Genetic Variants in Epigenetic Pathways and Risks of Multiple Cancers in the GAME-ON Consortium
Total Page:16
File Type:pdf, Size:1020Kb
Published OnlineFirst January 23, 2017; DOI: 10.1158/1055-9965.EPI-16-0728 Research Article Cancer Epidemiology, Biomarkers Genetic Variants in Epigenetic Pathways and & Prevention Risks of Multiple Cancers in the GAME-ON Consortium Reka Toth1,2, Dominique Scherer1,3, Linda E. Kelemen4, Angela Risch2,5,6,7, Aditi Hazra8,9,Yesilda Balavarca1, Jean-Pierre J. Issa10,Victor Moreno11, Rosalind A. Eeles12, Shuji Ogino13,14,15, Xifeng Wu16, Yuanqing Ye16, Rayjean J. Hung17,18, Ellen L. Goode19, and à Ãà Cornelia M. Ulrich1,20,21, on behalf of the OCAC, CORECT , TRICL, ELLIPSE , DRIVE, and GAME-ON consortiaa Abstract Background: Epigenetic disturbances are crucial in cancer ated with ER-negative breast, endometrioid ovarian, and overall initiation, potentially with pleiotropic effects, and may be influ- and aggressive prostate cancer risk (OR ¼ 0.93; 95% CI ¼ 0.91– enced by the genetic background. 0.96; q ¼ 0.005). Variants in L3MBTL3 were associated with Methods: In a subsets (ASSET) meta-analytic approach, we colorectal, overall breast, ER-negative breast, clear cell ovarian, investigated associations of genetic variants related to epigenetic and overall and aggressive prostate cancer risk (e.g., rs9388766: mechanisms with risks of breast, lung, colorectal, ovarian and OR ¼ 1.06; 95% CI ¼ 1.03–1.08; q ¼ 0.02). Variants in TET2 were prostate carcinomas using 51,724 cases and 52,001 controls. False significantly associated with overall breast, overall prostate, over- discovery rate–corrected P values (q values < 0.05) were consid- all ovarian, and endometrioid ovarian cancer risk, with ered statistically significant. rs62331150 showing bidirectional effects. Analyses of subpath- Results: Among 162,887 imputed or genotyped variants in 555 ways did not reveal gene subsets that contributed disproportion- candidate genes, SNPs in eight genes were associated with risk of ately to susceptibility. more than one cancer type. For example, variants in BABAM1 were Conclusions: Functional and correlative studies are now need- confirmed as a susceptibility locus for squamous cell lung, overall ed to elucidate the potential links between germline genotype, breast, estrogen receptor (ER)–negative breast, and overall pros- epigenetic function, and cancer etiology. tate, and overall serous ovarian cancer; the most significant variant Impact: This approach provides novel insight into possible was rs4808076 [OR ¼ 1.14; 95% confidence interval (CI) ¼ 1.10– pleiotropic effects of genes involved in epigenetic processes. Cancer À 1.19; q ¼ 6.87  10 5]. DPF1 rs12611084 was inversely associ- Epidemiol Biomarkers Prev; 26(6); 816–25. Ó2017 AACR. 1National Center for Tumor Diseases and German Cancer Research Center Note: Supplementary data for this article are available at Cancer Epidemiology, (DKFZ), Heidelberg, Germany. 2Division of Epigenomics and Cancer Risk Fac- Biomarkers & Prevention Online (http://cebp.aacrjournals.org/). tors, German Cancer Research Center (DKFZ), Heidelberg, Germany. 3Institute ÃIncluding the Colon Cancer Family Registry (CCFR) of Medical Biometry and Informatics, University of Heidelberg, Heidelberg, Germany. 4Medical University of South Carolina and Hollings Cancer Center, R. Toth, D. Scherer, R.J. Hung, E.L. Goode, and C.M. Ulrich contributed equally to Charleston, South Carolina. 5Division of Cancer Research and Epigenetics, this article. Department of Molecular Biology, University of Salzburg, Salzburg, Austria. ÃÃThe following authors are linked to ELLIPSE: 6Cancer Cluster Salzburg, Salzburg, Austria. 7Translational Lung Research 1,2 1 3,4 5,6 Center Heidelberg (TLRC-H), Member of the German Center for Lung Research Rosalind A. Eeles ,Zsofia Kote-Jarai , Kenneth Muir , Ali Amin Ala Olama , 1,5 7,8 9 10 (DZL), Heidelberg, Germany. 8Brigham and Women's Hospital, Harvard Medical Sara Benlloch , David E. Neal , Jenny L. Donovan , Freddie C Hamdy , 11,12 13 13 School, Harvard T.H. Chan School of Public Health, Boston, Massachusetts. Graham G. Giles , Fredrik Wiklund , Henrik Gronberg , Christopher A. Hai- 14 15 17 18 19 9Channing Division of Network Medicine, Department of Medicine, Brigham man , Fredrick Schumacher , David J. Hunter , Peter Kraft , Ruth C. Travis , 20 21 22 21 and Women's Hospital, Harvard Medical School, Boston, Massachusetts. Elio Riboli , Stephen J. Chanock , Sonja I. Berndt , Demetrius Albanes 10 11 School of Medicine, Temple University, Philadelphia, Pennsylvania. Catalan 1The Institute of Cancer Research. 2Royal Marsden NHS Foundation Trust. Institute of Oncology, IDIBELL, L'Hospitalet de Llobregat, Barcelona, Catalonia, 3Institute of Population Health, University of Manchester 4Warwick Medical 12 13 Spain. The Institute of Cancer Research, London, United Kingdom. Depart- School, University of Warwick. 