This page has been archived and is no longer updated
A susceptibility locus for lung cancer maps to nicotinic acetylcholine receptor subunit genes on 15q25
Author: Rayjean J. Hung
Keywords
Keywords for this Article
Add keywords to your Content
Save
|
Cancel
Share
|
Cancel
Revoke
|
Cancel
Rate & Certify
Rate Me...
Rate Me
!
Comment
Save
|
Cancel
Flag Inappropriate
The Content is
Objectionable
Explicit
Offensive
Inaccurate
Comment
Flag Content
|
Cancel
Delete Content
Reason
Delete
|
Cancel
Close
Full Screen
"LETTERS Asusceptibilitylocus forlungcancermapstonicotinic acetylcholine receptor subunit genes on 15q25 Rayjean J. Hung 1,2 *, James D. McKay 1 *, Valerie Gaborieau 1 , Paolo Boffetta 1 , Mia Hashibe 1 , David Zaridze 3 , Anush Mukeria 3 , Neonilia Szeszenia-Dabrowska 4 , Jolanta Lissowska 5 , Peter Rudnai 6 , Eleonora Fabianova 7 , DanaMates 8 ,VladimirBencko 9 ,LenkaForetova 10 ,VladimirJanout 11 ,ChuChen 12 ,GaryGoodman 12 ,JohnK.Field 13 , TriantafillosLiloglou 13 , GeorgeXinarianos 13 , Adrian Cassidy 13 , JohnMcLaughlin 14 ,Geoffrey Liu 15 , StevenNarod 16 , Hans E. Krokan 17 , Frank Skorpen 17 , Maiken Bratt Elvestad 17 , Kristian Hveem 17 , Lars Vatten 17 , Jakob Linseisen 18 , Franc�oise Clavel-Chapelon 19 , Paolo Vineis 20,21 , H. Bas Bueno-de-Mesquita 22 , Eiliv Lund 23 , Carmen Martinez 24 , Sheila Bingham 25 , Torgny Rasmuson 26 , Pierre Hainaut 1 , Elio Riboli 20 , Wolfgang Ahrens 27 , Simone Benhamou 28,29 , Pagona Lagiou 30 , Dimitrios Trichopoulos 30 , Ivana Holca�tova� 31 , Franco Merletti 32 , Kristina Kjaerheim 33 , Antonio Agudo 34 , Gary Macfarlane 35 , Renato Talamini 36 , Lorenzo Simonato 37 , Ray Lowry 38 , David I. Conway 39 , ArianaZnaor 40 ,ClaireHealy 41 ,DianaZelenika 42 ,AnneBoland 42 ,MarcDelepine 42 ,MarioFoglio 42 ,DorisLechner 42 , Fumihiko Matsuda 42 , Helene Blanche 43 , Ivo Gut 42 , Simon Heath 43 , Mark Lathrop 42,43 & Paul Brennan 1 Lung cancer is the most common cause of cancer death worldwide, with over one million cases annually 1 . To identify genetic factors that modify disease risk, we conducted a genome-wide association study by analysing 317,139 single-nucleotide polymorphisms in 1,989 lung cancer cases and 2,625 controls from six central European countries. We identified a locus in chromosome region 15q25thatwasstronglyassociatedwithlungcancer(P59310 210 ). This locus was replicated in five separate lung cancer studies com- prising an additional 2,513 lung cancer cases and 4,752 controls (P55310 220 overall), and itwas found to account for 14% (attri- butable risk) of lung cancer cases. Statistically similar risks were observed irrespective of smoking status or propensity to smoke tobacco. The association region contains several genes, including three that encode nicotinic acetylcholine receptor subunits (CHRNA5, CHRNA3 and CHRNB4). Such subunits are expressed in neurons and other tissues, in particular alveolar epithelial cells, pulmonary neuroendocrine cells and lung cancer cell lines 2,3 ,and theybindtoN9-nitrosonornicotineandpotentiallungcarcinogens 4 . A non-synonymous variant ofCHRNA5thatinducesanaminoacid substitution(D398N)atahighlyconserved siteinthesecondintra- cellularloopoftheproteinisamongthemarkerswiththestrongest disease associations. Our results provide compelling evidence of a locusat15q25predisposingtolungcancer,andreinforceinterestin nicotinicacetylcholinereceptorsaspotentialdiseasecandidatesand chemopreventative targets 5 . Lung cancer is caused predominantly by tobacco smoking, with cessation of tobacco consumption being the primary method for prevention. The risk among those who quit smoking remains ele- vated (although less than those who continue to smoke), and former smokers makeupanincreasing proportion oflungcancer patientsin countries where tobacco consumption has declined 6,7 . Treatment strategies are of limited efficacy, with an overall 5-year survival rate of about 15% 8 . Lung cancer has an important heritable component 9 , and identifying genes that are involved may suggest chemopreven- tiontargetsorallowforidentificationofgroupsathighrisk.Despitea large number of studies including both sporadic and multi-case families, success in identifying genes that cause lung cancer has been extremely limited. Theavailability oftagging single-nucleotide polymorphism (SNP) panels across the whole genome allows for efficient and comprehen- sive analysis of common genomic variation to be conducted without apriorihypothesesbasedongenefunctionordiseasepathways.They *These authors contributed equally to this work. 1 InternationalAgencyforResearchonCancer(IARC),Lyon69008,France. 