Viral Quasi-Species Evolution During Hepatitis Be Antigen Seroconversion
2007; Elsevier BV; Volume: 133; Issue: 3 Linguagem: Inglês
10.1053/j.gastro.2007.06.011
ISSN1528-0012
AutoresSeng Gee Lim, Yan Cheng, Stéphane Guindon, Bee Leng Seet, Lay Yong Lee, Peizhen Hu, Shanthi Wasser, Frank Peter, Theresa May Chin Tan, Matthew Goode, Allen G. Rodrigo,
Tópico(s)Hepatitis Viruses Studies and Epidemiology
ResumoBackground & Aims: Although viral quasi-species evolution may be related to pathogenesis of disease, little is known about this in hepatitis B virus (HBV); consequently, we aimed to evaluate the evolution of HBV quasi-species in patients with well-characterized clinical phenotypes of chronic hepatitis B. Methods: Four cohorts of well-defined clinical phenotypes of chronic hepatitis B, hepatitis Be antigen (HBeAg) seroconverters (spontaneous seroconverters and interferon-induced seroconverters) and nonseroconverters (controls and interferon nonresponders) were followed during 60 months on average. Serum from 4 to 5 time points was used for nested polymerase chain reaction, cloning, and sequencing of the precore/core gene (20 clones/sample). Only patients with genotype B were used. Sequences were aligned using Clustal X, then serial-sample unweighted pair grouping method with arithmetic means phylogenetic trees were constructed using Pebble 1.0 after which maximum likelihood estimates of pairwise distances under a GTR + I + G model was assessed. Viral diversity and substitution rates were then estimated. Results: Analysis of 3386 sequences showed that HBeAg seroconverters had 2.4-fold higher preseroconversion viral sequence diversity (P = .0183), and 10-fold higher substitution rate (P < .0001) than did nonseroconverters, who had persistently low viral diversity (3.6 × 10−3 substitutions/site) and substitution rate (2.2 × 10−5 substitutions · site−1 · month−1). After seroconversion, there was a striking increase in viral diversity. Most seroconverters had viral variants that showed evidence of positive selection, which was seen mainly after seroconversion. Conclusions: The high viral diversity before a reduction in HBV DNA and before HBeAg seroconversion could either be related to occurrence of stochastic mutations that lead to a break in immune tolerance or to increased immune reactivity that drives escape mutations. Background & Aims: Although viral quasi-species evolution may be related to pathogenesis of disease, little is known about this in hepatitis B virus (HBV); consequently, we aimed to evaluate the evolution of HBV quasi-species in patients with well-characterized clinical phenotypes of chronic hepatitis B. Methods: Four cohorts of well-defined clinical phenotypes of chronic hepatitis B, hepatitis Be antigen (HBeAg) seroconverters (spontaneous seroconverters and interferon-induced seroconverters) and nonseroconverters (controls and interferon nonresponders) were followed during 60 months on average. Serum from 4 to 5 time points was used for nested polymerase chain reaction, cloning, and sequencing of the precore/core gene (20 clones/sample). Only patients with genotype B were used. Sequences were aligned using Clustal X, then serial-sample unweighted pair grouping method with arithmetic means phylogenetic trees were constructed using Pebble 1.0 after which maximum likelihood estimates of pairwise distances under a GTR + I + G model was assessed. Viral diversity and substitution rates were then estimated. Results: Analysis of 3386 sequences showed that HBeAg seroconverters had 2.4-fold higher preseroconversion viral sequence diversity (P = .0183), and 10-fold higher substitution rate (P < .0001) than did nonseroconverters, who had persistently low viral diversity (3.6 × 10−3 substitutions/site) and substitution rate (2.2 × 10−5 substitutions · site−1 · month−1). After seroconversion, there was a striking increase in viral diversity. Most seroconverters had viral variants that showed evidence of positive selection, which was seen mainly after seroconversion. Conclusions: The high viral diversity before a reduction in HBV DNA and before HBeAg seroconversion could either be related to occurrence of stochastic mutations that lead to a break in immune tolerance or to increased immune reactivity that drives escape mutations. See editorial on page 1031. See editorial on page 1031. Surprisingly, little is known about viral evolution in chronic hepatitis B, which has a worldwide disease burden of 350 million and is a leading cause of liver cirrhosis and liver cancer.1Lavanchy D. Hepatitis B virus epidemiology, disease burden, treatment, and current and emerging prevention and control measures.J Viral Hepat. 2004; 11: 97-107Crossref PubMed Scopus (2133) Google Scholar Viral evolution has been established as important in the pathogenesis of chronic viral disease in hepatitis C virus (HCV) and HIV.2Ross H.A. Rodrigo A.G. Immune-mediated positive selection drives human immunodeficiency virus type 1 molecular variation and predicts disease duration.J Virol. 