Generation of Stable Cell Lines Expressing Lamivudine-Resistant Hepatitis B Virus for Antiviral-Compound Screening

Abstract

Lamivudine [β-l-(−)-2′,3′-dideoxy-3′-thiacytidine] is a potent inhibitor of hepadnavirus replication and is used both to treat chronic hepatitis B virus (HBV) infections and to prevent reinfection of transplanted livers. Unfortunately, lamivudine-resistant HBV variants do arise during prolonged therapy, indicating a need for additional antiviral drugs. Replication-competent HBV constructs containing the reverse transcriptase domain L180M/M204V and M204I (rtL180M/M204V and rtM204I) mutations associated with lamivudine resistance were used to produce stable cell lines that express the resistant virus. These cell lines contain stable integrations of HBV sequences and produce both intracellular and extracellular virus. HBV produced by these cell lines was shown to have a marked decrease in sensitivity to lamivudine, with 450- and 3,000-fold shifts in the 50% inhibitory concentrations for the rtM204I and rtL180M/M204V viruses, respectively, compared to that for the wild-type virus. Drug assays indicated that the lamivudine-resistant virus exhibited reduced sensitivity to penciclovir [9-(4-hydroxy-3-hydroxymethyl-but-1-yl) guanine] but was still inhibited by the nucleoside analogues CDG (carbocyclic 2′-deoxyguanosine) and abacavir {[1S,4R]-4-[2-amino-6-(cyclopropylamino)-9H-purin-9-yl]-2-cyclopentene-1-methanol}. Screening for antiviral compounds active against the lamivudine-resistant HBV can now be done with relative ease.

Over 350 million people worldwide are chronically infected with hepatitis B virus (HBV) despite the availability of an effective vaccine. Chronic HBV infection is associated with both cirrhosis and hepatocellular carcinoma (4, 5, 33). Until recently, the only licensed treatment available was alpha interferon, which has a response rate of only 30 to 40% in selected patients (51). The (−) enantiomer of 2′-dideoxy-3′-thiacytidine, also known as lamivudine, is a potent inhibitor of HBV replication (11, 12, 16, 41). Lamivudine is phosphorylated by cellular deoxycytidine kinases. This triphosphate form of lamivudine then competes with cellular dCTP and is incorporated by the viral polymerase, causing premature chain termination during both the reverse transcription and DNA synthesis steps of HBV DNA synthesis (45).

Chronically infected individuals treated with long-term lamivudine therapy show histological improvement and a significant reduction of viral load (15, 27, 28). Lamivudine has also been shown to be beneficial in both the prevention and treatment of reinfection of livers transplanted into HBV carriers (3, 7, 8, 19, 37). A drawback to lamivudine therapy is that the template for hepadnavirus replication, covalently closed circular DNA (cccDNA), is relatively unaffected by lamivudine (36). Although the half-life of HBV cccDNA in humans has not been established, the half-lives of both duck HBV and woodchuck HBV cccDNA have been shown to be quite long. Both are estimated to be between 30 and 57 days (1, 56). Because of the apparent stability of cccDNA, continuous suppression of HBV replication will likely require long-term antiviral therapy. However, extended lamivudine monotherapy does result in selection of lamivudine-resistant HBV variants. The incidence of lamivudine resistance was 16 to 32% in chronically infected individuals treated with lamivudine for 1 year (M. Atkins, C. Hunt, and N. Brown, Hepatology, abstr. 319A, 1998). Longer-term studies have shown incidence rates as high as 49 and 58% after 104 weeks and 3 years, respectively, of lamivudine therapy (30). Lamivudine resistance in liver transplant patients has also been demonstrated (6, 32, 48), with the incidences of resistance posttransplant as high as 60% at 11 months (34).