2 SchoolofPublicHealth,UniversityofCaliforniaatBerkeley,Berkeley,California94720,USA. 3 Instituteof Carcinogenesis, Cancer Research Centre, Moscow 115478, Russia. 4 Department of Epidemiology, Institute of Occupational Medicine, Lodz 90950, Poland. 5 M. Sklodowska-Curie Memorial Cancer Center and Institute of Oncology, Warsaw 02781, Poland. 6 National Institute of Environmental Health, Budapest 1097, Hungary. 7 Specialized Institute of Hygiene and Epidemiology, Banska Bystrica 97556, Slovakia. 8 Institute of Public Health, Bucharest 050463, Romania. 9 Charles University in Prague, First Faculty of Medicine, Institute of Hygiene and Epidemiology, Prague 2 12800, Czech Republic. 10 Department of Cancer Epidemiology and Genetics, Masaryk Memorial Cancer Institute, Brno 65653, Czech Republic. 11 Palacky University,Olomouc 77515, Czech Republic. 12 Fred Hutchinson Cancer Research Center, Seattle, Washington 98109, USA. 13 Roy Castle Lung Cancer Research Programme, University of Liverpool Cancer Research Centre, Liverpool L3 9TA, UK. 14 Cancer Care Ontario, and the Samuel Lunenfeld Research Institute, Toronto M5G 2L7, Canada. 15 Princess Margaret Hospital, Ontario Cancer Institute, Toronto M5G 2M9, Canada. 16 Women?s College Research Institute, Toronto M5G 1N8, Canada. 17 Norwegian University of Science and Technology,Trondheim 7489, Norway. 18 Division of Cancer Epidemiology,German CancerResearch Centre (DKFZ), Heidelberg 69120, Germany. 19 INSERMERI20,Institut Gustave Roussy,Villejuif94805,France. 20 DepartmentofEpidemiologyandPublicHealth,ImperialCollege,LondonW21PG,UK. 21 InstituteforScientificInterchange(ISI),Torino10133,Italy. 22 Centre for Nutrition and Health, National Institute of Public Health and the Environment, Bilthoven 3710 BA, The Netherlands. 23 Institute of Community Medicine, University of Tromso�, Tromso� 9037, Norway. 24 Andalusian school of Public Health and Ciber Epidemiology y Salud Publica, Granada 18011, Spain. 25 MRC Centre for Nutrition and Cancer, University of Cambridge, Department of Public Health and Primary Care and MRC Dunn Human Nutrition Unit, Cambridge CB2 0XY, UK. 26 Department of Radiation Sciences, Oncology,UmeaUniversity,Umea90187,Sweden. 27 EpidemiologicalMethodsandEtiologicResearch,BremenInstituteforPreventionResearchandSocialMedicine,Bremen28359, Germany. 28 INSERM U794, Fondation Jean Dausset-CEPH, Paris 75010, France. 29 CNRS FRE2939, Institute Gustave Roussy, Villejuif 94805, France. 30 Department of Hygiene and Epidemiology,UniversityofAthensSchoolofMedicine,Athens11527,Greece,andDepartmentofEpidemiology,HarvardSchoolofPublicHealth,Boston,Massachusetts02115,USA. 31 Institute of Hygiene and Epidemiology, Prague 2 12800, Czech Republic. 32 University of Turin, Turin 10126, Italy. 33 Cancer registry of Norway, Oslo 0310, Norway. 34 Institut Catala` d?Oncologia, Barcelona 08907, Spain. 35 University of Aberdeen School of Medicine, Aberdeen AB25 2ZD, UK. 36 Aviano cancer center, Aviano 33081, Italy. 37 Department of EnvironmentalMedicineandPublicHealth,UniversityofPadua,Padua35131,Italy. 38 UniversityofNewcastleDentalSchool,NewcastleNE24BW,UK. 39 UniversityofGlasgowDental School, Glasgow G2 3JZ, UK. 40 Croatian National Cancer Registry, National Institute of Public Health, Zagreb 10000, Croatia. 41 Trinity College School of Dental Science, Dublin 2, Ireland. 42 Centre National de Genotypage, Institut Genomique, Commissariat a` l?e�nergie Atomique, Evry 91000, France. 