INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Hepatitis B virus (HBV) is a small, partially double-stranded DNA (dsDNA) virus that causes acute and chronic hepatitis in humans. More than 350 million people are persistently infected with HBV, making it a global public health concern. HBV is a leading cause of death in many parts of the world, as chronically infected individuals are at significantly increased risk for developing potentially fatal cirrhosis or hepatocellular carcinoma (5, 35). Until recently, the only approved therapy for chronic HBV infection was treatment with the cytokine alpha interferon. Only 30 to 40% of chronically infected individuals with low-level viremia and evidence of active liver disease (generally indicated by high alanine aminotransferase levels) respond to interferon with sustained elimination of HBV (46, 68). Apart from its limited efficacy, interferon is expensive, requires administration by subcutaneous injection, and may cause dose-limiting side effects. More effective therapies are needed to treat the majority of chronically infected individuals, who do not respond to interferon. The recent approval of the deoxycytidine analog lamivudine has provided an alternative treatment option (16, 25, 31, 55). Other nucleoside and nucleotide analogs, including famciclovir and adefovir, have recently progressed to phase III clinical trials and may soon provide additional options (9).

Lamivudine is a dideoxycytidine derivative which is active against human immunodeficiency viruses (HIV) (8) and hepadnaviruses (17, 25). It has several advantages as an antiviral agent, including high oral bioavailability, low toxicity, and potent and specific anti-HBV and anti-HIV activities (25). Lamivudine was shown to suppress HBV replication very effectively in vitro (17), as well as in short-term preclinical (59) and clinical trials (16, 25, 31, 55). Lamivudine is phosphorylated intracellularly by cellular kinases which generate lamivudine triphosphate, a dCTP analog that competes with dCTP for recognition by viral DNA polymerases. The addition of lamivudine to the 3' end of nascent HBV DNA strands causes immediate chain termination (57), which secondarily prevents the maturation and release of infectious viral particles, thus limiting the spread of the virus.

Famciclovir is a prodrug for the deoxyguanosine analog penciclovir, a compound originally developed and approved for the treatment of herpesvirus infections (65). Penciclovir was subsequently shown to be an effective inhibitor of hepadnavirus replication in vitro (27, 58) and in duck and woodchuck models of HBV infection (59). Although clinical trials showed that famciclovir treatment caused only modest suppression of HBV replication in chronically infected patients (15a, 36), this translated into significant histological improvement (15a). The use of famciclovir to treat chronic HBV infection and as a prophylactic agent during liver transplantation is currently being explored (2, 3, 54), as is its use in combination with lamivudine and other agents (60).

Adefovir is an acyclic dAMP analog which shows broad-spectrum antiviral activity (12). The potent antihepadnaviral activity of adefovir was identified initially in in vitro assays (73) and in animal models (21, 45) and later confirmed in phase II and III clinical trials when patients who were chronically infected with HBV were treated with an orally available prodrug, adefovir dipivoxil (19).

Short-term monotherapy with either lamivudine, famciclovir, or adefovir causes a rapid decrease in viremia, which usually returns to pretreatment levels after therapy is stopped (16, 31, 43, 55). The most likely reason for this relapse is persistence of the covalently closed circular (CCC) form of viral DNA (47, 64) which exists as a minichromosome in the nuclei of infected cells (44). Viral mRNAs, which are both necessary and sufficient to initiate HBV replication in the cell cytoplasm, continue to be transcribed from HBV CCC DNA even during aggressive antiviral therapy (40, 47). Therapy with antiviral deoxynucleoside or deoxynucleotide analogs inhibits cytoplasmic HBV replication but has no direct effect on HBV CCC DNA, which is transcribed by cellular RNA polymerase II (42). For this reason, it is expected that long-term therapy will be necessary to maintain suppression of HBV replication.

