Diabetes Publish Ahead of Print, published online September 9, 2009 SIRT1 gene variation, BMI, and obesity SIRT1 genetic variation is related to body mass index and risk of obesity Running title: SIRT1 gene variation, BMI, and obesity M. Carola Zillikens MD1, Joyce B.J. van Meurs PhD1, Fernando Rivadeneira MD, PhD1,2 , Najaf Amin MSc 2, Albert Hofman MD2, PhD, Ben A Oostra PhD3, Eric J.G. Sijbrands MD, PhD1, Jacqueline C.M. Witteman PhD2, Huibert A.P. Pols MD, PhD1,2 , Cornelia M. van Duijn PhD2, André G. Uitterlinden PhD1,2 1Department of Internal Medicine and 2Department of Epidemiology and 3Department of Clinical Genetics, Erasmus Medical Center, Rotterdam, The Netherlands. Correspondence: M.C. Zillikens Email: [email protected] Submitted 10 April 2009 and accepted 27 August 2009. This is an uncopyedited electronic version of an article accepted for publication in Diabetes. The American Diabetes Association, publisher of Diabetes, is not responsible for any errors or omissions in this version of the manuscript or any version derived from it by third parties. The definitive publisher-authenticated version will be available in a future issue of Diabetes in print and online at http://diabetes.diabetesjournals.org. Copyright American Diabetes Association, Inc., 2009 SIRT1 gene variation, BMI, and obesity Objective SIRT1 has pleiotropic metabolic functions. We investigated whether SIRT1 genetic variation is associated with obesity. Research design and methods In 6251 elderly subjects from the prospective, population-based Rotterdam Study (RS), three Single Nucleotide Polymorphisms (SNPs) in the SIRT1 gene were studied in relation with body mass index (BMI) and risk of obesity (BMI > 30kg/m2) and prospectively with BMI-change after 6.4 years follow-up. We used cross-sectional data from 2347 participants from the Erasmus Rucphen Family (ERF) study for replication. Results Minor alleles of rs7895833 (G=20.2%) and rs1467568 (A=36.8%) were associated with lower BMI in the RS (P=0.02 and 0.04) and in the replication cohort ERF (P=0.03 and 0.008) and in studies combined (P=0.002 for both SNPs) with 0.2-0.4 kg/m2 decrease in BMI per allele copy. Carriers of these alleles had 13-18% decreased risk of obesity; odds ratio (95%CI) for rs7895833 the RS: 0.79 (0.67-0.94), P=0.007; in ERF: 0.93 (0.73-1.19), P=0.37 and in the studies combined 0.87 (0.77-0.97), P=0.02; for rs1467568 in the RS: 0.80 (0.68-0.94), P=0.007; in ERF: 0.85 (0.72-0.99), P=0.04 and in the studies combined: 0.82 (0.73-0.92), P=0.0009. In the RS, the two variants were also associated with lower BMI increase during 6.4 years of follow-up (P=0.01 and 0.08). Conclusions Two common variants in SIRT1 are associated with lower BMI in two independent Dutch populations. Carriers of these variants have 13-18% decreased risk of obesity and gain less weight over time. The availability of SIRT1 stimulators makes these findings relevant in light of the growing obesity epidemic. 2 SIRT1 gene variation, BMI, and obesity SIRT1 belongs to the Sirtuin protein family of and related diseases. However, in humans, nicotinamide adenine dinucleotide (NAD+)- effects of SIRT1 stimulation have not been dependent histone deacetyleases conserved in investigated in vivo and there is concern that evolution from bacteria to humans. In lower generalized SIRT1 activation may also have organisms like yeast, flies, and worms, the pro-ageing or adverse health effects, due to silent information regulator Sir2 protein is potential pleiotropic and tissue dependent related to longevity (1; 2) and to life span physiological functions (26-29). extension after caloric restriction (3-5). Based on the above findings we Humans have 7 sirtuins (SIRT1-7) (6) of hypothesized that genetic variation in SIRT1 which SIRT1 has the highest homology to may influence BMI and the risk of obesity in yeast Sir2. SIRT1 controls numerous humans. Previous smaller studies, which have physiological processes and protects cells examined the relation between variants in against stress (2),(7-13). SIRT1 also has an SIRT1 and obesity, have led to inconsistent important function in endocrine signaling, findings. A recent case control study of 1.068 specifically in glucose and fat metabolism obese patients and 313 normal weight (14-18). Increased hepatic SIRT1 activity controls, found a SIRT1 SNP associated to enhances gluconeogenesis and inhibits obesity risk. Unexpectedly, male but not glycolysis (15),(16). In the pancreas, SIRT1 female carriers of the allele associated with stimulates insulin secretion in response to lower obesity risk had increased visceral glucose (17),(18). In adipose tissue, SIRT1 adiposity on CT-scans. This was observed interacts with PPARγ to repress its only after adjustment for BMI (30). Another transcriptional activity, leading to inhibition recent small study investigating associations of adipogenesis during fasting and activation of SIRT1 SNPs with metabolic response to of lipolysis(14). This results in fat