43 Fondation Jean Dausset-CEPH, Paris 75010, France. Vol 452|3 April 2008|doi:10.1038/nature06885 633 Nature Publishing Group�2008 require very large series of cases and controls to ensure adequate statistical power, and multiple subsequent studies to confirm the initial findings. We conducted a genome-wide association study of lung cancer using the Illumina Sentrix HumanHap300 BeadChip containing 317,139 SNPs and estimated to tag approximately 80% of common genomic variation 10 . We initially genotyped 1,989 cases and 2,625 controls from the International Agency for Research on Cancer (IARC) central Europe lung cancer study. This was con- ducted in six countries between 1998 and 2002 and each centre fol- lowed an identical protocol to recruit newly diagnosed cases of primary lung cancer, as well as a comparable group of population or hospital controls (Supplementary Methods). We excluded sam- ples that failed one of several quality control criteria (Supplementary Methods) or because they showed evidence of admixture with Asian ethnicity(SupplementaryFig.1);wealsoexcluded7,116problematic SNPs. This resulted in a comparison of 310,023 SNPs between 1,926 cases and 2,522 controls. We analysed each SNP individually by calculating P-values for trend in a logistic regression model and incorporating additional parameters including country, age and sex (Supplementary Methods). The distribution of the bottom 90% of P-values was similar to the expected distribution, and the genomic control para- meter was 1.03, implying that there was no systematic increase in false-positivefindingsowingtopopulationstratificationoranyother form of bias (Fig. 1a). However, there was a marked deviation between the observed and expected P-values among the top 10% (Fig. 1b). In particular, two SNPs on chromosome 15q25, rs1051730 and rs8034191, were strongly associated with disease (P55310 29 and P59 310 210 , respectively), exceeding the genome-wide significance level of P55310 27 (Fig. 1c). Further analysis incorporating adjustment by principal components indi- cated that population stratification was unlikely to account for this observation (Supplementary Methods). Theoddsratio(OR)and95%confidenceinterval(CI)forcarrying onecopyofthemostsignificantmarker(rs8034191),adjustedbyage, sex and country, was 1.27 (1.11?1.44) and for carrying two copies of the allele was 1.80 (1.49?2.18); the allelic OR was 1.32 (1.21?1.45). When the data were analysed separately by country of origin, we found a significant association in all countries except Romania, which had the smallest sample numbers, although the trend in Romania was similar and the association was significant under a dominant model (data not shown). There was no evidence of heterogeneity by country of origin (P50.58). Further adjustment wasundertakenforvarioustobacco-relatedvariablesincludingdura- tion of smoking, pack years (average number of cigarette packs per day multiplied by years of smoking) and age at onset of smoking. Adjustment by duration of smoking provided the best-fitting model toaccountfortobaccousebasedontheAkaike?sinformationcriteria (Supplementary Methods), although the adjusted estimates with duration of smoking (allelic OR51.28 (1.16?1.42)) were similar to the estimates adjusted by age, sex and country only. We investigated further the association by genotyping 34 addi- tional 15q25 markers that were selected as follows. First, we used animputationmethod(seehttp://www.sph.umich.edu/csg/abecasis/ MACH/index.html) to identify additional genetic variants from the Centre d?Etude du Polymorphism Humain Utah (CEU) HapMap data that are likely to have a strong disease association, but are not present in the HumanHap300 panel. We attempted genotyping of SNPs from the 15q25 region with an association P-value of the imputed data of,10 26 . Second, we included SNPs of CHRNA5 and CHRNA3 that had been included in a previous study of these genes in nicotine dependence 11 . Third, we attempted genotyping of all non-synonymous SNPs in dbSNP from the six genes within or near the association region. The results for all markers tested in the 15q25 region, including those in the HumanHap300 panel, are shown in Supplementary Table 1. Twenty-three of the additional genotyped markers showed evidence of association exceeding the genome-widesignificancelevelof5310 27 (Fig.2).Thesespanmore than 182kilobases (kb) but are in strong linkage disequilibrium (pairwise D9.0.8 and r 2 .0.6) with two predominant haplotypes accounting for more than 85% of