A systematic analysis highlights multiple long non-coding RNAs associated with cardiometabolic disorders

Abstract

Genome-wide association studies (GWAS) have identified many susceptibility loci for cardiometabolic disorders. Most of the associated variants reside in non-coding regions of the genome including long non-coding RNAs (lncRNAs), which are thought to play critical roles in diverse biological processes. Here, we leveraged data from the available GWAS meta-analyses on lipid and obesity-related traits, blood pressure, type 2 diabetes, and coronary artery disease and identified 179 associated single-nucleotide polymorphisms (SNPs) in 102 lncRNAs (p-value < 2.3 × 10^-7). Of these, 55 SNPs, either the lead SNP or in strong linkage disequilibrium with the lead SNP in the related loci, were selected for further investigations. Our in silico predictions and functional annotations of the SNPs as well as expression and DNA methylation analysis of their lncRNAs demonstrated several lncRNAs that fulfilled predefined criteria for being potential functional targets. In particular, we found evidence suggesting that LOC157273 (at 8p23.1) is involved in regulating serum lipid-cholesterol. Our results showed that rs4841132 in the second exon and cg17371580 in the promoter region of LOC157273 are associated with lipids; the lncRNA is expressed in liver and associates with the expression of its nearby coding gene, PPP1R3B. Collectively, we highlight a number of loci associated with cardiometabolic disorders for which the association may act through lncRNAs.

Fig. 1

Identification of lncRNAs associated with cardiometabolic disorders. This flowchart shows our approach to identify lncRNA SNPs that are associated with cardiometabolic disorders and to provide additional evidence supporting the identfied associations. GWAS genome-wide association studies, SNP single-nucleotide polymorphism, eQTL expression quantitative trait loci, MAF minor allele frequency, LD linkage disequilibrium, miRNA microRNA, lncRNA long non-coding RNA, CMD cardiometabolic disorders.

Fig. 2

Heat map of the enrichment of lncRNA SNPs in the potential regulatory regions. This heat map was constructed by calculating the percentage of 55 lncRNA SNPs and their proxies in strong LD (R²>0.8) located in DNA regulatory regions as defined by the chromatin states using the Roadmap epigenomes data set across tissues relevant to cardiovascular disease (e.g., aorta, fetal heart, atrium, and ventricle), metabolic disorders (e.g., liver, pancreas, and adipose), and also control tissues (e.g., lung and brain). The lncRNA SNPs which are top variants in their loci associated with cardiometabolic disorders are shown in bold. As shown, almost all SNPs are not over-represented in the regulatory regions related to cardiometabolic disorders. Rs4841132 in lncRP11–10A14.4 (LOC157273) was the only SNP that the (potential) regulatory region, in which the SNP is located, was enriched for the relevant cell type (liver).

Fig. 3

Genetic and functional evidence suggesting LOC157273 to be a lipid susceptibility gene. a The regional association plots show the association of variants in LOC157273 with HDL and LDL cholesterol. b A figure generated by the UCSC Genome Browser on 8p23.1 locus hosting LOC157273. The position of SNP rs4841132 within the second exon of LOC157273 and CpG site cg17371580 within the conserved promoter region of lncRNA are depicted by red arrows. The figure also illustrates the expression of LOC157273 across different tissues using the Human Body Map catalog, indicating that the lncRNA is expressed in liver. c A schematic showing the lncRNA structure, the nearest protein coding gene PPP1R3B, the position of SNPs and a CpG site in the lncRNA associated with lipid traits.