Mutations associated with lamivudine resistance occur most often in the conserved tyrosine-methionine-aspartic acid-aspartic acid (YMDD) motif of the nucleotide-binding site of the viral polymerase (2, 17, 32, 35). Single nucleotide changes at codon 204 of the reverse transcriptase domain of the polymerase result in the substitution of either valine or isoleucine for methionine (rtM204V or rtM204I). The valine substitution, and occasionally the isoleucine substitution, is accompanied by an additional upstream mutation at codon 180, where a methionine is substituted for a leucine (rtL180M) (2). The development of resistance to lamivudine in the human immunodeficiency virus (HIV) has been shown to involve similar mutations at the YMDD motif, both in vitro and in vivo (10, 42, 49). The appearance of lamivudine-resistant HBV suggests that lamivudine therapy must be combined with other antiviral drugs to delay the emergence of resistant mutants in patients undergoing long-term therapy for chronic HBV infection. Thus, treatment of chronic HBV infection appears to involve challenges similar to those faced with treatment of HIV infection. Combination therapy, termed highly active antiretroviral therapy, is very effective in the treatment of HIV infection and prevents the emergence of drug-resistant variants. A similar combination antiviral therapy may also be effective in preventing drug resistance in chronically infected HBV patients.

The stable HBV-producing human hepatoblastoma cell line 2.2.15 (44), which carries HBV DNA stably integrated into the genome of HepG2 cells, has been used to successfully evaluate the effects of antiviral drugs on HBV replication (12, 16). Primary hepatocytes from ducks congenitally infected with duck HBV have also been used to screen drugs for activity against hepadnaviruses (9, 47). However, neither of these systems can be used to study the effects of antivirals on the lamivudine-resistant HBV. Currently, analysis of lamivudine-resistant HBV involves tedious transient transfection or recombinant polymerase systems (2, 38). A tetracycline-inducible stable cell line expressing HBV containing the rtM204V mutation has been reported; however, HBV with this mutation alone is not seen clinically and so is not ideal for use in screening antiviral compounds (26). Fu and Cheng have also described stable cell lines expressing lamivudine-resistant HBV for use as an antiviral screening system (18). However, this study did not include the clinically relevant rtM204I mutation associated with lamivudine resistance. Recently, an in vitro system using recombinant baculoviruses to deliver the HBV genome into cells has been shown to be useful for testing antiviral compounds (14). Although this system is more efficient than transient transfections, it requires the generation of recombinant baculoviruses and subsequent infection of cultured cells prior to the drug analysis. The purpose of this study was to produce stable cell lines expressing lamivudine-resistant HBV which could be used to easily screen drugs for activity against the lamivudine-resistant HBV. The cell lines produce the lamivudine-resistant viruses with either the rtM204I mutation or the rtL180M/M204V mutations.

MATERIALS AND METHODS

Plasmids and constructs.

An HBV genomic DNA construct was derived from pKS-HBV1 (43) which contains the HBV subtype adw. A 1.7-mer HBV genome was cloned into pcDNAI/Amp (Invitrogen, Carlsbad, Calif.) to generate the plasmid pCMV-HBVwt. An XhoI/SalI fragment containing a neomycin resistance gene from pMCIneoPolyA (Stratagene, La Jolla, Calif.) was cloned into the NheI site of pCMV-HBVwt to generate pCMV-HBVwt-neo. Standard cloning techniques as described by Sambrook et al. were used for these steps (40).

Lamivudine-resistant mutations were introduced into the HBV genome by using the Altered Sites in vitro mutagenesis kit (Promega, Madison, Wis.). A monomer HBV genome from SphI-digested pKSV-HBV1 was cloned into pAlter-1 (Promega) to generate pAlt-HBVwt. This construct was then used for site-directed mutagenesis to generate pAlt-HBV-rtL180M, pAlt-HBV-rtM204V, and pAlt-HBV-rtM204I. The corresponding mutations in the HBV reverse transcriptase domain of the polymerase are L180M, M204V, and M204I, respectively, based on the consensus nomenclature for HBV polymerase mutations (46). The mutations created in the overlapping surface antigen reading frame are S171S, I195M, and W196L, respectively. All mutations were confirmed by DNA sequencing. The mutagenic oligonucleotides used for the site-directed mutagenesis were as follows: L180M, 5′-AGTCCGTTTCTCATGGCTCAGTTTAC-3′; M204V, 5′-CAGCTATGTGGATGATGTGG-3′; and M204I, 5′-CAGCTATATTGATGATGTGG-3′.

