A Genome-Wide Association Scan of RR and QT Interval Duration in 3 European Genetically Isolated PopulationsCLINICAL PERSPECTIVE
The EUROSPAN Project
Background— We set out to identify common genetic determinants of the length of the RR and QT intervals in 2325 individuals from isolated European populations.
Methods and Results— We analyzed the heart rate at rest, measured as the RR interval, and the length of the corrected QT interval for association with 318 237 single-nucleotide polymorphisms. The RR interval was associated with common variants within GPR133, a G-protein–coupled receptor (rs885389, P=3.9×10−8). The QT interval was associated with the earlier reported NOS1AP gene (rs2880058, P=2.00×10−10) and with a region on chromosome 13 (rs2478333, P=4.34×10−8), which is 100 kb from the closest known transcript LOC730174 and has previously not been associated with the length of the QT interval.
Conclusion— Our results suggested an association between the RR interval and GPR133 and confirmed an association between the QT interval and NOS1AP.
Received November 21, 2008; accepted April 13, 2009.
Quantitative electrocardiographic measurements have been shown to be valuable and noninvasive predictors of cardiovascular morbidity and mortality. In particular, increases in resting heart rate and in the length of the QT interval have been associated with increases in cardiovascular risk.
Clinical Perspective on p 322
Heart rate, often more accurately measured as the length in milliseconds of the RR interval (heart rate, 60 000/RR), is a well-known risk factor for morbidity in cardiovascular disease1,2 and all-cause mortality.3 Heritability, linkage, and association studies have suggested that the RR interval is modified by common genetic variations,4,5 but no RR modifier variants have yet been consistently confirmed.
Cardiac repolarization can be strongly altered in Mendelian disorders because of mutations in genes coding for ion channel subunits as in both long-QT syndrome (Online Mendelian Inheritance in Man 192500) and short-QT syndrome (Online Mendelian Inheritance in Man 60962) (http://www.ncbi.nlm.nih.gov/omim). Mildly increased QT intervals have also been associated with increased cardiovascular morbidity and mortality compared with QT intervals in the normal range.6,7 A QT-modifying polymorphism near the NOS1AP gene has been identified and successfully replicated.8–13
To extend the knowledge of genetic determinants of the length of QT and RR intervals, we performed a genome-wide association (GWA) analysis on 3 isolated European populations and then pooled the results via a meta-analysis performed on data for 2325 subjects.
We chose to focus on isolated populations because of the lower genetic heterogeneity and longer span of LD in isolated populations compared with outbred populations14,15 and the advantages these confer for the study of complex traits. In addition, isolated populations tend to show a lower level of environmental heterogeneity than the general population, which again can favor the identification of variants affecting complex traits, as shown in 1 of the 3 populations included in this meta-analysis.16
To investigate fully the contribution of isolated populations to gene mapping, a network comprising 5 centers involved in the study of isolated populations was established (European Special Population Research Network [EUROSPAN]; http://homepages.ed.ac.uk/s0565445/index.html). Three of the participating study locations (Italy, Scotland, and The Netherlands) had electrocardiographic measurements available, and their data were, therefore, included in the present work. We present here results from a meta-analysis of GWA scans (http://www.genome.gov/26525384) performed on these 3 genetically isolated populations on the length of RR and QT intervals.
Genotypes and Phenotypes
Genotypes were available for 1175 subjects in the South Tyrolean population, 745 individuals from Orkney, and 800 in the population from The Netherlands. We genotyped 318 237 single-nucleotide polymorphisms (SNPs) for each individual, using the Illumina 300 HumanHap SNP Chip (http://www.ncbi.nlm.nih.gov/SNP). Subjects with genotypic call rate >97% were retained in the analysis. Patients with atrial fibrillation, pacemaker, and defibrillator implants as well as pregnant women were excluded from the study.
We excluded from meta-analysis SNPs in which at least one of the study populations (1) had minor allele frequency <0.01, (2) were out of Hardy-Weinberg equilibrium (P<10−3), or (3) had a call rate <97%.
All participants in each individual study gave informed consent. Individual studies were approved by the competent ethics committees.