the haplotypes in patients and controls (Supplementary Table 2). To confirm our findings we genotyped rs8034191 and rs16969968 (where rs16969968 is a second variant with a strong disease asso- ciation) in five further independent studies of lung cancer: the European Prospective Investigation in Cancer and Nutrition (EPIC) cohort study (781 cases and 1,578 controls), the Beta- CaroteneandRetinolEfficacyTrial(CARET)cohortstudy(764cases and 1,515 controls), the Health Study of Nord-Tr�ndelag (HUNT) and Troms� cohort studies (235 cases and 392 controls), the Liverpool lung cancer case-control study (403 cases and 814 con- trols), and the Toronto lung cancer case-control study (330 cases and 453 controls) (Supplementary Methods). We observed an increased risk for both heterozygous and homozygous variants of rs8034191 in all five replication samples (Table 1), with no evidence of any statistical heterogeneity between studies. After pooling across all six studies, the ORs (95% CI) were 1.21 (1.11?1.31) and 1.77 (1.58?2.00) for heterozygous and homozygous carriers, respectively, the allelic OR was 1.30 (1.23?1.37), and the P-value for trend was 5310 220 . Further adjustment for duration of tobacco smoking did not alter the estimates: allelic OR51.30 (1.22?1.40). The genotype- specific model that estimated the OR for heterozygous and homo- zygouscarriers separately wasasignificantly better fitthanthe model estimating theallelicOR (P50.025),suggesting apotential recessive effect. The prevalence of the variant allele was 34%, resulting in 66% of the control participants carrying at least one copy, and the percentage of lung cancer explained by carrying at least one allele (that is, the population attributable risk) was 15% in the combined data set. We obtained a similar attributable risk in the central European study (16%) and in the replication studies (14%). The second variant with strong disease association (rs16969968) that was genotyped in the five replicationstudiesgaveverysimilarresults,asexpectedfromthestrong linkage disequilibrium (D951.00, r 2 50.92) among the disease- associated markers (allelic OR51.30 (1.23?1.38); P51310 220 ). The large number of patients in the combined data set allowed us to examine the association in different smoking categories and with respecttodifferenthistologicalsubtypes(SupplementaryTable3and 123456789101121314161715 19 21 X Chromosome c 246 810 b ?log 10 observed P -value 0 0.2 0.4 0.6 0.8 1.0 ?log 10 observed P -value 2 4 6 8 10 a 0 0.2 0.4 0.6 0.8 1.0 ?log 10 expected P-value ?log 10 expected P-value ?log 10 P -value 2 4 6 8 10 1 0 3 5 7 9 Figure 1 | Genome-wide association results in the central Europe study. a?c, Quantile?quantile plot for bottom 90% of P-values (a) and top 10% of P-values (b), as well as scatter plot (c)ofP-values in 2log scale from the trendtestfor310,023genotypedvariantscomparing1,926lungcancercases and 2,522 controls. LETTERS NATURE|Vol 452|3 April 2008 634 Nature Publishing Group�2008 Supplementary Discussion). Increased risks were seen for former smokers (P54310 27 ) and current smokers (P53310 210 ), as well as a potential increased risk for people who had never smoked (P50.013).Noappreciablevariationoftheriskwasfoundacrossthe main histological subtypes of lung cancer. We observed a similar risk after stratifying by age at diagnosis, and a slightly greater risk for women compared to men (P50.06) (Supplementary Table 3). Analysis of the susceptibility locus in additional lung cancer studies wouldbedesirabletoobtainfurtherinformation onthesepatternsof risk,particularly withrespecttosmoking status,cumulativecigarette consumption,ageandsex.Notably,theriskhaplotypeisrareinAsian (JapaneseandChinese)andnotobservedinAfrican(Yoruba)datain the HapMap database 12 and many of the risk alleles have markedly varied allele frequencies in different populations (Supplementary Table 1). Thus, future examination of the association of these mar- kers with lung cancer in different populations might contribute to refined mapping of the locus. We further investigated whether the locus was associated with cancersoftheheadandneckincludingthoseoftheoralcavity,larynx, pharynx and oesophagus. We analysed