Potential conflict of interest: Nothing to report.
Chronic hepatitis B virus (HBV) infection is a major health issue, especially in Asia. A recent genome-wide association study (GWAS) implicated genetic variants in the human leukocyte antigen (HLA)-DP locus associated with chronic hepatitis B in Japanese and Thai populations. To confirm whether the polymorphisms at the HLA-DP genes are associated with persistent chronic HBV infection in Han Chinese, we conducted an independent case-control study using 521 persistent chronic HBV carriers and 819 controls that included 571 persons with HBV natural clearance and 248 never HBV-infected (healthy) individuals. Eleven single nucleotide polymorphisms (SNPs) in a region including HLA-DPA and HLA-DPB and an adjacent SNP in strong linkage disequilibrium (LD) with a neighboring HLA-DR13 locus were genotyped using the TaqMan SNP genotyping assay. Eleven variants at HLA-DP showed a strong association with persistent chronic HBV carrier status (P = 1.82 × 10−12 to 0.01). We also stratified the analysis by HBV clearance status to test the association between these polymorphisms and HBV natural clearance; similar results were obtained (P = 2.70 × 10−11 to 0.003). Included SNPs define highly structured haplotypes that were also strongly associated with HBV chronic infection (block 1: odds ratio [OR] = 0.54, P = 8.73 × 10−7; block 2: OR = 1.98, P = 1.37 × 10−10). These results further confirm that genetic variants in the HLA-DP locus are strongly associated with persistent HBV infection in the Han Chinese population. (HEPATOLOGY 2011;53:422-428)
Hepatitis B virus (HBV) infection is a worldwide health problem that frequently leads to acute, fulminant, chronic hepatitis, liver cirrhosis, and hepatocellular carcinoma. More than 2 billion people have been infected with HBV worldwide, of whom 400 million are chronic carriers. HBV infection accounts for 600,000 to 1,200,000 deaths each year. The prevalence of HBV varies greatly, but it is endemic in all countries. In China, Southeast Asia, the Western Pacific, and sub-Saharan Africa, where HBV infection is usually acquired perinatally or in early childhood, the prevalence is high and the carrier rate exceeds 8%.1, 2 In China about 120 million people are HBV chronic carriers, and 50% to 80% of cirrhosis patients are infected with HBV.3 Chronic infection with HBV has become a key cause of hepatocellular carcinoma. In North America and Europe the prevalence of chronic HBV infection is low and primarily results from immigration from endemic areas, sexual transmission, injection drug use, or nosocomial infection.1, 2
Persistent HBV infection is influenced by a complex combination of viral, environmental, and genetic components including HBV genomic variability, host age, sex, plus concurrent infection with hepatitis C virus, hepatitis D virus, and HIV.4-7 However, segregation analysis and twin studies strongly support the role of host genetic components in determining the chronicity of HBV infection.8, 9 Several studies revealed that variants in several host genes, including IFNG, TNF, MBP, VDR, and ESR1 were associated with persistent HBV infection or HBV clearance.10-14 The human leukocyte antigen (HLA) class II loci also have been reported to be associated with HBV infection.15-18 Recently, a genome-wide association study (GWAS) demonstrated that genetic variants in the HLA-DP locus were strongly associated with chronic hepatitis B in Japanese and Thai populations.19 Han Chinese constitute about 92% of the population of China, 98% of Taiwan, 78% of Singapore, and about 20% of the world population.20 In our study we screened 11 single nucleotide polymorphisms (SNPs) within the HLA-DP genes and one SNP in strong linkage disequilibrium (LD) with a neighboring HLA-DR13 locus for association with persistent HBV chronic infection in Han Chinese from Hebei and Henan Provinces of northern China.
anti-HBc, antibody to hepatitis B core antigen; anti-HBs, antibody to hepatitis B surface antigen; CI, confidence interval; GWAS, genome-wide association study; HBeAg, hepatitis e antigen; HBsAg, hepatitis B surface antigen; HBV, hepatitis B virus; HIV, human immune deficiency virus; HLA, human leukocyte antigen; HWE, Hardy-Weinberg equilibrium; LD, linkage disequilibrium; MAF, minor allele frequency; OR, odds ratios; SNPs, single nucleotide polymorphisms.