South Tyrol, Italy
Subjects were sampled in the framework of the MICROS study, carried out in 3 isolated villages in Val Venosta (South Tyrol, Italy) in 2001 to 2003.17 Because of geographical, historical, and political reasons, the entire region experienced prolonged isolation from surrounding populations. The investigated population is characterized by an old settlement, a small number of founders, high endogamy rates, slow or null population expansion, and negligible immigration.18 Information on a participant’s health status was collected through a standardized questionnaire and clinical examinations, including electrocardiographic measurements. Twelve-lead resting electrocardiographic measurements were recorded using a digital recording system (Mortara Portrait; Mortara, Milwaukee, Wis). The Mortara portrait machine determines QT interval by the proprietary XL-electrocardiographic algorithm that has not been fully published but has shown to be in good accordance with other published electrocardiographic measurement algorithms.19 Laboratory data were obtained from standard blood analyses. Joint genotype and phenotype information was available for 970 subjects (409 men and 561 women).
Orkney Islands, Scotland (United Kingdom)
The Orkney Complex Disease Study (ORCADES) is an ongoing family-based cross-sectional study taking place in the isolated Scottish archipelago of Orkney. Genetic diversity is decreased compared with Mainland Scotland, consistent with the high levels of endogamy historically. Data from 745 participants aged 18 to 100 years from a subgroup of 10 islands were used in this analysis. Blood samples were taken from fasting participants and over 200 health-related phenotypes, and environmental exposures were measured in each individual. Digital 10-second electrocardiograms were taken after at least 10 minutes of supine rest, using a PC link with QT and RR intervals calculated using CardioView software (NUMED cardiac diagnostics; Sheffield, United Kingdom). Joint genotype and phenotype information was available for 679 subjects (315 men and 364 women).
Rucphen, The Netherlands
The Erasmus Rucphen Family (ERF) study was carried out on a Dutch isolated population located in the Southwest of The Netherlands.20 The population is characterized by rapid growth and minimal inward migration and has now expanded up to 20 000 inhabitants. Within this population, a specific subpopulation based on 20 couples (selected on the basis that they had at least 6 children baptized in the community church between 1850 and 1900) has been defined. All living descendants of the selected couples and their spouses (n=3000) have been recruited. All participants have been invited to the research center and were screened for quantitative traits, assessing cardiovascular, neuropsychiatric, endocrine, ophthalmologic, and musculoskeletal functions. A 10-second 12-lead electrocardiogram (on average, 8 to 10 beats) was recorded with an ACTA electrocardiogram (Esaote; Florence, Italy) with a sampling frequency of 500 Hz. All electrocardiograms were processed by the Modular Electrocardiogram Analysis System to obtain electrocardiographic measurement and interpretation. The Modular Electrocardiogram Analysis System determines common onsets and offsets for all 12 leads together on 1 representative averaged beat, with the use of template matching techniques, has been evaluated extensively and shown to have excellent correlation with diagnosis performed by cardiologists.21 Joint genotype and phenotype information was available for 676 subjects (252 men and 424 women).
Between-population homogeneity of study variables was assessed with the Kruskal-Wallis rank sum test. To ensure a better adaptation of models residuals to normality, QT and RR were transformed to normal distribution using rank transformation to normality. Multiple linear regression models were fitted to the normalized QT and RR, adjusting for age, sex, and RR, and age, sex, and body mass index, respectively. Sex-stratified models were also estimated. GWA scans under an additive model were performed separately for each population. The genomic control method was used to correct the distribution of P values that could be skewed in inbred populations.22,23
λ was estimated to be 1.6 in MICROS, 1.1 in ERF, and 1.1 in Orkney. λ for the pooled sample was 1.3. Association analyses were performed using the R package GenABEL (http://mga.bionet.nsc.ru/∼yurii/ABEL).24,25
Evidence from single studies was pooled together using a fixed effect meta-analysis based on inverse variance weighting.26 We set the global α for significance to 0.05. Under this constraint, using the Bonferroni correction based on the conservative assumption that the 318 237 tests performed were independent, the genome-wide significance threshold was defined as P≤1.57×10−7.