Cloning of fragments from pAlt-HBV-rtL180M, pAlt-HBV-rtM204V, and pAlt-HBV-rtM204I into pCMV-HBVwt-neo generated the plasmids pCMV-HBV-rtL180M/M204V-neo and pCMV-HBV-rtM204I-neo. The nomenclature of the plasmids is such that the amino acid substitution and the position of the substitution are indicated. For example, pCMV-HBV-rtL180M/M204V contains a leucine-to-methionine change at amino acid 180 of the reverse transcriptase domain as well as a methionine-to-valine change at amino acid 204. Plasmids were linearized with PvuI, phenol-chloroform extracted, and precipitated with a 1/10 volume of 3 M ammonium acetate and 2.5 volumes of 95% ethanol prior to use in transfection experiments.

Cell culture, transfections, and clone selection.

HepG2 cells, a human hepatoma cell line (catalogue number HB8065; American Type Culture Collection), were cultured at 37°C and 5% CO₂ in minimal essential medium (ICN Biomedicals, Costa Masa, Calif.) supplemented with 10% heat-inactivated fetal bovine serum (Life Technologies, Gaithersburg, Md.), 2 mM l-glutamine, 100 μg of streptomycin/ml, and 50 IU of penicillin G/ml and buffered with sodium bicarbonate (complete medium). The cell line 2.2.15 was cultured in complete medium supplemented with 300 to 500 μg of G418 (Life Technologies)/ml as a selective agent.

HepG2 cells were seeded into six-well plates or 60-mm-diameter culture dishes at approximately 60 to 70% confluence 12 to 24 h prior to transfections. Linearized plasmid DNA was used to transfect the HepG2 cells by using either the calcium phosphate method (40) or Lipofectin reagent (Life Technologies) according to the manufacturer's instructions. Cell monolayers were trypsinized 24 h after transfection and transferred into T25 flasks or 100-mm-diameter culture dishes. Clones were selected by growth in complete medium containing 500 μg of G418/ml. Isolated clones were then cultured and expanded. Culture supernatant was subsequently assayed for production of viral antigens and extracellular viral (ECV) DNA to select for HBV-producing clones.

Characterization of individual clones.

Cell culture supernatant was assayed for both HBV surface antigen (HBsAg) and HBV e antigen (HBeAg) (Heprofile HBsAg and Heprofile HBeAg, ADI Diagnostics, Willowdale, Ontario, Canada). HBV ECV DNA in the cell culture supernatant was quantitated by using a PCR-based antigen capture system as previously described (22).

Integration of HBV DNA.

The integration of HBV genomic sequences into the host-cell DNA was examined by Southern transfer analysis of cellular genomic DNA. Genomic DNA was prepared by lysing cell monolayers with 10 mM Tris-HCl (pH 8), 1 mM EDTA, and 0.1% Sarkosyl and digesting them with 500 μg of proteinase K/ml at 42°C overnight. The sample was then deproteinated by extraction with an equal volume of phenol-chloroform (1:1) followed by a final extraction with chloroform alone. The DNA was precipitated by using 0.2 M NaCl and 2 volumes of 95% ethanol. For Southern analysis, approximately 7 μg of genomic DNA was digested with NsiI and the resulting fragments were separated on a 1% agarose gel. The gel was depurinated for 15 min in 0.25 M HCl and transferred by capillary action onto Hybond N+ membranes (Amersham, Buckinghamshire, England) by using 0.4 M NaOH. Membranes were prehybridized overnight in 5× SSPE (1× SSPE is 0.18 M NaCl, 10 mM NaH₂PO₄, and 1 mM EDTA [pH 7.7]), 2% sodium dodecyl sulfate (SDS), 1× Denhardt's solution, and 50 μg of herring sperm DNA/ml at 65°C. HBV sequences were detected by hybridization with a nick-translated, ³²P-labeled 3.2-kb HBV probe. Membranes were washed twice with 2× SSPE-0.1% SDS and twice with 0.2× SSPE-0.1% SDS. Washes were 15 min each and were done first at room temperature and then at 65°C with each solution.