2002; 76: 11715-11720Crossref PubMed Scopus (102) Google Scholar, 3da Silva J. The evolutionary adaptation of HIV-1 to specific immunity.Curr HIV Res. 2003; 1: 363-371Crossref PubMed Scopus (20) Google Scholar, 4Pawlotsky J.M. Hepatitis C virus genetic variability: pathogenic and clinical implications.Clin Liver Dis. 2003; 7: 45-66Abstract Full Text Full Text PDF PubMed Scopus (189) Google Scholar Because hepatitis B virus (HBV) uses reverse transcriptase in replication, the polymerase also lacks fidelity,5Locarnini S. McMillan J. Bartholomeusz A. The hepatitis B virus and common mutants.Semin Liver Dis. 2003; 23: 5-20Crossref PubMed Scopus (117) Google Scholar thus establishing the basis for development of a multitude of genetic variants within the host, already described in the literature.6Alexopoulou A. Karayiannis P. Hadziyannis S.J. Aiba N. Thomas H.C. Emergence and selection of HBV variants in an anti-HBe positive patient persistently infected with quasi-species.J Hepatol. 1997; 26: 748-753Abstract Full Text PDF PubMed Scopus (35) Google Scholar However, HBV quasi-species have not been well characterized, especially with regard to clinical outcomes and events. Whether HBV viral "quasi-species" have a relation to the complex nature of clinical disease in HBV,7Chen D.S. Natural history of chronic hepatitis B virus infection: new light on an old story.J Gastroenterol Hepatol. 1993; 8: 470-475Crossref PubMed Scopus (70) Google Scholar particularly hepatitis Be antigen (HBeAg) seroconversion, an immunologically mediated event, is unclear. HBV-related liver disease is immunologically mediated, and control of HBV is thought to occur mainly through HBV-specific cytotoxic T cells.8Ferrari C. Missale G. Boni C. Urbani S. Immunopathogenesis of hepatitis B.J Hepatol. 2003; 39: S36-S42Abstract Full Text Full Text PDF PubMed Google Scholar To study the role of viral behavior in chronic hepatitis B, we selected patients belonging to well-characterized clinical phenotypes and compared their viral diversity and substitution rates: spontaneous HBeAg seroconverters (SSs) with untreated nonseroconverters controls (CCs) and interferon (IFN)–treated seroconverters (IRs, IFN responders) with interferon-treated nonseroconverters (NRs, IFN nonresponders) by cloning and sequencing the precore/core gene (the target of the immune response during seroconversion) of such patients taken during 4 to 5 time points that spanned a period of 60 months. Sera from patients with chronic hepatitis B defined as patients with HBsAg >6 months, with well-characterized clinical follow-up for >5 years were selected from a chronic hepatitis B database. These patients were presumed to have acquired HBV perinatally, because almost all had a family history of chronic hepatitis B. None had cirrhosis based on annual ultrasound screening. Only patients with genotype B were used to ensure that differences found in viral evolution and behavior were not due to genotypic differences. Patients who received IFN did so as part of a clinical trial of lymphoblastoid IFN (Wellferon; GlaxoWellcome, Beckenham, United Kingdom) and received between 3 and 5 million international units 3 times weekly for 24 weeks. Ten patients in each group were randomly selected if they fulfilled the following criteria and had sufficiently long follow-up. The index group was patients with documented HBeAg seroconversion and abnormal liver functional tests (LFTs) preceding seroconversion (spontaneous seroconverters), with serum at the following time points relative to HBeAg seroconversion in months: time point I (−22.7 ± 5.6 months), time point III (−4 ± 0.6 months), time point IV (12.6 ± 1.0 months), and time point V (30.2 ± 4.1 months). Control patients included those who were followed for a similar period of time and were persistently HBeAg positive with normal LFTs (untreated controls). Controls were matched for age and sex, and time point intervals. A second index group of patients with IFN-induced HBeAg seroconversion and abnormal LFTs before therapy (IFN responders), with serum at the following time points relative to HBeAg seroconversion: time point I (−29.4 ± 5.6 months), time point II (−11.1 ± 2.7 months), time point III (−5.2 ± 0.3 months), time point IV (9.7 ± 1.1 months), and time point V (29.4 ± 3.5 months). Patients were receiving IFN during time point II. Control patients were persistently HBeAg-positive nonseroconverters despite IFN therapy (nonresponders). Controls were matched for age and sex and time point intervals as well as IFN therapy. HBeAg seroconversion was defined as the loss of HBeAg and the development of anti-HBe antibody on at least 2 consecutive occasions 3 months apart. None of the patients lost HBsAg during the follow-up period. The serial serum samples in this study were taken from 4 to 5 time points for each patient, as listed