The contribution of non-coding regulatory elements to cardiovascular disease

Abstract

Cardiovascular disease collectively accounts for a quarter of deaths worldwide. Genome-wide association studies across a range of cardiovascular traits and pathologies have highlighted the prevalence of common non-coding genetic variants within candidate loci. Here, we review genetic, epigenomic and molecular approaches to investigate the contribution of non-coding regulatory elements in cardiovascular biology. We then discuss recent insights on the emerging role of non-coding variation in predisposition to cardiovascular disease, with a focus on novel mechanistic examples from functional genomics studies. Lastly, we consider the clinical significance of these findings at present, and some of the current challenges facing the field.

Keywords: GWAS; cardiovascular; epigenetics; functional genomics; gene regulation.

Figure 1.

Summary of genome-wide association signals and proximal genes for major cardiovascular diseases and traits. (a) Total number of associated single nucleotide polymorphisms (SNPs) for coronary artery disease (CAD), atrial fibrillation (AF), QT interval duration (QT), idiopathic dilated cardiomyopathy (DCM) and congenital heart disease (CHD). For each disease or trait, top solid-outline boxes correspond to SNPs at genome-wide significance with p-value lower than 1 × 10⁻⁸, and bottom dashed-outline boxes to sub-threshold SNPs (p-value < 1 × 10⁻⁴). The dashed red line denotes the genome-wide significance threshold. SNPs located in coding regions of the genome are represented as black bars, and with those in non-coding segments in light grey. Numbers on top of each bar indicate the total number of associated SNPs at a genome-wide level of significance (p-value < 1 × 10⁻⁸), and percentages the non-coding fraction across all associated SNPs. Data from [5]. (b) Genes proximal to the associated SNPs in (a). (p-value < 1 × 10⁻⁸) are represented as a Venn diagram for CAD, AF and QT (centre). Numbers indicate gene counts in each region of the plot, including genes associated with two traits/diseases. The boxes linked to each region summarize example genes proximal to GWAS signals, and the functional categories they belong to.

Figure 2.

Experimental approaches for functional investigation of non-coding elements associated with cardiovascular disease. (a) Epigenomic annotation of the SCN5A/SCN10A cardiac disease locus in human and mouse left ventricle. Human genome tracks show epigenome signals for three histone marks associated with regulatory activity (H3K4me3 in blue; H3K27ac in orange and H3K4me1 in green) [82]; GWAS lead SNPs in this locus (black bars); and putative promoters (purple) and enhancers (orange). Mouse genome tracks below show ChIP-seq data for cardiac transcription factors GATA4 (blue), NKX2–5 (green) and TBX3 (purple) [88]. Orthologous promoters and enhancers in the human and mouse loci are connected by light purple and orange guides, respectively. (b) Epigenomic annotation of the KCNH2 QT interval locus in the human left ventricle (epigenome signals as in (a)). Bottom tracks show genetic variants associated with QT interval duration (black bars), and long range interactions between the KCNH2 promoter (blue) and enhancer elements (red) [89]. (c) Genomic location of the lead genetic variant rs2595104, associated with atrial fibrillation (AF) and located upstream of the PITX2c transcript annotation. The grey inset shows the sequence context of the variant, the minor risk allele (0.31 frequency), the major protective allele (0.69 frequency) and the p-value of the AF association. The regulatory effect of this variant on PITX2c expression was analysed in iPSC-derived cardiomyocytes by CRISPR/Cas9 deletion of a 100 bp sequence encompassing rs2595104 (left barplot, 54% reduction); and also by CRISPR genetic editing producing isogenic cardiomyocytes carrying the major non-risk allele and the minor risk allele (right barplot, 27% difference). Adapted with permission from [59].