Antiviral compounds.

The nucleoside analogues lamivudine, penciclovir [9-(4-hydroxy-3-hydroxymethyl-but-1-yl) guanine], abacavir {[1S,4R]-4-[2-amino-6-(cyclopropylamino)-9H-purin-9-yl]-2-cyclopentene-1-methanol}, and CDG (carbocyclic 2′-deoxyguanosine) were obtained from Glaxo Wellcome, Research Triangle Park, N.C. Stock solutions of drugs were stored at −80°C in dimethyl sulfoxide. Lamivudine was used at final concentrations of 3.2, 16, 80, 400, and 2,000 nM for wild-type virus (2.2.15 cells) and 3.2, 16, 80, 400, and 2,000 μM for the mutant viruses (rtM204I and rtL180M/M204V). The final concentrations of penciclovir used were 2.4, 12, 60, 300, and 1,500 μM for both wild-type and mutant viruses. The final concentrations of abacavir used were 0.32, 1.6, 8.0, 40, and 200 μM for the wild-type virus and 1.6, 8.0, 40, 200, and 1,000 μM for the mutant viruses. Two different series of final CDG concentrations were used: 3.2, 16, 80, 400, and 2,000 μM for both wild-type and mutant viruses (see Table 2) and 0.128, 0.64, 3.2, 16, and 80 μM for both wild-type and mutant viruses (see Table 2, footnote b).

TABLE 2.

Effect of nucleoside analogues on wild-type and lamivudine-resistant HBV^a

Cell line	IC50 (μM)
	Lamivudine	Penciclovir	Abacavir	CDGb
2.2.15 (wild type)	0.0072 ± 0.0029	45.9 ± 32	3.4 ± 1.4	0.0009 ± 0.0003
rtM204I	3.3 ± 2.2	394.8 ± 213	3.6 ± 0.7	<0.0032
rtL180M/M204V	23 ± 6.0	236 ± 93	4.1 ± 2.2	<0.0032

^aValues are given as means and standard deviations of results from between two and six independent experiments done in triplicate.

^bIn separate experiments done under slightly different conditions from those for which results are shown in Table 2 (see Materials and Methods), IC₅₀s of CDG of 0.002, 0.003, and 0.002 μM were obtained for the wild-type, rtM204I, and rtL180M/M204V viruses, respectively.

Drug assay.

The 2.2.15 cell line and the rtM204I and rtL180M/M204V cell lines (the cell lines producing lamivudine-resistant viruses with either the rtM204I mutation or the rtL180M/M204V mutations) were seeded in 96-well plates at a concentration of 75,000 cells/well in 150 μl of complete medium plus 300 to 500 μg of G418 and allowed to grow to confluence. Four days after seeding, medium was replaced with 150 μl of compete medium containing reduced concentrations of serum (2%). Serum-reduced medium was used for the remainder of the drug study. One week after seeding, on day 1, medium was replaced with either medium alone or medium containing the desired concentration of antiviral compound. On days 3 and 5, cells were fed fresh medium alone or medium containing the antiviral compound. On day 7, medium was harvested from the cells and HBV ECV DNA present in the culture supernatant was quantitated as previously described (22). Medium samples were also analyzed for the production of HBsAg and HBeAg.

RESULTS

Production and analysis of stable transformants.