above. This study was approved by the Institutional Review Board of the National University Hospital. Biochemical tests were performed out with routine automated methods. HBsAg, HBeAg, anti-HBe, and anti-HBs were measured using microparticle enzyme immunoassay (Abbott Laboratories, North Chicago, IL). All patients were negative for antibodies to HCV (QuantiplexTM 2.0 Assay; Chiron Corp, Emeryville, CA), hepatitis D virus, and HIV (Abbott Laboratories, North Chicago, IL). Serum HBV DNA levels were quantified by an Amplicor HBV Monitor Test (Roche Molecular Systems, Pleasanton, CA) following the manufacturer's instructions. The dynamic range of this polymerase chain reaction (PCR)–based quantitative assay is from 4 × 102 to 4 × 107 copies/mL serum. Total DNA was extracted from 200 μL of each serum sample using QIAamp DNA blood mini kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions and eluted in 50 μL distilled water. Because HBeAg seroconversion is associated with the decrease in HBV DNA levels, nested PCR was performed for all samples. The first and second round PCR primers of amplifying precore/core gene were (first round) forward, 5′-AGA TTA GGT TAA AGG TCT TTG-3′; (first round) reverse, 5′-AGT TTC CCA CCT TAT GAG TC-3′; (second round) forward, 5′-TGT ACT AGG AGG CTG TAG GCA-3'; and (second round) reverse, 5′-TCC AAG GGA TAC TAA CAT TGA-3′. Amplification was performed out with DYNAzyme (Finnzymes, Espoo, Finland) and 5 μL of DNA template (extracted DNA from serum samples for the first PCR and the first PCR product for the second round PCR) in 50 μL reaction under the following conditions: initial 3 minutes of activation at 93°C, 35 cycles of 94°C denaturation for 1 minute, annealing at either 50°C or 55°C for 1 minute in the first and second round, respectively, and 72°C extension for 1 minute 30 seconds. The last cycle was followed by a final extension at 72°C for 10 minutes. A 700-base pair fragment (nt 1769–2469) containing the entire precore/core gene (639 base pairs) was amplified. To avoid resampling problems,9Rodrigo A.G. Hanley E.W. Goracke P.C. Learn G.H. Sampling and processing HIV molecular sequences: a computational evolutionary biologis's perspective.in: Rodrigo A.G. Learn G.H. Computational and evolutionary analyses of HIV sequences. Kluwer Academic, Boston, MA2001: 1-17Crossref Google Scholar 10 PCR reactions and cloning per sample were done for those with HBV DNA levels lower than 104Pawlotsky J.M. Hepatitis C virus genetic variability: pathogenic and clinical implications.Clin Liver Dis. 2003; 7: 45-66Abstract Full Text Full Text PDF PubMed Scopus (189) Google Scholar copies/mL. One PCR reaction and cloning were performed for samples with higher HBV DNA levels. To confirm that clones from high DNA samples were representative, in a subpopulation of 4 baseline high DNA samples, serial dilution was performed, and PCR reactions were performed followed by cloning. All PCR products were purified with QIAquick PCR Purification Kit (Qiagen, Hilden, Germany), cloned into pGEM-T vector (Promega, Madison, WI) and transformed into Escherichia coli JM 109 cells. Plasmid DNA was prepared with the Wizard Plus Minipreps DNA purification System (Promega). The presence of inserts was ascertained by digestion with the restriction enzymes NdeI and SacII (New England Biolabs Inc, Ipswich, MA) and subsequent electrophoretic size separation on 1.5% agarose tris-borate-EDTA gels. Two positive clones per cloning for those samples with 10 PCR reactions each and 20 to 25 positive clones for those with 1 PCR reaction each were sequenced using Big Dye Terminator and 3730 xl sequencer (Applied Biosystems, Foster City, CA). The sequencing primers were the flanking sequence of the insert on the pGEM-T vector: forward-pGEM-f, 5′-TTG TAA TAC GAC TCA CTA-3′, and reverse-pGEM-r, 5′-GGA TAA CAA TTT CAC ACA-3′. The significance of the differences among clinical groups and time points were tested using analysis of variance (ANOVA). Frequency variables were tested with Fisher's exact test. All graphical data were presented as means ± standard error. In all tests, P values < .05 were considered statistically significant. The statistical analysis was performed with SPSS 13.0 (2004; SPSS Inc, Chicago, IL) and JMP5.1 (2003; SAS Institute, Cary, NC). The region selected for analysis in the precore/core sequence (nt 1839–2306) comprises only the nonoverlapping region because the overlapping regions may be under different evolutionary constraints. To determine whether this region evolves differently from other regions in the HBV genome, we compared frequencies of nucleotide differences of all sequences of the HBV genome from Genebank, and examined particularly the X open reading frame, the precore signal sequence, and the overlapping region of core and polymerase. We found that the precore signal sequence (control region 1) and the last 82 nucleotides of the overlapping region of core