Linkage and association analyses of longitudinally measured lipid phenotypes in adolescence

Abstract

The genetic basis of cardiovascular disease (CVD) is complex and still largely elusive. Plasma lipid concentrations are well-established risk factors for cardiovascular disease (CVD), and have adult heritabilities ranging from 0.48 to 0.87. Estimates for adolescents are slightly higher (range 0.71 to 0.82). To identify loci affecting lipid concentrations across adolescence, we analyzed longitudinal lipid data in a sample of 134 monozygotic and 626 dizygotic twin pairs at ages twelve, fourteen and sixteen, and their siblings, from 760 Australian families. Univariate linkage analysis for each phenotype and time point was supplemented by multivariate analysis across the time points. A genome-wide association scan was also performed on a subset of the subjects (N = 441). The strongest linkage was seen for triglycerides on chromosome 6p24.3 (multivariate -log10 p = 6.81; equivalent LOD = 6.13; p = 1.55 × 10-7). Significant linkage was also found for LDL cholesterol on chromosome 2q35 (multivariate -log10p = 5.59; equivalent LOD = 4.53; p = 2.57 × 10-6). In the association analysis, rs10503840 on 8p21.1 was significantly associated with total cholesterol levels at age fourteen (p = 8.24 × 10-7, estimated significance threshold 2.45 × 10-6). Association at p < 2.25 × 10 -6 was also found between triglycerides at age 12 and rs10507266, in an intron of THRAP2 (MIM 608771) on 12q24.21; and between HDL-C at age 14 and rs10506325 in an intergenic region of 12q13.13. Suggestive evidence of association at ages twelve and fourteen was found between HDL-C and rs10492859 on 16q23 (p = 2.42 × 10-5 and 2.77 × 10-4, respectively). Further longitudinal genetic studies of cardiovascular risk factors, focused on critical periods of development or change, are needed.