Hepatitis B virus mutations associated with fulminant hepatitis induce apoptosis in primary Tupaia hepatocytes

Authors


  • Conflict of interest: Nothing to report.

Abstract

Hepatitis B virus (HBV) core promoter mutations have been implicated in the pathogenesis of fulminant hepatitis B. Due to the limited availability of primary human hepatocytes, the functional characterization of HBV mutants has been performed predominantly in transformed cells, which may not represent ideal model systems for studying virus–cell interactions. We and others have shown that primary hepatocytes of the tree shrew Tupaia belangeri support HBV infection and replication. In this study, we used primary Tupaia hepatocytes to analyze the phenotype of two HBV core promoter mutations that have been associated with a clinical outbreak of fatal fulminant hepatitis. Similar to previous findings in human hepatoma cells, the HBV core promoter mutations resulted in enhanced viral replication and core expression. Surprisingly, however, the presence of the mutations had a marked effect on hepatocyte viability not previously observed in hepatoma cells. Reduced cell viability was found to be due to the induction of apoptosis, as evidenced by caspase-3 activation and nuclear fragmentation. In conclusion, HBV mutants exhibit a novel phenotype in primary hepatocytes distinctly different from previous findings in hepatoma cell lines. This phenotype may have important implications for the understanding of the fulminant clinical course associated with HBV mutations. Supplementary material for this article can be found on the HEPATOLOGY website (http://interscience.wiley.com/jpages/0270-9139/suppmat/index.html). (HEPATOLOGY 2005.)41:247–256

Infection with hepatitis B virus (HBV) leads to a wide spectrum of clinical presentations ranging from an asymptomatic carrier state to self-limited acute or fulminant hepatitis to chronic hepatitis with progression to cirrhosis and hepatocellular carcinoma.1, 2 Viral factors such as HBV genotypes,3 as well as the host immune response, have been implicated in the pathogenesis and clinical outcome of HBV infection.2 Furthermore, evidence has been accumulating that certain HBV mutants lead to particular clinical manifestations, influence the natural course of infection, and modulate the response to antiviral treatment.1, 4–7

Several independent studies have identified distinct mutations clustered in enhancer II of the HBV core promoter in association with fulminant8–12 and chronic severe hepatitis B.6, 13–16 A common hallmark of core promoter mutations is the phenotype of enhanced viral replication in hepatoma cell lines transfected with replication-competent HBV constructs.6, 10, 11, 13, 15 The most prevalent variant comprises a double mutation (A to T at nucleotide 1764 and G to A at nucleotide 1766, nucleotide numbering according to Raney et al.17 located at the 3′ end of enhancer II of the basal core promoter).

We have previously identified two mutations in the HBV core promoter (C to T at nucleotide 1768 and T to A at position 1770) of a viral strain associated with a fatal outbreak of fulminant hepatitis B (FH strain18, 19). This nosocomial outbreak had an unusual clinical course, with fulminant hepatic failure leading to the death of 5 patients within a few days. Functional characterization of these mutations in human hepatoma cell lines had demonstrated a markedly enhanced viral replication compared with wild-type HBV.10 The phenotype of enhanced replication was the result of enhanced encapsidation of pregenomic RNA into HBV nucleocapsids by co- and post-transcriptional effects of the core promoter mutations.10, 20

Because virus–cell interactions in transformed hepatoma cells may not accurately reflect the phenotype of variants in natural host cells, functional studies in primary hepatocytes are needed for understanding the impact of defined HBV mutations on disease. However, functional studies in primary host cells have been hampered by the limited availability of primary human hepatocytes of sufficient quality. Therefore, we have developed an alternative model to address this important question: primary hepatocytes of the tree shrew Tupaia belangeri. Tupaia belangeri has been shown to be susceptible to infection with a variety of human viruses, including rotavirus, herpes simplex virus, and hepatitis B and C viruses.21–24 We and others have demonstrated that primary Tupaia hepatocytes (PTHs) represent a convenient and suitable in vitro model to study HBV infection and replication.25–27

In this study, we demonstrate that HBV core promoter mutants associated with fulminant hepatitis exhibit a phenotype in primary hepatocytes distinctly different from previous findings in hepatoma cell lines. This phenotype may have important implications for the understanding of the fulminant clinical course associated with mutations.

Abbreviations:

HBV, hepatitis B virus; FH, fulminant hepatitis; PTH, primary Tupaia hepatocyte; PBS, phosphate-buffered saline; GFP, green fluorescent protein; anti-HBs, anti–hepatitis B surface antigen antibody; HBsAg, hepatitis B surface antigen; sHBsAg, small HBsAg; MT, mutant; CH, chronic hepatitis; ADV, adefovir dipivoxil.

