Emergence of hepatitis B virus S gene mutants in patients experiencing hepatitis B surface antigen seroconversion after peginterferon therapy§

Authors

  • Chao-Wei Hsu,

    1. Liver Research Center, Chang Gung Memorial Hospital, Taipei, Taiwan
    Search for more papers by this author
  • Chau-Ting Yeh

    Corresponding author
    1. Liver Research Center, Chang Gung Memorial Hospital, Taipei, Taiwan
    2. Molecular Medicine Research Center, Chang Gung University, Taoyuan, Taiwan
    • Liver Research Unit, Chang Gung Memorial Hospital, 199 Tung Hwa North Road, Taipei, Taiwan
    Search for more papers by this author
    • fax: 886-3-3282824


  • Potential conflict of interest: Nothing to report.

  • This study was supported by grant CMRP371302 from the Chang Gung Medical Research Program.

  • §

    Chao-Wei Hsu contributed to the drafting of the article, the acquisition of data, and the analysis and interpretation of data. Chau-Ting Yeh contributed to the study concept and design, the drafting of the article, the critical revision of the article for important intellectual content, the acquisition of funding, and study supervision.

Abstract

With anti–hepatitis B virus (anti-HBV) therapy using peginterferon, the seroconversion of hepatitis B surface antigen (HBsAg), which is considered a cure of the disease, can be achieved in a small percentage of patients. Eight of 245 consecutive patients (3.27%) with chronic hepatitis B who received peginterferon therapy at our center achieved HBsAg seroclearance. Surprisingly, two of the eight patients remained viremic according to standard HBV DNA assays. The coding regions of the HBV pre-S/S gene, which were derived from serial serum samples, were analyzed. Site-directed mutagenesis experimentation was performed to verify the phenotypic alterations in Huh-7 cells. In patient 1, an sT125A mutant developed during the HBsAg-negative stage and constituted 11.2% of the viral population. The HBV DNA level was 2.73 × 104 IU/mL at the time of detection. This mutant was not detectable in the HBsAg-positive stages. A phenotypic study of Huh-7 cells showed a significant reduction of antigenicity. In patient 2, an sW74* truncation mutation was found during the HBsAg-negative stage and constituted 83.1% of the viral population. The HBV DNA level was 4.12 × 104 IU/mL at the time of detection. A phenotypic study of Huh-7 cells showed a complete loss of antigenicity. Patient 2 subsequently experienced an episode of hepatitis relapse 7 months after the end of treatment and was negative for HBsAg throughout the hepatitis flare. Conclusion: During antiviral therapy with peginterferon, the achievement of HBsAg seroconversion does not necessarily indicate viral eradication. The emergence of S gene mutants is another possibility, and a relapse with HBsAg-negative hepatitis can occur. (Hepatology 2011;)

It has been estimated that there are currently 350 million patients infected with hepatitis B virus (HBV) worldwide. A chronic HBV infection can lead to severe sequelae such as liver cirrhosis and hepatocellular carcinoma. Presently, there are two major therapeutic strategies for treating chronic hepatitis B: oral nucleoside/nucleotide analogues for HBV polymerase inhibition and interferon-based therapy for immune modulation. Lamivudine is the first clinically approved antiviral nucleoside analogue with a potent inhibitory effect on the RNA-dependent DNA polymerase of HBV, and it has been widely used in the past decade.1-6 Although the suppression of viral replication can be achieved rapidly in most patients, hepatitis B e antigen (HBeAg) clearance is observed in only a minority of patients with short-term treatment. The rapid relapse of HBV replication occurs after drug withdrawal, and this has been attributed to the persistence of HBV covalently closed circular DNA in hepatocytes.5, 7, 8 Prolonged use of lamivudine has thus been proposed, but this leads to the emergence of drug resistance.9-12 Several other oral antiviral agents, including adefovir dipivoxil, entecavir, telbivudine, and tenofovir, have subsequently been approved. Although these agents are all very effective in inhibiting HBV reverse transcriptase, their long-term use also leads to the development of drug resistance.13

Long-term lamivudine therapy may lead to the clearance of hepatitis B surface antigen (HBsAg), although this is not commonly observed. This clearance is generally interpreted as the eradication of the virus. However, it has been reported that HBV DNA is still detectable in some patients after HBsAg seroclearance.14 In a recent study,15 the mutation hot spot sP120A was identified in 6 of 11 patients who experienced HBsAg seroclearance but remained viremic after lamivudine therapy.