and polymerase (control region 2) were relatively conserved; hence, they were suitable to be used as controls (data not shown). Precore/core sequences, excluding HBx, and polymerase overlapping reading frames (nt 1839–2306) were aligned using Clustal X version 1.810Thompson J.D. Gibson T.J. Plewniak F. Jeanmougin F. Higgins D.G. The CLUSTAL_X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools.Nucleic Acids Res. 1997; 25: 4876-4882Crossref PubMed Scopus (36039) Google Scholar and edited manually using BioEdit version 5.07 (Tom Hall, Department of Microbiology, North Carolina State University). Serial-sample unweighted pair grouping method with arithmetic means (sUPGMA)11Drummond A. Rodrigo A.G. Reconstructing genealogies of serial samples under the assumption of a molecular clock using serial-sample UPGMA.Mol Biol Evol. 2000; 17: 1807-1815Crossref PubMed Scopus (71) Google Scholar phylogenetic trees were then constructed using Pebble 1.012Rodrigo A.G. Goode M. Forsberg R. Ross H.A. Drummond A. Inferring evolutionary rates using serially sampled sequences from several populations.Mol Biol Evol. 2003; 20: 2010-2018Crossref PubMed Scopus (13) Google Scholar from maximum likelihood estimates of pairwise distances estimated under a general time reversible (GTR) model of nucleotide evolution, which permitted nonuniform equilibrium frequencies of nucleotides and 6 different relative instantaneous rates of substitutions. In addition, a proportion of invariant sites and heterogeneity of rates across sites modeled by a γ distribution, were incorporated into the model (such a model is typically abbreviated GTR + I + G). Relative rates of change, the proportion of invariant sites, and the shape parameter of the γ distribution were estimated using the program PHYML when the unconstrained phylogenetic trees were constructed, and these rates were used in the sUPGMA analysis. The sUPGMA method allows clock-constrained phylogenetic trees, which take account of different sampling times, to be estimated. Evolutionary rates of substitution were also estimated from sUPGMA trees. Here, branch lengths of sUPGMA trees were optimized using a maximum likelihood approach similar to the one described by Rambaut et al13Rambaut A. Estimating the rate of molecular evolution: incorporating noncontemporaneous sequences into maximum likelihood phylogenies.Bioinformatics. 2000; 16: 395-399Crossref PubMed Scopus (314) Google Scholar and implemented in Pebble 1.0.14Goode M. Rodrigo A.G. Using PEBBLE for the evolutionary analysis of serially sampled molecular sequences.in: Baxevanis A. Current protocols in bioinformatics. J Wiley, New York2004Crossref Scopus (2) Google Scholar Using the GTR + I + G model, average pairwise genetic distances of HBV sequences for each time point within each patient were obtained. Diversity was analyzed using a mixed model ANOVA with JMP v.5.1 (SAS Institute), with time (time points I, III, V) as an ordinal effect, status (seroconverters or nonseroconverters) and treatment (IFN or no IFN) as fixed effects, and patient identification as a random effect nested within status and treatment. Interaction effects specified were time × status, time × treatment, status × treatment, and time × status × treatment. Diversity at each time point was also analyzed independently using a 2-factor ANOVA, with treatment and status as fixed factors and a treatment × status interaction effect. To confirm that the average diversities of nonseroconverters are not changing over time, we performed a mixed model ANOVA using only diversities from nonseroconverters, with time (time points I, III, and V) as an ordinal effect, treatment as a fixed effect, patient identification as a random effect nested within treatment, and time × treatment as an interaction effect. The pairwise DNA distances within time points and the evolutionary rates of substitution (between time point sequences) were also estimated by Mega 315Kumar S. Tamura K. Nei M. MEGA3: integrated software for molecular evolutionary genetics analysis and sequence alignment.Brief Bioinform. 2004; 5: 150-163Crossref PubMed Scopus (10689) Google Scholar with Tamura 3 parameter (γ) distance. The JTT distances with γ rates in Mega3 were estimated for the within time point amino acid sequences. The substitution rate(s) for each patient was estimated from the sequences collected at different time points. Two methods were used. The first assumes that the substitution rate is constant during the course of evolution.13Rambaut A. Estimating the rate of molecular evolution: incorporating noncontemporaneous sequences into maximum likelihood phylogenies.Bioinformatics. 