INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Lamivudine is a potent inhibitor of the hepatitis B virus (HBV) DNA polymerase and can quickly reduce liver injury in HBV carriers (34), apparently by suppressing virus replication. However, the majority of carriers are not cured by lamivudine, and drug-resistant virus emerges in most, often in association with an increase in virus titers towards pretreatment levels (3, 5, 9, 18, 25, 33, 34, 38). Difficulty in completely eliminating HBV stems directly from the mechanism by which this virus reproduces. When a hepadnavirus infects a cell, the incoming viral genome matures into a single covalently closed circular DNA (cccDNA). This cccDNA, located in the nucleus, serves as the template for the transcription of the larger-than-unit-length pregenomic RNA and of the subviral RNA species (8). A virus-encoded reverse transcriptase converts the pregenomic RNA into a partially double-stranded DNA genome in a series of reactions that take place inside virus nucleocapsids (36, 41), which are found in the cytoplasm of the infected cell. The virus nucleocapsids are subsequently enveloped and, after processing of the envelope glycoproteins (7, 30), are released from the cell as mature virions. In a pathway that is negatively regulated by the viral envelope proteins, a fraction of the virus nucleocapsids are transported to the cell nucleus to produce additional copies of cccDNA (37). Estimates of cccDNA copy number range from 5 to 50 or more per hepatocyte (21, 22, 31).

Because the infected state of a hepatocyte is defined by the presence of cccDNA, its stability is important in any consideration of antiviral therapies employing inhibitors of viral DNA synthesis. Cell culture studies with primary hepatocytes, which do not divide, indicated a high degree of cccDNA stability ((32); however, see reference 12). This stability may be a major reason why infections are harder to eliminate by polymerase inhibitors in "healthy" carriers with a slower rate of hepatocyte death and compensatory regeneration than in individuals with active hepatitis. However, it is not known whether cccDNA is lost during mitosis, as proposed, for example, for the Epstein-Barr virus (EBV) plasmid in the absence of EBNA1 (27), or whether it is distributed to daughter cells along with host chromosomes. Loss during mitosis would lead to a rate of cccDNA decline, in the absence of viral DNA synthesis, that would equal approximately twice the rate of infected-cell death. That is, cccDNA would be lost through cell death as well as through division of an infected cell that divided to replace the cell that died. In contrast, retention of cccDNA through the mitotic event would lead to a rate of loss equal to the rate of infected-cell death. However, in the latter case, because of the initially high cccDNA copy number, the fraction of infected hepatocytes would not begin to decline detectably until the average cccDNA copy in the liver dropped to ~1 to 2 (provided that all cccDNA molecules have the potential to be transcriptionally active). Thus, a long period of treatment, in comparison to the average life time of the infected hepatocyte, would be needed in order to facilitate complete cccDNA loss from an infected liver. Current estimates place the half-life of infected hepatocytes in HBV carriers at between 2 weeks and many months, depending on the severity of hepatitis and, thus, the extent of hepatocyte destruction by the immune system (34).

The present study was carried out in order to determine how cccDNA levels changed in the hepatocyte population during antiviral therapy. For our experiments, we used woodchucks chronically infected with woodchuck hepatitis virus (WHV). As an antiviral agent, we employed L-FMAU [1-(2-fluoro-5-methyl--L-arabinofuranosyl) uracil] (11, 34a), which is highly effective against WHV replication in vivo.

We found that L-FMAU produced up to 95 to 99% loss of cccDNA after 30 weeks. The average half-life of the cccDNA over the course of the in vivo experiments that would produce the loss observed after 30 weeks was ~33 to 50 days, similar to the half-life of hepatocytes estimated from proliferating-cell nuclear antigen (PCNA) staining indices of liver biopsy specimens collected during therapy. Immunoperoxidase and in situ hybridization of liver tissue sections revealed that the fractional loss of infected hepatocytes was ~5 to 10-fold less than the loss of cccDNA, consistent with the hypothesis that cccDNA is not lost but distributed to daughter hepatocytes during mitosis. Evidence in support of this hypothesis was also recently obtained using primary cultures of woodchuck hepatocytes infected with WHV and induced to proliferate by addition of epidermal growth factor (13).