rs8034191 in two separate studies of head and neck cancer conducted in Europe, the first being conducted in five countries of central Europe and overlapping with the lung cancer controls from five of the six countries included in the present genome-wide association study (726casesand 694 controls), and the second study being conducted in eight countries of Europe (the ARCAGE study) and including 1,536 cases and 1,443 controls. We observed no effect in either of the two studies separately or combined or in any of the cancer subgroups (Supplementary Fig. 2), implying that this association was specific for lung cancer. Similar results were also observed for rs16969968 (data not shown). The disease-associated markers span six known genes, including the nicotinic acetylcholine receptor subunits CHRNA5, CHRNA3 and CHRNB4, the IREB2 iron-sensing response element, PSMA4, which is implicated in DNA repair, and LOC123688, a gene of rs8034191 rs2656052 Centromere Telomere rs8031948 rs16969968 rs2036527 rs1317286 a b c rs17484235 rs10519203 rs931794 rs951266 CHRNB4CHRNA5 CHRNA3 PSMA4LOC123688IREB2 rs8034191rs2656052 rs8031948 rs16969968rs2036527 rs1317286rs17484235 rs10519203 rs931794 rs95126676.49 Mb 76.73 Mb 11 10 9 8 7 6 5 4 3 2 1 0 ?log 10 ( P -value) 76.4 Mb 76.49 Mb 76.73 Mb 76.8 Mb Figure 2 | Lung cancer area of interest across 15q25. a, P-values for SNPs genotypedin the 15q25 region(76.4?76.8Mb). Thedotted line indicatesthe genome-wide threshold of P,5310 27 . Points labelled with rs numbers haveaP,1310 29 .Pointsinredaregenotypedinthe317KIlluminapanel; pointsinblueindicateadditionalgenotypedSNPs(Taqman).b,Positionsof the six known genes. c, Pairwise r 2 estimates for 46 common SNPs from 76.49Mb to 76.73Mb in controls from the central Europe IARC study, with increasing shades of grey indicating higher r 2 values. The majority of pairwise D9 estimates for these SNPs exceed 0.8. Table 1 | Lung cancer risk and rs8034191 genotype T/C versus T/T genotype C/C versus T/T genotype Co-dominant model Cases* Controls* OR 95%CI OR 95%CI OR 95%CI P-values P-heterogeneity Overall 4,435 7,272 1.21 1.11?1.31 1.77 1.58?2.00 1.30 1.23?1.37 5310 220 By study 0.951 Central Europe 1,922 2,520 1.27 1.11?1.44 1.80 1.49?2.18 1.32 1.21?1.45 9310 210 Toronto 330 453 1.20 0.85?1.68 1.84 1.14?2.97 1.32 1.05?1.65 0.017 EPIC 781 1,578 1.18 0.97?1.43 1.68 1.29?2.19 1.27 1.12?1.44 2310 24 CARET 764 1,515 1.31 1.08?1.58 1.77 1.34?2.34 1.33 1.16?1.51 2310 25 Liverpool 403 814 1.04 0.80?1.34 1.65 1.11?2.44 1.20 1.00?1.44 0.047 HUNT/ Tromso� 235 392 1.09 0.77?1.54 2.02 1.21?3.37 1.32 1.04?1.68 0.022 Odds ratio (OR) and 95% confidence interval (CI) for lung cancer comparing heterozygous (T/C) and homozygous (C/C) genotypes of rs8034191 to homozygous (T/T) genotype, overall and separately for each of the six studies. ORs are standardized by age, sex and country. P-values are derived from the co-dominant model. *Subjects with valid call for rs8034191. NATURE|Vol 452|3 April 2008 LETTERS 635 Nature Publishing Group�2008 unknown function (Fig. 2). It is not possible to identify likely causal alleles or genes based on the differences in the strength of the statistical association because of the strong linkage disequilibrium. However, the nicotinic acetylcholine receptor subunits are strong candidate genes. CHRNA5 was the only gene found to contain a non-synonymous variant (rs16969968 in exon 5) with strong disease association (P53310 29 ). CHRNA3 contained a synonymous variant in exon 5 (rs1051730) that was also strongly associated with disease (P55310 29 ); the r 2 between these two variants being 0.99. Although the other markers with a strong disease association either resided in introns or were inter-genic, we cannot exclude the pos- sibility that they could have a biological effect on one or more of the genes from the region. However, other lines of evidence support a possible role for the nicotinic acetylcholine receptor subunit genes. Nicotinic acetylcholine receptor subunit genes code for proteins that form receptors present in neuronal and other tissues, in particu- laralveolarepithelialcells,pulmonaryneuroendocrinecells,andlung cancer cell lines 2,3 , and they bind to nicotine