Myocardial DNA damage combined with genotype predicts heart failure outcome in various underlying diseases

Abstract Background Reliable predictors of treatment efficacy in heart failure have been long awaited. DNA damage has been implicated as a cause of heart failure. Purpose To investigate the association of DNA damage in myocardial tissue with treatment response and prognosis of heart failure from various etiologies; and to elucidate the relationships between genotypes and DNA damage, as well as their additive effect in prognosis prediction. Methods We performed immunostaining of DNA damage markers, poly(ADP-ribose) (PAR) and γ-H2A.X on biopsy specimens from 175 patients with heart failure with reduced ejection fraction (HFrEF) of various underlying etiologies. We calculated the percentage of nuclei positive for each DNA damage marker (%PAR and %γ-H2A.X). The primary outcome was left ventricular reverse remodeling (LVRR) at ∼1 year. The secondary outcome was a composite of cardiovascular death, heart transplantation, and ventricular assist device implantation. Whole-exome data from consenting patients were further analyzed to classify pathogenic/likely pathogenic (P/LP) variants in cardiomyopathy-related genes. We compared the degree of DNA damage among different genotypes, and examined the additive effect of combining DNA damage markers and genotypes in predicting treatment response. Results The participants had a mean age of 56 years, of whom 72% were men. Among them, 55% achieved LVRR at ∼1 year. Patients who did not achieve LVRR after the optimization of medical therapies presented with significantly higher %PAR and %γ-H2A.X when compared with those who did [%PAR, 11 (6–19) % vs. 2 (1–4) %, P < 0.001; %γ-H2A.X, 22 (11–40) % vs. 5 (2–9) %, P < 0.001] (Figure, left bottom). The receiver-operating characteristic (ROC) analysis demonstrated good performance of both %PAR and %γ-H2A.X for predicting LVRR (area under the curve 0.87 and 0.85, respectively). Patients with lower %PAR or %γ-H2A.X had reduced long-term clinical events [PAR, hazard ratio 1.63 (95%CI 1.31- 2.01), P < 0.001; γ-H2A.X, 1.48 (1.27-1.72), P < 0.001] (Figure, left bottom). In addition, there was a strong negative correlation between the mean proportion of DNA damage marker positive nuclei and the probability of LVRR, across different underlying diseases (Figure, middle). Genotypes were identified in 105 patients (Figure, right top). Non-sarcomeric mutation (LMNA, etc) carriers presented with more pronounced DNA damage than sarcomeric mutation carriers (TTN, MYBPC3) (Figure, right bottom). Combination of genotypes and %γ-H2A.X augmented the area under the ROC for predicting 1-year LVRR to 0.90, from that of %γ-H2A.X alone at 0.85. Conclusions The degree of DNA damage accumulation in myocardial tissue differs by underlying diseases and genotypes, and determines the consequences of human heart failure. Assessment of DNA damage, especially when combined with genotype, is useful to predict treatment efficacy and prognosis of HFrEF patients with various underlying etiologies.graphical abstract