Materials and Methods
Cases and controls were recruited from Zhengding County in Hebei Province and Luohe City in Henan Province of northern China from May to September 2006. In 1983 Zhengding County established a database for epidemiological study of hepatitis B and assessment of hepatitis B vaccine. Specific details on HBV infection, liver function, disease outcome (including death related with HBV infection), hepatitis B vaccination, education, socioeconomic status, etc., were collected each year in several areas across the county. Luohe City's database was established in 2004, and the HBV markers were screened in several communities from 2004 to 2005. The individuals who were hepatitis B surface antigen (HBsAg)-positive were tested again 1 year later in 2006; similar to the Zhengding database, other relevant information was also collected on persons in the Luohe City database. About two-thirds of cases were identified from the Zhengding database and one-third of cases were from records of the Luohe database. Cases were persistent chronic HBV carriers who had been positive for both HBsAg and antibody to hepatitis B core antigen (anti-HBc), or positive for HBsAg only for at least 1 year. Among chronic HBV carriers, 97% were anti-HBc positive, 4% anti-HBs positive only, and about 11% had alanine aminotransferase levels (ATL) of more than 40 IU (mean 105 IU, range 41-403 IU; see Table 1). Controls were identified from the Zhengding database. Controls were at least 30 years of age with normal ATL and no history of hepatitis B vaccination (note: HBV vaccine was not available 30 years ago) including HBV natural clearances and healthy individuals. Clinical criteria for HBV natural clearance were: negative for HBsAg, plus positive for both antibody to hepatitis B surface antigen (anti-HBs) and anti-HBc, or plus anti-HBs positive without history of hepatitis B vaccination. About 70% HBV natural clearances were anti-HBc positive in our cohort (Table 1). Healthy controls were negative for HBsAg, anti-HBs, and anti-HBc without hepatitis B vaccination.
Table 1. Characteristics of Participants in a Study of Persistent Chronic HBV Carriers, Han Population From Northern China
Age was the age of enrollment for all groups.
SD, standard deviation.
All participants self-identified as Han Chinese and self-reported 6 or more months of residency in Zhengding County of Hebei Province or Luohe City of Henan Province, China. Persons with blood relatives enrolled in the study were excluded. HBV markers including HBsAg, anti-HBs, and anti-HBc were confirmed by solid radioimmunoassay at the time of study enrollment. Plasma ATLs were measured by the Reitman-Frankel method using a commercial kit. Blood samples were obtained from 521 persistent chronic HBV carriers (268 males and 253 females) and 819 controls (335 males and 484 females). The mean age was 41 years ± 14 for HBV chronic carriers and 49 years ± 11 for controls. The controls included 571 persons with HBV natural clearance and 248 never HBV-infected (healthy) individuals (see Table 1). Institutional Review Board approval was obtained from all participating institutions and informed consent was obtained from each study participant.
Genomic DNA was extracted from whole blood using phenol/chloroform with MaXtract high-density tubes. Four SNPs on HLA-DPA1, seven SNPs on HLA-DPB1, and one SNP in strong LD with HLA-DR13 were genotyped using a commercially available TaqMan SNP genotyping assay and GeneAmp PCR System 9700 (Applied Biosystems, Foster City, CA), in accordance with the manufacturer's instructions. The sequence detection software was used for allelic discrimination. For quality control, eight template-free controls, one Han Chinese family sample from another study cohort,21 and 5% of duplicate samples with a different extraction were included in each 384-well plate. The rates of successful genotyping calls were 96.3% to 99.4% for the 12 SNPs.
Hardy-Weinberg equilibrium (HWE) assumptions were independently tested for each polymorphism. Chi-square tests were used for allele case-control comparisons, which test for additive allele effects on the disease penetrance. Odds ratios (ORs) were calculated based on the 2 × 2 table of allele-by-trait counts. Differences between the theoretical binomially distributed genotypes and observed genotype frequencies were tested using a χ2 goodness of fit test. The ORs and 95% confidence intervals (CIs) were computed by logistic regression and all results shown are adjusted for age and gender.
Haplotype frequencies were estimated with the expectation-maximization (EM) algorithm, as implemented in SAS PROC Haplotype. Genotypes with all missing alleles were dropped from calculations for haplotype frequencies. Omnibus tests of haplotype association and haplotype-specific ORs were calculated by haplotype replacement regression, assuming an additive model using the probability of carrying each pair of haplotypes provided by PROC Haplotype. The most common haplotype or joint haplotype were used as the references; all the haplotypes with a frequency of less than 1% were combined with others. All statistics were calculated in the statistical package SAS and SAS Genetics v. 9.1.3. The LD map, blocks, and haplotypes were generated by Haploview software.22
Twelve SNPs in the HLA class II loci were genotyped in 521 persistent chronic HBV carriers and 819 controls (571 persons with HBV natural clearance and 248 healthy individuals). The genotype frequencies for 12 polymorphisms conformed to HWE expectations for controls. Table 2 lists the SNP ID, risk allele, number of subjects, OR with 95% CI, and P value for cases versus controls; P value for cases versus HBV natural clearances; and gene information for 12 SNPs. ORs and P values were adjusted for age and sex. Allele frequencies of 11 SNPs located within HLA-DPA1 and HLA-DPB1 gene complexes were significantly different between cases and controls (P = 1.82 × 10−12 to 0.01). The six most significant variants were located at HLA-DPB1 (P = 1.82 × 10−12 to 1.65 × 10−6) (Table 2; Fig. 1). The polymorphism of rs11752643 in strong LD with HLA-DR13 was not significantly associated with persistent chronic HBV infection (Table 2). To investigate the association of these polymorphisms with HBV clearance, we further analyzed the data using only the 571 HBV clearance individuals as controls. As shown in Table 2, the P values were quite similar (P = 2.70 × 10−11 to 0.003) for 11 SNPs located at HLA-DP, while rs11752643 remained nonsignificant.