loss, which lifestyle intervention found no association of is an important component of the effect of four SIRT1 SNPs in 917 overweight subjects caloric restriction on longevity in with baseline BMI or with BMI change after mammals(14). It is not known whether caloric 9 months follow-up (31). No population- restriction can extend lifespan in humans, but based studies with large sample size and clear beneficial effects on cardiovascular risk replication are available on the association factors have been described (19),(20). Fasting between SIRT1 genetic variation, BMI, BMI leads to up-regulation of SIRT1 in adipose change in time and risk of obesity. We tissue of mice, pigs and humans (21-23). therefore assessed the association of variation Based on findings in lower organisms, it has in the SIRT1 gene with BMI and the risk of been hypothesized that stimulation of SIRT1 being overweight or obese in the large cohort by agonists may mimic the beneficial effects of elderly subjects of the Rotterdam study of caloric restriction in mammals. (RS). For replication we used subjects with a Resveratrol, a naturally occurring SIRT1 wide age range from a genetically isolated agonist present in red wine and grapes, was population in the Netherlands, participating in indeed recently shown to prevent diet-induced the Erasmus Rucphen Family (ERF) study. In obesity and insulin resistance in mice and to the RS we also assessed prospectively the improve their survival on a high-calorie diet relation of SIRT1 variants with BMI change (24),(25). If activation of SIRT1 can result in on follow-up. loss of body fat without decreasing caloric intake, this could open the door for novel treatment and prevention strategies for obesity 3 SIRT1 gene variation, BMI, and obesity (cm) and weight (kg) were measured at the RESEARCH DESIGN AND METHODS initial examination and in the Rotterdam The Rotterdam study. The study also at follow-up examinations, in Rotterdam study is a prospective, population- standing position wearing indoor clothes based cohort study among 7,983 persons aged without shoes. Body mass index (BMI) was 55 years and older from a district of computed as weight in kilograms divided by Rotterdam, The Netherlands. The study was height in meters squared (kg/m2). Genotyping designed to investigate the incidence and Three tagging single nucleotide determinants of chronic disabling diseases. polymorphisms (SNPs) rs7895833, rs1467568 Rationale and design have been described and rs497849 were selected from the HapMap previously (32; 33). Informed consent was database (http://www.hapmap.org) that, obtained from each participant, and the together with constructed haplotypes, covered Medical Ethics Committee of the Erasmus 100% of the common (minor allele frequency Medical Center Rotterdam approved the >10%) variation of the SIRT1gene in study. Caucasians. Genotyping of the SIRT1 SNPs At baseline (1990–1993), all participants were was performed by Taqman on genomic DNA interviewed and underwent extensive physical isolated from peripheral leucocytes by examination. At the second follow-up (1997– standard salting-out procedures. Results were 1999) these examinations were performed analysed by the ABI Taqman 7900HT using according to the same protocol. the sequence detection system 2.22 software The Erasmus Rucphen Family (Applied Biosystems Inc., Foster City, CA, (ERF) study. This study is a family-based USA). To confirm the accuracy of genotyping cohort study that is embedded in the Genetic results, 332 (5%) randomly selected samples Research in Isolated Populations (GRIP) were re-genotyped with the same method. No program in the South West of the inconsistencies were observed. All used Netherlands. The aim of this program was to primers and probes are available on request. identify genetic risk factors in the Statistical Analyses. Hardy-Weinberg development of complex disorders (34-36). equilibrium of the three SIRT1 SNPs was For the ERF study, 22 families that tested with the GENEPOP-package (37). had at least five children baptized in the Subjects were grouped according to genotype community church between 1850-1900 were for individual SNP alleles, and by allele copy identified with the help of genealogical number of haplotype alleles. We inferred records. All living descendants of these multimarker haplotypes in the Rotterdam couples and their spouses were invited to take study only from these SNPs using the part in the study. Data collection started in program Phase (38). In the ERF study June 2002 and was finished in February 2005. haplotypes were not determined because they In this study, we focused on 2347 participants could not be inferred with high certainty due for whom complete phenotypic, genotypic to the complex pedigree structure. and genealogical information was available. Haplotype alleles were numbered in The Medical Ethics Committee of order of decreasing frequency in the Erasmus Medical Center Rotterdam approved population (Figure1).