Search

Background— Investigation of disease pathogenesis confined to protein-coding regions of the genome may be incomplete because many noncoding variants are associated with disease. We aimed to identify novel predictive markers for the course of enterovirus (CVB3) cardiomyopathy by screening for noncoding elements influencing the grossly different antiviral capacity of individual patients. Methods and Results— Transcriptome mapping of CVB3 cardiomyopathy patients revealed distinctive cardiac microRNA (miR) patterns associated with spontaneous virus clearance and recovery (CVB3-ELIM) versus virus persistence and progressive clinical deterioration (CVB3-PERS). Profiling of protein-coding genes and 754 miRs in endomyocardial biopsies of test cohorts was performed at their initial presentation, and those spontaneously eliminating the virus were compared with those with virus persistence on follow-up. miR profiling revealed highly significant differences in cardiac levels of 16 miRs, but not of protein-coding genes. Evaluation of this primary distinctive miR pattern in validation cohorts, and multivariate receiver operating characteristic curve analysis, confirmed this pattern as highly predictive for disease course (area under the curve, 0.897±0.071; 95% confidence interval, 0.758–1.000). Eight miRs were strongly induced in CVB3-PERS (miRs 135b, 155, 190, 422a, 489, 590, 601, 1290), but undetectable in CVB3-ELIM or controls. They are predicted to target multiple immune response genes, and 2 of these were confirmed by antisense-mediated ablation of miRs 135b, 190, and 422a in the monocytic THP-1 cell line. Conclusions— An immediate clinical application of the data is cardiac miR profiling to assess the risk of virus persistence and progressive clinical deterioration in CVB3 cardiomyopathy. Patients at risk are eligible for immediate antiviral therapy to minimize irreversible cardiac damage.