Association of estrogen receptor α polymorphisms with susceptibility to chronic hepatitis B virus infection

Authors


Abstract

Several studies have demonstrated that estrogen receptor α (ESR1) participates in the pathogenesis of persistent hepatitis B virus (HBV) infection. To examine whether polymorphisms at the ESR1 gene locus are associated with persistent HBV infection, we resequenced ESR1 genomic region for single nucleotide polymorphisms (SNPs) in 27 unrelated Chinese. Two haplotype-tagged SNPs (htSNP), T29C and A252966G, were selected for genotyping in 1,277 persistent HBV-infected cases, 748 spontaneously recovered controls, and 293 nuclear families using polymerase chain reaction (PCR)-restriction fragment length polymorphism (PCR-RFLP) analysis. We observed that the subjects bearing ESR1 29T/T genotype had an increased susceptibility to persistent HBV infection compared to those bearing at least one 29C allele (odds ratio 1.41; 95% CI, 1.17-1.71, P < .001). Consistent with the results of population-based association study, a significantly greater than expected transmission of the 29T allele (56.4%) from heterozygous parents to offspring with persistent HBV infection was observed (χ2 = 4.60, P = .033) using the transmission-disequilibrium test (TDT) in 293 nuclear families. Linkage disequilibrium (LD) mapping analysis indicated that the T29C polymorphism contained within a LD block located from promoter region to intron 3 of ESR1, suggesting that the strong association detected with T29C in ESR1 originated from ESR1 itself. In conclusion, our results suggest that the genetic variation at the ESR1 locus influences susceptibility to persistent HBV infection in a Chinese population. (HEPATOLOGY 2004;40:318–326.)

Hepatitis B virus (HBV) infection is one of the major infectious diseases with more than 350 million chronic carriers worldwide; this causes a broad spectrum of liver diseases ranging from asymptomatic carrier, fulminant hepatitis, chronic hepatitis, and liver cirrhosis to hepatocellular carcinoma.1 Persistent HBV infection has been considered a multifactorial and polygenic disorder with viral, environmental, and genetic components,2, 3 including HBV genomic variability,2 host age, sex, concurrent infection with the hepatitis C virus, hepatitis D virus, and human immune deficiency virus.3 However, segregation analysis and twin studies strongly support the role of host genetic components in determining the chronicity of HBV infection.4–6 Genetic association analyses based on Gambia, European, and Asian cohorts have implicated the human leukocyte antigen allele DRB1*1302 in the clearance of HBV infection.7–9 Several population studies have also revealed that some non-human leukocyte antigen loci, including interferon γ (IFN-γ),10 tumor necrosis factor α (TNF-α),11 mannose binding protein (MBP),12 and vitamin D receptor (VDR),13 are associated with persistent HBV infection or HBV clearance. Disease susceptibility for infectious diseases is considered to be determined at different functional levels such as cytokine production, antigen presentation, and receptor recognition, as has been shown for malaria. Therefore, an unknown number of other unidentified genes are likely to modify the susceptibility to persistent HBV infection.

HBV genome possesses a glucocorticoid response element.14 Estrogen has been shown to suppress HBV replication in male athymic mice transplanted with HBV-transfected HepG2 cells.15 The level of estrogen receptors (ESRs) in the cytosol of peripheral blood mononuclear cells was significantly lower in asymptomatic HBV carriers and patients with chronic hepatitis than in healthy controls. In addition, when peripheral blood mononuclear cells from patients with chronic hepatitis B were incubated with interferon α (IFN-α) in vitro and in vivo, the level of ESRs increased by increasing the concentration of IFN-α.16, 17 Such findings suggest that one cause of the deficiency in virus elimination in HBV carriers may be the unresponsiveness of the immune system to sex hormones as a result of the low level of ESRs. On the basis of in vivo and in vitro functional relevance of the estrogen/estrogen receptors axis in the pathogenesis of chronic persistent hepatitis, we hypothesize that the ESRs may be excellent biological candidate susceptibility genes for persistent HBV infection. There are two known ESRs: estrogen receptor α (ESR1) and estrogen receptor β (ESR2). It is expected that the genetic variation within ESR1 and ESR2 could influence the effects of estrogens, which in turn results in genotype-dependent differences in susceptibility to persistent HBV infection.

