Abstract 387: Weighted Gene Co-Expression Network Analysis Identification and Experimental Validation of Genetic Determinants of Right Ventricular Function

Christoph Rau,Katie Lee,Jeffrey Chukwuneke,Liu Yanan,Li Qing,Wang Yibin,Emily J. Tsai

DOI: https://doi.org/10.1161/res.121.suppl_1.387

IF: 23.213

2017-01-01

Circulation Research

Abstract:Right ventricular dysfunction (RVD) portends worse outcomes in chronic heart failure (HF), irrespective of left ventricular function. Despite the significance and prevalence of RVD, its molecular pathobiology is poorly understood. Gene networks can identify and explore complex relationships between genes, gene pathways, and phenotypes. We used weighted gene co-expression network analysis (WGNA) of human heart samples to identify genetic determinants of RV function and validated candidate genes in animal models of RV versus LV dysfunction. We isolated and sequenced total RNA from explanted left and right ventricles of ischemic cardiomyopathy (ICM) patients (n=10) who underwent cardiac transplantation. Non-failing donor hearts that were declined for transplant were used as controls (n=5). Half the ICM patients had severe RV failure (RVF) based upon hemodynamic indices (mean±SEM) with mean RA 25.8±1.7 mmHg, RA:PCWP ratio 1.05±0.11, and PA pulsatility index 1.6±0.2. Using direct gene membership comparison and z-score analysis, we found that more than half the modules in RV-specific networks are shared by combined RV-LV gene networks. We used principle component analysis to examine modules that correlated with RVF-related phenotypes. Two modules unique to the RV network identified selenium-binding protein gene Selenbp1 as a potential driver of RVF. We also identified a strong association between anti-apoptotic gene Mif and vesicle fusion gene Snap47 with RA:PCWP and PA pulsatility indices, suggesting potential important roles of these genes in RVF. Finally, peroxiredoxin 5 PRDX5 and E3 ubiquitin-protein ligase SHPRH were associated with protection against RVF. We sought to validate these candidate gene drivers of RV function in mouse models of RV versus LV-specific dysfunction induced by pulmonary artery banding (PAB) and transverse aortic constriction (TAC), respectively. PRDX5 expression was markedly reduced in RV of PAB versus Sham but unchanged in LV of TAC versus Sham. Selenbp1 was diminished in both RV of PAB and LV of TAC mice, relative to Sham. Expression of other candidate genes did not vary significantly with PAB or TAC. Our preliminary findings suggest that cardiac PRDX5 may play an important role in regulating RV function.