Transfection of HepG2 cells with pCMV-HBV-rtL180M/M204V-neo or pCMV-HBV-rtM204I-neo, depicted in Fig. 1, resulted in the establishment of individual stable clones. Potential HBV-producing clones were first assayed for the ability to produce the viral antigens HBsAg and HBeAg by using an enzyme-linked immunosorbent assay-based system. Antigen-producing clones were then analyzed for production of ECV by using an antigen capture assay (22).

FIG. 1.

Map of pCMV-HBV-neo construct used to generate stable cell lines rtL180M/M204V and rtM204I. (A) Schematic diagram of the replication-competent HBV construct shows the cytomegalovirus immediate early (CMV-IE) promoter, HBV sequences, NsiI recognition site, ampicillin resistance (amp R) marker and neomycin resistance (neomycin R) marker. (B) Diagram depicting organization of open reading frames in HBV sequences. CMV, cytomegalovirus; Pol, polymerase; X, X gene; S, surface. (C) Nucleotide sequence of wild-type and mutant viruses. Altered nucleotides are depicted in bold.

Table 1 shows the levels of viral antigen and of ECV HBV DNA produced by the established cell lines. The control used in this study was the 2.2.15 cell line, which stably expresses wild-type HBV. HBsAg and HBeAg were produced at comparable levels in all cell lines. The 2.2.15 cells produced an average of 40 pg of HBV DNA/ml over 48 h in the culture supernatants of untreated (no drug) wells 14 days postplating. The levels of HBV DNA present in the culture supernatants of the cell lines expressing the mutant viruses were consistently lower than that produced by the 2.2.15 cells. The rtM204I cell line produced an average of 3.5 pg of viral DNA/ml. The rtL180M/M204V cell line produced an average of 6.5 pg of viral DNA/ml. Analysis of the intracellular viral replicative intermediates indicated that active viral replication was occurring in the cell lines, as indicated by the presence of the relaxed circular, linear, and single-stranded forms of the viral DNA (data not shown). Again, the levels were consistently lower for the mutant viruses than for wild-type HBV.

TABLE 1.

Production of HBsAg, HBeAg, and ECV DNA^a

Cell line	HBsAg (ng/ml)	HBeAg (ng/ml)	HBV DNA (pg/ml)
2.2.15 (wild type)	82.0 ± 28.6	116.3 ± 51.7	40.6 ± 39
rtM204I	69.6 ± 26.4	79.2 ± 30.0	3.5 ± 0.5
rtL180M/M204V	40.3 ± 15.5	126.9 ± 34.7	6.5 ± 2.5

^aValues are expressed as means and standard deviations of results from a minimum of two independent experiments done in triplicate. Values represent accumulation in culture media from confluent cell monolayers over 48 h.

Stable integration of the HBV genome in each of the cell lines was confirmed by isolating genomic DNA, digesting it with NsiI, and analyzing the resulting fragments by Southern hybridization. There is a single recognition site for NsiI present within the HBV sequence but absent from the vector sequence. Hybridization with a probe consisting of only HBV sequences should result in two bands for every copy of integrated HBV sequences, assuming the copies are not in tandem. Figure 2 shows the results of Southern analysis of genomic DNA of the stable cell lines. Both the rtM204I and rtL180M/M204V cell lines, shown in lanes 2 and 3, respectively, showed more than two bands of HBV-specific DNA sequences. They both also contained intense bands of 11 and 9.5 kbp, respectively. This is approximately equivalent to the size of the pCMV-HBV-neo constructs, which likely indicates that both cell lines contain multiple integrated copies of the plasmid arranged in tandem. The experiment has been repeated with other restriction enzymes that cut in either vector or HBV sequences (data not shown). The results indicate the presence of both head-to-head and head-to-tail repeats of the HBV construct in the cellular genome of the stable cell lines. Both cell lines contain a band of approximately 3.9 kbp (the band is evident in the rtM204I cell line upon a longer exposure; data not shown). This band was also seen in other stable HBV cell lines that were not used in this study (data not shown). It likely represents the relaxed circular form of the viral DNA, which also runs at 3.9 kbp, and so was not considered when examining the integration of HBV sequences. As expected, HepG2 cellular DNA, shown in lane 1, contains no HBV-specific sequences.

