Analysis of lipoprotein lipase haplotypes reveals associations not apparent from analysis of the constituent loci.

Abstract

Simultaneously analysing genotype effects at several closely-linked loci may be preferable to analysing them separately, but can be difficult, due to multiple genotype classes, small class sizes, and non-independence induced by associations among loci. Analysis of haplotype effects offers an alternative approach. We studied effects of haplotypes comprising 3 loci (5' to 3': PvuII, HindIII, and Ser 447 -Stop) in the lipoprotein lipase (LPL) gene on plasma lipid levels and LPL activity, in 807 Dutch males with coronary atherosclerosis. We analysed haplotype effects in individuals for whom haplotypes could either be determined unequivocally or inferred with high probability, using contrasts suggested by likely evolutionary relationships among the haplotypes. One haplotype was associated with significantly higher total cholesterol, while another was associated with significantly lower triglyceride levels. Though these two haplotypes had generally opposite effects on lipids, both were associated with significantly higher LPL activity. In genotype analyses, the HindIII (-) allele was associated with higher LPL activity; however, one haplotype bearing it had no significant effect on LPL activity. Haplotypes thus provided more information than genotypes alone would have. The two haplotypes with consistently different effects on lipid levels despite similar effects on LPL activity, provide further evidence that aspects of LPL biology, apart from its catalytic function in lipolysis, may mediate its effects on plasma lipids at least in coronary artery disease patients.