Genetic analysis of 103 candidate genes for coronary artery disease and associated phenotypes in a founder population reveals a new association between endothelin-1 and high-density lipoprotein cholesterol

Abstract

Coronary artery disease (CAD) is a major health concern in both developed and developing countries. With a heritability estimated at ~50%, there is a strong rationale to better define the genetic contribution to CAD. This project involves the analysis of 884 individuals from 142 families (with average sibships of 5.7) as well as 558 case and control subjects from the Saguenay Lac St-Jean region of northeastern Quebec, with the use of 1,536 single-nucleotide polymorphisms (SNPs) in 103 candidate genes for CAD. By use of clusters of SNPs to generate multiallelic haplotypes at candidate loci for segregation studies within families, suggestive linkage for high-density lipoprotein (HDL) cholesterol is observed on chromosome 1p36.22. Furthermore, several associations that remain significant after Bonferroni correction are observed with lipoprotein-related traits as well as plasma concentrations of adiponectin. Of note, HDL cholesterol levels are associated with an amino acid substitution (lysine/asparagine) at codon 198 (rs5370) of endothelin-1 (EDN1) in a sex-specific manner, as well as with a SNP (rs2292318) located 7.7 kb upstream of lecithin cholesterol acyl-transferase (LCAT). Whereas the other observed associations are described in the current literature, these two are new. Using an independent validation sample of 806 individuals, we confirm the EDN1 association (P<.005), whereas the LCAT association was nonsignificant (P=.12).

Figure 1.

Overview of the association analysis. The family and case-control samples were combined to perform an association analysis (i.e., the joint analysis). Twelve associations proved to be statistically significant. Of these, 10 are either known in the literature or in LD with known functional SNPs. The two remaining associations are new and thus were tested in the validation sample. The association between rs5370 and HDL cholesterol concentrations was confirmed in the validation sample.

Figure 2.

Association between HDL cholesterol concentrations and rs2292318 (in LCAT). Age- and sex-adjusted, log-transformed HDL cholesterol concentrations are shown as a function of rs2292318 genotype (for family and case-control samples combined). A box-and-whisker plot is used; by graphically displaying the median and the end of the 1st and 3rd quartiles, these plots provide a robust assessment of the associations observed. Overall, 1,106 G/G, 307 A/G, and 19 A/A individuals are represented (P=.00002).

Figure 3.

Association between the HDL cholesterol concentrations and rs5370 (in EDN1). Age- and sex-adjusted, log-transformed HDL cholesterol concentrations are shown as a function of rs5370 genotype with both sexes combined (“Total”). Overall, 894 G/G, 474 T/G, and 64 T/T individuals are represented using a box-and-whisker plot (P=.00001). Data are also shown for both sexes separately—525 G/G, 256 T/G, and 28 T/T males are represented (P=.14), and 369 G/G, 218 T/G, and 36 T/T females are represented (P=.00001). All three box-and-whisker plots were made by combining the family and case-control samples. By graphically displaying the median and the end of the 1st and 3rd quartiles, these plots provide a robust assessment of the associations observed.