Unfortunately, long-term monotherapy with nucleoside analogs, including lamivudine and famciclovir, has engendered another problem: the frequent emergence of drug-resistant strains of HBV (1-4, 23, 32, 33, 41, 52-56, 70, 71). Drug resistance arises as a consequence of the inherently low fidelity of HBV replication and the high viral load and turnover which characterize chronic HBV infection. In addition, immune pressure generates hypervariability in the major hydrophilic loop of hepatitis B surface antigen (the main target for neutralizing antibodies) and consequently causes variability in the viral polymerase, which is encoded by the same genome sequence in an overlapping reading frame (34, 42, 66).

The incidence of lamivudine resistance increases with treatment duration; it typically develops in 16 to 32% of patients within 1 year of therapy and rises to more than 50% within 2 years (16, 32, 33, 55, 61). Sequencing of HBV DNA derived from individuals who exhibit viral breakthrough during nucleoside analog therapy has revealed a number of specific mutations (1-4, 24, 33, 62, 63). The majority of lamivudine-resistant HBV strains contain mutations in the C (catalytic) domain of the viral DNA polymerase which map to the sequence that encodes the YMDD (tyrosine-methionine-aspartate-aspartate) motif. Based on homology to other polymerases, this tetrapeptide is believed to be involved in nucleotide binding and catalysis of the DNA polymerization reaction. Both YVDD (M550V) and YIDD (M550I) changes are frequently observed during lamivudine breakthrough; the former (M550V) almost invariably coexists with a second change (L526M) in the B (template binding) domain of the polymerase (24). The L526M change and several others located in or near the B domain have been clinically correlated with resistance to famciclovir (2, 3, 70). The emergence of such drug-resistant HBV strains emphasizes the need for new antivirals and therapeutic strategies. The phenotypes of drug-resistant HBV strains also warrant further investigation both in vitro and in vivo, since they appear to differ substantially from wild-type (wt) HBV in replication competence (29, 38, 48, 53) and this presumably affects pathogenesis in the host.

Characterization of both wt and drug-resistant HBV strains has been hampered by the virus' host specificity in vivo and poor infectivity in vitro. Woodchuck hepatitis virus provides the best opportunity for studying the selection of drug-resistant virus populations in vivo. However, the mutations that confer lamivudine resistance on woodchuck hepatitis virus in the woodchuck are different from those which confer resistance on HBV (37). Therefore, results obtained from woodchuck studies must be interpreted with caution as they may not be predictive for HBV. Whether recently developed transgenic mice (20) can be used for such studies remains to be established. Convenient in vitro cell culture systems are likewise required. Although viral replication may occur following delivery of full-length genomic HBV DNA to hepatic cell lines by transfection, most transfection methods are inefficient, making accurate quantitation of replication-deficient viral mutants extremely difficult. We have previously reported the development of a novel in vitro system for studying HBV replication which utilizes recombinant baculoviruses to deliver replication-competent HBV genomes to HepG2 cells (14). In contrast to other methods of DNA delivery, baculovirus-mediated transduction of HBV is very efficient and leads to high levels of HBV replication in the majority of cells. The increased levels of HBV replication which can be achieved by this method make the HBV-baculovirus model well suited for a variety of molecular studies, including antiviral testing. Recently, we have used this system to perform a detailed characterization of the effects of lamivudine on wt HBV replication and CCC DNA accumulation in vitro (15). Based on the utility of the HBV-baculovirus system for studying wt virus, we sought to expand the model to include novel baculovirus vectors to transfer HBV genomes encoding mutations in the polymerase gene that are commonly associated with clinical resistance to lamivudine and famciclovir. The studies reported here therefore had two main goals, namely, (i) to construct HBV-baculovirus vectors that encode the L526M, M550I, and L526M M550V polymerase mutations and (ii) to use these vectors to compare the sensitivities of wt and mutant HBV to three antivirals of current clinical importance: lamivudine, penciclovir (the active form of famciclovir), and adefovir.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Cell culture. Sf21 insect cells (Invitrogen, Groningen, The Netherlands) were maintained in a nonhumidified incubator at 28°C without CO₂ and were grown in Grace's insect medium (Gibco BRL Life Technologies, Grand Island, N.Y.) containing yeastolate, lactalbumin hydrolysate, and 10% heat-inactivated fetal bovine serum (15). HepG2 cells were maintained in humidified 37°C incubators at 5% CO₂ and grown in minimal essential medium supplemented with 10% heat-inactivated fetal bovine serum.