2000; 16: 395-399Crossref PubMed Scopus (314) Google Scholar The second permits the rates to vary across different sampling intervals.16Drummond A. Forsberg R. Rodrigo A.G. The inference of stepwise changes in substitution rates using serial sequence samples.Mol Biol Evol. 2001; 18: 1365-1371Crossref PubMed Scopus (41) Google Scholar Maximum-likelihood phylogenies were also obtained for each patient under a model in which each branch was allowed to have its own length (unconstrained model). Constrained and unconstrained trees were then compared, assuming an asymptotic χ2 distribution of the likelihood ratio statistic to test the hypothesis of constant evolutionary rates as described in Drummond et al.16Drummond A. Forsberg R. Rodrigo A.G. The inference of stepwise changes in substitution rates using serial sequence samples.Mol Biol Evol. 2001; 18: 1365-1371Crossref PubMed Scopus (41) Google Scholar Maximum likelihood phylogenies were also estimated for each patient dataset under the GTR + G + I model, allowing both heterogeneity in evolutionary rates across sites and a proportion of invariant sites, using PHYML.17Guindon S. Gascuel O. A simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood.Syst Biol. 2003; 52: 696-704Crossref PubMed Scopus (14600) Google Scholar The trees were further optimized under various codon-based models of substitution to analyze the selection processes that act at the HBeAg and HBcAg proteins level. These analyses were performed using a program written by one of the authors (S.G.). We analyzed each dataset assuming either (1) that sites evolved under strong negative selection or neutrally (Model M1) or (2) that sites evolved under strong negative selection, neutrally or under positive selection (Model M2).18Nielsen R. Yang Z. Likelihood models for detecting positively selected amino acid sites and applications to the HIV-1 envelope gene.Genetics. 1998; 148: 929-936Crossref PubMed Google Scholar, 19Yang Z. Nielsen R. Goldman N. Pedersen A.M. Codon-substitution models for heterogeneous selection pressure at amino acid sites.Genetics. 2000; 155: 431-449Crossref PubMed Google Scholar The likelihood of the 2 models were compared to a test for positive selection. Detection of amino acid positions at which positive selection probably occurred was also performed using a standard empirical Bayesian approach.18Nielsen R. Yang Z. Likelihood models for detecting positively selected amino acid sites and applications to the HIV-1 envelope gene.Genetics. 1998; 148: 929-936Crossref PubMed Google Scholar The clinical and laboratory characteristics of all the patients can be seen in Table 1. The level of alanine aminotransferase (ALT) and HBV DNA over time in IFN-treated patients is illustrated in Figure 1 (spontaneous seroconverters and untreated controls [data not shown] had a similar pattern to IFN-treated patients). Viral diversity, phylogenetic trees, substitution rates, and selection patterns of 3386 HBV sequences from clones of the precore/core gene (the target of the immune response during seroconversion) were analyzed by seroconverter status and by subgroup (with or without IFN treatment). For high DNA samples, serial dilution, cloning, and sequencing and analysis for viral diversity showed that these had similar viral diversity to undiluted samples that had only one PCR reaction and cloning.Table 1Baseline Clinical Features of PatientsSpontaneous seroconverters (SS)IFN seroconverters (IR)IFN nonresponders (NR)Nonseroconverter controls (CC)Age (y), mean ± SD32 ± 2.134 ± 4.228.4 ± 323.6 ± 3.3Ratio of males to females6:48:28:25:5ALT (IU/L) , mean ± SD148 ± 70300 ± 19454 ± 1038 ± 15HBV DNA (log10 copies/mL), mean ± SD8.017 ± 0.288.393 ± 0.328.481 ± 0.178.825 ± 0.08NOTE: Seroconverters were generally older than nonseroconverters, but baseline HBV DNA is similar between clinical groups. Baseline ALT levels were higher in SSs and IRs at baseline. Open table in a new tab NOTE: Seroconverters were generally older than nonseroconverters, but baseline HBV DNA is similar between clinical groups. Baseline ALT levels were higher in SSs and IRs at baseline. Our analyses revealed striking differences in the sequence diversities of HBV sequences among seroconverters (SS and IR) and nonseroconverters (CC and NR). Average pairwise viral nucleotide sequence diversity in nonseroconverters (CC and NR) was persistently low in contrast to the 2.4-fold significantly higher levels in seroconverters (SS and IR) particularly before seroconversion (3.6 × 10−3 ± 1.4 × 10−3 and 8.5 × 10−3 ± 1.4 × 10−3 substitutions/site, respectively), irrespective of IFN treatment (data not shown). This difference is statistically significant (P = .0183). After seroconversion, although viral load decreased by 3 log (Figure 1), there was a concurrent striking increase in viral diversity (P < .0001; Figure 2) in SSs (1.83 × 10−2 ± 3.4 × 10−3 substitutions/site) and IRs (1.7 × 10−2 ± 2.2 × 10−3 substitutions/site) 1 year after seroconversion. Amino acid sequence diversity (data not shown) had an almost identical pattern to that of DNA sequence diversity (Figure 2). To determine whether the nonoverlapping region of precore/core had different substitution rates compared with the