A report of a young adult male with a loss-of-function mutation of the ESR1 gene suggests that the estrogen exerts its effect primarily via ESR1.18 In the present study, we screened single nucleotide polymorphisms (SNPs) systematically in the ESR1 gene, then examined the relationship between SNP markers in this gene and susceptibility to persistent HBV infection in a Chinese population.

Abbreviations

HBV, hepatitis B virus; IFN, interferon; ESR, estrogen receptor; ESR1, estrogen receptor α; SNP, single nucleotide polymorphism; HBsAg, hepatitis B surface antigen; anti-HBc, antibody to hepatitis B core antigen; anti-HBs, antibody to hepatitis B surface antigen; IgG, immunoglobulin G; kb, kilobase; PCR, polymerase chain reaction; htSNP, haplotype-tagged single nucleotide polymorphism; RFLP, restriction fragment length polymorphism; bp, base pair; LD, linkage disequilibrium; SYNE1, spectrin repeat containing, nuclear envelope 1; TDT, transmission-disequilibrium test.

Patients and Methods

Study Subjects.

The case-control population contained 2,025 unrelated adult Chinese who were recruited from the outpatient clinics and hospitalization wards at Southwest Hospital (Chongqing), Quanzhou Hospital (Quanzhou), and Infectious Disease Hospital (Beijing) between February 2000 and May 2003. Of these, 1,277 subjects who had been positive for both hepatitis B surface antigen (HBsAg) and antibody to hepatitis B core antigen (anti-HBc) immunoglobulin G (IgG) for at least 12 months were defined as the persistent HBV-infected group. In this group, 421 cases were asymptomatic carriers with normal serum levels of alanine aminotransferase and aspartate aminotransferase throughout the study, without previous history of hepatitis B; 544 were subjects with chronic hepatitis B who had previous history of hepatitis B and elevated levels of alanine aminotransferase/aspartate aminotransferase or total bilirubin during our recruitment; and 312 were subjects with liver cirrhosis who were diagnosed with positive ultrasonography/computed tomography imaging or histopathological findings on examination of liver tissue during the study period. The remaining 748 subjects, who were negative for HBsAg and positive for antibody to hepatitis B surface antigen (anti-HBs) and anti-HBc IgG, were defined as spontaneously recovered controls. All the individuals who were positive for anti-HBs and negative for anti-HBc IgG (may have vaccination history) were excluded. All the recruited subjects had no serological evidence for hepatitis C virus, hepatitis D virus, and human immune deficiency virus coinfection.

A total of 293 Chinese nuclear families were recruited in Southwest Hospital (Chongqing). All the probands, including 152 asymptomatic carriers, 137 with chronic hepatitis B, and 4 with liver cirrhosis, were positive for both HBsAg and anti-HBc IgG for at least 12 months. All the probands had no serological evidence for hepatitis C virus, hepatitis D virus, and human immune deficiency virus coinfection and had no vaccination history. Among the 293 nuclear families, mothers in 129 were positive for HBsAg, whereas mothers were negative in the 164 remaining families.

Written informed consent was obtained from all the subjects and families, and the study was performed with the approval of the ethical committee of the Chinese Human Genome.

SNP Discovery.

