interest as biomarkers in heart failure (HF). Most studies in HF focus on single extracellular RNAs related to phenotypes and outcomes, and few describe their functional roles. We hypothesized that clusters of plasma microRNAs (miRNAs) associated with left ventricular (LV) remodeling in human HF would identify novel subsets of genes involved in HF in animal models.

Methods and results: We prospectively measured circulating miRNAs in 64 patients with systolic HF (mean age, 64.8 years; 91% men; median LV ejection fraction, 26%) with serial echocardiography (10 months apart) during medical therapy. We defined LV reverse remodeling as a 15% reduction in LV end-systolic volume index. Using principal components analysis, we identified a component associated with LV reverse remodeling (odds ratio=3.99; P=0.01) that provided risk discrimination for LV reverse remodeling superior to a clinical model (C statistic, 0.58 for a clinical model versus 0.71 for RNA-based model). Using network bioinformatics, we uncovered genes not previously widely described in HF regulated simultaneously by >2 miRNAs. We observed increased myocardial expression of these miRNAs during HF development in animals, with downregulation of target gene expression, suggesting coordinate miRNA-mRNA regulation. Target mRNAs were involved in autophagy, metabolism, and inflammation.

Conclusions: Plasma miRNAs associated with LV reverse remodeling in humans are dysregulated in animal HF and target clusters of genes involved in mechanisms implicated in HF. A translational approach integrating human HF, bioinformatics, and model systems may uncover novel pathways involved in HF.

Clinical trial registration: URL: https://www.clinicaltrials.gov. Unique identifier: NCT00351390.