Interferon-α has been used in the treatment of chronic hepatitis B for more than 2 decades.16 Although the clinical use of interferon has been limited by its extensive adverse effects, this therapeutic strategy has several advantages, including a definite course of therapy, no known drug resistance, and a more sustained therapeutic response. Regular interferon has a shorter half-life and has to be given three times per week. A meta-analysis of 15 randomized controlled trials showed that patients who were positive for HBeAg and were treated with regular interferon for more than 3 months demonstrated HBV DNA inhibition (37%) and HBeAg loss (33%).17 Peginterferon alfa-2a has a half-life of approximately 77 hours and can be given once every week. The results from a clinical trial (n = 814) indicated that after 6 months of treatment, significantly more patients who received peginterferon alfa-2a therapy with or without lamivudine achieved HBeAg seroconversion in comparison with patients who received lamivudine only.18 Intriguingly, 16 patients (2.9%) receiving peginterferon alfa-2a (alone or in combination with lamivudine) experienced HBsAg seroconversion and were considered cured.

This dogma was challenged when we discovered two patients who experienced HBsAg seroconversion after they had been treated with peginterferon but continued to be viremic. Strikingly, one of the patients subsequently experienced an episode of hepatitis relapse, which was found to be HBsAg-negative hepatitis. Here we analyze the genetic and phenotypic changes in the S gene sequences of these two patients.

Abbreviations

ALT, alanine aminotransferase; anti-HBs, antibody to hepatitis B surface antigen; CMV, cytomegalovirus; DAPI, 4′,6-diamidino-2-phenylindole; GP27, glycoprotein 27; HBeAg, hepatitis B e antigen; HBsAg, hepatitis B surface antigen; HBV, hepatitis B virus; MAHBs, monoclonal antibody against HBsAg; mRNA, messenger RNA; P24, protein 24; PCR, polymerase chain reaction; PEG-IFN, peginterferon; Tris-HCl, trishydroxymethylaminomethane hydrochloride.

Patients and Methods

Patients.

From May 2002 to November 2009, 245 patients received anti-HBV therapy with peginterferon at the Liver Research Center of Chang Gung Memorial Hospital (Taipei, Taiwan). HBsAg seroclearance was documented in eight patients (3.27%). Two remained viremic according to standard HBV DNA assays. These two patients were included in this study. Patient 1 was a 57-year-old male who was negative for HBeAg. A liver biopsy sample showed an Ishak histology activity index of 8 and a fibrosis score of 4. Immunohistochemistry revealed tissue positive for HBV core antigen and HBsAg. He had genotype B HBV. Peginterferon alpha-2a (180 μg/week) was given to the patient. The treatment course is plotted in Fig. 1. After 48 weeks of treatment, the alanine aminotransferase (ALT) level was 44 U/L; the patient was negative for HBsAg and positive for antibody to hepatitis B surface antigen (anti-HBs) according to a radioimmunoassay at the end of the treatment. However, the HBV DNA level remained 2.73 × 104 IU/mL. After informed consent was obtained, serum samples were used for quantitative HBsAg assays and HBV S gene sequence analysis.

Figure 1.

Clinical course of patient 1. The progressive changes in the ALT, HBV DNA, and HBsAg levels, which were tested with the Abbott Laboratories Architect assay (see the Patients and Methods section), are depicted. HBsAg and anti-HBs levels were assessed with a radioimmunoassay. Abbreviation: PEG-IFN, peginterferon.

Patient 2 was a 20-year-old male who was positive for HBeAg. A liver biopsy sample showed an Ishak histology activity index of 7 and a fibrosis score of 2. Immunohistochemistry revealed tissue positive for HBV core antigen and HBsAg. He also had genotype B HBV. The serum HBV DNA level was 1.21 × 107 IU/mL, and the ALT level was 706 U/L before the treatment. Peginterferon alfa-2b (120 μg/week) was given to the patient. The clinical course is plotted in Fig. 2. HBeAg seroclearance was not achieved during the clinical course. However, he became negative for HBsAg and subsequently became positive for anti-HBs according to a radioimmunoassay at the end of treatment. Notably, the HBV DNA level remained 4.12 × 104 IU/mL. After informed consent was obtained, serum samples were used for quantitative HBsAg assays and HBV S gene sequence analysis.

Figure 2.

