| [Background]Gastric cancer?GC?is one of the malignant tumors with high morbidity and mortality all over the world.Both environmental and genetic factors contribute to occurrence of GC.The known environmental risk factors were H pylori infection,Epstein-Barr virus infection,Low intake of fruits and vegetables and high consumption of pickled foods,smoking and obesity.Heritability of GC estimated in a large twin cohort accounts for 22%.Recently,Genome-wide association studies?GWAS?have been a powerful tool to discover genetic variations associated with multiple traits and diseases.Taking advantage from GWAS,11 genetic variants were detected to be associated with GC.However,few distinguished GC susceptibility loci can explain only the tip of an iceberg of the heritability.This“missing heritability”phenomenon may be due to the stringent threshold used in traditional GWAS,ignoring many potential GC susceptibility loci.Thus,more effort needs to be put in exploring GC susceptibility loci on a genome-wide scale.Although quantities of disease-and trait-related loci were discovered from GWAS,most of them were located in noncoding sequence.The Encyclopedia Of DNA Elements?ENCODE?Project found that non-coding region sequences are so vital that they played great roles in gene regulation.Subsequently,more and more studies discovered such noncoding variants can influence the occurance of dieases through long-range regulation of targeted gene expression.Deoxyribonuclease I hypersensitive sites?DHS?are sensitive and precise markers of human cis-regulatory elements including promoters,enhancers,silencers,insulators,representing regulatory regions in genome.Researchers found that genetic variations in non-coding regions associated with diseases or traits identified from GWAS were significantly enriched in the DHS region,suggesting that genetic variation in the DHS region were important genetic mechanisms for the regulation of complex diseases or traits.Therefore,we conducted this analysis to systemically explore the relationship between genetic variations in DHS and susceptibility to GC.[Methods]In this study,a case-control study was designed to investigate the correlation between genetic variations in genomic regulatory regions and the risk of GC.First of all,we identified GC risk loci through four GC GWAS:1)NJ-GWAS:detecting genotypes of 565 stomach body cancer cases and 1,162 healthy controls with Affymetrix 6.0 chips.2)BJ-GWAS:detecting genotypes of 468 stomach body cancer cases and 1,123 healthy controls with Affymetrix 6.0 chips.3)JS-GWAS:detecting genotypes of 1,190 GC cases and 1,073 healthy controls with Infinium chips.4)NCI-GWAS:detecting genotypes of 1,625 GC cases and 2,100 healthy controls with Illumina chips.We performed strict quality control on the four GWAS,and excluded unqualified genetic variants?located in sex chromosomes,low call rate,minor allele frequency<0.01,deviating from the Hardy-Weinberg equilibrium?and samples?low call rate,repeat,relatives,high heterozygosity and obvious stratified populations?.Then,IMPUTE2 was used to perform imputation.Ultimately,6,056,523 genetic variants from 3,771 cases and 5,426 controls passed the quality control.On the basis of combined DHS data from 125 tissues and cell lines obtained from ENCODE,we analyzed the correlation between genetic variants located in DHS and GC risk through logistic regression.Then,meta-analysis was conducted by GWAMA to combine the effect of genetic variants from 4 GWAS.We conducted bioinformatics analysis about expression quantitative trait loci?e QTL?and differential expression using genotype and mRNA data from GTEx?170 normal stomach tissues?and TCGA?413 GC tissues and 32 normal tissues?databases.We screened candidate GC risk loci located in DHS with the following criteria:P value from meta-analysis<0.001,P value from eQTL analysis<0.01,targeted gene was differentially expressed in GC cancer and normal tissues?P<0.001?.Then,we conducted validations among 6,222 GC cases and 7,510 controls by Sequenom and TaqMan assays.As for the candidate gene,we first preformed co-expression analysis on the basis of GTEx and TCGA databases.After that,the functional enrichment analysis among co-expressed genes of the candidate gene was carried out by using the R software package"clusterprofiler".[Results]In this study,the correlations between 943,254 genetic variations located in the DHS and the risk of GC were assessed in 3,771 GC cases and 5,426 controls.Our results showed that there were 2,106 SNPs associated with GC risk?P<0.001?.Among those SNPs,284 SNPs showed eQTL and differential expression simultaneously,and we selected 16 independent loci to conduct population verifications.Finally,among 9,993 GC cases and 12,936 controls,we found rs7811798 in 7q21.11 was significant associated with GC susceptibility,and G allele of rs7811798 could reduce GC risk(OR=0.85,95%CI=0.81-0.90,P=2.56×10-8).Stratified analysis suggested that there was no obvious heterogeneity among subgroups,and rs7811798 was related to risk of GC in populations younger than 60years old,age 60 or older,male and female.Rs7811798 is located about 87.6 kb downstream of SEMA3D,and the G allele of rs7811798 was associated with the increased expression of SEMA3D(P=7.16×10-3).Differential expression analysis found that the expression of SEMA3D was down-regulated in GC tissues(paired samples:P=1.50×10-5,unpaired samples:P=8.97×10-5).Pathway enrichment analysis revealed that genes co-expressed with SEMA3D were mainly enriched in tumor-related pathways such as calcium signaling pathway,cAMP signal pathway,Rap1 signal pathway and Ras signal pathway.[Conclusion]Our study systematically evaluated the relationship between genetic variations in DHS and GC risk.We identified a genetic variation on chromosome7q21.11 might influence susceptibility to GC through regulating expression of SEMA3D,which further highlighted the vital role of human regulatory DNA variations and development of GC.Our results can help understand the biologic mechanism of occurance of GC,perform individual risk prediction,and conduct individualized prevention of GC. |
PDF Full Text Request