Materials and Methods

Constructs and Recombinant Adenoviruses.

Replication-competent constructs of wild-type HBV (adwR9), core promoter mutant strains, and pCDLacZ have been previously described.20 The HBV constructs contained a 1.2 × genomic length of HBV ( Supplementary Fig. 1 ) and contained the identical genetic background (adwR9 strain10, 20). All mutant constructs were sequenced previously to confirm the expected sequence.10, 20 Control plasmids pGEM7 and pEGFP were obtained from Promega Corporation (Madison, WI) and Clontech Corporation (Palo Alto, CA), respectively. Recombinant chimeric HBV adenoviruses were generated as described previously.26, 28 Purified parental adenovirus vector was obtained from M. Bartolomé and L. Mohr (Department of Medicine II, University of Freiburg, Freiburg, Germany).

Isolation and Culture of PTHs.

Tupaia belangeri specimens were bred and maintained at the animal facilities of the University Hospital Freiburg in accordance with institutionally approved protocols and the National Institutes of Health guidelines for the use of experimental animals.29 Primary hepatocytes were prepared from adult animals and maintained as described.25 For inhibition of HBV replication, PTHs were incubated in the presence or absence of adefovir dipivoxil (Gilead Sciences, Foster City, CA) as described previously.30

Transfection of Recombinant DNA.

PTHs and Huh-7 cells grown on 100-mm or six-well plates were transfected with 4 μg plasmid DNA using liposomes (Lipofectin or Lipofectamin Plus; Invitrogen, Carlsbad, CA) in serum-free William's E medium or OPTI-MEM (Invitrogen). Huh-7 cells were maintained as described.10, 20 Transfection efficiency was monitored via cotransfection with construct pCDLacZ, expressing β-galactosidase under the control of the cytomegalovirus promoter.20LacZ transcription was monitored via Northern blot analysis using a LacZ-specific probe. LacZ expression was monitored via analysis of LacZ enzymatic activity in cell lysates using a chemiluminescent assay (Galacto-Light, Tropix, Bedford, MA) according to the manufacturer's protocol.

Adenoviral Transduction.

Primary hepatocytes or Huh-7 cells were infected with recombinant or parental adenoviruses. Virus stock aliquots containing the appropriate number of plaque-forming units to obtain a desired multiplicity of infection of 1 were mixed with 1 mL culture medium per well in a six-well plate and added to the cells for 6 hours. After the infection period, the viral inoculum was removed and the cells were washed extensively in phosphate-buffered saline (PBS) and incubated again in medium. At various time points, cells and medium were harvested and processed as described in Analysis of HBV Replication and Transcription.

Analysis of HBV Replication and Transcription.

PTHs or Huh-7 cells were harvested for viral RNA and DNA analysis 3 or 4 days after transfection. RNA was prepared using the RNeasy System (Qiagen, Hilden, Germany), analyzed via formaldehyde agarose gel electrophoresis, and hybridized with a HBV-specific probe as described.10, 31 Viral replicative DNA intermediates associated with intracellular core particles were isolated via ultracentrifugation of cell lysate through a 30% sucrose cushion and then analyzed via Southern blot hybridization.10, 19

Analysis of HBV Protein Expression.

Three to 5 days after transfection or transduction of cells with replication-competent plasmids or adenoviruses, cells were lysed, and HBV core and surface proteins were analyzed via SDS-PAGE and immunoblot of cell lysates using anticore (H800; dilution 1:3,000) or anti–hepatitis B surface antigen antibodies (anti-HBs) (032-A and 4F7; dilution 1:1,000; ViroGen, Watertown, MA) as described recently.20, 26 For control of transduction efficiency of adenoviruses, blots were reprobed using anti–green fluorescent protein (GFP) antibody (dilution 1:4,000; Clontech). Hepatitis B surface antigen (HBsAg) synthesis was analyzed in the cell culture medium using commercially available enzyme immunoassays (IMXHBsAg microparticle enzyme immunoassay, Abbott, North Chicago, IL, or Enzygnost HBsAg 5.0 immunoassay, Dade Behring, Marburg, Germany).