SNP screening of all exons, relevant exon-intron boundaries, and the approximately 2-kilobase (kb) promoter region of ESR1 (referred to as GenBank accession no. NT_023451) was performed by polymerase chain reaction (PCR) direct sequencing as described previously.19, 20 The screening panel included 27 unrelated individuals selected randomly and regardless of disease status from the 2,025 individuals. The sample size gave us 95% probability of detecting alleles with a minimal frequency of 5.4%.21 Briefly, the primers for the target regions were designed using Web-based software Primer3.0 (Whitehead Institute, Cambridge, MA). DNA samples from 27 Chinese individuals were amplified and purified. The PCR products were then sequenced using ABI PRISM Dye Terminator Sequencing Kit with Amplitaq DNA polymerase (ABI, Foster City, CA) and loaded onto an ABI 3700 sequencer. SNP candidates were identified by the PolyPhred program (University of Washington, Bothell, WA) and inspected by two observers. SNP positions and individual genotypes have been confirmed by reamplifying and resequencing the SNP site from the opposite strand.

Haplotype Construction and Haplotype-Tagged SNP (htSNP) Determination.

ESR1 haplotypes from the unrelated Chinese samples were assigned by the PHASE program.22 PHASE is an implementation of the Stephens-Donnelly method of haplotype reconstruction, which uses a Bayesian approach incorporating a priori expectations of haplotypic structure from population genetics and coalescent theory. Program htSNP2 (Wellcome Trust, Cambridge) was used to determine the htSNPs as described previously.23

Genotyping.

We extracted genomic DNA from peripheral blood leukocytes of 5 mL whole blood using standard phenol/chloroform protocols. DNA samples were diluted to 8 ng/μL and distributed to 96-well plates (DNA panels), each of which contained 94 samples and 2 DNA-free control water. Then, 2 htSNPs (T29C polymorphism in exon 1 and A252966G polymorphism in intron 5) were selected for genotyping in the case-control populations and nuclear families using PCR-restriction fragment length polymorphism (RFLP) analysis.

For the T29C polymorphism, an amplification using forward primer 5′-GACCATGACCCTCCACACCAAAGGATC-3′ and reverse primer 5′-ACCGTAGACCTGCGCGTTG-3′ was performed. A BamH I recognition site was introduced by 1-base mismatch (underlined base) in the forward primer. PCR conditions were identical to those for SNP discovery,19, 20 except for an annealing temperature of 61°C and a total reaction volume of 12.5 μL. The reaction yielded a 220-base pair (bp) amplicon. PCR product of 3 μL was digested with 4 U of BamH I (NEB, Beverly, MA), and separated on a 3% agarose gel. The presence of the 29C allele created a BamH I restriction site. Digested amplicons from the homozygotes for 29C allele appeared as 197-bp and 23-bp bands on the agarose gel electrophoresis, while homozygotes for 29T allele appeared as a 220-bp band. Heterozygotes had all 3 of these bands.

For the A252966G polymorphism, an amplification of a 311-bp fragment using forward primer 5′-GCCCTGTTTTTATTCCCAGAA-3′ and reverse primer 5′-CATAAAAATCCACTGGTTCTAAGTC-3′ was performed. A Alw26 I recognition site was introduced by a 1-base mismatch (underlined base) in the reverse primer. The PCR conditions were identical to T29C polymorphism genotyping except for an annealing temperature of 55°C. The 311-bp amplicon was digested with 4 U of Alw26 I (NEB) and separated on the 3% agarose gel. The homozygotes for 252966A allele yielded 2 restriction fragments of 281 bp and 30 bp after Alw26 I digestion; homozygotes for 252966G allele remained uncut (311-bp band), and heterozygotes yielded all 3 of these bands.

Genotyping was performed in a blind manner so that the performers did not know subjects' case/control status. The accuracy of genotyping data for each SNP obtained from PCR-RFLP analysis was validated by direct sequencing of a 15% masked, random sample of cases and controls.

LD Block Mapping.