After about 36 weeks of L-FMAU administration, virus titers in the serum, which had dropped below the limits of detection of our assays (~10⁶ per ml), began to rise. Prior to this, mutant WHVs were detected with a distribution of nucleoside changes in the active site of the DNA polymerase, some of which were identical to those found after long-term lamivudine therapy (29, 45). The delayed spread of some of these mutants into apparently uninfected hepatocytes suggested that they may have a slower in vivo growth rate in the presence of L-FMAU than wild-type WHV in the absence of drug.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Woodchucks. Adult uninfected and WHV-infected woodchucks (Marmota monax), trapped in New York state and in Delaware, respectively, were purchased from Northeastern Wildlife (South Plymouth, N.Y.) and housed in the laboratory animal facility of the Fox Chase Cancer Center (FCCC). Experiments with woodchucks were reviewed and approved by the FCCC Institutional Animal Care and Use Committee. L-FMAU (10 mg per kg of body weight) was administered orally once a day (between 6 and 10 a.m.) in Dyets Woodchuck Control Diet (Dyets, Inc., Bethlehem, Pa.) at a concentration of 10 mg/ml. A total of 11 chronically infected woodchucks were used in this study. Four (343, 344, 345, and 346) had not been previously treated. Another seven, used to evaluate cross resistance between lamivudine and L-FMAU, had received a 14-month dose of lamivudine (44), which had ended 5 months prior to the treatment with either L-FMAU or placebo (Dyets Woodchuck Control Diet without L-FMAU). The sera of this latter group contained WHV genetic variants characteristic of the resistance to lamivudine that had developed in these woodchucks. Serum collection and liver biopsy during the course of these studies were carried out as previously described (22).

L-FMAU treatment of woodchuck primary hepatocytes. Primary woodchuck hepatocyte cultures were prepared and maintained at 37°C on 60-mm tissue cultures dishes coated with rat tail collagen in a serum-free L15 medium supplemented with insulin, hydrocortisone, and phosphonoacetic acid, as previously described (2, 32). Culture fluids (3 ml) were changed daily. WHV infection of hepatocytes was established by adding 50 µl of serum (ca. 10⁸ virions) from a chronically infected woodchuck at 2 days postseeding. Starting 4 days after WHV infection, L-FMAU was present in culture medium at a 10 µM concentration. Hepatocyte monolayers were harvested at 4, 8, 16, 24, 32, and 40 days after WHV infection and stored at 80°C for subsequent extraction of viral nucleic acids.

Analysis of WHV DNA. Total DNA and cccDNA isolations from primary hepatocyte cultures and subsequent analyses were performed as described previously (32). Isolation of total DNA and cccDNA from woodchuck liver biopsy specimens was also carried out following a previously described procedure (21). Briefly, approximately 0.05 to 0.1 g of liver tissue was disrupted with a loose-fitting Dounce homogenizer in 1.5 ml of 10 mM Tris-HCl (pH 7.5)-10 mM EDTA. The number of nuclei in each homogenate was determined following staining of aliquots with ethidium bromide and counting in a hemacytometer under fluorescent illumination. Each homogenate was divided into two aliquots, one for extraction of non-protein-bound cccDNA, and the other for total DNA isolation (21, 37). Either 10 µg of total DNA or the cccDNA extracted from 10⁶ liver cells was subjected to electrophoresis through 1.5% agarose gels. (The genome of the woodchuck was assumed to weigh 5 pg per diploid cell.) Known amounts of full-length, cloned viral DNA were electrophoresed as a control. The DNAs were then transferred to a nitrocellulose filter (Schleicher & Schuell, Keene, N.H.) following partial depurination and fragmentation with alkali, similar to the procedure described by Wahl et al. (40), and hybridized with a ³²P-labeled WHV DNA probe representing the complete genome. Radioactive signals were quantified using a Fuji phosphoimager (Fuji Corporation, Tokyo, Japan). The amount of viral DNA present in the liver samples was estimated by comparison to the hybridization control and, for the purpose of copy number estimations, is presented as equivalents of full-length, double-stranded WHV DNA.