For metabolic labeling of small HBsAg (sHBsAg), PTHs (day 5 postinfection) were starved for 1.5 hours in methionine and cystein-free DMEM medium and labeled for 30 minutes with 250 μCi of [35S]methionine and [35S]cystein (Revidue PRO-MIX L-[35S] In Vitro Cell Labeling Mix, Amersham, Buckinghamshire, England in methionine and cystein-free DMEM medium) as described.20 The cells were then washed and lysed, and sHBsAg was immunoprecipitated using monoclonal anti-HBs (4F7) as described.20 Immunoprecipitated proteins were analyzed with a phosphoimager (Fuji, Tokyo, Japan). Signals were quantified using MacBas V2.4 software (Fuji).

Analysis of HBV Virion Synthesis and Secretion.

PTHs were infected with recombinant adenoviruses as described. Five days following infection, cell culture medium was pooled from one six-well plate of infected hepatocytes. Secreted virions were concentrated via centrifugation of 12 mL medium to a final volume of 0.4 mL using Amicon Ultracentrifugation Devices (Millipore, Bedford, MA). Concentrated virions were subjected to CsCl gradient ultracentrifugation and sedimented virions were detected as described recently.26

Analysis of Hepatocyte Apoptosis.

Two to seven days following infection with recombinant adenoviruses, PTHs were removed from culture dishes, resuspended in PBS, and fixed in 4% paraformaldehyde/PBS at 4°C. Apoptosis was assessed via nuclear staining of fixed cells using DNA-binding fluorochrome Hoechst-3325832 and detection of active caspase-3 using immunofluorescence.33–35 Following fixation, cells were permeabilized (0.1% Triton X-100 in PBS) and stained with a phycoerythrin-conjugated rabbit antihuman active caspase-3 antibody (Ab 67345X BD Pharmingen, San Diego, CA; dilution 1:500 in PBS33–35). Apoptosis was quantified by counting the average number of cells with nuclear fragmentation or positive staining for active caspase-3 per total cells using a Zeiss Axiophot microscope (Carl Zeiss, Jena, Germany).

Results

Functional Analysis of FH Core Promoter Mutations Associated With Fulminant Hepatitis in PTHs.

To identify the optimal conditions for transduction of PTHs, we compared different methods of viral gene transfer using a GFP expression construct. Although calcium phosphate–mediated transfection resulted only in low transfection efficiency and high toxicity (data not shown), liposome-mediated gene transfer allowed reasonable transfection efficiency at low toxicity ( Supplementary Fig. 2 ). The use of recombinant adenoviruses allowed transgene expression in all hepatocytes. A multiplicity of infection of 1 was sufficient to transduce all PTHs without cell death (see Supplementary Fig. 2 ).

After having defined the optimal conditions for hepatocyte transfection and transduction, we characterized the biological phenotype of the two previously well-characterized core promoter mutations associated with a fatal outbreak of fulminant hepatitis (FH mutant [MT]; see Supplementary Fig. 110) in PTHs. Using liposome-mediated gene transfer, we first analyzed the replication levels of terminally redundant wild-type and mutant HBV constructs in PTHs. FH MTH resulted in an approximately 10-fold increase of HBV replication (Fig. 1A) as quantified via phosphoimaging. To compare the biological phenotype of the FH MT with another core promoter mutation, we transfected PTHs with a construct containing a second pair of core promoter mutations associated with chronic hepatitis (CHMT; see Supplementary Fig. 115). CH MT exhibited a 1.5- to twofold increase in replication (see Fig. 1A) compared with wild-type HBV. Enhanced replication of mutants was reflected by an increase of relaxed circular and double-stranded linear HBV DNA species (see Fig. 1A). Background strains from the wild-type, FH MT, and CH MT were identical (adwR9 strain10, 20), ruling out the possibility that genotype- or subtype-specific factors were responsible for the observed differences.

Figure 1.