All locations on a physical map are referred to build 36 of the human genome reference map. LD was computed using the R package genetics27 and plots obtained with the R package LDheatmap.28 All analyses were performed using R.25
Characteristics of each population sample are given in Table 1. There were significant differences between populations for all study variables (P<0.0001). The MICROS population was significantly younger than those from ERF (P<0.0001) and Orkney (P<0.0001), whereas mean ages for ERF and Orkney populations were not significantly different from each other (P=0.73). The Orkney sample had the highest body mass index, whereas South Tyrol showed the lowest value; all pairwise tests were significantly different (P<0.001). The RR interval was shorter in the MICROS population than in the ERF and Orkney samples, and all pairwise tests were significantly different; average heart rate (60 000/RR) for the 3 populations ranged from 60 bpm (Orkney) to 67 bpm (MICROS). The QT interval was significantly longer in Orkney sample than in MICROS, whereas no significant difference in QT interval duration was detected in the remaining pairwise comparisons. In addition, aggregated data were available for larger cohorts recruited for each study (a subset of which was then genotyped). Such data include the prevalence of hypertension, diabetes, and myocardial infarction, together with information on the use of β-blocking drugs. All data are based on self-reporting. Data are shown in the online-only Data Supplement (Table I).
Table 2 shows population-specific and pooled results of the test of association between the length of RR interval (after adjusting for age, sex, and body mass index) and the 25 SNPs with lowest P values. For SNPs located within genes, gene names are reported as well. Two SNPs (rs885389 and rs1725789) located in the GPR133 gene on chromosome 12 exceeded the threshold of genome-wide significance (P=3.88×10−8 and 1.48×10−7, respectively). Two SNPs are located in FRMD4A, 2 in AKT3, and 1 in RASGRF1, but none of them is genome-wide significant. All of the above-mentioned SNPs are intronic.
Figure 1 shows a detailed view of all SNPs present in GPR133. Boundaries of GPR133 are shown as dotted vertical lines. A lack of significant association is evident for most of the gene sequence, with the exception of the 3′ portion of the gene with the 2 SNPs reaching genome-wide significance and 2 more SNPs that emerged from background noise.
The rare allele of both SNPs is associated with a shortening of the length of RR interval (and increased heart rate). For SNP rs885389, each risk allele confers a decrease of 14 ms in the length of RR interval, whereas for SNP rs1725789, each risk allele causes a decrease of 16 ms; this roughly corresponds to an increase of 1 bpm per risk allele in terms of heart rate.
A figure showing the results of genome-wide association is available in the Data Supplement (Figure I). A graph of LD structure in GPR133 in each population is available in the Data Supplement (Figure II). Although no strong LD blocks were detected, the 2 genome-wide significant SNPs are in strong LD (r2=0.68 in MICROS and Orkney, r2=0.69 in ERF).
When analyzing the data from men separately, 1 SNP on chromosome 1 reached genome-wide significance (rs17706439, P=2.82×10−8), but it was not in proximity (<100 kb) to any known gene. No SNPs in GPR133 reached significance. In the analysis of the data from women, we found 2 genome-wide significant SNPs: 1 on chromosome 7 (rs1874326, P=9.79×10−8) and 1 on chromosome X (rs4610880, P=1.44×10−7). rs1874326 is located in the TRIM24 gene, known to mediate transcriptional control; rs4610880 is located in the open reading frame CXorf36. No SNPs in GPR133 were significant in women alone.
The 25 SNPs with lowest P values after meta-analysis, sorted by ascending the P value, are reported in Table 3 (for a graph of the GWA scan, see supplementary Figure III). Five SNPs located in or around NOS1AP reached genome-wide significance. The most significant SNP (rs2880058, P=2×10−10) was located 25 kb upstream of NOS1AP. A genome-wide significant result was also observed for rs10494366, through which the association between QT and NOS1AP polymorphisms was first identified.8
Association results in NOS1AP gene are shown in Figure 2. LD structure in NOS1AP gene is available as Data Supplement (Figure IV).
One SNP on chromosome 13 was also above the genome-wide threshold for significance (rs2478333, P=4.34×10−8). It is located >100 kb from the closest putative locus (LOC730174) and 300 kb from the nearest known gene, succinate-CoA-ligase (SUCLA2).