Serum virus titers. To measure WHV titers in serum, 50 µl of serum was centrifuged through a 4-ml, 10 to 20% sucrose step gradient containing 150 mM NaCl and 20 mM Tris-HCl (pH 7.5) for 3 h at 50,000 rpm in a Beckman SW60 rotor. The virus pellet was then resuspended in 50 µl of 100 mM NaCl-10 mM Tris-HCl (pH 7.5)-10 mM EDTA-0.1% sodium dodecyl sulfate (SDS)-2 mg of pronase per ml and incubated at 37°C for 1 h. The samples were then subjected to 1.5% agarose gel electrophoresis, transferred to nitrocellulose membranes, hybridized with ³²P-labeled WHV probe, and quantified as described above.

Detection of WHV core antigen- and nucleic acid-positive hepatocytes, PCNA-positive hepatocytes, and infiltrates of CD3-positive cells in liver tissue sections. Biopsy specimens for immunoperoxidase assays and in situ hybridizations were fixed in a 3:1 mixture of ethanol and glacial acetic acid for 20 min at 4°C and then overnight at 4°C in 100% ethanol, followed by dehydration and embedding in paraffin wax. Immunoperoxidase staining for WHV core antigen, PCNA, and CD3-positive leukocytes and in situ hybridization for detection of WHV nucleic acids were carried out as previously described (17, 29).

Genotyping of WHV DNA. Direct sequencing of PCR products and of cloned PCR products, as well as determination of restriction site polymorphisms, was used for genotyping (45). Virions were collected from 50 µl of woodchuck serum by centrifugation, and viral DNA was released by digestion with a mixture of SDS and pronase, as described above. The pronase-treated mixture was then extracted twice with phenol-chloroform-isoamyl alcohol (25:24:1), and viral DNA was precipitated by the addition of 2 volumes of ethanol. cccDNA for genotyping was extracted from liver tissue as described above, subsequently purified using an alkaline extraction protocol (42), and digested with EcoRI. PCR amplification of a region spanning the active site of the viral DNA polymerase was carried out as described (45). The primers were 5'-AGATTGGTGGTGCACTTCTCTCAGG-3' (WHV nucleotides 385 to 408) and 5'-CCACGGAATTGTCAGTGCCCAACC-3' (nucleotides 1474 to 1451), using the numbering system of Kodama et al. (23). The PCR products were purified using a QIAquick kit (Qiagen Inc., Hilden, Germany) according to the manufacturer's instructions, and both strands were sequenced. Selected products were also cloned prior to sequence analysis.

Transfection of HepG2 cells. For transfection, we used plasmids containing WHV DNA in which transcription of the viral pregenome is directed by a cytomegalovirus immediate-early promoter. The wild-type construct and variants of the wild type containing type I, II, or III mutations of the polymerase active site were reported previously (45). Twenty hours before transfection, HepG2 cells were seeded at 3 × 10⁶ cells per 6-cm tissue culture dish in F-12 minimal essential medium-containing 10% fetal calf serum. L-FMAU at the indicated concentrations was added at this time and maintained at the same concentration in the medium thereafter. Transfection was carried out 1 to 2 days postseeding using a CaPO₄ coprecipitation protocol (39). At 4 days posttransfection, the cells were harvested, and WHV core DNA was extracted (45). One quarter of each sample was subjected to Southern blot analysis, as described above.

SDH assay. Sorbital dehydrogenase (SDH) in serum was determined by Anilytics, Inc., Gaithersburg, Md. Concentrations are expressed in international units per liter.

