Positively Selected PPARG and GIP Variants Impart Phenotypic Variations after a Glucose Challenge Test in Pregnant Women.

Abstract

Recent studies have revealed that variants within or surrounding more than three dozen genes are associated with risk of diabetes, obesity, and/or a number of metabolic parameters. However, genetic variants responsible for variations in energy-balance regulation during pregnancy are largely unknown. To uncover common single nucleotide polymorphisms (SNPs) that may contribute to phenotypic variations in energy-balance regulation, or impaired glucose metabolism during human pregnancy, we hypothesized variants that confer adaptation in energy-balance regulation are associated with a partial selection because selection pressures associated with such variants likely fluctuated over time in response to recent changes in human culture. To identify SNPs that were subject to moderate environmental selection, we screened for variants that have moderately divergent allele frequencies and extended haplotypes in select human population(s) using the HapMap II and III dataset. To investigate whether the differentially selected variants contribute to phenotypic variations in energy-balance regulation during pregnancy, we analyzed levels of sugar and serum hormones in 123 unrelated pregnant women who underwent a 1-hr oral glucose tolerance test for gestational diabetes during the 23rd-29th weeks of pregnancy. Based on the screening of >1000 candidate genes, we identified three dozen SNPs in CDKAL1, CYB5R4, GAD2, GIP, and PPARG as prime candidates. Among these SNPs, GIP and PPARG variants were positively selected in European and Asian populations. Based on functional testing, we showed that more than two dozen GIP variants within a selected haplotype impart phenotypic variations in GIP response and glucose metabolism after the 1-hr OGTT (p<0.05). In addition, we found that PPARG variants (rs2920500 and rs2920502) are associated with serum insulin levels (p<0.05), but not levels of GIP or C-peptide. Our study demonstrated that genetic variants stemmed from prior gene–environmental interactions in GIP and PPARG are associated with glucose and/or insulin metabolism during pregnancy. Because PPARG has been associated with the development of T2DM and that PPARG modulates GIP receptor expression in beta cell, these data indicate that the enteroinsular axis in humans and the treatment of T2DM by PPARG ligands (e.g., thiazolidinedione) could be subject to heterogeneous regulation by multiple adaptive variants.