To refine the boundary of the susceptibility region, we constructed a linkage disequilibrium (LD) block map at the genomic region surrounding the ESR1 gene, which includes LOC345839 and spectrin repeat containing, nuclear envelope 1 (SYNE1; referred to as GenBank accession no. NT_023451). Several fragments in LOC345839 and SYNE1 were resequenced for screening SNPs in the panel with 27 unrelated Chinese individuals, and 17 SNPs were identified. A total of 16 SNPs, including 10 common SNPs in ESR1 locus (Table 2) and 6 common SNPs with minor allele frequencies greater than 10% (estimated from the 27 unrelated individuals) screened from LOC345839 and SYNE1, were selected for genotyping in a DNA panel with 192 individuals randomly selected from the 2,025 subjects using direct sequencing. The pairwise LD measure D′ (Lewontin's D′) calculation was performed using the Arlequin package (University of Geneva, Geneva). The values of pairwise D′ were plotted using the GOLD program.24

Table 2. Positions and Frequencies of SNPs Within the Human ESR1 Gene
No.SNPPositionFrequency*Region
  • NOTE. The position of the SNPs is relative to the first nucleotide of the open reading frame of the ESR1 gene, referred to as accession number NT_023451.

  • *

    Minor allele frequency.

1C-1888T200617.0195′Flank
2T-1612C200893.1115′Flank
3C-417G202088.1305′Flank
4T29C (Ser10Ser)202534.407Exon 1
5G194A (Ala65 Ala)202699.037Exon 1
6C436A (Pro146Gln)202941.019Exon 1
7G554A203059.019Intron 1
8G698T203203.167Intron 1
9A34957G237462.019Intron 2
10G72576A275081.056Intron 2
11C72827T (Arg243Arg)275332.019Exon 3
12T72959C275464.222Intron 3
13C136474G (Pro325Pro)338979.444Exon 4
14G136611A339116.037Intron 4
15G204056A406561.370Intron 5
16T252950C455455.074Intron 5
17A252966G455471.389Intron 5
18G253263T455768.389Intron 6
19G290929A (Arg555His)493434.019Exon 8
20G291047A (Thr594Thr)493552.167Exon 8

Data Analysis.

Allele frequencies for each SNP were determined by gene counting. The significance of deviations from Hardy-Weinberg equilibrium was tested using the random-permutation procedure implemented in the Arlequin package. Unconditional logistic regression analysis was performed to adjust risk factors, with persistent HBV infection as a dependent variable and age and sex as independent variables. Each genotype was assessed with the use of dominant, recessive, and additive genetic models, and the P value, odds ratio (OR), and 95% confidence interval (CI) were calculated. An association was considered significant at a P value of less than .05 on unconditional logistic regression analysis involving recessive, dominant, or additive genetic models. These analyses were performed using SPSS software (version 9.0, SPSS Inc., Chicago, IL).

Transmission disequilibrium test (TDT) was conducted using the program TRANSMIT.25 Significance measurements were verified within TRANSMIT using 10,000 bootstrap samples. The attributable fraction was computed by the formula as described previously.26

Results

Table 1 shows the clinical and demographic characteristics of the case-control population and nuclear families. There was no significant difference between case patients and control subjects in terms of mean age distribution. However, although an effort was made to obtain a frequency match on sex between cases and controls, more women were presented in the case group than in controls (29.5% vs. 22.9%; χ2 = 10.6, P = .001).

Table 1. Clinical and Demographic Characteristics of the Subjects in the Case-Control Population and Nuclear Families
CharacteristicCase-Control PopulationNuclear Family
PI (N = 1277)SR (N = 748)Proband (N = 293)Parents (N = 586)
  1. NOTE. Values of age, alanine aminotransferase (ALT) and total bilirubin (TBIL) are expressed as means ± SD.

  2. Abbreviations: PI, persistent HBV-infected subjects; SR, spontaneously recovered subjects.

Sex    
 Male900 (70.5%)577 (77.1%)191 (65.2%)293 (50.0%)
 Female377 (29.5%)171 (22.9%)102 (34.8%)293 (50.0%)
Age (y)35.9 ± 13.133.6 ± 14.419.5 ± 8.745.7 ± 10.4
HBsAg statusAll +All −All +201/385
Anti-HBs statusAll −All +All −277/309
Anti-HBc IgG statusAll +All +All +354/232
Serum ALT (IU/L)150.5 ± 309.722.3 ± 11.2118.3 ± 265.635.3 ± 62.2
Serum TBIL (μmol/L)64.1 ± 112.411.7 ± 5.237.0 ± 91.717.2 ± 37.4

