Genetic variants in LPL, OASL and TOMM40/APOE-C1-C2-C4 genes are associated with multiple cardiovascular-related traits

Abstract

Background: Genome-wide association studies (GWAS) have become a major strategy for genetic dissection of human complex diseases. Analysing multiple phenotypes jointly may improve both our ability to detect genetic variants with multiple effects and our understanding of their common features. Allelic associations for multiple biochemical traits (serum alanine aminotransferase, aspartate aminotransferase, butrylycholinesterase (BCHE), C-reactive protein (CRP), ferritin, gamma glutamyltransferase (GGT), glucose, high-density lipoprotein cholesterol (HDL), insulin, low-density lipoprotein cholesterol (LDL), triglycerides and uric acid), and body-mass index, were examined.

Methods: We aimed to identify common genetic variants affecting more than one of these traits using genome-wide association analysis in 2548 adolescents and 9145 adults from 4986 Australian twin families. Multivariate and univariate associations were performed.

Results: Multivariate analyses identified eight loci, and univariate association analyses confirmed two loci influencing more than one trait at p < 5 × 10-8. These are located on chromosome 8 (LPL gene affecting HDL and triglycerides) and chromosome 19 (TOMM40/APOE-C1-C2-C4 gene cluster affecting LDL and CRP). A locus on chromosome 12 (OASL gene) showed effects on GGT, LDL and CRP. The loci on chromosomes 12 and 19 unexpectedly affected LDL cholesterol and CRP in opposite directions.

Conclusions: We identified three possible loci that may affect multiple traits and validated 17 previously-reported loci. Our study demonstrated the usefulness of examining multiple phenotypes jointly and highlights an anomalous effect on CRP, which is increasingly recognised as a marker of cardiovascular risk as well as of inflammation.

Figure 1

Manhattan plots for multivariate QTL analysis in unrelated-subject data (N = 4986) for the 13 traits. Genomic position is on the x-axis and the -log₁₀ of the association p-value is on the y-axis. Points with p-value of 5 × 10^-8 are shown in red.

Figure 2

Radar chart of polymorphisms on chromosome8 (a), chromosome 12 (b) and chromosome 19 (c). Each dot on the plot represents the standardized beta (1-unit change per copy increment of the minor allele) of each trait from univariate testing.

Figure 3

Q-Q plot of multivariate analysis. Black points correspond to SNPs included in the analyses. The 45° line refers to no significant association. The dotted line corresponds to p-value of 5 × 10^-8. "Excluded" line is where SNPs that were found in significant regions (genes) in univariate analyses were removed.