In the meta-analysis performed on the female subsample (n=1349), we identified 2 SNPs achieving genome-wide significance, rs2880058 (P=1.03×10−9) and rs6670339 (P=1.24×10−7); both of them are in or around NOS1AP and also reached genome-wide significance in the pooled analysis (Table 3). When analyzing the male subsample (n=976), we did not identify any genome-wide significant hits.
Meta-analysis of GWA scans for the RR interval allowed us to identify 2 SNPs reaching genome-wide significance, located in the GPR133 gene.
To our knowledge, only 1 previous GWA study was performed to investigate the genetic contribution to the length of the RR interval,29 without identifying any genome-wide significant signals. The authors made results available through the dbGAP database (http://www.ncbi.nlm.nih.gov/sites/entrez?Db=gap); however, we were not able to compare the region in which we found the strongest signal because the published results lack SNPs between 129 840 347 and 130 588 160 bp, with GPR133 spanning 130 004 790 to 130 189 786 bp (distances of the closest SNPs to GPR133 being ≈160 and 400 kb, respectively).
Ours is the first study to identify significant SNPs associated with the length of the RR interval. Two SNPs, rs885389 and rs1725789, reached genome-wide significance. These are located in the 3′ region of GPR133 and are in strong but not complete LD.
We replicated the association of NOS1AP with the length of the QT interval adjusted for age, sex, and RR interval. Additional confounding factors, such as QT-prolonging drugs and cardiovascular disease status, were not included in the analysis. Both previous studies8 and our results consistently identified the association between SNPs in the 5′ region of NOS1AP and upstream of NOS1AP coding sequence, and the length of QT interval. In addition, we identified a genome-wide significant hit with SNP rs2478333. This SNP is located >100 kb from the closest putative locus and 300 kb from the nearest characterized gene. Our findings on the genetic determinants of the length of QT interval allow us to verify the importance of NOS1AP and to identify a SNP on chromosome 13 reaching genome-wide significance although far away from any known gene.
One limitation of this study is the lack of independent replication. In addition, functional demonstration of the effect of GPR133 on the RR interval and of NOS1AP on the QT interval still needs to be provided.
GPR133 is a gene coding for a G-protein–coupled receptor (GPCR).30 Expression data retrieved on Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo) showed that GPR133 is expressed in atria, ventricles, and septal myocardial tissue (Gene Expression Omnibus accession numbers: GDS651, GDS1557, GDS1559, and GDS2206).
At present, GPCRs belonging to the rhodopsin GPCR family, in particular adrenergic receptors, have been studied extensively for their influence on heart activity and widely used as pharmacological targets.31
We undertook a search of an available database (GLIDA; http://pharminfo.pharm.kyoto-u.ac.jp/services/glida) of GPCRs and their ligands to characterize GPR133 better.32 We performed a similarity search (based on sequence) to identify GPCRs that could have structural similarities with GPR133. Among GPCRs stored in the GLIDA database, those demonstrating the highest similarity were ELTD1, CELSR1, EMR1, and LPHN1. Several members of the families of cadherins, latrophilins, and ETL receptors were shown as being similar to GPR133. A previous work showed that ELTD1 (previous name, ETL) is developmentally regulated in heart.33 We checked for association of ELTD1, CELSR1, EMR1, and LPHN1 with the length of the RR interval but found no evidence of association; no P value <10−4 was identified in any of the genes or in the regions 100 kb up- and downstream.
In conclusion, we propose a role for GPR133 in affecting the length of the electrocardiographic RR interval and heart rate. β-Adrenergic receptors are members of GPCRs targeted by β-blocker drugs for the management of cardiac arrhythmias; GPR133 could, therefore, represent an ideal novel target for a pharmacological approach. Assigning a ligand to this orphan receptor and identifying the causal variant are among the priorities to confirm a role of GPR133 in determining the heart rate.
We thank all participants of the MICROS, ERF, and ORCADES studies. For the MICROS study in South Tyrol, we thank the primary care practitioners Raffaela Stocker, Stefan Waldner, Toni Pizzecco, Josef Plangger, Ugo Marcadent, and the personnel of the Hospital of Silandro (Department of Laboratory Medicine) for their participation and collaboration in the research project. We thank Matthias Wjst for contributing to the article with useful discussion and Daniela Grazio for valuable help in collecting data. For the ORCADES study, we acknowledge the invaluable contributions of Lorraine Anderson and the research nurses in Orkney and the administrative team in Edinburgh.