Resequencing of 6,313 bp of ESR1 genomic regions in the 27 samples revealed 20 SNPs (Table 2). Although 2 SNPs (6 and 19) can result in nonsynonymous amino acid change, they are rare SNPs with very low minor allele frequency and unfit for genotyping analysis. Haplotypes were constructed on the basis of the genotype data from the 20 SNPs using PHASE.22 A 95% phase assignment was made with more than 91% certainty. Thirty haplotypes were identified, among which were 13 with frequency more than 3% (Table 3). Four common SNPs (4, 8, 13, and 17), which capture more than 94% of the haplotype diversity observed within gene regions, were determined as htSNPs; 2 of them (T29C polymorphism in exon 1 and A252966G polymorphism in intron 5) were used as markers for subsequent genotyping analysis.

Table 3. Estimated Frequencies of Haplotypes and htSNPs Observed at ESR1 Gene
HaplotypeSNPsFrequency (%)
123inline image567inline image9101112inline image141516inline image181920
  1. NOTE. The number of SNPs is referred to in Table 2. Boxed SNPs represent the htSNPs that can capture the common haplotypes segregating in the Chinese population. Dots represent the allele that is found on the most common haplotype.

1CTCTGCGGAGCTGGATGTGG14.8
2...C........C.G.AG.A7.4
3..............G.AG..7.4
4...........C........5.6
5...........CC.......3.7
6...........CC.G.....3.7
7.C.C...T....C.GCAG..3.7
8..GC.......CC...AG..3.7
9..GC................3.7
10..GC........C.......3.7
11............C.G.AG..3.7
12.C.C...T....C.G.AG..3.7
13.......T......G.AG..3.7
Others                    31.5

On the basis of unconditional logistic regression analysis with adjustment for age and sex, a statistically significant association with susceptibility to persistent HBV infection was observed with the T29C polymorphism under both the recessive and additive genetic models (Table 4). Subjects bearing the 29T/T genotype had an increased susceptibility to persistent HBV infection compared to those bearing the 29T/C genotype (OR 1.39; 95% CI, 1.14-1.70, P = .001), and compared to those bearing at least 1 29C allele (OR 1.41; 95% CI, 1.17-1.71, P < .001). When we limited the regression analyses to men or women, the results remained significant (Table 4). No significant association was found between A252966G polymorphism and susceptibility to persistent HBV infection. The genotype distributions for both SNPs were in Hardy-Weinberg equilibrium in both case and control populations (Table 4).

Table 4. Association Between Persistent HBV Infection and Polymorphisms in the ESR1 Gene
PolymorphismGenotypePatientsDominant ModelRecessive ModelAdditive Model
PISRP ValueOR (95% CI)P ValueOR (95% CI)P ValueOR (95% CI)
  • NOTE. For all subjects, the associations were performed by unconditional logistic regression analysis adjusted for age and sex. For men and women only, the associations were performed by unconditional logistic regression analysis adjusted for age.

  • Abbreviations: PI, persistent HBV infected subjects; SR, spontaneously recovered subjects.

  • *

    Value is for comparison of the T/T genotype with the T/C plus C/C genotype.

  • Value is for comparison of the T/T genotype with the T/C genotype.

  • Value is for comparison of the A/A plus A/G genotype with the G/G genotype.

  • §

    Value is for comparison of the G/A genotype with the G/G genotype.