Histopathology. Histopathology was done on formalin-fixed liver sections stained with hematoxylin and eosin. Liver injury was graded on a subjective scale. Inflammation was a major determinant, with the other factors, hepatocyte necrosis, vacuolization, biliary hyperplasia, Kupffer cell activation, and variation in hepatocyte nuclear size, influencing the degree of injury. Scoring was as follows: 0, no evidence of liver injury; ±, scant numbers of inflammatory cells in portal tracts with minimal inflammation in the liver parenchyma; 1, mild accumulations of lymphocytes in portal areas and focal accumulations in the parenchyma, with individual hepatocyte necrosis, Kupffer cell aggregates, and variation in hepatocyte nuclear size also present; 2, moderate inflammation of portal tracts with sites of extension into the terminal distributing vasculature; and 3, moderate to extensive inflammatory infiltrate extending from the portal tract into adjacent parenchyma or portal inflammation accompanied by moderate to extensive parenchymal inflammation.

Computational model of redistribution and loss of cccDNA during cell division when viral DNA synthesis is inhibited. For the purpose of computation, the liver was divided into compartments, each containing cells with a particular number of cccDNA copies. The model was initialized to contain a particular distribution of copies; for example, a truncated Poisson distribution with a mean of 30 copies and truncation limits of 20 and 40 copies can be specified. In the truncated Poisson, 100% of the cccDNA is initially present at a copy number of 20 to 40. Liver size was held constant by adjusting the cell replication rate to compensate for the cell killing rate. Each replicating cell binomially and symmetrically distributes its cccDNA to its two daughters. The sequence of events is: cells are killed, cells divide to repopulate the liver, and cccDNA is redistributed among replicated daughters. In the model, a specified fraction of cccDNA may also be lost during this cell division. In that case, this fraction is removed prior to redistribution of the remaining cccDNA copies. At completion, the program supplies the resulting distribution of cccDNA copies. Other program outputs included graphs of the fraction of cells infected over time, fraction of cccDNA left over time, total cell deaths over time, and the daily cell death rate. (Scripts for the MacIntosh are available upon request from S. Litwin at S_Litwin{at}fccc.edu.)

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

L-FMAU treatment did not facilitate loss of cccDNA from WHV-infected primary hepatocyte cultures. L-FMAU acts on the hepadnavirus reverse transcriptase to inhibit viral DNA synthesis (1). By analogy to EBV, this may occur through noncompetitive binding to the viral polymerase rather than through incorporation into the growing DNA chains (24). The objective of the present study was to characterize the correlation between the decline in infected hepatocytes and cccDNA loss during L-FMAU therapy of chronically infected woodchucks. A pilot study was first carried out to determine if L-FMAU had any unexpected effects on cccDNA stability in primary hepatocyte cultures, which are composed of nondividing hepatocytes. In a previous study with the antiviral agent lamivudine, no loss of cccDNA was observed over that explainable by cell loss from the cultures (32).

View larger version (66K): [in this window] [in a new window]   FIG. 1.   L-FMAU does not induce loss of cccDNA from primary cultures of WHV-infected woodchuck hepatocytes. Hepatocyte cultures were infected and treated with L-FMAU, beginning 4 days postinfection, as described in Materials and Methods. Total DNA (A) and cccDNA (B) were extracted from the cells at the indicated times and subjected to Southern blot analysis. Each lane contained one-quarter of the DNA extracted from a 6-cm-diameter tissue culture dish. RC, relaxed circular 3.3-kbp DNA.

Oral administration of L-FMAU inhibited WHV replication and induced a progressive loss of cccDNA in chronically infected woodchucks. Chu et al. (10) reported that treatment of chronic WHV carriers with L-FMAU at a daily dose of 10 mg per kg of body weight produced a >1,000-fold reduction in serum virus titers within 2 weeks. This dose was used to study the effects of L-FMAU therapy on cccDNA levels in the liver. In our study, four woodchucks chronically infected with WHV (343, 344, 345, and 346) were each treated with L-FMAU by daily oral administration. (Woodchuck 344 died accidentally after 4 weeks of treatment.)