and nicotine derivatives including N9-nitrosonornicotine. An association of CHRNA3 and CHRNA5 variants with nicotine dependence has been reported 11,13 . Theassociatedmarkersincludethenon-synonymousCHRNA5SNP, rs16969968,whichisoneofourmarkersoflungcancerrisk.ThisSNP introduces a substitution of aspartic acid (D) to asparagine (N) at aminoacidposition398(D398N)oftheCHRNA5protein,locatedin the central part of the second intracellular loop. Although the func- tion of the second intracellular loop and the possible biological con- sequences of the D398N alteration remain to be elucidated, this aminoacidishighlyconservedacrossspecies,suggestingthatitcould have functional importance (Supplementary Fig. 3). A T529A sub- stitution in the second intracellular loop of a4 nAchR, another nico- tinic acetylcholine receptor subunit, is known to lead to altered responses to nicotine exposure in the mouse 14 . WithintheARCAGEstudy(seeabove),allparticipantswereasked a series of questions relating to tobacco addiction based on the Fagerstrom tolerance questionnaire 15 , and we used these to examine whether the chromosome 15q25 locus might be implicated in lung cancer through involvement in tobacco dependence. Two of these questions(?timetofirstcigarette?and?numbersofcigarettesperday?) have been shown to be particularly strongly associated with nicotine dependence, and responses toboth questions result in a ?heaviness of smoking index (HSI)? with a score of between 0 and 6 (ref. 16). We did not observe an association in the ARCAGE controls between rs16969968 and any of the individual Fagerstrom indices of nicotine addiction,orwhencomparingcontrolswithaHSIof0tothosewitha HSIof3ormore (Supplementary Table4).Almost identicalpatterns wereobservedforrs8034191(datanotshown).Thus,ourdatadonot support an important role for the locus in nicotine addiction. However, a previous study of a large number of candidate gene markers (4,309 SNPs) identified a possible association between rs16969968 and addiction (uncorrected P-value56.4310 24 ) using contrasting extreme phenotypes as measured by the Fagerstrom test for nicotine dependence (FTND) 11 . Asecond study also identified an associationbetweenvariantsintheregionofchromosome15q25and numbers of cigarettes smoked per day, although it did not assess directly rs16969968 13 . The FTND and HSI measures of nicotine dependence are highly correlated together, and with cigarettes per day 17 ,andadditional studiestoclarifytherelationshipbetweenchro- mosome 15q25 variants and tobacco dependence are warranted in light of these results. Our observation of an increased risk with the chromosome 15q25 locus and lung cancer in non-smokers, as well as the lack of an associationwithsmoking-relatedheadandneckcancers,wouldindi- cate that the disease mechanism with lung cancer is unlikely to be explained by an association with tobacco addiction. Independent biological data also suggest that nicotinic acetylcholine receptors could be involved in lung cancer through other mechanisms. It has been suggested that N9-nitrosonornicotine and nitrosamines may facilitate neoplastic transformation by stimulating angiogenesis and tumour growth mediated through their interaction with nico- tinic acetylcholine receptors 18?20 . The expression of these receptors can also be inhibited by nicotine receptor antagonists, which, if con- firmedtobeinvolvedindiseaseaetiologythroughsuchamechanism, implies possible chemoprevention opportunities for lung cancer 5 . No markers outside of those on chromosome 15q25 exceeded the genome-wide significance level for association with lung cancer, although a further 29 had a significance level of P,5310 25 (Supplementary Table 5). Although most were isolated markers, tenwerefoundtobeclusteredinasegmentofapproximately1mega- base (Mb) on chromosome 6p (28.5?29.5Mb) within an extended region of high linkage disequilibrium around the major histocom- patibility complex. Genotyping of the most significant SNP from the 6pregion(rs4324798)intheotherfivestudiesprovidedindependent evidenceofassociation(P54310 23 ).In thecombineddataset,the trend test reached genome-wide significance (P54310 27 ; see Supplementary Fig. 4). The