HBV infectious clone and lamivudine-resistant mutants. A wt HBV clone of 1.28 times genome length (genotype A, subtype adw2) previously shown to be infectious in vitro was excised from the plasmid pHBV1.2 using the restriction enzymes PvuII and BamHI and ligated into the baculovirus transfer vector pBlueBac4.5 (Invitrogen) using SmaI and BamHI restriction sites (6). The resulting plasmid, pBBHBV1.28 (Fig. 1), was used to sequence both strands of the HBV construct to ensure that no sequences outside the published chronic-carrier consensus sequence for wt HBV were present in the clone (3). Lamivudine resistance mutations were introduced into the wt pBSHBV1.28 plasmid by site-directed mutagenesis using the Quick change kit (Stratagene, La Jolla, Calif.). Introduced mutations caused the following amino acid changes in the polymerase protein: (i) a change from leucine to methionine at codon 526 (L526M) in the B domain, (ii) a change from methionine to isoleucine at codon 550 in the C domain (M550I), and (iii) both an L526M change and a change from methionine to valine at codon 550 (L526M M550V). Figure 2 illustrates these changes, as well as the concomitant changes in the envelope protein (34).

View larger version (17K): [in this window] [in a new window]   FIG. 1.   Diagrammatic representation of the HBV construct used to generate HBV-baculovirus recombinants. HBV-baculovirus recombinants were constructed from a 1.28 times unit length HBV genome of genotype A, serotype adw2. This HBV clone has been fully sequenced and contains no nucleotides which lie outside a published wt consensus sequence (3). Note that there is only one copy each of the precore, core, pre-S1-pre-S2-S, polymerase, and X open reading frames within the construct. The positions of the core (C), pre-S, S, and X promoters, enhancer I (EI), and enhancer II (EII), as well as the baculovirus polyhedron promoter (PH), are indicated.

View larger version (35K): [in this window] [in a new window]   FIG. 2.   Polymerase mutations in drug-resistant HBV. Drug resistance mutations in the 1.28 times unit length wt HBV construct (Fig. 1) were generated by site-directed mutagenesis. The following amino acid changes in the polymerase gene product resulted: (i) an L-to-M amino acid substitution at position 526 (L526M) in domain B of the polymerase, (ii) an M-to-I amino acid substitution at position 550 (M550I) in domain C of the polymerase, and (iii) a double mutation with the L526M amino acid substitution and a second, M-to-V, amino acid substitution at position 550 (L526M M550V). The M550I and M550V mutations also resulted in changes to the deduced amino acid sequence of the surface antigen gene, as indicated. TP, terminal protein; R., reverse; ORF, open reading frame; HBsAg, hepatitis B surface antigen.

Generation of recombinant baculovirus. HBV-baculovirus recombinants were generated by cotransfection of pBBHBV1.28 containing either wt HBV or the respective lamivudine- and famciclovir-resistant HBV and linear Bac-N-Blue Autographa californica nuclear polyhedrosis virus DNA (Invitrogen) into Sf21 cells. Recombinant baculoviruses were isolated from the cotransfection supernatant by plaque purification in Sf21 cells. Baculovirus from individual plaques was amplified by inoculation into 100-mm-diameter dishes of Sf21 cells. One week postinoculation, baculovirus was concentrated from the medium of infected cells by centrifugation (14, 15) and viral DNA was extracted by incubation in sodium dodecyl sulfate buffer with proteinase K (Roche Diagnostics Australia Pty. Ltd., Rose Park, S.A., Australia), phenol-chloroform extraction, and ethanol precipitation as previously described (50). DNA purified from recombinant virus was analyzed by PCR amplification of the HBV polymerase region using the forward primer PCRBACF2 (5'-ATTCTTTGTCCATTGATCGAAGCGAG-3') and the reverse primer OS1 (5'-GCCTCATTTTGTGGGTCACCATA-3'). PCR products were purified using the JetQuick DNA purification kit (Stratagene), and both strands of the product were then sequenced using the primers TTS2 (5'-TGCACGATTCCTGCTCAA-3' [plus strand]) and OS2 (5'-TCTCTGACATACTTTCCAAT-3' [minus strand]) to ensure that the desired resistance mutations were retained in each recombinant. Restriction enzyme digestion, agarose gel electrophoresis, and Southern blotting using a ³²P-labeled HBV probe were also performed as described previously (14) to verify that each HBV-baculovirus recombinant contained an entire HBV insert.