control regions, we analyzed genetic diversity of these regions (Figure 3A and B). In nonseroconverters (Figure 3A), no difference was observed in viral genetic diversity of the precore/core nonoverlapping region compared with the control regions, but in IFN seroconverters (data not shown), significant differences (P < .05) at almost every time point were observed. In spontaneous seroconverters (Figure 3B), significant differences (P < .01) were observed only after seroconversion. IFN had a weak association with increased nucleotide diversity in IRs before seroconversion (from 7.4 × 10−3 ± 1.2 × 10−3 to 8.5 × 10−3 ± 1.6 × 10−3 substitutions/site; P =.073) while concurrently reducing viral load (from 8.4 ± 0.32 to 7.1 ± 0.58 log10copies/mL; data not shown). We next sought to determine the substitution rate in the precore/core gene, defined as the number of nucleotide substitutions per site and per month (Figure 4). Substitution rate was one log higher in seroconverters, SS and IR (2.4 × 10−4 substitutions · site−1 · month−1 or 2.8 × 10−3 substitutions · site−1 · year−1) (0.28%/year), compared with nonseroconverters, CC and NR (2.2 × 10−5 substitutions · site−1 · month−1 or 2.6 × 10−4 substitutions · site−1 · year−1) (0.026%/year) (P < .0001). In most patients, no significant difference in rates was observed over time. In 17/20 nonseroconverters (CC and NR), HBV DNA sequence phylogenies (Figure 5) were starlike, and viral sequences obtained at different time points were strongly intermingled. In contrast, phylogenies estimated from seroconverters (SS and IR) showed a clear tendency for viral sequences from different time points to cluster separately (Figure 5). The branch lengths of these trees were consistently longer than those reconstructed by viral sequences from nonseroconverters (P < .001). Finally, we sought evidence for positive selection among protein sequences, using a codon-based model. Positive selection in HBeAg and HBcAgs protein were found in 1/20 nonseroconverters (CC and NR) and 14/20 seroconverters (SS and IR) (Figure 6;P < .0001). An empirical Bayes approach identified 9 positively selected amino acid positions in seroconverters, compared with 1 in nonseroconverters (Figure 6). In particular, positions 13 and 135 of the core protein appear to be under positive selection in most seroconverters. Most of the positive selection was detected in sequences after seroconversion or combined sequences before and after seroconversion (Figure 6). In summary, our study has shown that long-term viral evolutionary behavior in chronic hepatitis B is different in HBeAg seroconverters than in nonseroconverters. Seroconverters have 2.4 times higher viral diversity before seroconversion and one log higher substitution rate. The high viral diversity was present at baseline before a noticeable decrease in HBV DNA before seroconversion. Amino acid diversity paralleled the findings of viral genetic diversity. Phylogenetic analysis showed a complexity of sequences before HBeAg seroconversion, with loss of populations of related sequences and new populations arising from different clades. Positive selection was found to be present particularly after seroconversion. The nonoverlapping region of the core protein shows a high level of sequence diversity, particularly after seroconversion compared with the control sequences, the precore signal sequence, and the nucleic acid binding domain of core. There have been no in-depth characterizations of HBV quasi-species with the exception of a recent study by Osiowy et al,20Osiowy C. Giles E. Tanaka Y. Mizokami M. Minuk G.Y. Molecular evolution of hepatitis B virus over 25 years.J Virol. 2006; 80: 10307-10314Crossref PubMed Scopus (118) Google Scholar although a few studies have attempted to define long-term mutation rates of HBV.21Bozkaya H. Akarca U.S. Ayola B. Lok A.S. High degree of conservation in the hepatitis B virus core gene during the immune tolerant phase in perinatally acquired chronic hepatitis B virus infection.J Hepatol. 1997; 26: 508-516Abstract Full Text PDF PubMed Scopus (44) Google Scholar, 22Bozkaya H. Ayola B. Lok A.S. High rate of mutations in the hepatitis B core gene during the immune clearance phase of chronic hepatitis B virus infection.Hepatology. 1996; 24: 32-37Crossref PubMed Google Scholar, 23Hannoun C. Horal P. Lindh M. Long-term mutation rates in the hepatitis B virus genome.J Gen Virol. 2000; 81: 75-83Crossref PubMed Scopus (145) Google Scholar The findings from the latter studies are consistent with our findings, showing that in the "immunotolerant" phase, there are few mutations but these increase during the "immunoreactive" phase. The study by Osiowy et al20Osiowy C. Giles E. Tanaka Y. Mizokami M. Minuk G.Y. Molecular evolution of hepatitis B virus over 25 years.J Virol. 