All subjects (N = 2,025)NN = 1277NNN = 748      
 T29CT/T528 (41.5%)246 (32.9%).0951.26 (0.96–1.64)<.0011.41 (1.17–1.71)*.0011.39 (1.14–1.70)
 T/C595 (46.8%)388 (51.9%)      
 C/C148 (11.7%)108 (14.2%)      
 A252966GA/A500 (39.2%)299 (40.0%).6650.96 (0.80–1.16).0780.79 (0.60–1.03).0820.78 (0.59–1.03)§
 A/G576 (45.2%)351 (46.9%)      
 G/G199 (15.6%)94 (13.1%)      
Men only (N = 1479) N = 901N = 578      
 T29CT/T376 (41.7%)193 (33.4%).0331.40 (1.03–1.90).0071.35 (1.09–1.69)*.0311.29 (1.02–1.63)
 T/C418 (46.4%)290 (50.2%)      
 C/C103 (11.9%)91 (16.4%)      
 A252966GA/A355 (39.4%)239 (41.3%).3740.91 (0.73–1.13).1080.78 (0.57–1.06).1590.79 (0.57–1.10)§
 A/G403 (44.7%)265 (45.8%)      
 G/G142 (15.9%)72 (12.9%)      
Women only (N = 546) N = 376N = 170      
 T29CT/T152 (40.4%)53 (31.0%).6810.88 (0.48–1.62).0281.57 (1.05–2.34)*.0161.67 (1.10–2.52)
 T/C177 (47.1%)98 (59.0%)      
 C/C45 (12.5%)17 (10.0%)      
 A252966GA/A145 (38.6%)60 (35.4%).3591.20 (0.81–1.77).4480.81 (0.47–1.40).2710.72 (0.41–1.29)§
 A/G173 (46.0%)86 (51.1%)      
 G/G57 (15.4%)22 (13.5%)      

We then tested the association between the T29C polymorphism and susceptibility to persistent HBV infection, using 293 Chinese nuclear families for the TDT statistics. When all families were considered, a significantly greater than expected transmission of the 29T allele (56.4%) from heterozygous parents to offspring with persistent HBV infection was observed (χ2 = 4.60, P = .033; Table 5), consistent with the results of population-based association study. Because the vertical transmission would show much higher risk than the horizontal transmission for persistent infection, we further divided these families into 2 groups: families with mothers positive for HBsAg and those with mothers negative for HBsAg. The TDT results of both groups were not significant, with a 55% transmission rate of the T allele (P = .21) in families with mothers positive for HBsAg and a 57% transmission rate (P = .11) in families with mothers negative for HBsAg, indicating that chronic HBV infection among the probands is not related to maternal HBsAg status (Table 5). With regard to the insignificant result in the families with mothers negative for HBsAg, this may be due to the reduced power of the TDT test because of the reduced number of families after stratification.24

Table 5. Results of the TDT for ESR1 T29C Polymorphism
Families*Genotype and Allele Distribution in ProbandsTDT
C/CT/CT/TCTAlleleTransmittedNontransmittedχ2P Value
  • *

    Group 1 indicates nuclear families with mothers positive for HBsAg; group 2 indicates nuclear families with mothers negative for HBsAg.

All (N = 293)33 (11.4%)141 (48.6%)116 (40.0%)207 (35.7%)373 (64.3%)29T1591234.596.033
      29C1231594.596.033
Group 1 (N = 129)14 (11.0%)66 (52.0%)47 (37.0%)94 (37.0%)160 (63.0%)29T68511.381.207
      29C51681.381.207
Group 2 (N = 164)19 (11.7%)75 (46.0%)69 (42.3%)113 (34.7%)213 (65.3%)29T91722.642.108
      29C72912.642.108

The T29C polymorphism showed significant association with susceptibility to persistent HBV infection and disease severity. However, such association can be caused by LD between the causative SNP(s) and the primary associated SNP. We then refined the location of the susceptibility sites by constructing a LD block map. Figure 1 indicates that there are 2 LD blocks covering the ESR1 gene. The T29C and A252966G polymorphisms remain the htSNPs for each LD block. We observed that the T29C polymorphism was contained within block A, which was located from promoter region to intron 3 of ESR1, spanning about 140 kb, in the approximate 390-kb LOC345839-ESR1-SYNE1 region (Fig. 1). We therefore conclude that the strong association detected with T29C in ESR1 originates from ESR1 itself.

Figure 1.