region contains up to 20 documented genes and identification of causal variants is complicated by strong linkagedisequilibriumbetweenvariantswithinneighbouringhuman leukocyte antigen (HLA) and non-HLA genes 21 . Further analyses in multiple diverse populations will be required to confirm this locus and to identify additional lung cancer susceptibility variants. To aid in this, we have made our genome-wide association results available through a publicly accessible website (http://www.ceph.fr/cancer). METHODS SUMMARY A detailed description of the component studies can be found in the Supplementary Methods. The genotyping of the IARC central Europe study was conducted using Illumina Sentrix HumanHap300 BeadChip. We excluded variants with a call rate of less than 95% or whose allele distributions deviated strongly from Hardy?Weinberg equilibrium among controls. We also excluded subjects with a completion rate less than 90% or whose reported sex did not matchwiththeinferredsexbasedontheheterozygosityratefromtheXchromo- somes. Unexpected duplicates and unexpected first-degree relatives were also excluded from the analysis. Additional quality control measures were applied as described in the Supplementary Methods. Population outliers were detected using STRUCTURE 22 with HapMap subjects as internal controls, and were subsequently excluded from the analysis. Additional analyses for population stratification were undertaken with EIGENSTRAT 23 . Odds ratios (OR) and 95% confidence intervals (CI) were calculated using multivariate unconditional logistic regression models. CEU HapMap SNPs were imputed using MACH (http://www.sph.umich.edu/csg/abecasis/MACH/index.html). Genotyping of additional markers was undertaken with Taqman or Amplifluor assays. Genotyping for all five replication studies was conducted for rs8034191 and rs16969968, and effect estimates from all six lung cancer studies were combined using a fixed-effect model. All P-values are two-sided. Received 30 November 2007; accepted 7 March 2008. 1. Ferlay, J., Bray, F., Pisani, P. & Parkin, M. GLOBOCAN 2002. IARC CancerBase No 5, version 2.0 (IARC, Lyon, 2004). 2. Minna,J.D.Nicotineexposureandbronchialepithelialcellnicotinicacetylcholine receptor expression in the pathogenesis of lung cancer. J. Clin. Invest. 111, 31?33 (2003). 3. Wang, Y. et al. Human bronchial epithelial and endothelial cells express a7 nicotinic acetylcholine receptors. Mol. Pharmacol. 60, 1201?1209 (2001). 4. Schuller, H. M. Nitrosamines as nicotinic receptor ligands. Life Sci. 80, 2274?2280 (2007). 5. Russo, P., Catassi, A.,Cesario, A. &Servent, D. Development ofnovel therapeutic strategies for lung cancer: targeting the cholinergic system. Curr. Med. Chem. 13, 3493?3512 (2006). 6. International Agency for Research on Cancer. Reversal of risk after quitting smoking. IARC Handbooks of Cancer Prevention Vol. 11, 15?27 (IARC, Lyon, 2007). 7. International Agency for Research on Cancer. Tobacco smoke and involuntary smoking. IARC Monographs Vol. 83, 33?47 (IARC, Lyon, 2004). 8. Coleman,M.P.etal.EUROCAREWorkingGroup.EUROCARE-3summary:cancer survival in Europe at the end of the 20th century. Ann. Oncol. 14 (suppl. 5), v128?v149 (2003). 9. Matakidou, A., Eisen, T. & Houlston, R. S. Systematic review of the relationship between family history and lung cancer risk. Br. J. Cancer 93, 825?833 (2005). 10. Barrett, J. C. & Cardon, L. R. Evaluating coverage of genome-wide association studies. Nature Genet. 38, 659?662 (2006). LETTERS NATURE|Vol 452|3 April 2008 636 Nature Publishing Group�2008 11. Saccone, S. F. et al. Cholinergic nicotinic receptor genes implicated in a nicotine dependence association study targeting 348 candidate genes with 3713 SNPs. Hum. Mol. Genet. 16, 36?49 (2007). 12. International HapMap Consortium. A second generation human haplotype map of over 3.1 million SNPs. Nature 449, 851?861 (2007). 13. Berretini, W. et al. a-5/a-3 nicotinic receptor subunit alleles increase risk for heavy smoking. Mol. Psychiatry advance online publication doi:10.1038/ sj.mp.4002154 (29 January 2008). 14. Tritto, T., Stitzel, J. A., Marks, M. J., Romm, E. & Collins, A. C. Variability in responsetonicotineintheLSxSSRIstrains:potentialroleofpolymorphismsina4 and a6 nicotinic receptor genes. Pharmacogenetics 12, 197?208 (2002). 