Pleiotropic Associations of Risk Variants Identified for Other Cancers With Lung Cancer Risk: The PAGE and TRICL Consortia

S Lani Park,Steven Buyske,Joachim Heinrich,David C Christiani,Charles Kooperberg,Ewa Deelman,Paul Brennan,Wei Chen,Younghun Han,Megan D Fesinmeyer,Rayjean J Hung,Iona Cheng,Mala Pande,Robert Goodloe,Richard S Houlston,Yufei Wang,Logan Dumitrescu,William S Bush,Maria Timofeeva,Lynne R Wilkens,Timothy Eisen,Gerardo Heiss,Angela Risch,Maria Teresa Landi,Yi Lin,Kari E North,Jay H Fowke,John K Field,Alicia Young,Dana C Crawford,Christopher I Amos,Christopher A Haiman,Ulrike Peters,Michele L Coté ,Demetrius Albanês ,Lucia A Hindorff ,Fredrick Schumacher ,Anne M Butler ,Michael J Thun ,Cara L Carty ,Holli H Dilks ,Loı̈c Le Marchand ,Heike Bickebller ,Jonathan Kocarnik ,Neil E Caporaso ,Christian Caberto ,Shelly Ann Love

doi:10.1093/jnci/dju061

Abstract

BackgroundGenome-wide association studies have identified hundreds of genetic variants associated with specific cancers. A few of these risk regions have been associated with more than one cancer site; however, a systematic evaluation of the associations between risk variants for other cancers and lung cancer risk has yet to be performed.MethodsWe included 18023 patients with lung cancer and 60543 control subjects from two consortia, Population Architecture using Genomics and Epidemiology (PAGE) and Transdisciplinary Research in Cancer of the Lung (TRICL). We examined 165 single-nucleotide polymorphisms (SNPs) that were previously associated with at least one of 16 non–lung cancer sites. Study-specific logistic regression results underwent meta-analysis, and associations were also examined by race/ethnicity, histological cell type, sex, and smoking status. A Bonferroni-corrected P value of 2.5×10–5 was used to assign statistical significance.ResultsThe breast cancer SNP LSP1 rs3817198 was associated with an increased risk of lung cancer (odds ratio [OR] = 1.10; 95% confidence interval [CI] = 1.05 to 1.14; P = 2.8×10–6). This association was strongest for women with adenocarcinoma (P = 1.2×10–4) and not statistically significant in men (P = .14) with this cell type (P het by sex = .10). Two glioma risk variants, TERT rs2853676 and CDKN2BAS1 rs4977756, which are located in regions previously associated with lung cancer, were associated with increased risk of adenocarcinoma (OR = 1.16; 95% CI = 1.10 to 1.22; P = 1.1×10–8) and squamous cell carcinoma (OR = 1.13; CI = 1.07 to 1.19; P = 2.5×10–5), respectively.ConclusionsOur findings demonstrate a novel pleiotropic association between the breast cancer LSP1 risk region marked by variant rs3817198 and lung cancer risk.

Highlights

We examined 165 single-nucleotide polymorphisms (SNPs) that were previously associated with at least one of 16 non–lung cancer sites
Our findings demonstrate a novel pleiotropic association between the breast cancer LSP1 risk region marked by variant rs3817198 and lung cancer risk
Our study included a collaboration between two large consortia [11,12], in which we examined the consistency of associations by race/ ethnicity, tumor histology, sex, and smoking status

Summary

Methods

Study Participants Two consortia contributed data to this study: the Population Architecture using Genomics and Epidemiology (PAGE) [12] and the Transdisciplinary Research in Cancer of the Lung (TRICL) [11], which is part of the Genetic Associations and MEchanisms in ONcology (GAME-ON) consortium, and is associated with the International Lung Cancer Consortium (ILCCO). This collaboration provided information on 18 023 patients with lung cancer and 60 543 control subjects from 13 studies (Supplementary Table 1, available online). Participants’ informed consent and institutional review board approval was obtained for all studies except Epidemiologic Architecture for Genes Linked to Environment, which accesses the Vanderbilt University biorepository (EAGLE-BioVU), which is considered nonhuman subjects research due to sample de-identification [13]

Results

Discussion

Conclusion