Gene–Gene Interaction between PPARδ and PPARγ Is Associated with Abdominal Obesity in a Chinese Population

The peroxisome proliferator-activated receptors (PPARs) -α, -δ/β and -γ are the ligand-activated transcription factors that function as the master regulators of glucose, fatty acid and lipoprotein metabolism, energy balance, cell proliferation and differentiation, inflammation, and atherosclerosis. The objective of the current study was to examine the main and interactive effect of seven single nucleotide polymorphisms (SNPs) of PPARδ/γ in contribution to abdominal obesity. A total of 820 subjects were randomly selected and no individuals were related. The selected SNPs in PPARδ (rs2016520 and rs9794) and PPARγ (rs10865710, rs1805192, rs709158, rs3856806, and rs4684847) were genotyped. Mean difference and 95% confident interval were calculated. Interactions were explored by the method of generalized multifactor dimensionality reduction. After adjustment for gender, age, and smoking status, it was found that the carriers of the C allele (TC + CC) of rs2016520 were associated with a decreased risk of abdominal obesity compared to the carriers of the TT genotype (mean difference = −2.63, 95% CI = −3.61–−1.64, P < 0.0001). A significant two-locus model (P = 0.0107) involving rs2016520 and rs10865710 and a significant three-locus model (P = 0.0107) involving rs2016520, rs9794, and rs1805192 were observed. Overall, the three-locus model had the highest level of testing accuracy (59.85%) and showed a better cross-validation consistency (9/10) than two-locus model. Therefore, for abdominal obesity defined by waist circumference, we chose the three-locus model as the best interaction model. In conclusion, the C allele in rs2016520 was significantly associated with a lower abdominal obesity. Moreover, an interaction among rs2016520, rs1805192, and rs9794 on incident abdominal obesity could be demonstrated.