Single Nucleotide Polymorphisms of the Peroxisome Proliferator–Activated Receptor-α Gene (PPARA) Influence the Conversion From Impaired Glucose Tolerance to Type 2 Diabetes: The STOP-NIDDM Trial

Laura Andrulionyte 1 Teemu Kuulasmaa 1 Jean-Louis Chiasson 0 Markku Laakso 1 for the STOP-NIDDM Study Group 2 0 Research Centre, Centre Hospitalier de l'Universite de Montre al, Ho tel-Dieu, and Department of Medicine, Univer- sity of Montreal , Quebec , Canada. Academy Professor, Department of Medicine, University of Kuopio and Kuopio University Hospital , 70210 Kuopio , Finland 1 Department of Medicine, University of Kuopio and Kuopio Univer- sity Hospital , Kuopio , Finland; and the 2 A complete list of STOP-NIDDM trial members can be found in ref. 11. 2007 by the American Diabetes Association. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact Peroxisome proliferator-activated receptor (PPAR) , a transcription factor of the nuclear receptor superfamily, regulates fatty acid oxidation. We evaluated the association of single nucleotide polymorphisms (SNPs) of the PPAR- gene (PPARA) with the conversion from impaired glucose tolerance to type 2 diabetes in 767 subjects of the STOP-NIDDM trial in order to investigate the effect of acarbose in comparison with placebo on the prevention of diabetes. In the placebo group, the G (162V) allele of rs1800206 increased the risk for diabetes by 1.9-fold (95% CI 1.05-3.58) and was associated with elevated levels of plasma glucose and insulin. The effect of this allele on the risk of diabetes in the placebo group was enhanced by the simultaneous presence of the risk alleles of the PPAR-2, PPAR- coactivator 1, and hepatic nuclear factor 4 genes (odds ratios 2.2, 2.5, and 3.4, respectively). In the acarbose group, subjects carrying the minor G allele of rs4253776 and the CC genotype of rs4253778 of PPARA had a 1.7- and 2.7-fold increased risk for diabetes. Our data indicate that SNPs of PPARA increase the risk of type 2 diabetes alone and in combination with the SNPs of other genes acting closely with PPAR-. Diabetes 56:1181-1186, 2007 - Kunder the transcriptional control of peroxisome ey proteins involved in lipid metabolism are proliferatoractivated receptor (PPAR) (1). Among the three PPAR subtypes, PPAR- is found predominantly in the liver, kidney, and heart, where it upregulates the expression of genes involved in fatty acid metabolism (1), particularly when activated by PPAR- coactivator 1 (PGC-1) (2). PPAR- agonists lower plasma lipid levels, decrease intrahepatic and skeletal muscle lipid accumulation and adiposity, and normalize glucose and insulin concentrations, therefore markedly reducing insulin resistance and the risk of type 2 diabetes (35) in various rodent models of type 2 diabetes and insulin resistance, whereas gemfibrozil and fenofibrate can improve insulin sensitivity in humans (6,7). Along with regulation of lipid and glucose metabolism, PPAR- is as an attractive candidate gene for type 2 diabetes. Among previous studies (8,9), only one crosssectional study has reported an association of haplotype of PPARA with the age of diabetes diagnosis (10). Therefore, we evaluated the association of single nucleotide polymorphisms (SNPs) of PPARA with the conversion from impaired glucose tolerance (IGT) to type 2 diabetes in subjects of the STOP-NIDDM trial. Furthermore, the effects of SNPs of PPARA, along with the SNPs of PPARcoactivator 1 (PGC-1A), PPAR-2, (PPARG2), and hepatic nuclear factor 4 (HNF4A) on the conversion to diabetes were investigated. RESEARCH DESIGN AND METHODS The STOP-NIDDM trial was a double-blind, placebo-controlled study that randomized 1,429 subjects with IGT to either acarbose or placebo groups (11). Annual oral glucose tolerance tests were performed to evaluate the conversion to diabetes. The entire population was followed up on an average of 3.3 1.2 years. DNA was available from 767 subjects from seven countries (385 men and 382 women; 354 treated with acarbose and 413 with placebo). Their mean BMI was 30.8 4.1 kg/m2 and age 54.7 7.9 years. All participants signed an informed consent form, and the study was approved by appropriate institutional review boards. DNA analyses. We screened three SNPs (rs135559, rs1800206, and rs4253778) from the study of Flavell et al. (10). In addition, eight SNPs were selected using the Tagger software (12) (http://www.broad.mit.edu/mpg/tagger/faq. FIG. 1. Schematic map of the 83.7-kb genomic region showing the approximate location of 11 SNPs genotyped in PPARA. A: PPARA messenger RNA. Arrows indicate the start and finish sites of translation (modified from 18). B: LD statistics (r2) among the SNPs of PPARA and the minor allele frequency (MAF). SNPs are coded by National Centre for Biotechnology Information dbSNP accession numbers. html) and data available from the HapMap project (13) (http://www.hapmap.org; Data Release #20/phase II, January 2006). In the genomic region of 100 kilo base pairs (kbp) including the 93.2-kbp PPARA locus, 10 of all 11 selected SNPs captured up to 70% of common variants with minor allele frequency 5%; r2 0.8 (rs135539 not genotyped in HapMap). Screening of SNPs of PPARA was performed with TaqMan Allelic Discrimination Assays (Applied Biosystems). Genotyping reaction was amplified on a GeneAmp PCR system 2700 (95C for 10 min, followed by 40 cycles of 95C for 15 s and 60C for 1 min), and fluorescence was detected on an ABI Prism 7,000 sequence detector (Applied Biosystems). The genotyping success rate was 100%, and rescreening of 7% of subjects gave 100% identical results. TaqMan Allelic Discrimination Assays were also used to genotype 8 SNPs of HNF4A (14), of which rs4801424 G/C (C risk allele), rs2425637 G/T (G risk allele), rs2071197 G/A (G risk allele), and rs3818247 G/T (T risk allele) have been associated with the conversion to type 2 diabetes (14). Screening methods for polymorphisms of PGC-1A (rs8192678 G/A: Gly482Ser; 482Ser risk allele) and PPARG2 (rs1801282 C/G: Pro12Ala; Pro12Pro risk genotype) have been previously described (15). Statistical analysis. The SPSS program (SPSS, Chicago, IL) version 11.0 for Windows was used in data analysis. Results are expressed as either means SD or percentages. Not normally distributed parameters were logarithmically transformed. Categorical variables were compared using the 2 and Fishers exact tests under the dominant and recessive models and continuous variables with the two-tailed Students t test, ANOVA, or nonparametric tests when appropriate. Logistic regression analyses and further adjustment for appropriate covariates were applied to evaluate whether the SNPs predicted the development of type 2 diabetes. P values 0.05 were considered statistically significant. The Haploview program (16) (available at http://www.broad. mit.edu/mpg/haploview/) was used to calculate and visualize the linkage disequilibrium (LD) statistics and haplotype blocks among the SNPs. Haplotype estimation from unrelated individual (...truncated)