Table 2. Association of SNP Variant Alleles Within HLA-DP and HLA-DR13 Regions With Chronic Hepatitis B, Han Population From Northern China
See Figure 1 for map position and LD relationships.
Odds ratios (OR) were calculated basis on the 2×2 table of allele counts from cases vs. controls. CI: confidence interval.
P value of Armitage's trend test from cases vs. controls (controls included HBV natural clearance and healthy individuals).
P value of Armitage's trend test from cases vs. clearances (clearances were HBV natural clearances).
The definition of HBV clearance and healthy individuals included in Materials and Methods for clinical criteria and Table 1.
For 11 significant SNPs, we examined the association of genotype frequencies between cases and controls (both clearance and healthy combined), and also between cases and clearance controls only. Table 3 presents the genotype distribution in each group: OR with 95% CI and P values for carriers versus controls, and carriers versus clearances. As illustrated in Fig. 1, the first five SNPs showed minor alleles (four in HLA-DPA1 and one adjacent within HLA-DPB1) associated with decreasing risk/protection of HBV chronic infection (Table 3; OR = 0.33 to 0.66, P = 6.7 × 10−7 to 0.045 for homozygote, OR = 0.50 to 0.77, P = 4.6 × 10−7 to 0.036 for heterozygote). The first four SNPs located in HLA-DPA1 formed haplotype block 1 (Fig. 1). The last six variants located on gene HLA-DPB1 had minor alleles significantly associated with increasing risk/susceptibility of HBV chronic infection (OR = 2.46 to 3.34, P = 5.7 × 10−12 to 7.0 × 10−7 for homozygote, OR = 1.56 to 2.36, P = 6.0 × 10−9 to 0.004 for heterozygote). These six SNPs with susceptibility minor alleles formed haplotype block 2 (Fig. 1). Similar significant associations were observed when we compared HBV carriers with HBV clearances (Table 3; columns 8, 9).
Table 3. Genotype Distribution and Association with Chronic Hepatitis B, Han Population from Northern China
Carriers vs Controls
Carriers vs Clearances
Next we examined haplotype association for block 1, block 2, and the two blocks combined. Table 4 lists the haplotype frequencies in cases and controls, OR with 95% CI and P values for block 1 and block 2. The haplotype AACT, which retains all rare protective alleles of block 1, was significantly associated with decreasing risk of chronic hepatitis B infection (OR = 0.54, P = 8.73 × 10−7). The haplotype GAGATT (which retains all rare susceptible alleles of block 2) and GGGGTC (which retains three rare susceptible alleles of block 2) were significantly associated with increased the risk of chronic hepatitis B infection (OR = 1.98, P = 1.37 × 10−10 for GAGATT; OR = 1.7 P = 0.002 for GGGGTC).
Table 4. Results of Haplotype Association Analysis, Han Population from Northern China (N=412 Cases; N=680 Controls)
The most common haplotype as the reference.
Bolded are risk alleles.
OR, odds ratio; CI, confidence interval.
SNPs order of block 1: rs2395309, rs3077, rs2301220 and rs9277341 within the HLA-DPA1 gene.
SNPs order of block 2: rs9277535, rs10484569, rs3128917, rs2281388, rs3117222 and rs9380343 within the HLA-DPB1 gene.
Table 5 presents a combination of haplotype block 1 and block 2 considered together. The combined protective haplotypes of block 1 (AACT) and block 2 (AGTGCC) were very strongly associated with decreased risk of chronic hepatitis B (OR = 0.36, P = 3.0 × 10−11). The protective haplotype of block 2 (AGTGCC) combined with other haplotypes of block 1 were also significantly associated with decreased risk of chronic hepatitis B infection (OR = 0.56 to 0.65, P = 0.002 to 0.0002).