Preparative baculovirus amplification and purification. Individual baculovirus isolates containing either wt HBV or the L526M, M550I, or L526M M550V mutant were amplified by infecting suspension cultures of Sf21 cells in log phase at a multiplicity of infection (MOI) of 0.5 PFU per cell. Infections were allowed to proceed until most of the cells showed visible signs of infection (after approximately 4 to 5 days). Baculovirus particles were concentrated and purified from infected Sf21 medium as described previously (7, 50). Purified virus was titrated in quadruplicate in Sf21 cells by endpoint dilution (50). Baculovirus DNA was purified from a small aliquot of each high-titer stock, and PCR and subsequent DNA sequencing were repeated as described above to confirm that each stock retained the intended mutations.

Baculovirus transduction and drug treatments. HepG2 cells were seeded into six-well plates at a density of 2 × 10⁶ to 3 × 10⁶ cells per well and were transduced with 50 PFU of HBV-baculovirus per cell 16 h postseeding as previously described (14). HepG2 cells were fed medium containing the indicated concentrations of antiviral drugs immediately after transduction. Each antiviral was tested against wt HBV and the L526M, M550I, and L526M M550V mutants in parallel using the same stocks of drug-containing medium to ensure that different recombinants were exposed to identical drug concentrations. Transduced HepG2 cells were fed every other day with fresh drug for a total of 1 week, with the last drug treatment beginning 24 h prior to the analysis of HBV DNA.

Analysis of HBV DNA. Intracellular HBV replicative intermediates were isolated from cytoplasmic core particles essentially as previously described (22). Briefly, cell monolayers were lysed by incubation at 4°C for 20 min in phosphate-buffered saline containing 0.5% Nonidet P-40. Cell lysates were transferred to microcentrifuge tubes and spun for 5 min to pellet nuclei. Supernatants were transferred to clean tubes, adjusted to 10 mM MgCl₂, and incubated with 10 U of DNase I (Roche Diagnostics) at 37°C. After 1 h, the digestion mixture was adjusted to 50 mM EDTA-30 mM Tris-0.5% sodium dodecyl sulfate-500 µg of proteinase K per ml, and incubation at 37°C was continued for an additional 4 h. Sequential extractions with Tris-saturated phenol and chloroform were performed, and nucleic acids were then recovered by precipitation with 1 volume of isopropanol. Pellets containing viral DNA were redissolved in 10 mM Tris-10 mM EDTA and digested with 20 U of DNase-free RNase (Roche Diagnostics) before analysis by electrophoresis and Southern blotting as described above.

Quantification of HBV DNA and data analysis. Intracellular HBV DNA replicative intermediates were quantified by densitometery of suitable autoradiograms using a Bio-Rad GS-670 scanning densitometer and Molecular Analyist software (Bio-Rad). Densitometric data were analyzed using TableCurve2D, a graphics-statistics software package from Jandel Scientific (San Rafael, Calif.) as described previously (10, 11). Dose-dependent inhibition of the accumulation of intracellular HBV replicative intermediates, when it occurred, could be expressed mathematically in nearly all cases by the following logistic dose-response equation: y = a/(1 + [x/b]^c), which describes a sigmoid curve where y is the amount of HBV DNA detected compared to that in untreated controls (defined as 100%) and x is the drug concentration; a represents the curve's amplitude, b is the x value at its transition center, and c is a parameter which defines its transition width (Fig. 3). Values and confidence intervals for each parameter were estimated from individual dose-response equations. When dose-dependent inhibition of intracellular HBV accumulation was incomplete over the range of drug concentrations used, an alternative procedure was adopted as described in footnote d to Table 1.