Gene content and linkage disequilibrium (LD) in LOC345839-ESR1-SYNE1 region. (A) Genomic structure of genes in this region and the distribution of genotyped SNPs. Genes are shown by arrows pointing to the direction of transcription. The arrows point to the SNPs investigated. (B) A GOLD plot of color-coded pairwise disequilibrium statistics (D′) between SNPs. The locus extends from the bottom left of the figure to the top right. Red and yellow indicate areas of strong LD. LDs are approximately divided into 2 blocks, A and B. The region associated with susceptibility to persistent HBV infection is in block A, which locates from promoter to intron 3 of ESR1, spanning about 140 kb. Underlined SNPs indicate those genotyped in the present study. (C) One set of haplotypes each for LD blocks A and B. Boxed SNPs represent the htSNPs that can capture the common haplotypes segregating in the Chinese population for each LD block. Qtel, q telomere; Cen, centromere.

Discussion

In this study, we found that the 29T/T genotype was associated with significantly increased risk of susceptibility to persistent HBV infection in comparison with the T/C and C/C genotypes. By LD block mapping, we refined the associated area to the region from promoter to intron 3 of ESR1, spanning about 140 kb, and thereby suggested that the strong association detected with T29C in ESR1 originated from ESR1 itself. To our knowledge, this is the first report of the genetic association between ESR1 and susceptibility to persistent HBV infection.

The design and results of our study include many of the features that are considered desirable components of an ideal association study.27 These characteristics include a large sample size, small P values, and an association that makes biological sense. Furthermore, the consistency of our family-based TDT investigation, which excluded the possibility of spurious genetic association due to hidden population stratification,24, 28 is an additional strong point of this work.

The associated T29C, which is a synonymous polymorphism, may not have functional consequences. To our knowledge, there is no nonsynonymous SNP (which results in amino acid change) with frequency above 5% in the ESR1 gene. However, the results of this study clearly show strong LD from promoter region to intron 3 of the ESR1 locus (Fig. 1). It is very likely that there exist(s) as-yet-unidentified causative regulatory polymorphism(s) that is (are) in LD with the observed associated T29C in the 140-kb block A region. Indeed, there are a (TA)n variable number of tandem repeat (VNTR) located approximately 1 kb upstream of the first exon and a PvuII RFLP in intron 1, about 400 bp upstream of exon 2.29 A regulatory element resembling a steroid response element has been identified in the 5′ flanking region of the human ESR1 gene, approximately 220 bp downstream of the (TA)n repeat.30 Maybe the variable length of the (TA)n repeat affects the usage of this estrogen-responsive element or other known cis-acting elements, thereby affecting gene transcription. Previous studies have shown that a VNTR in proximity to a promoter can have a significant influence on transcriptional regulation.31 Recently, a report showed that the PvuII polymorphism is located within a potential bMyb binding site, which is able to regulate transcription efficacy of a reporter gene.32 Since no functional effect of polymorphisms on expression or function of the ESR1 protein has been established conclusively so far, further studies are needed to identify all polymorphisms in this region of the gene and clarify which polymorphism(s) may possess the functional consequence(s) to the ESR1 and in turn provide mechanistic plausibility for the observed association between T29C polymorphism and susceptibility to persistent HBV infection.

Although the mechanism by which ESR1 polymorphisms may influence susceptibility to persistent HBV infection requires further investigation, there are several biologically plausible explanations. HBV genome possesses a glucocorticoid response element, and its replication is sensitive to estrogen.14, 15 The level of estrogen receptor in the cytosol of peripheral blood mononuclear cells is significantly lower in asymptomatic HBV carriers and patients with chronic hepatitis than in healthy controls, and administration of IFN-α increases its level in a dose-dependent manner in vitro and in vivo.16, 17 Together with the genetic effect of ESR1 polymorphism on several other clinical phenotypes, including breast cancer,33 bone mineral density,34 hypertension,35 lipid levels,36 coronary atherosclerosis,37, 38 and spontaneous abortion,39 these illustrate the pleiotropic nature of the ESR1 protein. The estrogen/estrogen receptors axis can be simultaneously involved in several different metabolic pathways, including the reproductive system, bone, cardiovascular function, immune regulation,40, 41 liver function,42, 43 and HBV replication.14, 15