View larger version (21K): [in this window] [in a new window]   FIG. 2.   Suppression and rebound of viremia during long-term treatment with L-FMAU. Virus titers during treatment with L-FMAU were quantified by Southern blot assays for virus particles in the serum, as described in Materials and Methods. A titer of 106 DNA equivalents per ml represents the lower limit of detection of our assays. Serum samples for genotype analysis were collected 9 weeks before and 6, 19, 24, 27, 33, 38, 42, 45, 51, 53, and 58 weeks after initiation of therapy. The changes in the sequences of the B and C domains of the viral DNA polymerase which define the mutant designations were determined by direct sequencing of PCR products and are illustrated in Fig. 5. wt, wild type.

View larger version (35K): [in this window] [in a new window]   FIG. 3.   Loss of replicating and cccDNA forms of viral DNA during antiviral therapy. Viral DNA was extracted from liver biopsy specimens for analysis of total and cccDNA forms of viral DNA by Southern blot analysis on 1.5% agarose gels, as described in Materials and Methods. Each lane contained either 10 µg of total DNA (A) or a cccDNA-enriched fraction recovered from 106 liver cells (based on nuclear counts in the liver lysates) (B). The average copy numbers of viral DNA per hepatocyte, shown at the bottom of panels A and B, were calculated as equivalents of full-length, double-stranded viral DNA and assume that the liver comprises 70% hepatocytes. The genome of the woodchuck was assumed to weigh 5 pg per diploid cell. The percent loss of cccDNA during treatment is summarized in panel C. The arrows indicate the losses of cccDNA after 30 weeks that would occur via a first-order decay with the indicated half-lives. Woodchuck 344 (wc344) died after 4 weeks of therapy.

cccDNA distribution among infected hepatocytes after antiviral therapy. Infected hepatocytes contained, on average, ~20 to 60 copies of cccDNA prior to L-FMAU administration (Fig. 3B). Once viral DNA synthesis is inhibited and the mature replication intermediates are depleted from the cytoplasm, no new cccDNA synthesis should occur. Thus, as cccDNA is lost through cell death, the cccDNA copy number in infected cells should decline through dilution as infected cells divide, provided that the cccDNA can survive through mitosis. The predicted effect on cccDNA distribution, assuming a Poisson distribution of between 20 and 60 copies per hepatocyte and a cccDNA half-life in the liver of 33 to 50 days, can be calculated by computational methods (Materials and Methods). After 30 weeks, 11 to 46% of the hepatocytes should still be infected, even though only 1.2 to 5.4% of the cccDNA remained. Most hepatocytes that remained infected would have fewer than 10 copies of cccDNA.

View larger version (254K): [in this window] [in a new window]   FIG. 4.   Progressive loss of hepatocytes with detectable levels of WHV core antigen and nucleic acids during antiviral therapy. All liver biopsy specimens were assayed for the fraction of hepatocytes with detectable levels of WHV core antigen and nucleic acids. The results of these analyses for serial liver biopsies from woodchuck 346 are illustrated. Core antigen was detected by an immunoperoxidase assay (A), and viral nucleic acids were detected by in situ hybridization (B). The percentage of positive hepatocytes detected with each assay is summarized in Table 1. Magnification: (A) ×200; (B) ×100. Uninfected liver tissue is shown as a negative control in panel A. In panel B, tissue hybridized with a plasmid-specific probe (no WHV insert) served as the negative control.

L-FMAU-resistant WHV emerged after prolonged treatment. Treatment of chronically infected woodchucks with lamivudine, a nucleoside analog, leads to a temporary suppression of virus replication. However, within a year, drug-resistant strains of WHV emerge, and virus production increases towards pretreatment levels (29, 45). A similar rebound in viremia was found after 8 to 9 months of treatment with L-FMAU, as summarized in Fig. 2. To determine if the increase in virus titers was associated with the emergence of virus with a mutation(s) in the active sites of the polymerase, direct sequencing of PCR products spanning this region was carried out.

