A directed network analysis of the cardiome identifies molecular pathways contributing to the development of HFpEF

Abstract

Aims

The metabolic syndrome and associated comorbidities, like diabetes, hypertension and obesity, have been implicated in the development of heart failure with preserved ejection fraction (HFpEF). The molecular mechanisms underlying the development of HFpEF remain to be elucidated. We developed a cardiome-directed network analysis and applied this to high throughput cardiac RNA-sequencing data from a well-established rat model of HFpEF, the obese and hypertensive ZSF1 rat. With this novel system biology approach, we explored the mechanisms underlying HFpEF.

Methods and results

Unlike ZSF1-Lean, ZSF1-Obese and ZSF1-Obese rats fed with a high-fat diet (HFD) developed diastolic dysfunction and reduced exercise capacity. The number of differentially expressed genes amounted to 1591 and 1961 for the ZSF1-Obese vs. Lean and ZSF1-Obese+HFD vs. Lean comparison, respectively. For the cardiome-directed network analysis (CDNA) eleven biological processes related to cardiac disease were selected and used as input for the STRING protein-protein interaction database. The resulting STRING network comprised 3.460 genes and 186.653 edges. Subsequently differentially expressed genes were projected onto this network. The connectivity between the core processes within the network was assessed and important bottleneck and hub genes were identified based on their network topology.

Classical gene enrichment analysis highlighted many processes related to mitochondrial oxidative metabolism. The CDNA indicated high interconnectivity between five core processes: endothelial function, inflammation, apoptosis/autophagy, sarcomere/cytoskeleton and extracellular matrix. The transcription factors Myc and Peroxisome Proliferator-Activated Receptor-α (Ppara) were identified as important bottlenecks in the overall network topology, with Ppara acting as important link between cardiac metabolism, inflammation and endothelial function.

Conclusions

This study presents a novel systems biology approach, directly applicable to other cardiac disease-related transcriptome data sets. The CDNA approach enabled the identification of critical processes and genes, including Myc and Ppara, that are putatively involved in the development of HFpEF.

Introduction

Nearly 50% of all heart failure (HF) patients suffer from heart failure with preserved ejection fraction (HFpEF). HFpEF is characterized by left ventricular (LV) diastolic dysfunction resulting due to impaired LV relaxation and/or filling. This form of HF is frequently seen in patients suffering from co-morbidities that are associated with the metabolic syndrome, namely obesity, hyperlipidaemia, diabetes mellitus and hypertension [1]. This association tentatively suggests that systemic metabolic derangements predispose to the development of HFpEF. Thus diastolic dysfunction as such has been attributed to disturbances in myocardial relaxation, calcium homeostasis abnormalities [2], as well as to increased passive stiffness either within [3] or outside the cardiomyocyte [4]. These changes are secondary to the changes in systemic metabolism and the chronic low-grade inflammation that accompany the metabolic syndrome [5]. One of the prevailing current hypotheses is that the low-grade systemic inflammation leads to coronary microvascular dysfunction, which eventually impacts cardiomyocyte stiffness [6]. However, the pathogenesis of HFpEF remains enigmatic and is subject of intense investigation.

The application of system biology approaches to ‘omics’ data has been proven useful in the identification of biological pathways and genes involved in disease processes [7,8]. The aim of the present study was to develop a novel, cardiome-directed network analysis approach to analyse genome-wide transcriptome data in order to gain insight into the molecular mechanisms underlying HFpEF. Thereto, we performed RNA sequencing on cardiac tissue of ZSF1 rats, an extensively characterized preclinical model of metabolic risk–mediated HFpEF [[9], [10], [11], [12]]. ZSF1-Obese rats are the F1 hybrid offspring of Zucker Diabetic Fatty (ZDF) rats and Spontaneously Hypertensive Heart Failure (SHHF) rats [13], each having a distinct leptin receptor mutation. ZSF1-Obese rats, carrying both leptin receptor mutations, are characterized by morbid obesity, diabetes mellitus and hypertension and have been shown to develop diastolic dysfunction and HFpEF [9]. In contrast, ZSF1-Lean rats only suffer from hypertension and do not develop diastolic dysfunction. A subgroup of ZSF1-Obese rats received a high-fat diet (HFD) to assess if an additional metabolic challenge would exacerbate HFpEF progression. Changes in systemic metabolism and cardiac geometry and function were monitored.

To interrogate these transcriptome data sets we developed a novel network analysis approach, specifically tailored for cardiome data sets, to identify transcripts and biological processes that are putatively involved in the development of HFpEF. To this end, eleven processes that have previously been linked to cardiac pathogenesis were predefined and selected for further analysis. The publicly accessible protein-protein interaction (PPI) database STRING [14] was used to construct a cardiac PPI network linked to these eleven processes. This network formed the basis to identify cellular processes and transcripts putatively involved in the development of HFpEF.