FIG. 2.

Integration of HBV sequences into genomic DNA of HepG2, rtM204I, and rtL180M/M204V cell lines. Genomic DNA from confluent monolayers was isolated, digested with NsiI, and analyzed by Southern blotting for the presence of HBV sequences.

Analysis of antiviral activity.

The cell lines producing the rtM204I and rtL180M/M204V viruses were used to screen for antiviral activity of lamivudine and several other antiviral agents to determine if this system would be appropriate for antiviral drug screening. Lamivudine was used to confirm that the mutations introduced conferred lamivudine resistance. Purine-based analogues, particularly guanosine analogues, have been shown to be more effective anti-HBV agents, possibly because they have the potential to target both the protein priming and DNA synthesis steps (21). As protein priming and reverse transcription are separate biochemical reactions involving different regions of the polymerase protein (50), it is possible that the activity of purine analogues may be less affected by mutations in the YMDD motif. The purine analogues penciclovir, abacavir, and CDG were used to determine whether the lamivudine-resistant mutations also conferred resistance to purine analogues. Each cell line was treated with the antiviral compounds for 1 week. Culture supernatant was then removed, and the HBV ECV DNA was quantitated. The 2.2.15 cell line was used as the wild-type HBV control.

Table 2 shows the effects of the various drugs on wild-type and mutant virus production. A 50% inhibitory concentration (IC₅₀) of lamivudine of 7.2 nM was obtained for the wild-type virus. IC₅₀s of lamivudine of 3.3 and 23 μM were obtained for the rtM204I and rtL180M/M204V viruses, respectively. These values are comparable with those found in previous studies and confirmed that viruses containing these mutations showed a marked insensitivity to lamivudine (2, 25, 39). An IC₅₀ of penciclovir of 45.9 μM was achieved for the wild-type virus and IC₅₀s of 394.8 and 236.7 μM were obtained for the rtM204I and rtL180M/M204V viruses, respectively. The high standard deviation values obtained may be due to the fact that the activity of penciclovir is affected by the growth rate of the cells, which likely differs slightly between experiments. The IC₅₀s of abacavir were 3.4 μM for the wild-type virus and 3.6 and 4.1 μM for the rtM204I and rtL180M/M204V viruses, respectively. Analysis of CDG resulted in an IC₅₀ of 0.9 nM for the wild-type virus and less than 3.2 nM for both the rtM204I and rtL180M/M204V viruses. To obtain more accurate IC₅₀s, lower concentrations of CDG were used in a second experiment (Table 2, footnote b). The IC₅₀s were 2, 2.98, and 2.4 nM for the wild-type, rtM204I, and rtL180M/M204V viruses, respectively. IC₅₀s of all drugs were well below previously published toxicity levels for the type of cells used in this study (13, 24, 26).

DISCUSSION

Transfection of HepG2 cells with pCMV-HBV-rtM204I-neo and pCMV-HBV-rtL180M/M204V-neo gave rise, respectively, to the stable cell lines rtM204I and rtL180M/M204V. These cell lines produced both intracellular and extracellular virus, although generally at lower concentrations than the control 2.2.15 cell line did.

The quantitation of ECV in this study involved the binding of virus to anti-HBsAg-coated plates (22). However, the lower levels of virus in cell lines expressing the mutant viruses were unlikely to be due to differences in the binding properties of the mutant viruses, as Melegari et al. (35) showed that the mutations introduced into the overlapping HBsAg reading frame did not affect the binding properties of HBsAg. This is further supported by the lower level of viral replicative intermediates found upon Southern analysis of intracellular virus.