Clinical course of patient 2. The progressive changes in the ALT, HBV DNA, and HBsAg levels, which were tested with the Abbott Laboratories Architect assay (see the Patients and Methods section), are depicted. HBsAg and anti-HBs levels were assessed with a radioimmunoassay. HBsAg-negative hepatitis is indicated by gray arrows. Abbreviation: PEG-IFN, peginterferon.

In patient 2, a hepatitis relapse occurred 7 months after the end of his treatment (Fig. 2). The ALT level was elevated to 828 IU/mL, and the HBV DNA level was elevated to 3.18 × 107 IU/mL. Interestingly, this patient's serum remained negative for HBsAg according to a radioimmunoassay throughout this exacerbation. Thus, this patient experienced an episode of HBsAg-negative hepatitis.

Serological Assays.

The HBV DNA concentration was quantified with the Roche TaqMan HBV monitor (Roche Diagnostics, Basel, Switzerland). The detection limit of this test was 69 copies/mL. In this test, 5.82 copies/mL was equivalent to 1 IU/mL. Serum hepatitis markers, including anti-HBs, HBeAg, and antibody to HBeAg (Ausria II and HBeAg radioimmunoassays, Abbott Laboratories, North Chicago, IL) and antibody to hepatitis D antigen (Formosa Biomedical Technology Corp., Taiwan), were assayed with commercially available kits. HBsAg was also measured with another enzyme immunoassay when this was necessary (Enzygnost HBsAg 5.0, Dade Behring Marburg GmbH, Marburg, Germany). Serum antibody to hepatitis C virus levels were assayed with third-generation enzyme immunoassay kits (HCV EIA III, Abbott Laboratories). The quantitative assessment of HBsAg was performed with an automated chemiluminescent microparticle immunoassay (Architect HBsAg, Abbott Laboratories) according to the manufacturer's instructions.

Extraction of HBV DNA and Polymerase Chain Reaction (PCR).

For HBV DNA isolation, serum (100 μL) was mixed with 300 μL of a buffer [13.3 mmol/L trishydroxymethylaminomethane hydrochloride (Tris-HCl), pH 8.0; 6.7 mmol/L ethylene diamine tetraacetic acid; 0.67% sodium dodecyl sulfate; and 133 mg/μL proteinase K] and incubated at 55°C for 4 hours. After phenol-chloroform extractions, DNA was precipitated with cold ethanol. The precipitate was dissolved in 20 μL of a Tris-HCl (10 mmol/L, pH 8.0)/ethylene diamine tetraacetic acid (1 mmol/L) buffer. PCR was performed for 30 cycles with a DNA thermal cycler (PerkinElmer Cetus, Norwalk, CT). The primers were called PS1 (5′-ATATTCTTGGGAACAAGAGC-3′, nucleotides 2828-2847, sense) and PS2 (5′-GGAATAACCCCATCTTTTTG-3′, nucleotides 867-848, antisense); all nucleotide sequences were numbered according to a reference sequence with GenBank accession number X02763. For the prevention of PCR-generated mutations, TaKaRa Ex Tag polymerase (Takara Shuzo Co., Shiga, Japan), which was capable of proofreading, was used with the PCR assay. A serum sample obtained from an HBsAg-negative normal subject and an aliquot of pure water were included as negative controls. The methods of cloning and sequencing were described previously.15 For each sample, seven clones with inserts were selected for sequence analysis with an automatic DNA sequencer (CEQ 2000, Beckman Instruments, Inc., Fullerton, CA).

Pyrosequencing for the Verification of the Identified S Gene Mutations.

To further verify our sequence data resulting from direct sequencing, pyrosequencing was also performed. For the detection and quantitative assessment of the percentages of the two target S gene mutants, two sequencing primers called pyroS1 (5′-TGTCCTCCGATTTGTCC-3′, nucleotides 350-366) and pyroS2 (5′-ATCAACCACCAGCACGG-3′, nucleotides 446-512) and a reverse primer called PR (5′-GTGCAGTTTCCGTCCGTAGG-3′, nucleotides 600-581, antisense) were designed. The reverse primer was biotinylated to allow immobilization of the PCR product on streptavidin-coated beads. Samples were prepared in a 96-well format with a PyroMark Q96 vacuum prep workstation (Qiagen GmbH, Hilden, Germany) and with Sepharose high-performance streptavidin beads (GE Healthcare Bio-Sciences Corp., Piscataway, NJ). Primer annealing was conducted at 50°C for 2 minutes. Pyrosequencing was performed in a PyroMark Q96 ID pyrosequencer (Qiagen) according to the manufacturer's instructions with a pyrosequencing reagent kit (PyroMark Gold Q96 reagents, Qiagen).

