A Genome-wide Association Study Of Chronic Hepatitis B Virus Infection In East China

Hepatitis B virus (HBV) infection is one of the most serious and prevalent healthproblems, affecting more than2billion people worldwide, and there are700to800million people infected with HBV in China. More than90%of immune-competentadults spontaneously resolve acute HBV infection; others—less than10%—becomeHBV carriers, however, the underlying mechanism is unknown. Chronic HBVinfection includes asymptomatic carriers and chronic hepatitis B, both of which areclosely related to hepatocellular carcinoma (HCC). The control of HBV infection willpromote the control of HCC incidence. The prevalence of chronic HBV infection isdifferent between different ethnics, with a higher rate among Asians, even after theymove to regions with low chronic HBV infection rate, the prevalence of chronic HBVinfection among them was still higher than locals, which indicates that genetic factorsmay influence the susceptibility of chronic HBV infection. Moreover, geneticevidence from twin studies and family-clustering studies has suggested that hostfactors are critical for determining the outcome of HBV infection. Thereafter, it is ofgreat importance to identify high-risk individuals and to adopt preventive measures ortreatment, which may reduce chronic HBV infection rate effectively.Single nucleotide polymorhphisms are the most common genetic variants inhuman, which refers to variation with frequency more than1%, including transition,inversion, deletion and insertion. Genome-wide association study (GWAS) is apowerful tool to identify the most significant susceptible single nucleotidepolymorphisms (SNPs) associated with certain disease based on high throughputgenotyping platform and testing millions of the whole genome SNPs in a largesample size of cases and controls, followed by several independent replications. Japanhas performed two GWASs in2009and2011respectively, and identified that SNPs atHLA-DP/DQ were associated with chronic Hepatitis B. However, in the two studies,the history of HBV exposure in the control groups was unknown, which may confound the interpretation of a strong association. In2012, Japan and Koreaperformed a GWAS on chronic HBV infection together and the controls were naturalHBV clearers. However, the sample size in the genome-wide scan phase was small,with less than200pairs of cases and controls, of which the statistical power was lowand may not be able to identify susceptible SNPs systematically. Besides, all threeloci identified from the previous three GWASs were at HLA-DP/DQ, which onlyexplain partial susceptible mechanism and it calls for new studies with better designand large sample size to systematically identify susceptible loci or gene associatedwith chronic HBV infection.The cases in this study were HBV carriers, defined as positive for HBsAg andHBV core antigen (HBcAg). The controls were HBV natural clearers, defined asnegative for HBsAg, plus positive for HBV surface antibody (HBsAb) and HBcAg.All study subjects were exclusive of the hepatitis C virus (HCV) infection, i.e.negative for the antibody of HCV. This GWAS comprises three phases. The1000cases in the discovery phase were from Shanghai, including500HBV carriers withHCC and500HBV carriers without HCC; the987controls were from Zhangjiagang.1248cases and1248controls from South Jiangsu (Zhangjiagang and Changzhou)were included in Phase I replication; and Phase II with1000cases and1803controlswere from Central Jiangsu (Taizhou and Nantong).In the discovery phase, the1000cases were genotyped with Illumina OmniExpress chips and the983controls were genotyped with Illumina Omni Zhonghuachips. SNPs in both chips (595,310SNPs) were flow into quality control. After thequality control of the study variables (SNPs) and subjects, a total of951cases(including478HCC HBV carriers and473non-HCC HBV carriers) and937controlswith490,610SNPs. We calculated the P values, odds ratio (OR) and95%confidenceintervals (CI) in the logistic regression with the additive model. SNPs met thefollowing criteria were selected in the Phase I replication:1) SNPs showing asignificant difference of P≤10-5between cases and controls;2) for the HCCstatus-stratified analysis, a P≤10-3threshold was used;3) SNPs with the lowest Pvalue when there were high linkage disequilibrium (LD)(r2>0.8). SNPs showing association significance less than0.05in Phase I replication were selected to replicatein Phase II.Eventually, we discovered two novel loci associated with chronic HBV infection:rs3130542at6p21.33(downstream of HLA-C, OR=1.39,95%CI=1.26-1.53, P=5.09×10-11) and rs4821116at22q11.21(at UBE2L3, OR=0.81,95%CI=0.75-0.87,P=1.69×10-9). Additionally, we replicated the previously identified association ofrs7453920at6p21.32(HLA-DQ) with chronic HBV infection (OR=0.54,95%CI=0.48-0.61, P=8.54×10-25). HLA-DQ belongs to type II HLA molecules, which isclosely related to HBV antigen presentation, and rs7453920may influence theadaptive immune response to HBV infection. HLA-C belongs to type I HLAmolecules, and rs3130542may influence the cell-mediated immunity. UBE2L3hasbeen reported associated with several auto-immune disorders, but not in HBVinfection. Thus UBE2L3may be a new susceptible gene in HBV infection, which mayoffer a new way to study the mechanism of chronic HBV infection.We further analyzed whether the3SNPs associated with HCC risk. In thediscovery phase of GWAS, the3SNPs were not associated with HCC risk in478HCC and473non-HCC HBV carriers. Besides, we tested the genotype distribution inan independent1300HBV-positive HCC cases and compared with the HBV carriersin the two replication phases. No association was detected, either. After combiningthe results of the upper two parts, still no association was detected (rs7453920: OR=0.90, P=2.39×10-1; rs3130542: OR=1.09, P=2.19×10-1; rs4821116: OR=1.03,P=6.54×10-1). Thus all3SNPs were associated with chronic HBV infection, butnot HCC risk, suggesting they may affect the process of chronic HBV infection, butnot in the carcinogenesis of HCC.In this study, we identified two new specific susceptibility loci for chronicHBV infection at6p21.33and22q11.21, and validated the association betweengenetic variants at6p21.32and chronic HBV infection. It’s valuable that our findingsin detecting more susceptibility factors and illustrating potential mechanisms onchronic HBV infection. Meanwhile, our newly identified susceptibility markers mightbe applied in screening and prevention for high risk HBV infectors.