Viral replication, transcription, and protein expression of core promoter mutants in PTHs. PTHs were transfected with terminal-redundant replication-competent HBV DNA ( Supplementary Fig. 1 ) or control DNA (pGEM7) as described in Experimental Procedures. (A) Viral replication. Viral replication was analyzed via Southern blot analysis of nuclease-resistant, encapsidated HBV DNA using a [32P-dATP]-labeled HBV specific probe 4 days after transfection. (B) Viral transcription. Viral transcription was assessed via Northern blot analysis of HBV RNA using a [32P-dCTP]-labeled HBV-specific probe 4 days after transfection (upper panel). To control for transfection efficiency and RNA loading, transcription of cotransfected plasmid pCDLacZ was monitored via Northern blot analysis using a LacZ-specific probe (lower panel). (C) Core expression. Core protein expression was analyzed via 15% SDS-PAGE and immunoblot of hepatocyte lysates using a polyclonal core-specific antibody (H800; dilution 1:3,000) and a horeseradish peroxidase–conjugated anti–rabbit immunoglobulin G secondary antibody 4 days after transfection. (D) HBsAg secretion. HBsAg secretion in tissue culture medium was analyzed via enzyme immunoassay 4 days after transfection as described in Experimental Procedures. Results show the mean and standard deviation of three experiments. Molecular weight markers (kd) are indicated on the left. WT, HBV wild-type construct; FH MT, fulminant hepatitis mutant; CH MT, chronic hepatitis mutant; RC, relaxed circular HBV DNA; DL, double-stranded linear HBV DNA; SS, single-stranded HBV DNA (replicative intermediates); HBV, hepatitis B virus; 3.5 kb, pregenomic HBV RNA; 2.4/2.1 kb, subgenomic HBV RNA; MW, molecular weight; core, HBV core protein; HBsAg, hepatitis B surface antigen; S/N, signal-to-noise ratio.

In two previous studies, we demonstrated that enhanced viral replication observed in the FH MT is due to a co- and post-transcriptional effect of the core promoter mutations on core protein synthesis. Increased core protein synthesis results in enhanced nucleocapsid assembly followed by enhanced viral encapsidation of pregenomic RNA into HBV nucleocapsids and enhanced viral replication.10, 20 In extensive functional studies using transcomplementation assays dissecting the contribution of core, HBx, polymerase, and the pregenomic RNA for enhanced replication observed for the FH MT, we have demonstrated that enhanced core protein expression, nucleocapsid assembly, encapsidation, and replication are largely independent of the pregenomic transcript level.10, 20 Therefore, in human hepatoma cells the phenotype of the FH MT is characterized by a markedly enhanced replication and core expression without a concomitant increase in the pregenomic RNA transcript level.10, 20

To study whether the FH MT exhibits a similar phenotype in PTHs, we analyzed viral transcription and core expression in PTHs transfected with terminally redundant R9 constructs containing a wild-type sequence, FH, or CH MT. RNAs were purified and analyzed via Northern blot hybridization. Two species of HBV RNA (3.5 and 2.4/2.1 kb, respectively) were synthesized from all three constructs (Fig. 1B). As shown in Fig. 1B, the FH MT demonstrated a twofold increase of 3.5-kb RNA compared with the adw wild-type (the densitometric reading of 3.5-kb RNA was 2,136 arbitrary units for wild-type and 4,167 arbitrary units for FH MT). There was no significant difference between the 2.1- and 2.4-kb transcripts (Fig. 1 B). Transfection efficiency was similar for all constructs as demonstrated by similar levels of LacZ RNA, transcribed from the cotransfected construct pCDLacZ (Fig. 1 B). Although this minor difference in the 3.5-kb RNA level between the wild-type and FH MT appeared small, it was confirmed by two other independent experiments. The construct containing the CH MT produced similar 3.5 kb RNA and 2.1/2.4 kb RNA levels compared to wild-type constructs.

To study whether FH MT results in a similar increase in core protein expression as previously observed in human hepatoma cells,10, 20 we studied core protein expression of the FH MT in PTHs. As shown in Fig. 1C, the FH MT exhibited a strong increase (tenfold after correction for transfection efficiency) of core protein expression compared with wild-type HBV, paralleling the observed increase in HBV replication. In contrast, the CH MT resulted in a marginal increase (1.5-fold) in core protein expression (see Fig. 1C).

A similar phenotype of FH MT–induced enhanced replication and core expression—which was largely independent of the pregenomic RNA transcript level—was observed when recombinant chimeric HBV adenoviruses were used as a delivery system (Fig. 2A,B).

Figure 2.

HBsAg expression in HBV adenovirus-transduced PTHs. PTHs were transduced with recombinant HBV adenoviruses as described in Experimental Procedures. (A) HBV replication, (B) transcription (top panel), core protein expression (bottom panel), and (C) HBsAg secretion were assessed 5 days after transduction as described in Fig. 3. (D) HBsAg expression was analyzed via immunoblot analysis of hepatocyte lysates using an anti-HBs–specific antibody (032-A). sHBsAg (p24 and p27), middle HBsAg (p33 and p36), and large HBsAg (p39 and p42) are indicated on the right. Molecular weight markers (kd) are indicated on the left. GFP, green fluorescent protein; WT, HBV wild-type construct; FH MT, fulminant heptitis mutant; RC, relaxed circular HBV DNA; DL, double-stranded linear HBV DNA; SS, single-stranded HBV DNA (replicative intermediates); 3.5 kb, pregenomic HBV RNA; 2.4/2.1 kb, subgenomic HBV RNA; core, HBV core protein; HBsAg, hepatitis B surface antigen; S/N, signal-to-noise ratio; Anti-HBs, anti–hepatitis B surface antigen antibody.

