| Background: Pancreatic cancer(PanC)is one of the most fatal cancers,which has the characteristics of early metastasis and high mortality.Globally,PanC is the seventh leading cause of cancer deaths.The ATM pathway-related genes play important roles in diverse biological processes and pathological disorders,such as DNA damage repair,autophagy regulation,metabolic disorders and cancers,including PanC.And the cadherin signaling pathway encompasses a large superfamily of membrane proteins that regulate calcium-dependent cell-cell adhesion and tissue morphogenesis,which plays an important role in tumor invasion and metastasis regulation.Based on the pathogenesis and biological behavior of PanC,we hypothesized that genetic variations in the ATM and cadherin pathways are associated with PanC risk.Objective: 1.To identify the associations between genetic variants in the ATM and the cadherin pathways and PanC risk;2.Screening the molecular biomarkers in the ATM and the cadherin pathways that are potentially associated with PanC risk,and initially establishing the PanC risk prediction model based on genetic variations in the ATM and the cadherin pathways;3.To explore the potential molecular biological functions of the PanC risk associated genetic variants and their corresponding genesMaterials and Methods: We included 15,423 European subjects for imputation and single-locus analyses using genotyping data from two previously published PanC genome-wide association studies(GWASs)of Pancreatic Cancer Cohort Consortium(Pan Scan)and Pancreatic Cancer Case-Control Consortium(PanC4).We assessed the associations of 31 499 single nucleotide polymorphisms(SNPs)in the 198 ATM pathway related genes and29 963 SNPs in the 109 cadherin pathway related genes with PanC risk.Multivariable logistic regression analysis,meta-analysis,multiple test correction and multivariable stepwise logistic regression analysis were used to identify SNPs which were independently and significantly associated with PanC risk.Furthermore,genetic model analysis,joint analysis,stratified analysis and the prediction model for PanC risk analysis were conducted for the independent PanC risk-associated SNPs.In addition,a variety of functional prediction software and websites were used to predict potential functions and perform the expression quantitative trait loci(eQTL)analysis for the PanC risk-associated SNPs.Results: 1.We found six SNPs independently and significantly associtated with PanC risk.Three SNPs in the ATM pathway: PIK3C3 rs76692125 G > A,with a meta-analysis derived odds ratio(OR)= 1.26,95% confidence interval(CI)=(1.12-1.43),P = 2.07×10-4;INSR rs11668724 G > A,0.89(0.84-0.94),P = 4.21×10-5 and MAP3K4 rs13207108 C > T,0.83(0.75-0.92),P = 2.26×10-4.Three SNPs in the cadherin pathway: KIF5 B rs211304 C > G,0.89(0.82–0.95),P = 6.93×10-4;FMN1 rs117648907 C > T,1.33(1.13–1.56),P = 2.11×10-4 and MGAT3 rs34943118 T > C,1.11(1.05–1.17),P = 8.10×10-5.2.The PIK3C3 A,the FMN1 T and the MGAT3 C alleles were associated with significantly increased risk of PanC(Ptrend < 0.001);while the INSR A,MAP3K4 T and KIF5 B G alleles were associated with significantly decreased risk of PanC(Ptrend < 0.001,0.001 and 0.001,respectively).3.We combined the risk genotypes of PIK3C3 rs76692125 GA+AA,INSR rs11668724 GG and MAP3K4 rs13207108 CC into one variable of the number of risk genotypes(NRG)in the ATM pathway and the risk genotypes of KIF5 B rs211304 CC,FMN1 rs117648907 CT+TT and MGAT3 rs34943118 CC into one variable of the NRG in the cadherin pathway,respectively.The joint analysis suggested that the NRG was significantly associated with risk of PanC in a dose-response manner in each of the pathways(Ptrend<0.001).4.We employed stratified analyses for risk modification by age and sex in the ATM pathway and in the caherin pathway,respectively.No interaction was observed between the subgroups in the ATM pathway(Pinter > 0.050);the stratified analysis of cadherin pathway age subgroups suggested that the risk associated with the NRG was more evident in age subgroups of 60–70 years(OR = 1.33,95% CI = 1.11–1.61,P= 0.003)and >70 years(OR = 1.47,95% CI = 1.21–1.79,P< 0.001)than in <60 years(OR = 1.08,95% CI = 0.87-1.34,P = 0.514),and the interaction between age subgroups and risk genotypes was statistically significant(Pinter = 0.034).However,there was no significant interaction between sex subgroups and risk genotypes(Pinter = 0.230).5.We combined the risk genotypes of PIK3C3 rs76692125 GA+AA,INSR rs11668724 GG,MAP3K4 rs13207108 CC,KIF5 B rs211304 CC,FMN1 rs117648907 CT+TT and MGAT3 rs34943118 CC into one variable of the NRG in the ATM and the cadherin pathways.The joint analysis suggested that the NRG was significantly associated with risk of PanC in a dose-response manner(Ptrend<0.001).All individuals were divided into a low-genetic risk score(GRS)group with an NRG of 0-3 and a high-GRS group with an NRG of 4-6.Compared with the low-GRS group,the high-GRS group had an increased PanC risk by 36%(P < 0.001).6.The eQTL analysis showed that the PIK3C3 rs76692125 A allele was significantly correlated with a decreased mRNA expression level;while the INSR rs11668724 A,the KIF5 B rs211304 G and the MGAT3 rs34943118 C alleles were significantly correlated with increased their corresponding mRNA expression levels.Conclusions: 1.PIK3C3 rs76692125,INSR rs11668724 and MAP3K4 rs13207108 in the ATM pathway,and KIF5 B rs211304,FMN1 rs117648907 and MGAT3 rs34943118 in the cadherin pathway were independently associated with risk of PanC.2.PIK3C3 rs76692125 GA+AA,INSR rs11668724 GG and MAP3K4 rs13207108 CC were served as risk genotypes in ATM pathway.The NRG was significantly associated with risk of PanC in a dose-response manner.3.KIF5 B rs211304 CC,FMN1 rs117648907 CT+TT and MGAT3 rs34943118 CC were served as risk genotypes in the cadherin pathway.The NRG was significantly associated with risk of PanC in a dose-response manner.4.The risk associated with the NRG was more evident in age subgroups of 60–70 years and >70 years than in <60 years in the cadherin pathway.5.Compared with the low-GRS group,the high-GRS group had an increased PanC risk by 36% in the risk prediction model for PanC.6.PIK3C3 rs76692125 A,INSR rs11668724 A,KIF5 B rs211304 G and MGAT3 rs34943118 C alleles were significantly correlated with their corresponding mRNA expression levels. |
PDF Full Text Request