The response to cardiac resynchronization therapy in heart failure patients is predicted and described by the blood transcriptome

Background Cardiac resynchronization therapy (CRT) improves outcomes in heart failure (HF) patients with reduced ejection fraction. However, approximately 30% of the patients do not respond to the treatment. More research is needed to better select patients for this treatment and to understand the underlying molecular mechanisms of treatment-response. Purpose Our aims were: 1) To define a pre-therapeutic blood gene expression profile that could predict response to CRT and 2) to identify gene expression changes occurring after CRT associated with different outcomes. Methods We enrolled 200 HF patients selected for CRT-defibrillator (CRT-D) implantation and 50 non-HF (control) patients. Coronary sinus and peripheral blood samples were obtained from HF patients during CRT-D implantation and from control patients during invasive cardiac electrophysiological examination. Additional peripheral blood samples were obtained from HF patients 1 year after implantation. Based on the change in ejection fraction (EF) at 1 year follow-up, three groups of HF patients were selected: Non-responders (NR; delta EF < 10 and EF at 1 year < 30%), Responders (R; delta EF > 10 and EF at 1 year > 40%), Super-responders (SR; delta EF > 20 and EF at 1 year > 50%). Whole genome expression profiling was performed on age- and gender-matched patients from each group, using 3’ RNA sequencing, followed by quantitative PCR. Results We identified 45 NR patients, 41 R patients and 23 SR patients. A total of 20 patients from each HF group and from the control group could be matched based on age and gender and were selected for 3’ RNA sequencing analysis. Weighted Gene Co-expression Network Analysis (WGCNA) of the data obtained from coronary sinus blood revealed a gene module positively associated with the SR phenotype, containing genes associated with inflammation. Another gene module, negatively associated with the SR phenotype, contained genes associated with oxidative phosphorylation. Direct comparison of SR and NR gene expression data also identified inflammation-related genes with a lower level of expression in NR compared both to SR and to control patients. Oxidative-phosphorylation-related genes were found to be upregulated in all HF patients compared to controls in both coronary sinus and peripheral blood at the start of the study. Analysis of peripheral blood samples obtained at the 1-year time point showed a reset of the expression level of these genes in SR patients, but not in NR and R patients. Conclusions Based on our study population, HF patients who did not respond to CRT (NR group) could be distinguished from patients who did respond to CRT (SR group), by a lower sinus coronary blood expression level of genes associated with inflammation at baseline. In addition, improvement of cardiac function was associated with a downregulation and reset of expression levels of genes involved in oxidative phosphorylation.