15. Fagerstrom, K.O. &Schneider, N.G. Measuring nicotine dependence: areview of the Fagerstrom Tolerance Questionnaire. J. Behav. Med. 12, 159?182 (1989). 16. Heatherton,T.F.,Kozlowski,L.T.,Frecker,R.C.&Fagerstrom,K.O.AFagerstrom test for nicotine dependence: a revision of the Fagerstrom tolerance questionnaire. Br. J. Addict. 86, 1119?1127 (1991). 17. Chabrol, H. et al. Comparison of the Heavy Smoking Index and of the Fagerstrom test for nicotine dependence in asample of749 cigarette smokers. Addict. Behav. 30, 1474?1477 (2005). 18. Lam, D. C. et al. Expression of nicotinic acetylcholine receptor subunit genes in non-small-cell lung cancer reveals differences between smokers and nonsmokers. Cancer Res. 67, 4638?4647 (2007). 19. West, K. A. et al. Rapid Akt activation by nicotine and a tobacco carcinogen modulates the phenotype of normal human airway epithelial cells. J. Clin. Invest. 111, 81?90 (2003). 20. Dasgupta, P. & Chellappan, S. P. Nicotine-mediated cell proliferation and angiogenesis: new twists to an old story. Cell Cycle 5, 2324?2328 (2006). 21. De Bakker, P. I. et al. A high resolution HLA and SNP haplotype map for disease associationstudiesintheextendedhumanMHC.NatureGenet.38,1166?1172(2006). 22. Falush, D., Stephens, M. & Pritchard, J. K. Inference of population structure using multilocus genotype data: linked loci and correlated allele frequencies. Genetics 164, 1567?1587 (2003). 23. Price, A. L. et al. Principal components analysis corrects for stratification in genome-wide association studies. Nature Genet. 38, 904?909 (2006). Supplementary Information is linked to the online version of the paper at www.nature.com/nature. Acknowledgements The authors thank all of the participants who took part in this research and the funders and support staff who made this study possible. We also thank R. Peto for his comments on the manuscript. Funding for the initial genome-wide study was provided by INCa, France. Additional funding for replication studies was provided by the US NCI (R01 CA092039) and the Ontario Institute for Cancer Research (OICR). AuthorContributionsP.B.andM.L.designedthestudy.R.J.H.,J.D.M.,A.B.andH.B. coordinated the preparation and inclusion of all biological samples. R.J.H., J.D.M., V.G. and S.H. undertook the statistical analysis. Bioinformatics analysis was undertakenbyF.M.,M.F.andS.H.,D.Z.andM.D.coordinatedthegenotypingofthe central Europe samples, and J.D.M, R.J.H. and V.G. coordinated the genotyping of the other studies. All other co-authors coordinated the initial recruitment and management of the studies. M.L. obtained financial support for genotyping of the central Europe study, and P.B. and R.J.H. obtained financial support for genotyping of the other studies. P.B. and M.L. drafted the manuscript with substantial contributions from R.J.H. and J.D.M. All authors contributed to the final paper. Author Information Reprints and permissions information is available at www.nature.com/reprints. Correspondence and requests for materials should be addressed to P.B. (brennan@iarc.fr). NATURE|Vol 452|3 April 2008 LETTERS 637 Nature Publishing Group�2008 "
Add Content to Group
|
Bookmark
|
Keywords
|
Flag Inappropriate
share
Close
Digg
Facebook
MySpace
Google+
Comments
Close
Please Post Your Comment
*
The Comment you have entered exceeds the maximum length.
Submit
|
Cancel
*
Required
Comments
Please Post Your Comment
No comments yet.
Save Note
Note
View
Public
Private
Friends & Groups
Friends
Groups
Save
|
Cancel
|
Delete
Please provide your notes.
Next
|
Prev
|
Close
|
Edit
|
Delete
Genetics
Gene Inheritance and Transmission
Gene Expression and Regulation
Nucleic Acid Structure and Function
Chromosomes and Cytogenetics
Evolutionary Genetics
Population and Quantitative Genetics
Genomics
Genes and Disease
Genetics and Society
Cell Biology
Cell Origins and Metabolism
Proteins and Gene Expression
Subcellular Compartments
Cell Communication
Cell Cycle and Cell Division
Scientific Communication
Career Planning
Loading ...
Scitable Chat
Register
|
Sign In
Visual Browse
Close
Comments
CloseComments
Please Post Your Comment