Table 5. Joint Effects of Haplotype Block 1 and Block 2 with Chronic Hepatitis B
In this study, 12 SNPs that were previously reported to be associated with chronic hepatitis B18, 19 were interrogated in 521 persistent chronic HBV carriers and 819 controls in a Han Chinese population from northern China. Eleven SNPs located within HLA-DPA1 and HLA-DPB1 were strongly significantly associated with persistent chronic HBV carrier status (Table 2). The genotype association analysis further confirmed this association (Table 3). The minor alleles of five SNP loci (four on HLA-DPA1 and one in the HLA-DPB1 region) were protective from risk of chronic hepatitis B. The minor alleles of six SNP loci in the HLA-DPB1 gene region were susceptible to chronic hepatitis B (Tables 2, 3). A closely adjacent SNP, rs11752643, which did not track HLA DPA1 or HLA-DPB2, but which did show strong LD with HLA-DR13, was not associated with chronic hepatitis B. Our results from an independent Han Chinese population replicated in SNP allele direction and statistical strength the reported Japanese/Thai GWAS association.19
The haplotypes directly inform how alleles are organized along the chromosome and may provide additional power for mapping disease genes; haplotypes also provide insight into factors influencing the dependency among genetic markers. The haplotype-based methods can potentially capture cis-interactions between two or more causal variants. The haplotypes should be more informative than individual genotypes for revealing disease-causing mechanisms at a candidate gene.23 Based on these assumptions, we explored the haplotype and joint haplotype association of significant SNPs with chronic hepatitis B infection. The dominant (major) alleles were risk alleles for rs2395309, rs3077, rs2301220, rs9277341, and rs3135021 in the Han Chinese population. The first four of these SNPs formed haplotype block 1; the haplotype GGTC combined by risk alleles was the most common haplotype (freq. = 0.577) in our cohort (Table 4, Fig. 2). The less common haplotype AACT (freq. = 0.209) combined by protective alleles was significantly associated with decreasing risk of chronic hepatitis B infection (Tables 2, 4). The dominant alleles were protective allele for rs9277535, rs10484569, rs3128917, rs2281388, rs3117222, and rs9380343 located on HLA-DPB1 in our study population. These six SNPs formed haplotype block 2, the haplotype AGTGCC (freq. = 0.499) combined by protective alleles was the most common haplotype (Table 4, Fig. 2). The haplotype GAGATT (freq. = 0.327) combined by risk alleles was significantly associated with increasing risk of HBV chronic infection (OR = 1.98, P = 1.37 × 10−10; Table 4). The haplotype GGGGTC with three risk alleles was also associated with significant susceptibility to HBV chronic infection. We also tested the joint effects for haplotype block 1 and block 2. The susceptible haplotypes GGTC of block 1 and GAGATT of block 2 comprised the most common joint haplotype (Table 5). The joint haplotype of two protective haplotypes from block 1 and block 2 (AACT*AGTGCC) exerted a strong protective effect against HBV chronic infection (OR = 0.36, P = 3.03 × 10−11; Table 5). The joint haplotypes including the protective haplotype of block 2 (AGTGCC) showed significant protective effects as well (Table 5). These findings supported the single-SNP association results.
Our analysis included 12 comparisons, but the 12 SNPs are not independent: 10 of them formed two haplotype blocks. A Bonferroni correction for multiple SNPs tested, which assumes the independence of all tests performed, is overly conservative. Nonetheless, after a correction for multiple SNP comparisons (current P value times 4), all 11 SNPs retain statistical significance (Tables 2, 3).
Genes involved in the immune response including HLA loci are among the most numerous and diverse in the human genome. Classical HLA loci spanning 4 Mb on the short arm of chromosome 6p2124 include the class I and class II molecules identified for their role in presentation of antigen to CD8+ and CD4+ T cells, respectively. The HLA class II molecules are expressed as cell surface glycoproteins that bind and present short peptide epitopes to CD4+ T cells. Each HLA subtype has a particular binding motif that dictates a specific range of peptides that can physically bind in a groove on the surface of the HLA molecule.25 Human HLA class II molecules are classified in three isotypes: HLA-DR, -DQ, and -DP. Compared to other class II molecules, very limited information is available concerning peptide interactions and the role of HLA-DP polymorphic positions both in peptide binding and T-cell recognition. Functional analysis has shown that HLA-DP plays a key role in T-cell allorecognition and peptide binding.26 There are no specific amino acids changes for 11 significant SNP variants, but these 11 SNPs located within or around the HLA-DPA1 and HLA-DPB1 locus, spanning a 52-kb region of chromosome 6, were in very strong LD with HLA-DP alleles. The 11 SNPs are likely the proxy markers for adjacent, yet to be identified, functional HLA-DP polymorphisms. Our finding suggests that variations in HLA-DP molecules would influence binding or presentation of viral peptides, perhaps regulating virus clearance and chronic hepatitis B pathogenesis. Further study should focus on how these variants impact gene expression and function.
In summary, our results further confirm that genetic variants in the HLA-DP locus are strongly associated with persistent HBV infection in the Han Chinese population of Northern China.
We thank all the participants in the cohorts. We thank Cheryl A. Winkler for invaluable discussion. We thank Michael Campsmith for review and editing the article. We thank Man-Huei Chang and Quanhe Yang for sharing SAS genetics software.