Site-Directed Mutagenesis Experiment.

Site-directed mutagenesis was performed with a PCR-based method as described previously.19 Two primers that were complementary to each other and contained the target mutation site were synthesized. For the creation of the sT125A mutant, the primers SS [5′-GCAAAACCTGCGCGACTCCTGC-3′ (the mutation site is underlined), nucleotides 519-540, sense] and AS (the antisense oligonucleotide of SS) were used. A plasmid named pCMV-HBV (CMV indicates cytomegalovirus), which contained one copy of a greater-than-unit length HBV genome (3.37 kb, adw subtype), was used as the PCR template. Another primer called S1 (5′-TCTCCGCGAGGACTGGGGAC-3′, nucleotides 126-145, sense) was synthesized. PCR was performed with S1/AS and PS2/SS as the primers for the generation of two DNA fragments containing the mutation site. After gel purification, PCR was performed for 10 cycles with a mixture of the two fragments in the absence of primers. Finally, S1 and PS2 were added to the reaction mixture, and PCR was performed for 20 more cycles. The PCR product was blunt-ended and was inserted into pRc/CMV (Invitrogen, San Diego, CA) to generate pCMV-sT125A. As a control, a wild-type surface gene sequence was also PCR-amplified with the plasmid pCMV-HBV as the template. The PCR product was also blunt-ended and was inserted into pRc/CMV to obtain pCMV-S. For the creation of the sW74* mutant (pCMV-sW74*), the procedure was the same, except that the SS primer was replaced [5′-TTGTCCTGGTTATCGCTGAATG-3′ (the mutation site is underlined), nucleotides 361-382, sense].

Cell Line and DNA Transfection.

Huh-7 cells were maintained in Dulbecco's modified Eagle's medium with 10% fetal bovine serum. Transfection was performed with Lipofectamine 2000 (Invitrogen).

Northern and Western Blot Analysis.

The sequence flanked by primers S1 and PS2 was amplified, labeled, and used as the probe for northern analysis.15 The medium (100 μL) from the cell culture plates was loaded directly onto the nitrocellulose membrane. The following monoclonal antibodies (1:1000 dilution) were tested: monoclonal antibody against HBsAg (MAHBs)1 (lot M-21853, Genzyme Diagnostics, San Carlos, CA), MAHBs2 (lot M-21737, Genzyme Diagnostics), and MAHBs3 (clone 3B52, Chemicon International, Temecula, CA). The following immunogens were used to generate these antibodies: recombinant HBsAg derived from yeast (MAHBs1 and MAHBs2) and a culture fluid of human culture cells expressing HBsAg (MAHBs3). All monoclonal antibodies were confirmed by the providers to be capable of detecting HBsAg in clinical samples with an enzyme immunoassay. For sodium dodecyl sulfate–polyacrylamide gel electrophoresis, cells were lysed with trishydroxymethylaminomethane-buffered saline (10 mM Tris-HCl, pH 7.2, and 150 mM sodium chloride) containing 0.5% Nonidet P-40 (Sigma, Saint Louis, MO) and were centrifuged at 1500g. The soluble fraction (the cytoplasmic fraction) was subjected to sodium dodecyl sulfate–polyacrylamide gel electrophoresis, which was followed by western blot analysis. For visualization of both the mutant and wild-type HBsAg, a rabbit polyclonal anti-HBs antibody (ViroStat, Portland, ME) was used. As a control, β-actin was detected with the Ab-5 anti-actin antibody (NeoMarkers, Inc., Fremont, CA).

Immunofluorescence Analysis.

Huh-7 cells were grown on cover slips and transfected with DNA plasmids. Forty-eight hours after the transfection, the cells were fixed in acetone at −20°C for 2 minutes. A rabbit polyclonal anti-HBs antibody (ViroStat; 1:100 dilution) and a fluorescein isothiocyanate–conjugated goat anti-rabbit antibody (Leinco Technologies, Inc., Saint Louis, MO; 1:150 dilution) were used as the primary and secondary antibodies, respectively. For the visualization of the nuclei, the cells were stained with 4′,6-diamidino-2-phenylindole (DAPI; 200 ng/mL).

Results

Detection of the Two S Gene Mutations During the HBsAg-Negative Stages.

