Serum exosomal microRNAs in patients with heart failure are correlated with myocardial DNA damage and future reverse remodeling

Abstract Background Reliable predictors of treatment response in heart failure have been long awaited. Assessment of myocardial DNA damage, which is involved in the pathogenesis of heart failure, has been shown to be able to precisely predict reverse remodeling (1). Myocardial DNA damage-related serum exosomal microRNAs, if any, may serve as a non-invasive predictor for reverse remodeling in heart failure. Methods We screened patients who were diagnosed as heart failure with reduced ejection fraction (HFrEF) regardless of etiologies at multiple centers and cryopreserved their sera. We next performed 1:1 matching for age, sex, etiology, and left ventricular ejection fraction (LVEF) to determine the analysis cohort. In this cohort, for those who underwent endomyocardial biopsy, we assessed myocardial DNA damage using immunostaining of γ-H2A.X on biopsy specimens as previously reported (1), and extracted RNAs from the exosomes isolated from their sera, on which we performed small RNA-sequencing. MicroRNAs whose read frequency were correlated with γ-H2A.X-positive nuclei proportion (%γ-H2A.X), as defined by a Spearman's ρ 0.4 or -0.4, were quantified in the entire matched analysis cohort using quantitative reverse transcription polymerase chain reaction (RT-qPCR) to validate the correlation. The expression level of the validated microRNAs were then compared between the patients who achieved left ventricular reverse remodeling (LVRR) at 1-year with optimal medical therapy and those who did not. The prediction performance was evaluated with receiver operating characteristic (ROC) analysis. Results We screened 202 patients with HFrEF, from whom 42 matched pairs (84 patients) were retrieved as analysis cohort. Within them, 23 underwent biopsy, and %γ-H2A.X was higher in those with 1-year LVRR (P = 0.01). Upon analyzing the sequencing data of their serum exosomal microRNAs, 33 out of the 2,883 microRNAs had their Spearman's ρ with %γ-H2A.X 0.4 or -0.4 (red dots in Figure A). Validation was performed using RT-qPCR in the entire matched analysis cohort which revealed that miR-x had a remarkable negative correlation with %γ-H2A.X and differed significantly between patients with and without LVRR at 1-year (Figure B). Expression level of serum exosomal miR-x was identified as a predominant predictor for 1-year LVRR with an area under the ROC curve of 0.81. Conclusions Measurement of expression levels of serum exosomal microRNAs offers a new approach for prediction of treatment response in HFrEF. Serum exosomal miR-x, which was negatively correlated with myocardial DNA damage, was indicated as a promising predictor for LVRR at 1-year. Further investigations focusing on the function of miR-x and its potential as a therapeutic target are ongoing.