View larger version (21K): [in this window] [in a new window]   FIG. 3.   Inhibition of wt HBV replication by lamivudine, adefovir, or penciclovir. Double-stranded HBV replicative intermediates were quantified by densitometry of appropriately exposed autoradiographs of Southern blots (see Fig. 4 to 6). Image densities were expressed as percentages of the mean density of untreated controls (defined as 100%). The resulting data, which describe the dose-dependent effects of drugs on wt HBV replication, were fitted to logistic dose-response equations with the aid of TableCurve2D software. The 95% confidence limits appear on either side of the fitted dose-response curves; the equation y = a/(1 + [x/b]c) accurately describes the data. With the exceptions noted in footnote d to Table 1 and the legend to Fig. 5, all data sets were also accurately described by this equation. As shown, the lowest concentrations of lamivudine and adefovir (0.001 and 0.01 µM, respectively) appeared to stimulate wt HBV replication. This is also reflected by >100 values for the a (amplitude) parameter listed in Table 1. The cause(s) of this phenomenon is not known but may include (i) allosteric effects of (d)NTP analogs on intracellular nucleotide pools and/or (ii) interference with virus secretion, which might result in accumulation of intracellular replicative intermediates.

Nucleotide sequence accession number. The sequence of pBBHBV1.28 has been submitted to the GenBank database and assigned accession number AF305422.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Construction of recombinant baculoviruses. Recombinant baculovirus containing wt HBV or the L526M, M550I, or L526M M550V HBV variant was successfully generated and amplified in Sf21 cells. Southern blotting was used to confirm that baculovirus stocks contained complete HBV constructs, the sequences of which were verified after PCR amplification of both strands (data not shown). No mutations (other than those intentionally introduced by site-directed mutagenesis) were detected at any stage during baculovirus amplification, indicating that the baculoviruses maintained the sequences and structural integrity of each HBV insert.

Sensitivity to lamivudine. To confirm that the point mutations which created the L526M, M550I, and L526M M550V polymerase mutants conferred phenotypic lamivudine resistance on the genotype A clone of HBV used in these experiments, the abilities of these viruses to replicate in the presence of increasing concentrations of lamivudine were examined. HepG2 cells were transduced with baculovirus which contained either wt or mutant HBV at an MOI of 50 PFU per cell. The cells were then exposed continuously for 7 days to 0, 0.001, 0.01, 0.1 1.0, or 50 µM lamivudine. HBV replicative intermediates were extracted from the cytoplasm of transduced cells at the end of the 7-day treatment period before analysis by Southern blotting and autoradiography. Double-stranded replicative intermediates were quantitated by densitometry of autoradiographs, and dose-response curves were fitted to the resulting data with the aid of TableCurve2D (Fig. 3 and 4). Drug concentrations which were required to reduce HBV replication by 50% (IC₅₀s) were estimated from the equations which described each curve (Table 1). The results verified that wt HBV was extremely sensitive to lamivudine, whose IC₅₀ was approximately 0.01 µM, as estimated from intracellular HBV dsDNA. In this experiment, the single L526M change conferred partial resistance to lamivudine, as demonstrated by a threefold increase in the estimated IC₅₀. Both the M550I and L526M M550V changes conferred much greater resistance to lamivudine, as the altered dose-response plots (Fig. 4) and the increases in the estimated IC₅₀s indicate. Table 1 presents the results of a single experiment in which the levels of inhibition of wt HBV and all of the mutants by all three analogs were compared in parallel. For logistical reasons, replicate experiments compared either the sensitivities of wt and mutant HBV to individual drugs or the sensitivity of wt HBV or specific mutants to all of the drugs. For three consecutive experiments, the estimated IC₅₀ of lamivudine as an inhibitor of wt HBV replication was 0.10 ± 0.10 µM (mean ± standard deviation; range, 0.01 to 0.20 µM); in these experiments, 1 µM lamivudine inhibited wt HBV replication by 96.4% ± 4.1% (mean ± standard deviation; range, 91.8 to 99.6%. Despite interexperimental variation in the IC₅₀s of the individual drugs, intraexperimental variation was less than 20%. More importantly, the ranking of IC₅₀s was reproducible, showing sensitivity to lamivudine consistently decreasing in the order wt > L526M > M550I > L526M M550V by factors of <10 (L526M), >20 (M550I), and >100 (L526M M550V), respectively.

