Interaction between genome-wide variants and physical activity on body mass index: a meta-analysis of 109,924 individuals. T. O. Kilpeläinen1,2, R. A. Scott2, A. Mahajan3, L. Xue4, F. Renström5, M. Graff6, D. Hadley7, T. Workalemahu8, M. den Hoed2,9, T. W. Winkler10, A. Y. Chu11, N. L. Heard-Costa12, T. Haller13, T. S. Ahluwalia1, J. V. van Vliet-Ostaptchouk14, P. J. van der Most14, M. L. Grove15, L. Quaye16, S. Snitker17, E. J. C. de Geus18, T. Lehtimäki19, L. Qi8, P. W. Franks5,8, I. B. Borecki20, K. Monda6, D. I. Chasman11, K. E. North6, L. A. Cupples4,12, R. J. F. Loos2,21, The GIANT Consortium 1) The Novo Nordisk Foundation Center for Basic Metabolic Research, Section of Metabolic Genetics, University of Copenhagen, Denmark; 2) MRC Epidemiology Unit, Institute of Metabolic Science, University of Cambridge, UK; 3) Wellcome Trust Centre for Human Genetics, University of Oxford, UK; 4) Department of Biostatistics, Boston University School of Public Health, MA; 5) Department of Clinical Science, Genetic and Molecular Epidemiology Unit, Lund University, Malmö, Sweden; 6) Department of Epidemiology, School of Public Health, University of North Carolina at Chapel Hill, NC; 7) Pediatric Epidemiology Center, University of South Florida, Tampa, FL; 8) Department of Nutrition, Harvard School of Public Health, Boston, MA; 9) Department of Medical Sciences, Molecular Epidemiology and Science for Life Laboratory, Uppsala University, Sweden; 10) Department of Genetic Epidemiology, Institute of Epidemiology and Preventive Medicine, University of Regensburg, Germany; 11) Division of Preventive Medicine, Brigham and Womens Hospital, Boston, MA; 12) Framingham Heart Study, Framingham, MA; 13) Estonian Genome Center, University of Tartu, Estonia; 14) Department of Epidemiology, University of Groningen, University Medical Center Groningen, The Netherlands; 15) School of Public Health, University of Texas Health Science Center at Houston, TX; 16) Department of Twin Research and Genetic Epidemiology, King's College London, UK; 17) Department of Medicine, Division of Endocrinology, Diabetes & Nutrition, University of Maryland School of Medicine, MD; 18) EMGO+ Institute for Health and Care Research, VU Medical Center & VU University, Amsterdam, The Netherlands; 19) Department of Clinical Chemistry, Fimlab Laboratories, Tampere University Hospital and University of Tampere, Finland; 20) Department of Genetics, Washington University School of Medicine, St. Louis, MO; 21) The Genetics of Obesity and Related Metabolic Traits Program, The Icahn School of Medicine at Mount Sinai, New York, NY.

   The global obesity epidemic underscores the importance of gaining a deeper understanding of the biology of weight regulation to develop better preventive strategies. Obesity has a strong genetic component, but lifestyle factors such as physical activity (PA), may attenuate the impact of genetic susceptibility. Recently, a candidate gene-based meta-analysis confirmed that the body mass index (BMI)-increasing effect of the FTO obesity locus is attenuated by PA. To identify novel loci that interact with PA on BMI, we performed a meta-analysis of 38 GWAS, including 109,924 individuals of European descent. We standardized PA across all participating studies by categorizing it into a dichotomous variable (physically inactive vs. active) in each study. Overall, 23% of individuals were categorized as inactive. To screen for interactions, we performed a joint test of the SNP main effect and the SNP x PA interaction term, adjusting for age and age2 in men and women separately. In addition, we performed a test for the SNP x PA interaction term alone, with the same adjustments. We pooled the results from individual studies using fixed effects inverse variance meta-analysis. The joint P-value reached genome-wide significance (P<5x10-8) for 19 loci in all individuals, or in men or women separately. However, for each of these loci, the association was driven by the SNPs main effect on BMI, rather than by its interaction with PA. When we examined the SNP x PA interaction effects of the 19 loci, separately from main effects, the interaction reached nominal statistical significance for FTO (P=6x10-6), GIPC2 (P=0.02), and ADCY3 (in women; P=0.03). For these 3 loci, the BMI-increasing effect of the risk allele was smaller in the active group than in the inactive group. In a genome-wide screen for the SNP x PA interaction term alone, no locus reached P<5x10-8. However, the interaction for 4 novel loci (SV2C, SLIT3, CGNL1, TEKT5) reached P<10-6. In this sample of 109,924 individuals, we confirm that the BMI-increasing effect of FTO is attenuated by PA, but find little evidence for other loci interacting with PA. To follow-up our initial findings and to extend our total sample size, we are currently collecting data from approximately 25 additional studies with either genome-wide or Metabochip data. Our study may reveal novel genetic loci whose effects on obesity-risk are modified by PA, and may give important biological insights into the regulation of body weight.

You may contact the first author (during and after the meeting) at