2006; 80: 10307-10314Crossref PubMed Scopus (118) Google Scholar was quite different from our study. They examined viral quasi-species in 8 HBeAg-negative patients at 2 time points 25 years apart and obtained a range of estimates for the average rate of substitution. The average rate we have obtained for our nonseroconverters falls within this range. In some patients from their study evidence was found of viral divergence and positive selection but, because this cannot be correlated with clinical and hepatitis B-related reactivation events over the intervening 25years, it is difficult to interpret those findings. When we examine HBV quasi-species in the context of viral quasi-species in general, a simple qualitative model of genetic diversity24Stumpf M.P. Pybus O.G. Genetic diversity and models of viral evolution for the hepatitis C virus.FEMS Microbiol Lett. 2002; 214: 143-152Crossref PubMed Google Scholar has been proposed to explain HIV and HCV genetic diversity, which may apply to HBV in the context of our findings. To paraphrase that hypothesis, viral load decreases as the immune response increases, thus reducing the ability of the virus to adapt. Stronger immune pressure leads to greater selection for escape mutants, which generate viral diversity. When immune responses are low, there is little selective pressure, so few adaptive mutations will be observed, despite a large viral population size. For high levels of immune response, the viral effective population size will become small; therefore, selection for escape mutants will be weak. Only at intermediate levels of population size and immune selection will there be a large number of escape mutants. In our study, nonseroconverters have a high viral load and low quasi-species diversity, and they typically have a weak immune response.25Webster G.J. Reignat S. Brown D. et al.Longitudinal analysis of CD8+ T cells specific for structural and nonstructural hepatitis B virus proteins in patients with chronic hepatitis B: implications for immunotherapy.J Virol. 2004; 78: 5707-5719Crossref PubMed Scopus (349) Google Scholar In seroconverters, high viral diversity is associated with a reduction in viral load before seroconversion, and this stage of HBV is typically characterized by increased T-cell responses.26Tsai S.L. Chen P.J. Lai M.Y. et al.Acute exacerbations of chronic type B hepatitis are accompanied by increased T cell responses to hepatitis B core and e antigens Implications for hepatitis B e antigen seroconversion.J Clin Invest. 1992; 89: 87-96Crossref PubMed Scopus (255) Google Scholar, 27Rehermann B. Lau D. Hoofnagle J.H. Chisari F.V. Cytotoxic T lymphocyte responsiveness after resolution of chronic hepatitis B virus infection.J Clin Invest. 1996; 97: 1655-1665Crossref PubMed Scopus (274) Google Scholar However, it is recognized that immune control of HBV requires a polyclonal and multispecific T-cell response.28Rehermann B. Immune responses in hepatitis B virus infection.Semin Liver Dis. 2003; 23: 21-38Crossref PubMed Scopus (100) Google Scholar Thus, it is enticing to propose that there were intermediate levels of immune response to HBV which are associated with escape mutations, but the presence of escape mutations typically require the demonstration that a naturally occurring mutated T-cell epitope does not trigger an immune response. This has been shown in patients with chronic hepatitis B,29Bertoletti A. Costanzo A. Chisari F.V. et al.Cytotoxic T lymphocyte response to a wild type hepatitis B virus epitope in patients chronically infected by variant viruses carrying substitutions within the epitope.J Exp Med. 1994; 180: 933-943Crossref PubMed Scopus (206) Google Scholar but exactly how common is this phenomenon is unclear30Rehermann B. Pasquinelli C. Mosier S.M. Chisari F.V. Hepatitis B virus (HBV) sequence variation of cytotoxic T lymphocyte epitopes is not common in patients with chronic HBV infection.J Clin Invest. 1995; 96: 1527-1534Crossref PubMed Scopus (111) Google Scholar because extensive study of this subject is still lacking. Supporting evidence of escape mutations comes from an evolutionary viewpoint in our study by the finding of positively selected amino acid mutations after seroconversion, suggesting that these mutations have an advantage in surviving seroconversion. Phylogenetically, it is observed that closely related populations of viral quasi-species appear to have been replaced by new populations arising from a different clade after seroconversion. A more fundamental issue is that of the high baseline viral genetic diversity in the seroconverter group and events that initiated the change in diversity from the low levels seen in nonseroconverters. It seems that alterations in viral genetic diversity are likely to be modulated by changes in viral polymerase fidelity or altered immune reactivity (or both). The immune reactivity of patients with chronic hepatitis B differ according to the clinical phenotype of disease. It is certainly observed that patients who are HBeAg positive and with normal LFTs have little T-cell reactivity. However, those who are undergoing HBeAg seroconversion have increased T-cell responses with increased T-cell proliferation and increased T-cell frequencies to HBeAg and HBcAg.26Tsai S.L. Chen P.J. Lai M.Y. et al.Acute exacerbations of chronic type B hepatitis are accompanied by increased T cell responses to hepatitis B core and e antigens Implications for hepatitis B e antigen seroconversion.J Clin Invest. 