View larger version (37K): [in this window] [in a new window]   FIG. 4.   Effects of lamivudine on wt and drug-resistant HBV. HepG2 cells were transduced at 50 PFU/cell with HBV-baculovirus which encoded either wt or mutant (L526M, M550I, or L526M M550V) polymerase. Immediately after transduction, cultures were treated with the indicated concentrations of lamivudine (3TC). After 7 days of treatment, HBV replicative intermediates were extracted from cell lysates and analyzed by Southern blotting and autoradiography. (A) autoradiographs of Southern blots. After equivalent exposure times, image densities in (drug-free) control lanes were such as to suggest that replication fitness decreased in the order wt > L526M > L526M M550V > M550I, assuming that transduction efficiencies were equal and that the polymerase changes did not differentially affect virus secretion. The autoradiographs shown were exposed for different times so that the image density in each control lane is approximately the same. (B) results of graphical analysis of data from panel A. rc, relaxed circular.

Sensitivity to penciclovir. The sensitivities of wt and mutant HBV to penciclovir were compared next. Initial experiments indicated that penciclovir was a relatively weak inhibitor of wt HBV replication in HepG2 cells and that it failed to inhibit wt HBV replication at concentrations as high as 500 µM in one HepG2 subclone (data not shown). However, it was possible to reproduce dose-response relationships for wt HBV using high concentrations of penciclovir (Fig. 3 and 5). We performed cross-resistance testing by treating HepG2 cells transduced with 50 PFU of baculovirus which encoded wt or drug-resistant HBV using penciclovir concentrations of 0, 1, 10, 50, 100, and 500 µM. HBV replicative intermediates were extracted and assayed after 7 days of exposure to penciclovir (Fig. 5). Densitometric data were generated and analyzed as described above. Due to the high degree of penciclovir resistance shown by all of the mutants, it was not possible to fit logistic dose-response curves to the corresponding sets of data (see Table 1, footnote d). These results indicate that wt HBV replication was inhibited only by relatively high concentrations of penciclovir, for which an IC₅₀ of approximately 11.5 µM (95% confidence limits, 8.2 to 16.2 µM) was estimated for the dsDNA species. The single L526M change, which has previously been associated with clinical HBV breakthrough during famciclovir therapy (2, 3, 70), conferred significant resistance to penciclovir in vitro, as indicated by the approximately almost 1-log increase in the estimated IC₅₀. The M550I and L526M M550V mutants were also penciclovir resistant. Interestingly, the degree of resistance exhibited by both was greater than that shown by the L526M mutant, suggesting that mutations which affect nucleotide binding (M550I and M550V) and template binding (L526M) may contribute independently to the resistance phenotype (Table 1).

View larger version (37K): [in this window] [in a new window]   FIG. 5.   Effects of penciclovir on wt and drug-resistant HBV. HepG2 cells were transduced at 50 PFU/cell with HBV-baculovirus which encoded either wt or mutant (L526M, M550I, or L526M M550V) polymerase. Immediately after transduction, cultures were treated with the indicated concentrations of penciclovir (PCV). After 7 days of treatment, HBV replicative intermediates were extracted from cell lysates and analyzed by Southern blotting and autoradiography. (A) autoradiographs of Southern blots (exposures were chosen to give approximately equal image densities, as noted in the legend to Fig. 4). (B) results of graphical analysis of data from panel A. As noted in Table 1, footnote d, it was not possible to fit logistic dose-response curves to data for the M550I single and L526M M550V dual mutants. rc, relaxed circular.