Taken together, these findings demonstrate that the phenotype of enhanced replication and core expression without a concomitant increase in pregenomic RNA transcription is similar in PTH (see Figs. 1, 2) and human hepatoma cells.10, 20 This result suggests that the previously characterized molecular mechanism of a combined co- and post-transcriptional effect of the FH core promoter mutations on core expression10, 20 is also responsible for the observed FH MT–induced enhanced replication in PTHs.

Decreased HBsAg Expression in PTHs Transduced With Mutant HBV.

Surprisingly, and in contrast to previous studies performed in hepatoma cells, the FH MT demonstrated a marked decrease of HBsAg in the supernatant of PTHs (Figs. 1D, 2C). This result was not due to differences in transfection efficiency as evidenced by analysis of LacZ messenger RNA produced by cotransfected plasmid pCDLacZ (see Fig. 1B) or analysis of GFP expression of GFP complementary DNA in recombinant adenoviruses (see Fig. 2D). Further, subgenomic messenger RNA levels in PTHs were similar for the wild-type and FH MT, ruling out a significant difference in transcription efficiency (see Figs. 1B, 2B). These results were confirmed by primer extension analysis of viral subgenomic RNA species (data not shown). To study whether decreased HBsAg levels in the supernatant of transduced cells were due to a defect in HBsAg secretion, HBsAg levels were analyzed via immunoblotting in lysates of PTHs that had been transduced by recombinant adenoviruses encoding replication-competent HBV genomes as well as the GFP marker (see Fig. 2D). Transduction efficiency of PTHs was similar for the mutant and wild-type virus as indicated by similar levels of GFP expression in an immunoblot analysis of hepatocyte lysates (see Fig. 2D). Surprisingly, the FH MT resulted in a marked decrease of sHBsAg expression in primary hepatocytes (see Fig. 2D). The steady-state level of both sHBsAg species—p24 and its glycosylated form p27—were strongly decreased in hepatocytes transduced with the FH MT. The decrease in sHBsAg expression was confirmed using two different types of anti-HBs (data not shown). In contrast, expression of middle (p33 and p36) and large (p39 and p42) HBsAg was not markedly altered by the presence of the FH MT (Fig. 2D).

To study whether FH MT–induced downregulation of sHBsAg expression was the result of altered protein synthesis, infected PTHs were pulse-labeled with [35S]-methionine and [35S]-cystein, and cell lysates were examined by immunoprecipitation with anti-HBs. HBV adenovirus containing the FH MT displayed a markedly decreased synthesis of sHBsAg as compared with adenovirus containing wild-type HBV DNA (sHBsAg synthesisFH MT/ sHBsAg synthesisWT ≈30%; Fig. 3A). Subsequent chase in the presence of excess nonradioactive methionine and cystein revealed little or no difference in the turnover rate of sHBsAg in the FH MT and wild-type strains (data not shown). These data clearly indicate that the reduced sHBsAg levels are due to reduced surface messenger RNA translation.

Figure 3.

SHBsAg synthesis and virus secretion in mutant and wild-type HBV. PTHs were transduced with either wild-type or mutant replication-competent HBV DNA construct as indicated. (A) SHBsAg synthesis. Hepatocytes were subjected to metabolic labeling 5 days after transduction. After pulse-labeling with [35S]-methionine and [35S]-cystein for 30 minutes, hepatocytes were lysed and subjected to immunoprecipitation using a surface-specific antibody (anti-HBs) or an isotype control antibody (immunoglobulin G). Immunoprecipitated proteins were analyzed via SDS-PAGE and autoradiography (left panel). Molecular weight markers (kd) are indicated on the left. The identity of the sHBsAg bands was established via a parallel immunoblot using anti-HBs 0-32A (right panel). The band below p24 did not react with anti-HBs and therefore probably represents a cellular or nonspecific protein co-immunoprecipitating with p24 and p27. (B) Virus secretion. Secreted virions were concentrated from cell culture medium 5 days after transduction. Concentrated virions were then subjected to CsCl gradient centrifugation. A Southern blot analysis of HBV DNA in CsCl fractions is shown. The presence of mature relaxed circular HBV DNA in enveloped HBV particles (fractions 6 and 7) is indicated on the left. Naked single-stranded HBV DNA in immature nonenveloped HBV particles (fractions 8 and 9) and chimeric adenovirus DNA (fractions 8 and 9) are indicated on the right. Presence of HBsAg (determined by an HBsAg-specific enzyme immunoassay) and density of the respective fractions is indicated on the bottom. Transduction efficiency (monitored by immunoblot analysis of GFP expression in transduced cells from the same experiment) was similar for wild-type and mutant HBV (data not shown). IP, immunoprecipitation; anti-HBs, anti–hepatitis B surface antigen antibody; IgG, immunoglobulin G; IB, immunoblot; GFP, green fluorescent protein; WT, HBV wild-type construct; FH MT, fulminant hepatitis mutant; Ad, adenovirus; RC, relaxed circular HBV DNA; HBV, hepatitis B virus; SS, single-stranded HBV DNA (replicative intermediates); HBsAg, hepatitis B surface antigen.