1992; 89: 87-96Crossref PubMed Scopus (255) Google Scholar After HBeAg seroconversion, levels of HBV specific CD8 T cells are higher, which could be related to the low levels of HBV DNA.25Webster G.J. Reignat S. Brown D. et al.Longitudinal analysis of CD8+ T cells specific for structural and nonstructural hepatitis B virus proteins in patients with chronic hepatitis B: implications for immunotherapy.J Virol. 2004; 78: 5707-5719Crossref PubMed Scopus (349) Google Scholar There is, however, little information on the relation between viral evolution and immune reactivity in chronic hepatitis B, but, based on our study and the known immune responses in these different phases of chronic hepatitis B infection, we postulate 2 possible scenarios. The first, a decreasing HBV DNA before seroconversion may increase HBV-specific T-cell responses and stimulate the selection of mutations. In this hypothesis, the immune system drives viral mutations, by selecting those that are able to evade detection and elimination. The high number of viral variants after HBeAg seroconversion seems consistent with this hypothesis. Alternatively, the stochastic appearance of mutations in itself may, in fact, lead to an increase in T-cell response because the emergence of new T-cell epitopes may potentially break T-cell tolerance.29Bertoletti A. Costanzo A. Chisari F.V. et al.Cytotoxic T lymphocyte response to a wild type hepatitis B virus epitope in patients chronically infected by variant viruses carrying substitutions within the epitope.J Exp Med. 1994; 180: 933-943Crossref PubMed Scopus (206) Google Scholar The presence of high viral diversity in the baseline samples before a reduction in HBV DNA in HBeAg seroconverters lends support to this hypothesis. The 2 hypotheses are not mutually exclusive, and both mechanisms may be acting at different stages of chronic hepatitis B. What is certain is that patients with HBeAg-positive disease and normal ALT levels have little viral evolution, and these patients are also known to have low HBV-specific T-cell reactivity25Webster G.J. Reignat S. Brown D. et al.Longitudinal analysis of CD8+ T cells specific for structural and nonstructural hepatitis B virus proteins in patients with chronic hepatitis B: implications for immunotherapy.J Virol. 2004; 78: 5707-5719Crossref PubMed Scopus (349) Google Scholar; hence, escape mutation is not the mechanism by which virus persists at this stage of chronic hepatitis B infection. Future work needs now to focus on exploring these hypotheses and their role in pathogenesis of HBeAg seroconversion and immune-mediated liver disease. With IFN treatment, a reduction in viral load is accompanied by a trend toward increased viral diversity before seroconversion and is opposite to that found in successful treatment of HCV.31Farci P. Strazzera R. Alter H.J. et al.Early changes in hepatitis C viral quasi-species during interferon therapy predict the therapeutic outcome.Proc Natl Acad Sci U S A. 2002; 99: 3081-3086Crossref PubMed Scopus (192) Google Scholar This would suggest that the mechanism of successful IFN action is different in HBV than in HCV. In conclusion, the novel findings of our study are that HBeAg seroconverters have distinctly higher viral diversity and substitution rates that preceded seroconversion than do nonseroconverters. Whether this finding is due mainly to increased immune reactivity that drives viral escape mutations or due to stochastic appearance of viral mutations that lead to mutations altering cytotoxic T-cell epitopes that could potentially break tolerance is unclear, and future studies are needed to address this issue. The authors thank El Elyon, Hong Wan Jin, and Antonio Bertoletti for their invaluable contributions to the study and manuscript. The Natural History and the Staging of Chronic Hepatitis B: Time for Reevaluation of the Virus–Host Relationship Based on Molecular Virology and Immunopathogenesis Considerations?GastroenterologyVol. 133Issue 3PreviewHepatitis B is a disease with both virologic and immunologic components. The 2 key virologic events in the life cycle of the hepatitis B virus (HBV) are the generation from genomic DNA of the covalently closed circular (ccc) DNA transcriptional template and reverse transcription of the viral pregenomic (pg) RNA to form the HBV DNA genome. Because the virus utilizes reverse transcription to copy its genome, mutant viral genomes, or quasispecies, are frequently found in the blood of HBV-infected patients. Full-Text PDF
Referência(s)