Additive Effects of Obesity and TCF7L2 Variants on Risk for Type 2 Diabetes Among Cardiac Patients
- Qing Ling Duan, BSC12,
- Marie-Pierre Dubé, PHD3,
- Nancy Frasure-Smith, PHD4567,
- Amina Barhdadi, PHD3,
- François Lesperance, MD4567,
- Pierre Théroux, MD3,
- Judith St-Onge, DEC1,
- Guy A. Rouleau, MD, PHD, FRCP(C)1 and
- Jeanne M. McCaffery, PHD8
- 1Department of Medicine, Université de Montréal and Centre Hospitalier de l'Université de Montréal, Montréal, Canada
- 2Department of Human Genetics, McGill University, Montréal, Canada
- 3Montreal Heart Institute and Department of Medicine, Université de Montréal, Montréal, Canada
- 4Department of Psychiatry and School of Nursing, McGill University, Montréal, Canada
- 5Research Center, Montreal Heart Institute, Montréal, Canada
- 6Department of Psychiatry, Université de Montréal, Montréal, Canada
- 7Research Center, Centre Hospitalier de l'Université de Montréal, Montréal, Canada
- 8Weight Control and Diabetes Research Center, Brown Medical School and The Miriam Hospital, Providence, Rhode Island
- Address correspondence and reprint requests to Jeanne M. McCaffery, PhD, Weight Control and Diabetes Research Center, 196 Richmond St., Providence, RI 02903. E-mail: jeanne_mccaffery{at}brown.edu
A microsatellite marker, DG10S478, in the transcription factor 7-like 2 (TCF7L2) gene was previously associated with type 2 diabetes in three Caucasian populations (1). This association followed earlier reports by the same group (2) and a separate team (3), which showed suggestive linkage to chromosomal 10q. Grant et al. (1) demonstrated that allele X (a composite of all but the shortest allele) of DG10S478 conferred an increased risk for type 2 diabetes of 45 and 141% among heterozygotes and homozygotes, respectively. Since this report, numerous groups have replicated the association in various populations and extended it to include two intronic single nucleotide polymorphisms (rs12255372 and rs7903146) (4–20). In this study, we investigated the combined effect of obesity and genotype at DG10S478 and rs12255372 in predicting type 2 diabetes risk in a sample of French Canadian cardiac patients.
RESEARCH DESIGN AND METHODS—
Patients of French Canadian descent with established coronary artery disease recruited in two earlier studies, Polymorphisme (n = 484) and the Epidemiological Study of Acute Coronary Syndromes and the Pathophysiology of Emotions (ESCAPE; n = 596) (21), were included. All participants were identified between November 1998 and April 2002 at the Montreal Heart Institute and Hôpital Sacré-Coeur and gave written informed consent. Protocols were approved by the ethics committees at both institutions.
Type 2 diabetes was defined as the use of diabetes medications or fasting blood glucose >126 mg/dl (7.0 mmol/l). BMI was calculated as the weight in kilograms divided by the square of height in meters. Obesity was defined as BMI ≥30 kg/m2.
Genotyping
DNA extraction from blood used a standard protocol (Gentra Systems). DG10S478 was genotyped by radiolabeled (α-35S-dATP) PCR using a standard protocol and primers designed from the genomic sequence of human TCF7L2 (NM_030756; http://www.ncbi.nlm.nih.gov/). Products were separated by electrophoresis on 6% denaturing polyacrylamide gels. Allele sizes and frequencies were obtained from the Centre d'Etude du Polymorphisme Humaine (CEPH) database (http://www.cephb.fr). Genotyping of rs12255372 used TaqMan assays designed by Applied Biosystems. Products were analyzed with a spectrophotometer (Applied Biosystems, Foster City, CA) equipped with sequence detector software (SDS 2.2.2.).
Statistical analysis
Markers were tested for Hardy-Weinberg equilibrium by using the exact test (22). Three genetic models were used for case-control association testing: the χ2 genotype test, the χ2 allele test, and the Cochran-Armitage trend test on genotypes, which tests for additive allele effects on disease risk. To obtain exact P value estimates, 100,000 Monte Carlo permutations were performed. Mantel-Haenszel statistics were used for stratified analyses by obesity. Two-sided P values are reported. Genotype by obesity interaction was evaluated by testing the significance of the interaction term in a full logistic regression model with additive genetic effects modeled as −1 (G/G), 0 (G/T), and 1 (T/T) for the three genotype categories. The association between rs12255372 and BMI was measured using a genotype trend genetic model for an additive allelic effect captured by a regression model for BMI. The model was built using the GLM procedure in SAS, which gives a Fisher's ANOVA and Student's t test for regression coefficients. All data were analyzed using SAS version 9.1.3 and SAS/Genetics (SAS Institute, Cary, NC).