In contrast to these observations in PTHs, we did not observe any significant differences in sHBsAg expression levels between wild-type and mutant constructs in human hepatoma cells (data not shown). Thus, our data reveal an unexpected distinct biological phenotype of the FH MT in nontransformed host cells different from the phenotype observed in human hepatoma cells.10, 20

Finally, we studied whether decreased sHBsAg translation resulted in intracellular retention of virions potentially contributing to enhanced viral replication in PTHs. Therefore, we assessed secretion of enveloped virions in PTHs transduced with wild-type or mutant HBV genomes. As shown in Fig. 3B, transduction of PTHs with construct containing the FH MT did not show any evidence of an impairment in virion secretion compared with the wild-type construct. Enhanced viral replication resulted rather in an increase in HBV virion secretion. These data indicate that enhanced HBV replication induced by FH MT was not the result of HBV DNA retention in the hepatocyte and decreased sHBsAg translation does not result in intracellular retention of HBV virions.

Induction of Apoptosis by Mutant HBV in Primary Hepatocytes.

Following the transduction of hepatocytes with FH MT we observed a marked cytopathic effect compared to hepatocytes transduced with HBV wild-type or control vectors (Ad-GFP or the parental adenovirus). Studying hepatocytes transduced with HBV containing the FH MT by light microscopy revealed time-dependent cell death resulting in the detachment of hepatocytes and a marked decrease in the number of adherent cells. This finding was absent in transduced hepatoma cells. To study whether the cytopathic effect was due to enhanced apoptosis, two assays were performed. First, apoptosis was monitored via detection of active caspase-3 in situ; caspase-3 is a key protease that is activated during the early stages of apoptosis.33–35 Second, apoptosis was detected using nuclear staining of fixed cells using DNA-binding fluorochrome Hoechst-33258.32 Nuclear fragmentation occurs at late stages of apoptosis and was observed using fluorescence microscopy.

Using these methods, we observed that hepatocytes transduced with HBV containing the FH MT exhibited a marked increase of caspase-3 activation compared with hepatocytes transduced with HBV wild-type, the control vector Ad-GFP, or the parental adenovirus (Fig. 4). Caspase-3 activation induced by mutant HBV was present already on day 2 (Fig. 4). This finding was confirmed by the analysis of nuclear fragmentation as a marker for late-stage apoptosis on day 5 following transduction (Fig. 5). Interestingly, we also detected an increase of apoptotic cells in wild-type transduced cells at later time points (day 5 following transduction), albeit to a lesser extent than the FH MT (see Fig. 5). Transduction efficiency of hepatocytes was similar as assessed by GFP expression on day 2 (data not shown). These findings were reproduced in several experiments using different hepatocyte preparations. Taken together, these observations suggest that HBV gene expression and replication can induce apoptosis in primary hepatocytes and that the presence of the FH MT markedly enhances the induction of apoptosis.

Figure 4.

Induction of caspase-3 activation in HBV-transduced hepatocytes. Apoptosis was assessed via cytoplasmic staining of fixed cells using a phycoerythrin-conjugated antibody interacting with activated caspase-3 2 days after infection with recombinant adenoviruses.33–35 Representative sections of (A) uninfected hepatocytes and hepatocytes transduced with (B) the parental adenovirus and (C) GFP, (D) HBV wild-type, and (E) FH MT adenoviruses are shown. (F) The percentage of active caspase-3–positive cells was determined by counting the number of cells with positive caspase-3 staining divided by the number of total cells (n = 200). Results show the mean and standard deviation of three experiments. Ad, adenovirus; GFP, green fluorescent protein; HBV, hepatitis B virus; WT, wild-type; FH MT, fulminant hepatitis mutant.