Sources of Funding
EUROSPAN (European Special Populations Research Network) was supported by European Commission FP6 STRP grant 018947 (LSHG-CT-2006-01947) and grant 108-1080315-0302 from the Croatian Ministry of Science, Education and Sport (to I.R.); the MICROS study was supported by the Ministry of Health of the Autonomous Province of Bolzano and the South Tyrolean Sparkasse Foundation; the ORCADES study was supported by the Scottish Executive Health Department, the Royal Society, and the Wellcome Trust Clinical Research Facility; the ERF study was supported by grants from the Netherlands Organization for Scientific Research (NWO, 91203014), the Russian Foundation for Basic Research (NWO-RFBR, 047.017.043), and the Center of Medical Systems Biology (CMSB), and grants from the Center of Medical Systems Biology and from Netherlands Genomics Initiative (to C.M.v.D.).
Dyer AR, Persky V, Stamler J, Paul O, Shekelle RB, Berkson DM, Lepper M, Schoenberger JA, Lindberg HA. Heart rate as a prognostic factor for coronary heart disease and mortality: findings in three Chicago epidemiologic studies. Am J Epidemiol. 1980; 112: 736–749.
Kristal-Boneh E, Silber H, Harari G, Froom P. The association of resting heart rate with cardiovascular, cancer and all-cause mortality: eight year follow-up of 3527 male Israeli employees (the CORDIS Study). Eur Heart J. 2000; 21: 116–124.
Singh JP, Larson MG, O'Donnell CJ, Tsuji H, Evans JC, Levy D. Heritability of heart rate variability: the Framingham Heart Study. Circulation. 1999; 99: 2251–2254.
Elming H, Holm E, Jun L, Torp-Pedersen C, Kober L, Kircshoff M, Malik M, Camm J. The prognostic value of the QT interval and QT interval dispersion in all-cause and cardiac mortality and morbidity in a population of Danish citizens. Eur Heart J. 1998; 19: 1391–1400.
Okin PM, Devereux RB, Howard BV, Fabsitz RR, Lee ET, Welty TK. Assessment of QT interval and QT dispersion for prediction of all-cause and cardiovascular mortality in American Indians: the Strong Heart Study. Circulation. 2000; 101: 61–66.
Arking DE, Pfeufer A, Post W, Kao WH, Newton-Cheh C, Ikeda M, West K, Kashuk C, Akyol M, Perz S, Jalilzadeh S, Illig T, Gieger C, Guo CY, Larson MG, Wichmann HE, Marbán E, O'Donnell CJ, Hirschhorn JN, Kääb S, Spooner PM, Meitinger T, Chakravarti A. A common genetic variant in the NOS1 regulator NOS1AP modulates cardiac repolarization. Nat Genet. 2006; 38: 644–651.
Raitakari OT, Blom-Nyholm J, Koskinen TA, Kahonen M, Viikari JS, Lehtimaki T. Common variation in NOS1AP and KCNH2 genes and QT interval duration in young adults: the Cardiovascular Risk in Young Finns Study. Ann Med. 2008; 10: 1–8.
Tobin MD, Kähönen M, Braund P, Nieminen T, Hajat C, Tomaszewski M, Viik J, Lehtinen R, Ng GA, Macfarlane PW, Burton PR, Lehtimäki T, Samani NJ. Gender and effects of a common genetic variant in the NOS1 regulator NOS1AP on cardiac repolarization in 3761 individuals from two independent populations. Int J Epidemiol. 2008; 37: 1132–1141.
Pfeufer A, Sanna S, Arking DE, Müller M, Gateva V, Fuchsberger C, Ehret GB, Orrú M, Pattaro C, Köttgen A, Perz S, Usala G, Barbalic M, Li M, Pütz B, Scuteri A, Prineas RJ, Sinner MF, Happle C, Gieger C, Najjar SS, Kao WHL, Mühleisen TW, Dei M, Möhlenkamp S, Crisponi L, Erbel R, Jöckel KH, Naitza S, Steinbeck G, Marroni F, Hicks AA, Lakatta E, Müller-Myhsok B, Pramstaller PP, Wichmann HE, Schlessinger D, Boerwinkle E, Meitinger T, Uda M, Coresh J, Kääb S, Abecasis GR, Chakravarti A. Common variants at ten loci modulate the QT interval duration in individuals of European ancestry: the QTSCD consortium. Nat Genet. 2009; 41: 407–414.