To determine whether the S gene mutations were present in the serum samples derived during the HBsAg-negative stage, we performed DNA extraction, PCR, and direct sequencing for patients 1 and 2. In patient 1, the following amino acid sequence variations were identified (in comparison with the GenBank EU306677 reference sequence): psL30S, psE54D, psA55T, psG145S, sL97P, sT125A, and sN207H. Among these mutations, sT125A was located in the “a” determinant region and, therefore, was chosen for further study. All mutations were found to mix with the wild-type sequences by direct sequencing. The PCR product was then cloned into the pCR2.1-TOPO vector, and seven clones with inserts were sequenced. sT125A could be identified in one of the seven clones (14.3%). Pyrosequencing was also performed to verify the presence of the sT125A mutant. The corresponding mutant constituted 11.2% of the viral population. Subsequently, two samples from HBsAg-positive stages were submitted for PCR and sequence analysis. The sT125A mutation was not present in the HBsAg-positive samples according to direct sequencing or cloning and sequencing.

In patient 2, the following mutations were identified in the HBsAg-negative serum sample: psE54D, psI68T, psP69L, psH71Q, psI84L, sA5T, sR73H, and sW74*. The sW74* mutation resulted in the deletion of the whole “a” determinant region and was, therefore, chosen for further study. After cloning and sequencing, sW74* was present in all seven clones (100%) with inserts. Pyrosequencing was also performed to verify the presence of the sW74* mutant. The corresponding mutant constituted 83.1% of the viral population. Two samples from HBsAg-positive stages were submitted for PCR and sequence analysis. The sW74* mutation was not present in the HBsAg-positive samples.

Phenotypic Characterization for the sT125A and sW74* Mutants.

For the study of the phenotypic changes in HBsAg in the sT125A mutant, 5 μg of a plasmid (pCMV-sT125A or pCMV-S) was transfected into Huh-7 cells. Northern blot analysis showed that a greater amount of S messenger RNA (mRNA) was detected in the pCMV-sT125A–transfected cells versus the pCMV-S–transfected cells (1.5- and 2.2-fold greater in two independent experiments; Fig. 3A). Immunofluorescence analysis using a polyclonal anti-HBs antibody detected both sT125A and wild-type surface antigens in the transfected Huh-7 cells (Fig. 3B). The apparent transfection efficiency was approximately 5% in both sets of experiments. However, western blot analysis detected wild-type surface proteins [glycoprotein 27 (GP27) and protein 24 (P24)] but not the sT125A mutant surface protein. To assess the antigenicity of HBsAg secreted into the medium, we performed a slot blot analysis with the culture medium. A small amount of the mutant HBsAg was detected by two monoclonal antibodies (MAHBs1 and MAHBs2), but it was not detected by the third one (MAHBs3) or the polyclonal antibody. A radioimmunoassay (Ausria II) and an enzyme immunoassay (Enzygnost HBsAg 5.0) were then used to detect HBsAg in the culture medium. The Ausria assay failed to detect the mutant HBsAg, but the Enzygnost assay detected the antigen, albeit in a low positive range (the signal/cutoff ratio was 13.37 for sT125A and 1380 for the wild type).

Figure 3.

Phenotypic analysis of the sT125A mutant in Huh-7 cells. (A) Northern blot analysis of the S mRNA levels in Huh-7 cells transfected by (1) the vector only (mock), (2) pCMV-sT125A, or (3) pCMV-S (wild type). As loading controls, 28S and 18S ribosomal RNAs were used. (B) Immunofluorescence analysis of surface protein expression in Huh-7 cells transfected with pCMV-S (wild type) or pCMV-sT125A. A polyclonal anti-HBs antibody (anti-poly-S) was used to detect HBsAg. Nuclei were demonstrated by DAPI staining. The two pictures were merged. (C) Western blot analysis of intracellular HBsAg in Huh-7 cells transfected with (1) pCMV-sT125A, (2) pCMV-S (wild type), or (3) the vector only (mock). Actin was used as a loading control. The solid triangle indicates a nonspecific band. (D) Slot blot analysis of secretory HBsAg in a medium of Huh-7 cells transfected with pCMV-S (wild type) or pCMV-sT125A. Three monoclonal antibodies (mo1-mo3) and a polyclonal antibody (poly) were used (see the Patients and Methods section).