Figure 5.

Nuclear fragmentation as a marker for late-stage apoptosis in HBV-transduced hepatocytes. PTHs were removed from the collagen-coated culture dishes and fixed in 4% paraformaldehyde 5 days after infection with recombinant adenoviruses. Apoptosis was assessed via nuclear staining of fixed cells using DNA-binding fluorochrome Hoechst-33258. (A-E) Fluorescence microscopy of nuclear fragmentation in apoptotic hepatocytes (arrows). Representative sections of (A) uninfected hepatocytes and hepatocytes transduced with (B) the parental adenovirus and (C) GFP, (D) HBV wild-type, and (E) FH MT adenoviruses are shown. (F) Hepatocyte apoptosis was quantified by counting the average number of cells with nuclear fragmentation (apoptotic cells) per total cells (n = 200). Results show the mean and standard deviation of three experiments. Ad, adenovirus; GFP, green fluorescent protein; HBV, hepatitis B virus; WT, wild-type; FH MT, fulminant hepatitis mutant.

To study whether FH MT–induced apoptosis was due to enhanced replication present in the viral genome containing the FH MT, we studied apoptosis in the presence or absence of adefovir dipivoxil (ADV), a nucleotide analogue that efficiently inhibits HBV replication (Fig. 6). Although ADV strongly inhibited viral replication of the HBV genome containing the FH MT (Fig. 6A), FH MT–induced apoptosis (as detected via in situ activation of caspase-3 and nuclear fragmentation) remained unchanged (Fig. 6B). These findings indicate that FH MT–induced apoptosis is mediated by a molecular mechanism independent of enhanced replication.

Figure 6.

FH MT–induced hepatocyte apoptosis and HBV replication. Hepatocytes transduced with the parental adenovirus (Ad) or adenovirus containing HBV FH MT (Ad-FH MT) were incubated with or without ADV in a concentration of 100 μmol/L. (A) HBV replication. Viral replication in the presence (lanes 1-3) or absence (lanes 4-6) of ADV was assessed on days 1, 2, and 5 following transduction as described previously.35 (B) FH MT–induced apoptosis and viral replication. Hepatocytes were trans duced and incubated in the presence or absence of ADV as described in panel A, and apoptosis was assessed via cytoplasmic staining of fixed cells using a phycoerythrin-conjugated antibody interacting with activated caspase-3 (day 2) and nuclear staining of fixed cells using DNA-binding fluorochrome Hoechst-33258 (day 5) as described in Figs. 5 and 6. To exclude that apoptosis was induced by the parental adenovirus or ADV, PTHs were incubated with the parental adenovirus and/or ADV in parallel experiments. The percentage of active caspase-3–positive cells was determined by counting the number of cells with positive caspase-3 staining divided by the number of total cells (n = 200). The percentage of apoptotic cells after staining with DNA-binding flourochrome was quantified by counting the average number of cells with nuclear fragmentation (apoptotic cells) per total cells (n = 200). Results show the mean and standard deviation of three experiments. ADV, adefovir dipivoxil; RC, relaxed circular HBV DNA; DL, double-stranded linear HBV DNA; SS, single-stranded HBV DNA (replicative intermediates); Ad, adenovirus; FH MT, fulminant hepatitis mutant; p.i., post inoculation.

Discussion

This study describes a detailed analysis of the biological phenotype of two core promoter mutations associated with fatal fulminant hepatitis in primary hepatocytes. Our results reveal two novel phenotypes associated with these mutations in primary hepatocytes but not in hepatoma cells. These phenotypes include mutant-induced apoptosis and downregulation of sHBsAg translation.

Mutant HBV–induced apoptosis was demonstrated by two independent methods assessing early (caspase-3 activation) and late (nuclear fragmentation) markers for apoptosis. Because apoptotic effects were observed at very low multiplicities of infection of recombinant adenoviruses, it is unlikely that apoptosis was mediated by adenovirus transfer vectors and its gene products. This hypothesis is supported by previous observations demonstrating that replication-deficient recombinant adenoviruses do not significantly induce cell death in PTHs or primary human hepatocytes at the multiplicities of infection used in this study.26, 36