Several association studies have revealed that the IFN-γ,10TNF-α,11MBP,12 and VDR13 genes are related to the susceptibility to persistent HBV infection or HBV clearance. Most of the results, however, couldn't be replicated in subsequent studies in other populations. Although the highly significant association between ESR1 and susceptibility to persistent HBV infection derives from a biologically based a priori hypothesis, our initial findings should be independently verified in populations of different ancestry, such as Caucasians and Africans.

In the present study, we selected the spontaneously recovered subjects (i.e., previously exposed patients) but not the unexposed subjects as controls, so our results may be more reliable than other reports in which blood donors were recruited as controls. Because unexposed subjects remain at risk of acquiring persistent HBV infection, the inclusion of such a control group limits the ability to compare the polymorphisms results to patients with established persistent HBV infection.

In reviewing the results of this study, however, one must keep one potential limit in mind. It is well known that the major mode of infection in China is maternal-infantile transmission (especially before the beginning of HBV vaccination in China in the 1990s), and age at infection is the predominant factor in determining chronicity of HBV infection.44 Indeed, it is very difficult to identify the age and transmission route of infection in a case-control population using our hospital-based collection because the patients are usually first tested for HBV markers and diagnosed as adults in China. However, it may be difficult to detect a significant difference in a host genetic factor if the majority of patients were infected at infancy (i.e., may result in a type II error).45 Thus, it seems unlikely that the association of T29C polymorphism with susceptibility to persistent HBV infection can result from HBV infection at infancy (i.e., type I error is unlikely). The fact that the frequency of risky 29T/T genotype in probands with mothers positive for HBsAg (with high probability for HBV infection at infancy) is lower than that in mothers negative for HBsAg (with low probability for HBV infection at infancy) may justify our hypothesis (37.0% and 42.3%, respectively; Table 5).

If the 29T/T genotype is regarded as a risk factor for persistent HBV infection, then the population-attributable fraction associated with this genotype, which is a parameter that combines the strength of the epidemiological influence (relative risk) and the frequency of genotype (population exposure rate), can be estimated by the formula

equation image

where AF, f, and RR represent attributable fraction, population exposure rate, and relative risk, respectively.26 The attributable fraction calculated by relative risk (OR 1.41; 95% CI, 1.17–1.71; Table 4, recessive model) combined with the 29T/T genotype frequency (38.2%; Table 4) indicates that the 13.5% (95% CI, 6.1%–21.3%) elevation in risk of persistent HBV infection can be attributed to the susceptible effect of the 29T/T genotype. If this association is real, this genotype is associated with a relative low fraction of persistent HBV infection among the Chinese, thereby suggesting that other genes are likely to modify the susceptibility to this infection. When many susceptibility genes are identified and gene-gene interactions among polymorphisms in these genes are taken into account, prediction of persistent HBV infection susceptibility may become more accurate.

In summary, our results reveal an association between common ESR1 29T/T genotype and increased susceptibility to persistent HBV infection, and they provide support for the importance of ESRs in HBV persistence. The causative loci seem to be regulatory polymorphisms located within the 76-kb proximal promoter region or the 72 kb of the first 2 intron regions. Nevertheless, more steps are required before the importance of the ESR1 gene in chronic HBV infection can be fully ascertained. First, data from other populations are needed to confirm our initial observation. Second, the causative polymorphism in the ESR1 gene and its functional molecular mechanisms for this association with chronic HBV infection should be identified.

Acknowledgements

The authors thank Yi Fan, Rong Fang (Department of Infectious Diseases, Southwest Hospital, Chongqing, China) and Xiaojia Dong and Xiumei Zhang (Chinese National Human Genome Center at Beijing, Beijing, China) for assistance in sample collection and DNA isolation.

Ancillary