View larger version (11K): [in this window] [in a new window]   FIG. 5.   WHV active-site mutations.

View larger version (43K): [in this window] [in a new window]   FIG. 6.   Evidence that mutations that confer lamivudine-resistant DNA synthesis on a laboratory strain of WHV also confer resistance to L-FMAU. Resistance of viral DNA synthesis to L-FMAU was assayed in transfected HepG2 cells, as described in Materials and Methods. The predicted amino acid sequences of the DNA polymerase B and C domains of the type I, II, and III mutations are illustrated in Fig. 5. Evidence that these mutations confer lamivudine resistance has already been published (45).

Evidence for cross-resistance of L-FMAU in woodchucks previously treated with lamivudine. Although the sequence analysis was limited to a small region of the WHV genome, the above results suggested that L-FMAU and lamivudine would show cross-resistance in vivo. To further test this idea, an in vivo study was carried out employing seven woodchucks, six of which had developed resistance to lamivudine during a 14-month trial (44). With woodchuck 326, only a small rise in virus titers was noted toward the end of lamivudine treatment. Six months after the end of lamivudine therapy, four of these woodchucks were administered L-FMAU and three received placebo. The WHV genotypes of each woodchuck in the 5 months between lamivudine and L-FMAU therapy are described in Fig. 5 and 7. All but woodchuck 326 had at least one prevalent virus population with mutations in the Pol active site that had previously been associated with lamivudine resistance in woodchucks (45). The consequences of the type IV to VII mutations on virus DNA replication are still under investigation.

View larger version (22K): [in this window] [in a new window]   FIG. 7.   Evidence that in vivo infections that are resistant to lamivudine may also be resistant to L-FMAU. WHV-infected woodchucks in which virus titers had rebounded during a previous treatment with lamivudine received a 7-week treatment with L-FMAU (A and B); control animals received the Dyets formula as a placebo (C and D). Virus titers and genotypes in the interval between the end of the lamivudine therapy and the beginning of L-FMAU or placebo administration are shown in panels A and C. In panels A and C, the first serum sample was collected on the day that lamivudine treatment was discontinued. Sequence determinations from direct sequencing of PCR products were validated by cloning and sequencing of PCR products from the zero time points for woodchucks (wc) 326, 331, 336, 342, 335, and 338, as well as the 2-week time point from woodchuck 335. The variants are listed according to their relative abundance in the PCR products, the first being the most abundant. The predicted amino acid sequences of the B and C domains of the DNA polymerase of the various mutants are summarized in Fig. 5. wt, wild type.

L-FMAU treatment was not associated with hepatotoxicity. The liver toxicity of L-FMAU treatment was evaluated for each woodchuck by monitoring the serum concentrations of SDH and by analyzing liver tissue biopsy sections for signs of histopathology, including an enhanced PCNA staining index. SDH is an abundant hepatocellular enzyme, and its elevated level in serum is an indication of liver injury (20). As shown (Fig. 8), SDH levels were not increased by L-FMAU administration for up to 19 weeks. Moreover, no signs of enhanced liver injury were observed, either histologically or by enhanced leukocyte infiltration (Table 1). PCNA staining of liver sections also failed to show any pattern of enhanced cell proliferation that might have indicated liver toxicity (Table 1) (e.g., as has been observed during therapy with 2'-carbodeoxyguanosine [14]). These results are thus consistent with the lack of liver injury observed during administration of L-FMAU to duck HBV-infected ducks (1).