Sensitivity to adefovir. The sensitivities of wt and mutant HBV to adefovir, the active metabolite of the prodrug adefovir dipivoxil, were compared next. HepG2 cells transduced as described above with baculovirus which contained either wt or mutant HBV were exposed for 7 days to 0, 0.01, 0.1, 1, 10, or 100 µM adefovir. Intracellular HBV replicative intermediates were extracted and analyzed as described above, with the results shown in Fig. 3 and 6. In contrast to the results obtained following exposure to lamivudine or penciclovir, exposure to adefovir produced a dose-dependent inhibition of replication of both wt and mutant HBV. Dose-response curves (Fig. 6) and IC₅₀s derived from them (Table 1) indicated that wt HBV was sensitive to adefovir, with an estimated IC₅₀ of 0.08 µM (95% confidence limits, 0.06 to 0.10 µM). The corresponding IC₅₀s of adefovir as an inhibitor of lamivudine-resistant mutants increased less than fourfold, suggesting that the L526M, M550I, and L526M M550V changes do not confer significant cross-resistance to adefovir (Table 1).

View larger version (36K): [in this window] [in a new window]   FIG. 6.   Effects of adefovir on wt and drug-resistant HBV. HepG2 cells were transduced at 50 PFU/cell with HBV-baculovirus which encoded either wt or mutant (L526M, M550I, or L526M M550V) polymerase. Immediately after transduction, cultures were treated with the indicated concentrations of adefovir (PMEA). After 7 days of treatment, HBV replicative intermediates were extracted from cell lysates and analyzed by Southern blotting and autoradiography. (A) autoradiographs of Southern blots (exposures were chosen to give approximately equal image densities, as in Fig. 4 and 5). (B) results of graphical analysis of data from panel A. rc, relaxed circular.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Many aspects of HBV research, including the investigation of potential antivirals, have been impeded by the lack of convenient and reliable in vitro culture systems that are capable of supporting complete cycles of HBV replication. High levels of in vitro HBV replication can be achieved using the recently described recombinant HBV-baculovirus system (14, 15), which makes it suitable for the study of HBV mutants that replicate poorly in other systems (29, 38, 48, 53). The higher level of replication and the ability to control the viral MOI facilitate otherwise difficult comparisons of the sensitivities of wt and mutant HBV to antiviral agents.

Data obtained in the present study confirm and extend conclusions from previous studies which used alternative assay systems; in addition, they demonstrate the utility of the HBV-baculovirus system as a tool for the in vitro study of clinically relevant drug-resistant variants of HBV. In summary, they show (i) that the L526M variant is resistant to penciclovir and partially resistant to lamivudine, (ii) that the M550I and L526M M550I variants are resistant to both lamivudine and penciclovir, and (iii) that all of the variants remain sensitive to adefovir.

The differences in lamivudine sensitivity among the HBV mutants utilized in this study (Fig. 3 and 4 and Table 1) suggest that each point mutation contributes independently to the resistance phenotype. The single L526M change in domain B of the HBV polymerase increased lamivudine resistance less than 10-fold, while both the M550I and L526M M550V changes conferred much greater resistance (greater than 20- and 100-fold, respectively), consistent with previous observations (18, 29, 30, 48).

Previous studies indicated that the anti-HBV activity of penciclovir is more variable than that of lamivudine. Penciclovir was initially found to be a potent inhibitor of wt HBV replication in one clone of HepG2.2.15 cells, with a reported IC₅₀ of 0.6 µM (27), but subsequent reports noted either lack of activity or IC₅₀s which are 1 or 2 orders of magnitude greater (18, 48, 53, 72). The reason(s) for the differences is not clear but may include poor and variable phosphorylation of penciclovir, which seems to characterize most clones of hepatic cell lines (T. Shaw and G. Civitio, unpublished data) and/or high intracellular concentrations of competing dGTP.

High concentrations of penciclovir produced reproducible dose-dependent inhibition of wt HBV replication (as reflected by concentration-dependent reductions in the accumulation of the intracellular HBV dsDNA replicative intermediate) in the HepG2 clone used here. This made it possible to demonstrate that the L526M change alone was sufficient to significantly increase (almost 10-fold) penciclovir resistance. Both the dual (L526M M550V) and M550I mutants were also found to be penciclovir resistant in this system. These results are consistent with recent conclusions based on alternative assays (18, 55). While our observation that M550I in isolation is sufficient to confer significant penciclovir resistance may seem inconsistent with a report (based on data obtained using HepAD38 and HepAD79 cells which express wt or M550V mutant HBV, respectively) that M550V in isolation is not (72), the latter probably has little relevance, since M550V has not been found in clinical HBV isolates in the absence of L526M.