Newton-Cheh C, Eijgelsheim M, Rice KM, de Bakker PI, Yin X, Estrada K, Bis JC, Marciante K, Rivadeneira F, Noseworthy PA, Sotoodehnia N, Smith NL, Rotter JI, Kors JA, Witteman JCM, Hofman A, Heckbert SR, O'Donnell CJ, Uitterlinden AG, Psaty BM, Lumley T, Larson MG, Stricker BH. Common variants at ten loci influence QT interval duration in the QTGEN study. Nat Genet. 2009; 41: 399–406.
Pattaro C, Marroni F, Riegler A, Mascalzoni D, Pichler I, Volpato CB, Dal Cero U, De Grandi A, Egger C, Eisendle A, Fuchsberger C, Gögele M, Pedrotti S, Pinggera GK, Stefanov SA, Vogl FD, Wiedermann CJ, Meitinger T, Pramstaller PP. The genetic study of three population microisolates in South Tyrol (MICROS): study design and epidemiological perspectives. BMC Med Genet. 2007; 8: 29.
de Bruyne MC, Kors JA, Hoes AW, Kruijssen DA, Deckers JW, Grosfeld M, van Herpen G, Grobbee DE, van Bemmel JH. Diagnostic interpretation of electrocardiograms in population-based research: computer program research physicians, or cardiologists? J Clin Epidemiol. 1997; 50: 947–952.
Aulchenko YS, Ripke S, Isaacs A, van Duijn CM. GenABEL: an R library for genome-wide association analysis. Bioinformatics. 2007; 23: 1294–1296.
R Development Core Team, R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0. Available at: http://www.R-project.org. Accessed June 15, 2009.
Woodward M. Epidemiology: Study Designs and Data Analysis. Washington DC: Chapman & Hall/CRC; 2005.
Warnes G, Gorjanc G, Leisch F, Man M. Genetics: Population Genetics. Available at http://cran.r-project.org/web/packages/genetics/index.html. Last accessed June 15, 2009.
Shin JH, Blay S, McNeney B, Graham J. LDheatmap: an R function for graphical display of pairwise linkage disequilibria between single nucleotide polymorphisms. J Stat Soft. 2006; Code Snippet: 3.
Okuno Y, Tamon A, Yabuuchi H, Niijima S, Minowa Y, Tonomura K, Kunimoto R, Feng C. GLIDA: GPCR–ligand database for chemical genomics drug discovery–database and tools update. Nucleic Acids Res. 2008; 36: D907–D912.
Nechiporuk T, Urness LD, Keating MT. ETL, a novel seven-transmembrane receptor that is developmentally regulated in the heart. ETL is a member of the secretin family and belongs to the epidermal growth factor-seven-transmembrane subfamily. J Biol Chem. 2001; 276: 4150–4157.
Higher heart rate at rest (which is equivalent to a shorter RR interval) has been associated with an increased risk of cardiovascular disease and with an increased mortality from cardiovascular disease. Given the importance of heart rate as an independent predictor of cardiovascular disease and death, it is important to identify genetic variants that influence resting heart rate. Such knowledge may provide new insights into physiological pathways involved in control of sinus rhythm and the etiology of heart disease itself, which could promote the development of appropriate prevention strategies and identification of targets suitable for drug development. Our findings suggest that GPR133, a member of the G-protein–coupled receptor family, plays a role in determining the heart rate. Because β-adrenergic receptors are members of this G-protein–coupled receptor family and are targeted by β-blocker drugs in the management of cardiac arrhythmias, GPR133 could represent a promising candidate for testing in future pharmacological approaches.
The online-only Data Supplement is available at http://circgenetics.ahajournals.org/cgi/content/full/CIRCGENETICS.108.833806/DC1.