View larger version (14K): [in this window] [in a new window]   FIG. 8.   L-FMAU therapy did not induce a rise in serum SDH levels. Serum samples collected during the first 16 weeks of therapy were assayed for SDH, expressed in international units (IU) per liter (20). The level of SDH in chronically infected woodchucks is generally below 80 IU. No significant elevation was observed except in one woodchuck (wc344) at the time of death, apparently from sepsis unrelated to drug administration.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Antiviral therapy of chronic HBV infections presents a novel problem. At the beginning of treatment, every hepatocyte is apparently infected by the virus (4, 6, 16, 17, 19, 21, 22, 35). Moreover, the rate of turnover of this cell population, even with active liver disease, is low (34) (t_1/2 > 1 week and, in "healthy" carriers, probably greater than 1 month). Thus, merely inhibiting virus replication would not readily eliminate the virus unless cccDNA, the template for viral RNA synthesis, had a short half-life within infected cells. However, this issue is still controversial. Some studies suggest that the DNA may have a high turnover rate (12, 15). On the other hand, data from the present and other studies suggest that this DNA is highly stable in vivo (14, 26, 43). In particular, these data suggest that if cccDNA has a finite life time, its half-life in the chronically infected liver is similar to that of infected hepatocytes.

One mechanism that would accelerate virus clearance is loss of cccDNA during cell division. In the present study, we sought indirect evidence for loss of cccDNA during mitosis by assaying for declines in the average cccDNA copy number in infected cells during therapy with the nucleoside analog L-FMAU. If this DNA is lost during mitosis and if it does not have any intrinsic instability in nondividing cells, then once virus DNA replication is blocked, the cccDNA copy number in infected cells should, ideally, remain fixed as the liver proliferates. That is, cells would either have lost cccDNA through the process of cell division or retained the original amount because they had not yet divided. The data suggest, however, that the cccDNA was distributed to daughter cells during proliferation of infected hepatocytes, producing the observed decline in the average copy number among cells that remained infected after prolonged therapy. Moreover, loss by mitosis requires that the infected-cell number decline virtually from the beginning of therapy, a possibility inconsistent with the experimental findings. The decline in copy number, by itself, could be explained by a model in which cccDNA is lost during mitosis and is also lost by decay in cells that have survived without division throughout the course of therapy. For instance, the observed results at 30 weeks of therapy could be modeled by an infected-cell death rate of 0.75% per day and a cccDNA half-life of 70 days. However, this model predicts that only 60% of the cells would remain infected after 6 weeks of therapy, a possibility at odds with the overall data.

An alternative possibility is that the low copy number was the result of new infections of cells that had lost existing cccDNA in the presence of L-FMAU. If so, the prevalent cccDNA in the liver might then have a drug resistance genotype. However, after 30 weeks of therapy, at which time the average cccDNA copy number among infected cells had declined at least 5- to 10-fold, the wild-type virus sequence was still prevalent in the cccDNA population. Our data thus favor but do not prove the hypothesis that cccDNA survives through mitosis and is distributed to each daughter cell, resulting in a decline in cccDNA copy number per cell. Data from a recent study (13) of WHV cccDNA survival in primary hepatocyte cultures that were induced to undergo limited proliferation by addition of epidermal growth factor were also consistent with this possibility.

Examination of the data in Fig. 2 and in a previous study (45) revealed an unexpected result. The type I mutation was sometimes detectable as a prevalent species in serum virus at early times in therapy, when virus titers were still declining. Since the same mutation may be associated with the later rebound of virus titers (Fig. 2) (45) and since the type I mutation confers L-FMAU resistance on a laboratory strain of WHV (Fig. 6), the reason for the continued decline at early times is not obvious. Several possibilities, not necessarily mutually exclusive, need to be considered. First, additional mutations outside the sequenced region of the polymerase may contribute to mutant fitness. This was not evident in a previous study, in which the complete pol gene of selected type I mutants was sequenced (45). However, the possibility has not been ruled out. Second, the type I mutant may be a common quasispecies that is generated as a result of errors during reverse transcription of a pregenomic RNA that was transcribed from wild-type cccDNA. In that case, it would be expected that virus titers would continue to decline until a significant fraction of this mutant virus could be converted to cccDNA. Third, the type I mutant may have a low replication rate, which, together with the need for coinfection with a virus producing the viral envelope proteins, may delay its spread to uninfected hepatocytes.