The impaired polymerase activity associated with the mutant viruses may, at least partially, account for the lower virus production in cell lines producing the mutant viruses (25, 31, 35, 39). HBV containing the rtM204I or rtM204V mutation alone replicates at a significantly lower rate than the wild-type virus. Virus containing the double rtL180M/M204I or rtL180M/M204V mutations replicates at a higher rate than virus containing the single mutation alone but not as efficiently as wild-type virus (35). This may also explain why the levels of virus production by the rtL180M/M204V cell line were consistently higher than those by the rtM204I cell line.

Another factor contributing to the differences in levels of virus produced by the cell lines may be related to the nature of the integration of the HBV sequences. Both the number of integrated copies of the HBV genome and the site of integration may influence virus production. Southern analysis showed that each clone likely has multiple tandem copies of the HBV sequences integrated into the host DNA. It also showed that for each clone, the integration positions were unique. Because different regions of the chromosome can have various influences on the transcription activity of the integrated HBV genome, the integration site may affect the level of virus production. Therefore, individual clones cannot be directly compared for levels of virus production, nor can differences in their levels of virus production be considered significant. With respect to the drug analysis, the levels of virus produced by each cell line are internally controlled and have no effect on the results because the IC₅₀s are calculated based on virus production from untreated cells of the same cell line.

A system of classifying the degree of change in the susceptibility of a virus to an antiviral agent has been previously described (23) and was used in the present study. A threefold or smaller decrease in sensitivity is classified as no change, a decrease between three- and 10-fold is classified as reduced sensitivity, and virus with a decrease of 10-fold or greater is classified as resistant (23).

Virus containing the mutations rtM204I and rtL180M/M204V exhibited resistance to lamivudine, consistent with previous observations (2, 14, 17, 26, 38). A 450-fold increase in IC₅₀, compared to that for the wild-type virus, was seen for virus containing the rtM204I mutation. A 3,000-fold increase in IC₅₀ was seen for virus containing the rtL180M/M204V mutations.

Ladner et al. used a construct containing a cDNA copy of the pregenomic RNA of an HBV genome carrying the single rtM204V mutation to produce a stable tetracycline-inducible cell line (26). The shifts in IC₅₀s observed by Ladner et al. were smaller than those seen in the present study. It has been shown that lamivudine resistance conferred by the single rtM204V mutation is increased when the mutation is combined with the upstream rtL180M mutation, from a 186-fold increase to a >10,000-fold increase over the IC₅₀ for wild-type virus (2). Furthermore, the rtM204V mutation alone is rarely seen in clinical isolates of lamivudine-resistant HBV (2). Screening for antivirals still active against lamivudine-resistant HBV should therefore be done using the more clinically relevant rtM204I or rtL180M/M204V mutations.

The presence of the mutations also caused a decreased sensitivity to penciclovir. A ninefold increase in IC₅₀ was seen for virus containing the rtM204I mutation. A fivefold increase in IC₅₀ was seen for virus containing the rtL180M/M204V mutations. This is consistent with the findings of a previous report in which rtL180M/M204V virus was shown to have reduced sensitivity to penciclovir (18). The rtL180M mutation has been associated with penciclovir resistance (35), and the lack of significant difference in the IC₅₀s for the rtM204I and rtL180M/M204V viruses is surprising. The rtL180M/M204V virus might be expected to exhibit a greater decrease in sensitivity to penciclovir than the rtM204I virus. Similar results were obtained in the study by Delaney et al. in which virus expressing the single rtM204I mutation also exhibited a higher resistance to penciclovir than virus expressing either the rtL180M mutation or the rt180M/M204V mutations (14). This observation suggests that the rtL180M mutation alone may be insufficient to cause penciclovir resistance. Indeed, the rtL180M mutation is just one of several mutations associated with penciclovir resistance (53). Despite the fact that the mutant viruses in the present study show only a decreased sensitivity to penciclovir in vitro, patients with lamivudine-resistant virus show poor response to penciclovir treatment (20; P. Shields, R. Ling, T. Harrison, E. Boxall, E. Elias, and D. Mutimer, Hepatology, abstr. 260A, 1997). In addition, penciclovir is unlikely to be a good choice for combination therapy with lamivudine since resistance to both drugs is associated with a common mutation.