5Centre for Cancer Genetic Epidemiology, Depart- ment of Pathology, Brigham and Women's Hospital, and Harvard Medical ment of Public Health and Primary Care, University of Cambridge. 6Department 14 School, Boston, Massachusetts. Department of Medical Oncology, Dana-Far- of Clinical Neurosciences, University of Cambridge. 7Department of Oncology, 15 ber Cancer Institute, Boston, Massachusetts. Department of Epidemiology, University of Cambridge. 8Cancer Research UK Cambridge Research Institute, Li 16 Harvard T.H. Chan School of Public Health, Boston, Massachusetts. Department Ka Shing Centre 9School of Social and Community Medicine, University of Bristol. of Epidemiology, The University of Texas MD Anderson Cancer Center, Houston, 10Nuffield Department of Surgical Sciences, Faculty of Medical Science, Univer- 17 Texas. Lunenfeld-Tanenbaum Research Institute of Mount Sinai Hospital, sity of Oxford, Oxford, United Kingdom.11Cancer Epidemiology Centre, The 18 19 Toronto, Ontario, Canada. University of Toronto, Toronto, Canada. Mayo Cancer Council Victoria. 12Centre for Epidemiology and Biostatistics, Melbourne 20 Clinic College of Medicine, Rochester, Minnesota. Fred Hutchinson Cancer School of Population and Global Health, The University of Melbourne. 13Depart- 21 Research Center, Seattle, Washington. Huntsman Cancer Institute, Salt Lake ment of Medical Epidemiology and Biostatistics, Karolinska Institute. 14 Depart- a City, Utah. Members of GAME-ON ELLIPSE. ment of Preventive Medicine, Keck School of Medicine, University of Southern 816 Cancer Epidemiol Biomarkers Prev; 26(6) June 2017 Downloaded from cebp.aacrjournals.org on September 24, 2021. © 2017 American Association for Cancer Research. Published OnlineFirst January 23, 2017; DOI: 10.1158/1055-9965.EPI-16-0728 Epigenetic-Related Variants and Risks of Multiple Cancers Introduction Materials and Methods Genetic and epigenetic alterations are hallmarks of cancer Study population initiation and progression and can influence each other to work Within the GAME-ON Network, 32 studies from North Amer- cooperatively (1). Dysfunction of epigenetic processes, such as ica and Europe participated in this investigation (12–21). Studies DNA methylation, chromatin remodeling, and covalent histone included frequency matched cases and controls on at least age, modifications, can be as important in carcinogenesis as the change and all subjects were of European descent based on ancestry of the genetic material itself (2). Since the first studies that analyses. The study characteristics are summarized in Table 1. In described the global hypomethylation of cancer genomes and total, 51,724 cancer patients (breast, colorectal, lung, ovarian, and the hypermethylation of the promoter sequence of mainly tumor prostate with respective subtypes) and 52,001 controls were suppressor genes, several "pan-cancer" DNA methylation patterns included in the analysis. (patterns across multiple cancer types) have been identified (reviewed in ref. 3). The CpG island methylator phenotype Gene and variant selection, pathway assignment (CIMP) was first described in colorectal cancer (4), and later Genes (n ¼ 634) involved in epigenetic processes were iden- similar patterns were observed in several other tumor types. tified using GO and GeneCards databases by searching for the Highlighting the interplay between genetic and epigenetic following keywords: DNA methylation, DNA demethylation, changes, CIMP subtypes usually present with characteristic genetic histone acetylation, deacetylation, methylation, demethylation, alterations. CIMP-H colorectal cancers are frequently character- and other histone modification, chromatin remodeling, chroma- ized by BRAF mutations, whereas CIMP-L tumors tend to harbor tin modification, and histones. The recent literature was also KRAS mutations (5). Non-CIMP colorectal cancer, the B-CIMP– reviewed. After excluding genes on sex chromosomes and those negative breast cancer, and the low methylated tumor group of not covered in all cancer sites, 555 genes were included in the serous ovarian cancers frequently acquire TP53 mutations (5–8). analysis, which were categorized into one or more of epigenetic Furthermore, somatic mutations in epigenetic regulatory genes subpathways (Supplementary Table S1). that are either carcinogenic driver or passenger mutations are We analyzed all SNPs residing within 50 kb of the largest known to exist. Important mutations have been shown, for transcript for each gene (for databases, see Supplementary Table example, in DNMTs, IDH1, IDH2 and TETs (as important players S2). Overall, 162,887 polymorphisms were included in the final of DNA methylation); in EZH2 and KDM1A (involved in histone analysis. In the combined dataset, the major alleles (according