To determine the phenotypic alterations of the sW74* mutant, we transfected 5 μg of pCMV-sW74* or pCMV-S into Huh-7 cells. Northern blot analysis showed similar expression levels of the S mRNA (Fig. 4A). However, the polyclonal anti-HBs antibody failed to detect the sW74* mutant in either immunofluorescence analysis (Fig. 4B) or western analysis. The mutant HBsAg could not be detected by either the Ausria assay or the Enzygnost assay.

Figure 4.

Phenotypic analysis of the sW74* mutant in Huh-7 cells. (A) Northern blot analysis of the S mRNA levels in Huh-7 cells transfected with (1) the vector only (mock), (2) pCMV-sW74*, or (3) pCMV-S (wild type). As loading controls, 28S and 18S ribosomal RNAs were used. (B) Immunofluorescence analysis of surface protein expression in Huh-7 cells transfected with pCMV-sW74* or pCMV-S (wild type). A polyclonal anti-HBs antibody (anti-poly-S) was used to detect HBsAg. Nuclei were demonstrated by DAPI staining. The two pictures were merged.

Discussion

The goal of anti-HBV treatment has changed significantly in the past decades. Before the clinical availability of interferon and oral antiviral agents, cytoprotective agents were considered effective because of their ability to normalize or reduce ALT levels.20, 21 Since the approval of regular interferon for anti-HBV treatment, HBeAg seroconversion has been used as an important endpoint for the evaluation of effective treatment.22 Although HBeAg seroclearance is usually accompanied by a significant reduction of the HBV DNA level, a significant proportion of patients continue to have high and fluctuating HBV DNA levels, and this results in HBeAg-negative hepatitis.23 Molecular analysis has revealed the selection of mutants that fail to secrete HBeAg (precore stop codon mutants).24 Because of the availability of nucleoside/nucleotide analogues in the past decade, long-term suppression of HBV replication has been achieved. Large-scale retrospective and prospective studies have confirmed the link between a low HBV DNA level and a reduced risk of liver cancer.25, 26 Therefore, the continuous or lifelong suppression of HBV DNA to levels less than 2000 IU/mL is now the goal of treatment according to several clinical guidelines.27-29

With the approval of peginterferon for anti-HBV treatment, a new goal is now being pursued. HBsAg seroconversion has been observed in approximately 3% of patients receiving peginterferon therapy.18 In patients with acute hepatitis B infections, the seroclearance of HBsAg and the appearance of anti-HBs are considered a cure of the disease because anti-HBs is believed to be a protective antibody. However, in patients with chronic hepatitis B, several previous observations have led to arguments against this concept. In some children who receive the HBV vaccine, escape mutants can develop in the presence of anti-HBs.30 Occult HBV infections have been repeatedly reported in patients who are negative for HBsAg, with some of these positive for anti-HBs.31 A recent study has indicated that patients who receive lamivudine treatment and experience HBsAg seroconversion can harbor an S gene mutant (sP120A), which can results in a failure to detect the surface antigen.15 Therefore, with the precedent of HBeAg-negative hepatitis, it may not be so surprising to discover the occurrence of HBsAg-negative hepatitis following the availability of effective antiviral therapies. Clinically, this study indicates that after peginterferon therapy, HBsAg seroconversion alone is insufficient evidence for a cure to be claimed. Careful monitoring of serum HBV DNA levels is advised.

Searching the literature, we found that the sT125A mutant was reported in a chronically infected Argentinean patient with anti-HBs antibodies.32 Notably, both the sT125A mutant and the sP120A mutant mimic the vaccine escape mutants.33 Furthermore, in patient 1, anti-HBs was detectable during the HBsAg-negative stage, and this led to the speculation that the vaccine-like selection pressure was derived from the emergence of anti-HBs after the significant suppression of HBV replication by peginterferon. On the other hand, various S truncation mutations similar to the one identified in patient 2 have been reported in lamivudine-treated patients with hepatocellular carcinoma. However, in those patients, the truncation points seemed to avoid the transactivity-on region (codons 25-150), whereas in patient 2, the nonsense mutation occurred in the middle of this region. In this patient, the mutant did not persist for a very long time, and no hepatocellular carcinoma has developed yet.

In summary, we have identified two S gene mutations responsible for the failure to detect HBsAg in patients who received peginterferon treatment and experienced HBsAg seroconversion. One of the patients subsequently experienced HBsAg-negative hepatitis. Our results show that after peginterferon treatment, HBsAg seroconversion does not necessarily indicate the eradication of the virus. The emergence of an HBsAg-negative mutant virus is another possibility.

Ancillary