There was no significant shift in the IC₅₀s of either abacavir or CDG for virus expressing either the rtM204I mutation or the rtL180M/M204V mutations compared to those for the wild-type virus. HBV carrying the rtM204V mutation alone has also previously been shown to remain sensitive to CDG (26). This indicates that HBV resistant to lamivudine would likely be sensitive to treatment with either drug. Both drugs are therefore good candidates to be used in combination therapy with lamivudine. Abacavir and CDG are purine-based derivatives. Purine-based analogues have been shown to be more effective inhibitors of hepadnaviral replication than pyrimidine-based analogues (29, 47). Guanosine-based analogues may be more effective as anti-HBV agents as they have the potential to target two separate processes in HBV replication. In addition to acting as a DNA chain terminator, these drugs can also target the priming step of reverse transcription because the first nucleotide covalently bound to the primer protein is a guanosine residue (57). This may partially explain why changes in the YMDD motif do not affect the anti-HBV activity of abacavir or CDG. It is interesting that HIV resistance to abacavir has been observed and involves the equivalent M184V mutation that is responsible for resistance to lamivudine. It is not clear why HBV with the corresponding mutation remains sensitive to abacavir in this system. It is possible that additional mutations not yet recognized in HIV are required for the resistance. Alternatively, it is possible that additional mutations not found in HIV are required for similar resistance to abacavir by HBV. An example of this is the upstream rtL180M mutation accompanying the rtM204V mutation associated with lamivudine resistance in HBV but not in HIV.

A recent study using recombinant human HBV polymerase demonstrated that the mutations associated with both penciclovir and lamivudine resistance do not confer resistance to the purine-based analogue adefovir (52). Lamivudine-resistant HBV has also been shown to remain sensitive to lobucavir (38). Different resistance profiles for the rtL180M/M204V and rtM204V viruses have been noted for L-FMAU. Ying et al. (54a) have shown that HBV with the rtM204V mutation remains sensitive to L-FMAU, while a separate study has shown that the rtL180M/M204V virus is cross-resistant (18). This discrepancy emphasizes the importance of using clinically relevant mutations when screening antivirals.

The production of stable cell lines that express lamivudine-resistant HBV eliminates the need for tedious transfections before each drug-screening experiment. Screening for activity against lamivudine-resistant viruses can now be done with relative ease. There are, however, a few drawbacks to this system. Long-term drug analysis to assess the durability of the compound's antiviral effect is difficult due to the fact that maximal virus production is obtained from stationary cells. This requires that the cells be cultured in a confluent state with reduced serum, which limits the length of the experiment to approximately 14 days. However, this drug assay system was designed primarily for use in preliminary screening. Long-term studies to assess the durability of antiviral effects are best done with the woodchuck or duck animal models. Analysis of drug sensitivity is also limited to the genotype used to generate the stable cell lines. As well, because virus production is driven from an integrated template, it is highly unlikely that this system can be used to select for HBV variants resistant to antiviral compounds. To date, there have been no reports of lamivudine-resistant HBV being selected in vitro. In addition, resistant HBV variants take months to emerge in vivo, and while drug resistance does emerge in vivo after long-term treatment in the woodchuck model (54-56), it has never been demonstrated in the duck model. The emergence of drug-resistant HBV variants appears to be a relatively uncommon event requiring long-term therapy.

The results of this study showed that although viruses containing the rtM204I mutation or the rtL180M/M204V mutations are resistant to lamivudine, they are still sensitive to abacavir and CDG. With the recent licensing of lamivudine for treatment of HBV infection, widespread use of the drug has resulted in significant lamivudine resistance in HBV-infected individuals. It is imperative that additional antivirals active against the YMDD mutants be identified, as there are potential benefits of combination therapy to prevent the emergence of lamivudine-resistant HBV isolates in patients undergoing monotherapy.