Our results suggest that development of the M550I and L526M M550V resistance mutations during lamivudine therapy would cause the failure of subsequent famciclovir therapy, a conclusion which is supported by the limited amount of clinical data published to date (41, 67). Conversely, although the L526M mutation which emerges during famciclovir treatment confers only partial resistance to lamivudine, it remains possible that preselection of L526M as a result of prior exposure to famciclovir may predispose to failure of subsequent lamivudine treatment. Although L526M alone appears insufficient to drive viral breakthrough during lamivudine treatment, its preexistence may affect replication fidelity in a way which increases the probability and/or rate of acquisition of an additional mutation(s) (such as M550V) which confers greater lamivudine resistance. This hypothesis is supported by at least two reports, which describe the rapid replacement of the L526M mutant by an L526M M550V mutant after lamivudine replaced or was added to famciclovir therapy (56, 62).

Adefovir was active against all three of the HBV mutants used in the present study, in agreement with previous work, which has used either cell-based or enzyme-based assays (48, 69, 70, 72). Differences between the adefovir sensitivities of wt and mutant HBVs were not significant (fourfold or less). This indicates (i) that the lamivudine and/or famciclovir resistance conferred by the L526M, M550I, and L526M M550V polymerase changes does not confer cross-resistance to adefovir and (ii) that HBV mutants which are selected for by lamivudine are not hypersensitive to adefovir, at least in the HBV-baculovirus assay system. Adefovir hypersensitivity is a phenomenon that has been observed for the lamivudine-resistant M184V HIV mutant, which is analogous to the M550V single mutant of HBV (39). These observations add to the already convincing evidence indicating that adefovir is active against lamivudine-resistant HBV strains in vitro, suggesting that adefovir dipivoxil treatment would be appropriate for patients in whom lamivudine therapy has failed. Indeed, two recent reports have described the successful use of adefovir to rescue patients who experienced relapses after failure of lamivudine treatment due to the emergence of resistant HBV (51, 52). While adefovir dipivoxil is an effective broad-spectrum antiviral that is potentially useful against lamivudine-resistant HBV, its clinical use has been associated with nephrotoxicity during trials against HIV. However, doses of adefovir which are sufficient to inhibit HBV replication in vivo (19) are lower than the doses used during clinical trials against HIV (13, 26). Clinical trials which are currently under way will determine whether lower doses of adefovir are sufficient to block HBV replication without causing nephrotoxicity.

Although lamivudine, famciclovir, and adefovir have proved to be safe and efficacious suppressors of in vivo HBV replication in the short term, there is little, if any, evidence that they affect HBV CCC DNA which persists in the nuclei of infected cells. Elimination of viral CCC DNA is likely to require very long-term therapyprobably longer than several half-lives of the infected cell population (9, 37, 40, 60). To achieve successful long-term control of HBV replication, additional antiviral agents will be required; fortunately, several potentially useful agents, some of which have progressed to clinical trials, are currently being developed (9, 60). In the future, these will probably be used in combination(s), rather than alone, to minimize the risk of development of viral resistance (10, 11, 28, 60). Unfortunately, the fact that HBV resistance to adefovir has not yet been observed is no guarantee that it will not occur.

The results presented here confirm and extend previous observations and further validate the usefulness of the recombinant HBV-baculovirus system, which offers several advantages, most notably the capacity to support high levels of replication of both wt and mutant HBV. We anticipate that it will become a valuable asset for research into HBV and that the recombinants described here and others like them will be invaluable not only for defining patterns of drug resistance but also for testing the efficacy of different drug combinations. The system also has potential for investigation of effects of antiviral agents on HBV CCC DNA amplification and maintenance, since this key replicative intermediate is amplified to readily detectable levels (14, 15). Data thus obtained should contribute to the rational design of new therapeutic strategies for the management of chronic HBV infection, especially in cases in which drug-resistant HBV strains are implicated.