| Background:Lung cancer is one of the most serious malignancies that endanger human health in the world.The latest statistics from WHO showed that the global incidence of lung cancer in 2020 ranks second among all malignancies,second only to breast cancer,and the mortality rate ranks first among all malignancies.In China,due to the huge population base,the incidence and mortality of lung cancer rank the second and the first place among all malignancies,respectively.As the major histological type of lung cancer,non-small cell lung cancer（NSCLC）accounts for about 85%of all lung cancer cases.Previous studies have shown that lung cancer is one of the complex diseases with multiple causes,and its occurrence is influenced by both environmental and genetic factors.Smoking is the most important environmental risk factor that affects lung cancer.However,less than 20%of smokers will finally develop into lung cancer,suggesting that different individuals have various susceptibility to lung cancer.As a mature molecular epidemiological research method,genome-wide association study（GWAS）have successfully identified hundreds of tumor-associated single nucleotide polymorphism（SNP）,providing new insights into the study of genetic mechanisms of complex diseases and traits.Currently,GWASs have identified 51 susceptibility loci for lung cancer in European and Asian populations.Similar to other chronic non-communicable diseases,as an age-related disease,the median age at diagnosis of lung cancer is about 65～70 years,and those diagnosed no more than 50 years were often defined as early-onset lung cancer cases.The percentage of early-onset lung cancer usually varies between 10 and 15%and has been shown to be increasing over time,which has attracted more and more attention from researchers.Although previous studies have found several genetic variants associated with the early-onset of lung cancer,these studies were based on candidate gene strategies,with small sample sizes and mainly European populations.Till present,no GWAS of early-onset lung cancer has been reported,and it is urgent to carry out this research based on a large-scale Chinese population.Aim:In this study,based on the GWAS data of large-scale non-small cell lung cancer in the Chinese population,we conducted a GWAS of early-onset lung cancer to identify novel early-onset lung cancer susceptibility loci,and further bioinformatics analyses were applied to explore the potential biological mechanism of these variants.Methods:We conducted both genome-wide case-control and case-case analyses that included 2,556 early-onset lung cancer cases,10,769 lung cancer cases older than 50years,and 13,327 controls.All participants were from three studies:（1）NJMU GSA Project,which included three studies（the Nanjing GSA GWAS:4,149 cases and3,198 controls;the Beijing GSA GWAS:2,155 cases and 2,035 controls;and the Guangzhou GSA GWAS:3,944 cases and 4,065 controls）;（2）NJMU GWAS,which included two studies（the Nanjing GWAS:1,473 NSCLC cases and 1,962 controls;and the Beijing GWAS:858 NSCLC cases and 1,115 controls）;and（3）the NJMU Onco Array GWAS with 953 NSCLC cases and 953 controls.Analyses on the above GWASs were performed respectively,according to the following steps:（1）Genome-wide association analysis of early-onset lung cancer cases and controls:logistic regression was performed to calculate the effect of each variant,with age,sex,pack-year or smoking status,and the first ten principal components（PCs）as covariables;（2）A meta-analysis was conducted to combine the estimate of each GWAS,with P<5×10-6as the genome-wide significance threshold.Cochran’s Q test was applied to exclude variants that showed heterogeneity among different studies(I2≥75%or P-value for Cochran’s Q statistic≤1×10-4).（3）Sensitivity analysis:we further performed a case-case analysis on early-onset lung cancer cases and cases older than 50 years to identify the specific variants of early-onset lung cancer.Cox model adjusted by sex,pack-year or smoking status,and the first ten PCs was performed to calculate the hazard ratio（HR）and 95%CI.A meta-analysis of fixed-effect model（with P≤0.05 as the significance threshold）and Cochran’s Q test(I2≥75%or P≤1×10-4)were also performed;（4）Subgroup analyses based on gender,smoking status,and histological type were carried out to identify whether the effects of these variants were heterogeneous between different subgroups;（5）Interaction analysis was performed to explore the interaction between variants and smoking status;（6）Functional annotation and expression quantitative trait locus（e QTL）analysis based on multiple bioinformatics databases were performed to further explore the biological mechanism of the identified variants.Results:The present study identified four SNPs significantly associated with the risk of early-onset lung cancer(P<1.00×10-6),of which three SNPs were novel:rs2055817(5p15.1,OR=1.24,95%CI:1.15-1.35,P=1.39×10-7);rs9403497(6q24.2,OR=1.24,95%CI:1.15-1.35,P=1.61×10-7);and rs4762093(12q14.3,OR=1.31,95%CI:1.18-1.45,P=1.90×10-7).Rs2853677(5p15.33,OR=1.48,95%CI:1.36-1.60,P=3.58×10-21)had been reported by lung cancer GWASs.Subgroup analysis found that both rs2853677（5p15.33）and rs4762093（12q14.3）showed significant heterogeneity among participants with different smoking status（I2≥75%,or P-heterogeneity≤0.05）.The further gene-environment interaction analysis showed negative interactions for rs2853677(P-interaction=5.54×10-3)and rs80176294(P-interaction=6.85×10-3)with smoking.e QTLs analysis found that rs4762093[T]was associated with a decreased expression of LEMD3(P=1.00×10-8),suggesting that these variations may affect the risk of early-onset lung cancer by regulating the expression of target genes.Conclusion:The present study identified three novel susceptibility loci for early-onset lung cancer,two of which were specific for early-onset lung cancer.Our results provide important clues for exploring the biological mechanism of early-onset lung cancer,and new genetic markers for early-onset lung cancer risk prediction and the screening of high-risk populations. |
