Branched chain amino acids (BCAA) plasma levels represent potential biomarkers for development of type 2 diabetes (T2D). Mendelian randomization and prospective studies indicated causal relationship between SNPs in PPM1K (protein phosphatase Mg2+/Mn2+ dependent 1K) gene (Chr 4q22.1) and increased risk of incident T2D while other studies, including ours suggested the contribution of MUT (metylmalonyl-CoA mutase) gene (Chr 6p12.3) in insulin resistance (IR). To define genetic markers we performed fine-scale haplotype mapping of PPM1K in two populations (French and Romanians) with metabolic syndrome (n = 465). SNPs were genotyped with Affymetrix and enriched by imputation at 1000 genome density (BEAGLE 4.1). Haplotype association was performed by sliding window (4 SNPs windowing) and analyzed in HAPLOVIEW and PHASE 2.1. Previous leader SNPs in PPM1K (rs7678928 and rs1440581) were not associated with IR in our population but haplotype mapping revealed two major signals downstream (GGGC) and upstream (ACAG), centered by rs71609524 and rs4423843 with P < 3.76 x 10-3 (OR 1.59, 95% CI [1.13 - 2.23]) and < 4.72x10-3 (OR 1.50, 95% CI [1.12 - 2.01]), respectively. Values were weaker that those from MUT gene (P-value of the most significant haplotype, centered by rs2167284 was < 10-5). Total BCAA plasma levels were associated with PPM1K by several SNPs, including rs7656367 and rs4693955 with P < 5 x 10-3, OR of 0.41 and 3.22, respectively. Interestingly, correlation trend test with leucine, isoleucine and valine were better correlated with SNPs in the MUT gene: rs10948396 was associated with P < 7.91 x 10-5 (Bonferroni 6 x 10-2) and confirmed by correlation with total BCAA levels (P < 1.60 x 10-6, Bonferroni 1.22 x 10-3). These data suggest synergistic effects and indicate that new valuable genetic markers can be defined for clinical trials using fine-scale haplotype mapping, thus revealing complex relationship between genes in the determinism of IR and BCAA metabolism.
S. Haydar: None. C. Lautier: None. F. Grigorescu: None.