Haplotype of growth hormone and angiotensin I-converting enzyme genes, serum angiotensin I-converting enzyme and ventricular growth: pathway inference in pharmacogenetics.

Objectives An insertion/deletion (I/D) polymorphism in the angiotensin I-converting enzyme (ACE) gene is associated with variations in circulating and tissue angiotensin I-converting enzyme activity, and differences in exercise-induced left ventricular hypertrophic response. A genetic marker (CSH1.01) in the syntenic GH-CSH gene cluster correlates with metabolic syndrome in adult life in males. Approximately 24% linkage disequilibrium between CSH1.01T and D alleles of ACE I/D has also been reported. The objective was to examine the hypothesis that effects ascribed to ACE genotype may reflect causality in the GH-CSH cluster. Methods The ACE I/D polymorphism and GH-CSH BgIII-B single nucleotide polymorphism (in strong linkage disequilibrium with CSH1.01) were determined in 847 British Army recruits. Serum angiotensin I-converting enzyme activity and left ventricular mass were measured before and after a 12-week physical training program. Genotype and haplotype analyses of both markers were performed in relation to these phenotypes. Results The ACE I/D polymorphism was in linkage disequilibrium with BgIII-B (D'=0.3). Strong association was seen between ACE I/D genotypes and serum angiotensin I-converting enzyme activity (P < 0.0001), but not left ventricular mass change. BgIII-B genotypes also associated significantly with serum angiotensin I-converting enzyme level (P < 0.0001). Haplotype analysis, however, showed that most of this association resulted from linkage disequilibrium between BgIII-B and ACE I/D. BgIII-B did not associate with left ventricular mass change. Conclusions GH-CSH BgIII-B genotype associates significantly with angiotensin I-converting enzyme levels, but only through linkage disequilibrium with ACE I/D. Every phenotype with which ACE I/D has been associated merits investigation of potential causal effects originating in the GH-CSH cluster (and vice versa), otherwise the chain of causality could be misinterpreted.