ADIPOQ, PPAR, LEP, TCF7L2, FTO及PTPN1基因多型性與代謝症候群之關聯與基因-基因交互作用: 社區為基礎之病例對照研究

Abstract

Background: A growing focus in genetic studies of complex diseases is how to detect the joint effects of susceptibility genes. A common method is generalized multifactor dimensionality reduction (GMDR), which is Single Nucleotide Polymorphism (SNP)-based and has good power in identifying high-order gene-gene interactions. However, the functional relevance among the SNPs in the same gene may result in false negatives in using the GMDR. Instead, the gene-based method has been proposed to capture interactions that may involve such functional variants within a gene. Objective: 1) To investigate whether there exist associations with the metabolic syndrome for individual SNPs in each of six genes that have been implicated because of their association with some components of the metabolic syndrome, including ADIPOQ, LEP, PPARG, FTO, TCF7L2, and PTPN1; and 2) to search for gene-gene interactions among these genetic loci on the metabolic syndrome by means of the GMDR as well as the gene-based analysis. Methods: A case-control study among participants of health check-up, aged 40 years or above, was carried out in Northern Taiwan, with a total of 611 cases with metabolic syndrome and 1117 controls. Among 19 SNPs of the six genes selected for this study, multivariable logistic regression analyses were used to estimate the association between individual SNPs and metabolic syndrome with adjustment for sex and age. Gene-gene interactions were then investigated using both the GMDR method and a gene-based approach that utilities the genotype-trait distortion score. Results: Individual SNPs did not exhibit significant association with metabolic syndrome. The gene-gene interaction analyses using the GMDR showed that a combination of SNPs in ADIPOQ, TCF7L2, and PTPN1 was the best model, with a prediction accuracy of 0.54 and a cross-validation consistency of 4/10 (permutation P = 0.01). Meanwhile, the gene-based analysis indicated that PTPN1 interacted with FTO, regardless of the type of ratio statistics (mean-ratio or quantile ratio) or the method used in assessing significance level (curve method or rank method), or with ADIPOQ if the curve method was used for the quantile ratio. Conclusions: Our results suggest that PTPN1 is an important gene that may interact with FTO, ADIPOQ, and TCF7L2 in exerting their influence on the metabolic syndrome. The underlying mechanism of the interaction warrants further investigation and may lead to new insights about the occurrence of the metabolic syndrome.