Although HBV is considered a noncytopathic virus,1 hepadna virus–induced apoptosis and cytopathic effects have been described in three model systems. First, a duck hepatitis B variant containing a single amino acid change in the large surface antigen resulting in accumulation of covalently closed circular DNA has resulted in a strong cytopathic effect in hepatocytes invitro and in vivo.37–39 In this system, the level of viral replication and covalently closed circular DNA formation correlated with cytopathic effects in infected hepatocytes.37 Second, intracellular retention of the HBV large surface protein has been shown to induce apoptosis in cell lines.40, 41 In this model, overexpression of the large surface antigen resulted in cellular vacuolization and apoptosis of transfected hepatoma cells.41 Third, the HBx protein has been suggested to induce apoptosis in both a p53-dependent and p53-independent manner.42–44 Exploring the mechanism of these previous observations, a recent study has elegantly demonstrated that HBx interacts with c-FLIP, a key regulator of the death-inducing signaling complex.43 Recruitment of c-FLIP to the death-inducing signalling complex is inhibited by HBx resulting in hyperactivation of caspase-8 and caspase-3 by death signals.43

Considering these previous findings, the following mechanisms may account for the observation of FH MT–induced apoptosis: First, it is conceivable that enhanced replication contributes to the observed cytopathic effects. However, because inhibition of viral replication by a nucleotide analogue did not reverse FH MT–induced apoptosis, it is unlikely that this mechanism plays a crucial role in mediating this process. The finding of replication-independent apoptosis rather suggests that a viral protein (expressed from the transduced complementary DNA) is responsible for virus-induced hepatocyte cell death. Second, intracellular retention of the large surface protein has been shown to induce apoptosis. Because the FH MT did not result in an impairment of virion secretion (see Fig. 3), and because decreased sHBAg in cell culture medium was not due to HBsAg retention (see Figs. 2, 3), an intracellular retention of surface or other viral proteins contributing to FH MT–induced apoptosis is unlikely. Third, mutations within the HBx open reading frame may contribute to FH MT–induced apoptosis. Because the HBx open reading frame overlaps with the core promoter, the two core promoter mutations result in two amino acid changes of the HBx protein (valine is replaced by isoleucine in position 132 and phenylalanine is changed to tyrosine in position 133). It is therefore conceivable, that the two mutations in the HBx open reading frame may alter the interaction of HBx protein with c-FLIP, thus resulting in enhanced cell death (including caspase-3 activation as shown in Figs. 4 and 6). Studies using recombinant adenoviruses containing wild-type and mutant HBx protein are in progress to answer this important question.

The second hallmark of the mutant phenotype in primary hepatocytes was downregulation of sHBsAg translation. Interestingly, decreased sHBsAg translation did not impair virion secretion, suggesting that the reduced amount of sHBsAg does not limit the synthesis and secretion of enveloped virions.

The precise mechanisms accounting for decreased sHBsAg translation have yet to be elucidated. It is possible that FH MT–induced apoptosis contributes to the alteration of sHBsAg translation. Apoptosis-mediated downregulation of sHBsAg translation may also explain why HBsAg expression was not altered in transformed Huh-7 cells. Thus, it has been described in detail that apoptosis can result in the alteration of translation of messenger RNA by interacting with translation initiation factors.45, 46 An example is eukaryotic translation initiation factor 4G being targeted for proteolytic cleavage by caspase-3 in apoptotic cells.47 It is therefore conceivable that apoptosis-induced modulation of defined transcription factors is responsible for the observed decrease in sHBsAg translation. Notably, in our experimental system, inhibition of translation appeared to be specific to defined viral RNA but did not affect LacZ or GFP translation. The factors determining the specificity of apoptosis-induced alteration of translation remain to be further defined.

Whether the observed findings play an important role in the fulminant course of HBV infection associated with the two core promoter mutations has yet to be determined. Interestingly, three out of five patients presenting with fatal fulminant hepatitis following nosocomial transmission of the FH strain had undetectable or borderline levels of HBsAg despite the presence of HBV DNA.18, 48 This finding suggests that FH MT–associated downregulation of sHBsAg expression and HBsAg secretion may occur in the infected human host liver in vivo. Furthermore, induction of apoptosis has been a hallmark of fulminant liver failure in various animal models.49, 50 Therefore, FH MT–induced apoptocic hepatocyte death, as described here, appears likely to contribute to the pathogenesis of fulminant hepatitis B.

In conclusion, our results demonstrate that HBV mutants exhibit a novel phenotype in primary hepatocytes that is distinctly different from previous findings in hepatoma cell lines. This phenotype may have important implications for the understanding of the fulminant clinical course associated with HBV mutations.

Acknowledgements

The excellent technical assistance of Sabine MacNelly and Christine Rösler is gratefully acknowledged.