RESULTS—
In contrast to nondiabetic patients, those with diabetes were significantly older, had fewer years of education, were more obese with significantly higher BMI, had greater systolic blood pressure, and were more likely to have undergone coronary artery bypass surgery (P < 0.05).
Genetic association tests
Of the total population, 1,037 (96.2%) were successfully genotyped for DG10S478 and 1,004 (93.1%) for rs12255372. Both markers were in Hardy-Weinberg equillibrium (P = 0.9731 and 0.8557, respectively) and in strong linkage disequilibrium (D′ = 0.977 and r2 = 0.970). Using the nomenclature of Grant et al. (1), allele 0 of DG10S478 was the smallest, most frequent (68.2%), and was strongly associated with allele G of rs12255372. All other alleles (X) of DG10S478 (31.8%) were strongly correlated with allele T of rs12255372.
The rate of diabetes increased with an increasing dose of allele X of DG10S478 or allele T of s12255372 and among obese compared with nonobese participants (Table 1). The association was consistent for both groups but stronger in nonobese coronary artery disease patients. Also, rs12255372 was more closely associated with type 2 diabetes, which may simply reflect the relative instability of the microsatellite.
Logistic regression modeling showed that the interaction between obesity and genotype did not approach statistical significance (P > 0.34). Table 1 depicts the joint effects of these risk factors for type 2 diabetes using the nonobese group without a risk allele as the reference sample. The highest risk was observed among obese individuals who carried at least one risk allele. Within each genotype group, obesity was strongly associated with an increased risk of diabetes (P < 0.001). The regression model also showed an association between rs12255372 and BMI (P = 0.0481) with a decrease of BMI toward the T/T genotype. This association was not detected using a dominant genetic model.
CONCLUSIONS—
Our study confirms that polymorphisms in TCF7L2 (P < 0.0001) and obesity (P < 0.001) are both associated with an increased risk for type 2 diabetes in our French Canadian sample of cardiac patients. We observed that the genetic association was stronger in nonobese patients and that genotype T/T at rs12255372 was associated with reduced BMI (P = 0.0481), confirming previous reports (4,10). Furthermore, we did not detect an interaction between obesity and genotype, suggesting that these are independent risk factors for type 2 diabetes. A recent study by the DECODE (Diabetes Epidemiology: Collaborative Analysis of Diagnostic Criteria in Europe) study group proposed that a second single nucleotide polymorphism, rs7903146, correlated with DG10S478 and rs12255372 and may be the risk variant or its closest known correlate (4). However, previous studies (1,10) showed that none of these associated polymorphisms explain the linkage of type 2 diabetes to this chromosome 10q region, suggesting that another variant(s) nearby accounts for this linkage, which may be functional. Screening mRNA of TCF7L2 might reveal splice variants or alternative exons that were undetected by screening the genomic DNA.
Rate of type 2 diabetes and odds ratios for each genotype stratified by obesity with corresponding genetic association test results
Acknowledgments
This work was funded by the National Institutes of Health (Grant HL077442 to J.M.M.). Q.L.D. is funded by the Heart and Stroke Foundation of Canada. G.A.R. is supported by the Canadian Institutes of Health Research. Collection of data for the ESCAPE study was supported by the Medical Research Council of Canada and an unrestricted grant from GlaxoSmithKline (POP-37744), the Dana Foundation, the Montreal Heart Institute Research Fund, the Pierre David Fund, and the Foundation du Centre Hospitalier de l'Université de Montréal.
We would like to thank all the participants and Dr. Patrick A. Dion for careful review of this manuscript.
Footnotes
-
Published ahead of print at http://care.diabetesjournals.org on 10 March 2007. DOI: 10.2337/dc06-2421.
A table elsewhere in this issue shows conventional and Système International (SI) units and conversion factors for many substances.
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.
-
- Accepted February 28, 2007.
- Received November 7, 2006.
- DIABETES CARE
