| Background:Gastric cancer is one of the most common cancers worldwide.About half of new cases and deaths of gastric cancer occur in China,making gastric cancer a heavy health burden of people in China.Therefore,constructing a gastric cancer risk prediction model to enable detection of high-risk groups remains a urgent scientific problem to be solved.The risk stratification could help identifying targets to apply Helicobactor pylori(H.pylori)eradication and participate in endoscopic screening to achieve personalized prevention.Epidemiological research evidence shows that environmental factors and genetic factors might independently and jointly contribute to the heavy burden of gastric cancer in the Chinese population.Althouth about 56%population in China are infected with H.pylori,only 1-3%of the infected individuals eventually develop gastric cancer,suggesting that genetic factors play an important role in gastric cancer susceptibility.Our lab and other teams have conducted Genome-wide association studies(GWASs)and discovered a number of susceptible loci that affect the risk of gastric cancer and clarified the biological mechanisms of some loci with functional experiments.However,previous studies have found limited genetic loci associated with gastric cancer risk.The application of previously reported loci has not been evaluated using prospective cohort study in a large population.Adhering to a healthy lifestyle has been reported to reduce the risk of gastric cancer.However,the extent to which an increased genetic risk of gastric cancer can be offset by a healthy lifestyle is unknown.Therefore,multi-center and large-samplesize GW AS is required to achieve better understanding of the genetic architecture for gastric cancer and construct the genetic risk prediction model.Subsequent prospective evaluation of the model in a large population is expected to further provide strong evidence for the application.Aim:We aimed to generate a polygenic risk score(PRS)for gastric cancer using GWAS datasets,and to test the utility and effectiveness of this score in predicting the risk of incidnet gastric cancer in a large prospective cohort of Chinese individuals.We also aimed to assess the extent to which adopting a healthy lifestyle might reduce the risk of incident gastric cancer in the same cohort,particularly in the subset of individuals with a high genetic risk,as determined by the PRS.Methods:Using a two-stage study design,we did a meta-analysis and prospective cohort study.The first stage was a case-control design study.In this study,we conducted two new GWASs of gastric cancer including 6,483 cases and 5,488 controls.Combining with 4 previously published GWAS databases,we did a metaanalysis of these 6 studies which including 10,254 cases and 10,914 controls totally.In the second stage,we included 100,220 individuals with genotype and phenotype data available from the China Kadoorie Biobank(CKB)to evaluate the associations between PRS,lifestyle and the risk of gastric cancer.In this study,we included the raw genotype data with genotype array from 6 GWASs of gastric cancer amd conducted unified quality control to the genotype data to exclude unqualified individuals and genetic variants.The qualified genotypes for each chromosome were phased with SHAPEIT v2.12 and imputation was performed using 1000 Genomes Project(the Phase ? database)with Impute2 v2.3.1.In each GWAS,per-allele odds ratios(ORs)and standard errors(SEs)were calculated using multivariable logistic regression with SNPTEST v2.5,with age,sex,and the top ten principal components as covariates.Shared genetic variants in 6 GWASs that met the quality control criteria were included for meta-analysis,which conducted with the inverse-variance weighted fixed effect model by METAL.We removed variants with high heterogeneity among studies(P?1×10-4 for Cochran's Q test,or I2?75%).We constructed 5 PRS with different significance thresholds(P?5×10-8,P?5×10-7,P?5×10-6,P?5×10-5,P?5×10-4)with an r2 threshold of 0.2 on the basis of their linkage disequilibrium patterns using LDpred v1.0.6.Genetic risk was catagorized into low(bottom quintile of PRS),intermediate(quintiles 2 to 4),and high(top quintile)risk categories.Based on previously reported risk factors for gastric cancer,we included 4 lifestyle factors to define a healthy lifestyle,including being a current smoker,no alcohol drinking,a low consumption of preserved foods and the frequent intake of fresh fruit and vegetables.Lifestyles were catagorized into favorable(adopting all 4 healthy lifestyle factors),intermediate(adopting 2 or 3 healthy lifestyle factors)and unfavourable(adopting none or one healthy lifestyle factor)lifestyle.Cox proportional hazards regression model adjusted by age,sex,type of residential area(urban or rural),and the first ten principal components of ancestry was used to assess associations between genetic and lifestyle factors and gastric cancer incidence.Within and across genetic risk groups,we compared the relative risk and absolute risk in different lifestyle populations.Results:The present study identified 6 susceptibility loci of gastric cancer,among which one was novel locus(6p22.1).At 6p22.1,the risk rs28749114-A allele was significantly associated with increased risk of gastric cancer(OR=1.14,95%CI:1.09-1.19,P=5.41×10-9).With a P value and linkage disequilibrium-driven procedure,we constructed 5 gastric cancer PRS in Chinese population:(1)PRS-12(P?5×10-8 and r2<0.2,including 12 SNPs);(2)PRS-18(P?5×10-7 and r2<0.2,including 18 SNPs);(3)PRS-38(P?5×10-6 and r2<0.2,including 38 SNPs);(4)PRS-112(P?5?10-5 and r2<0.2,including 112 SNPs);(5)PRS-539(P?5×10-4 and r2<0.2,including 539 SNPs).Among the 100,220 participants included in the CKB cohort,the median follow-up time was 10.4 years,and a total of 692 new cases of gastric cancer were found.All 5 PRS had an approximate normal distrubution and were significantly associated with gastric cancer risk.Among them,the PRS-112 showed the strongest association with gastric cancer risk(per SD HR=1.27,95%CI:1.17-1.37,P=7.56×10-10).When PRS-112 was used for genetic risk stratification,we found that compared with participants who had a low genetic risk,participants with an intermediate(HR=1.54,95%CI:1.22-1.94,P=2.67×10-4),and high(HR=2.08,95%CI:1.61-2.69,P<0.0001)genetic risk had a significantly greater risk of incident gastric cancer,and the results did not change significantly after adjusting for lifestyle factors.The relative risk of incident gastric cancer was higher in participants with an unfavorable lifestyle than in those with an favourable lifestyle(HR=2.03,95%CI:1.46-2.83,P<0.0001),and the result did not change significantly after adjusting for PRS-112.Compared with individuals with low genetic risk and adopting a favourable lifestyle,the risk of gastric cancer in those with high genetic risk and adopting an unfavourable lifestyle increased by 4.14 times(HR=5.14,95%CI:2.04-12.93,P=5.2×10-4).In the high genetic risk stratification,participants with a favourable lifestyle had a lower risk of gastric cancer than those with an unfavourable lifestyle(HR=0.53,95%CI:0.29-0.99,P=0.048),whith an absolute risk reduction of 1.12%(95%CI:0.62-1.56).Conclusions:Individuals at an increased risk of gastric cancer could be identified by the PRS generated in this study.Genetic and lifestyle factors were independently associated with the risk of gastric cancer.Among individuals with high genentic risk,adhering to a healthy lifestyle cound substantially reduce the risk of gastric cancer.These results have improved the understanding of the roles of genetic risk and lifestyles on gastric cancer,and provided support for the individualized prevention of gastric cancer. |
PDF Full Text Request