The chronically WHV-infected woodchuck is a good model (32) to further elucidate the role of IFN-α and IFN-γ in chronic hepadnavirus infection and the potential therapeutic properties of IFN. Important prerequisites for this study were the molecular cloning and characterization of woodchuck-specific IFN-α (wIFN-α) and wIFN-γ (20, 34). Treatment of primary hepatocytes from persistently infected woodchucks with wIFN-α in vitro resulted in an inhibition of virus replication and a reduction in viral proteins (34). Recently, wIFN-γ was shown to enhance MHC-I gene expression but failed to deplete WHV replication intermediates and mRNAs in hepatocytes from persistently infected animals (22, 37).

High-capacity or helper-dependent Ad (HD-Ad) vectors seem to be good candidates for the delivery of cytokine genes to the liver. HD-Ad vectors do not express any viral proteins, and thus safety is considerably improved; in addition, the immune response induced by the vector itself is significantly reduced compared to previous-generation Ad vectors (17). Gene delivery with HD-Ad vectors was also shown to induce a long-lasting expression of proteins. In a monogenic hyperlipidemia mouse model, a single injection of an HD-Ad vector expressing the missing protein resulted in lifelong protection of the mice from hyperlipidemia (16). Gene delivery of an HD-Ad vector expressing mouse-specific IFN-α resulted in an intrahepatic IFN-α expression that protected the liver in acute hepatitis mouse models (1). Recently, we demonstrated that an Ad vector-mediated gene delivery is feasible in the woodchuck model: transduction of hepatocellular carcinomas (HCCs) of woodchucks in vivo with an Ad vector expressing interleukin 12/B7.1 was sufficient to induce a regression of large tumors (28).

In this study, we investigated the effects of wIFN-α and wIFN-γ on chronic hepadnavirus infection in the woodchuck by a new approach: we wanted to achieve local and long-lasting expression of IFN-α or IFN-γ in the liver in vivo by delivery of the respective cytokine genes to hepatocytes via HD-Ad vectors. We demonstrated an efficient HD-Ad vector-mediated delivery of the woodchuck genes for cytokines wIFN-α and wIFN-γ and the local expression of these cytokines, which resulted in elevated levels of wIFN-α and wIFN-γ, which could be measured in serum. We showed a reduction of WHV replication after transduction with the wIFN-α-expressing vector but not after transduction with the wIFN-γ-expressing vector.

Construction of HD-Ad vectors. The cloning of wIFN-α and the construction of HD-Ad vector encoding wIFN-α (HD-ADwIFN-α) was described in detail previously (34). The complete coding region of the wIFN-α-3a gene was cloned under the control of the liver-specific transthyretin promoter TTR (1). Biologically active wIFN-γ was cloned and characterized as described previously (20, 22). The HD-AdwIFN-γ vector is based on the plasmid pSTK129 carrying a 26.5-kb insert in the multiple cloning site of Bluescript KSII that contains the left terminus of Ad type 5 (Ad5) (nucleotides [nt] 1 to 440), a 19,952-bp fragment of the human hypoxanthine-guanine phosphoribosyltransferase gene (nt 1777 to 21729), a 6,545-bp fragment of the C346 cosmid (nt 10205 to 16750), and the right terminus of Ad5 (nt 35818 to 35935). HD-AdwIFN-γ contains the complete coding sequence of wIFN-γ cDNA under control of the murine CMV promoter terminated by simian virus 40 poly(A) (Fig. 1). Vector construction was as follows. The 560-bp IFN-γ cDNA fragment was obtained from the plasmid peWIFN-γ and inserted into the EcoRI site of pMH4. The complete expression cassette was isolated from pMH4wIFN-γ by XbaI/BglII digestion and integrated into the NotI site of pSTK129. The control vector HD-AdGFP has been described previously (30). Adenoviral constructs were cleaved by PmeI and transfected into 293Cre4 cells (4). Amplification and purification of the HD viruses was performed as described previously (30, 34).

FIG. 1. Construction of HD-AdwIFN-γ. The wIFN-γ gene was cloned under the control of the murine CMV promoter terminated by simian virus 40 poly(A). The expression cassette was inserted in the pSTK129 backbone as described in Materials and Methods. (more ...)

Expression of IFN by HD-AdwIFN-α and HD-AdwIFN-γ. To measure the biological activity of IFN expressed by HD-AdwIFN-α and HD-AdwIFN-γ, a permanent liver woodchuck cell line (WH12/6, kindly provided by P. Banasch, Heidelberg, Germany) was transduced with the two vectors (10² physical particles per cell). Supernatants were harvested for 7 days and tested by a virus protection assay (murine encephalomyocarditis virus [EMCV] bioassay) (22). Briefly, WH12/6 cells were cultured in 96-well microtiter plates and incubated for 24 h with serial dilutions of the supernatants or woodchuck sera. EMCV was added to the cells for an additional incubation of 24 h. Cells were stained and fixed with 0.1% crystal violet in 20% ethanol. One unit of IFN per ml was defined by its ability to protect 50% of the cells in a well against EMCV-induced cell destruction.

Woodchucks. Adult woodchucks chronically infected with WHV that were trapped in the state of Delaware were purchased from North Eastern Wildlife (Ithaca, N.Y.). One naïve woodchuck was born at our facilities in April 2001. Experiments were conducted in accordance with the Guide for the Care and Use of Laboratory Animals (27) and were reviewed and approved by the local Animal Care and Use Committee (Animal Care Center, University of Essen, Essen, Germany, and the district government of Düsseldorf, Germany). WHV DNA, WHV surface antigen, and antibodies to WHV core antigen were detectable in the sera of chronically WHV-infected woodchucks, while antibodies to WHsAg were absent. In the naïve animal, all markers of WHV infection were negative.

Transduction of primary woodchuck hepatocytes with HD-ADwIFN-α. Primary woodchuck hepatocytes were prepared as described in detail previously (21). Briefly, the liver cells were harvested by perfusion of the liver with a collagenase medium. Cell suspensions were run through a 70-μm-pore-size filter to remove tissue fragments. Hepatocytes were separated from other cells by repeated centrifugation at 50 × g. Primary woodchuck hepatocytes of woodchucks chronically infected with WHV were seeded into six-well plates. Hepatocytes were maintained in Williams medium supplemented with 2 mM glutamine, 0.05% glucose, 20 mM HEPES [pH 7.4], hydrocortisone, 12.5 μg of inosine/ml, 12 IU of insulin/ml, 5 mM sodium pyruvate, 50 IU of penicillin-streptomycin/ml, and 1% dimethyl sulfoxide. After 24 h, the woodchuck cells were transduced with HD-AdwIFN-α with 10, 10², or 10³ physical particles per cell. Culture supernatants and cells were collected every 2 days. The amounts of wIFN-α were measured by the EMCV bioassay. The expression of Mx protein was detected by Western blotting (see below). Primary woodchuck hepatocytes were transduced with an Ad vector encoding β-galactosidase (AdβGal) (Quantum Biotechnology, London, United Kingdom) as a control vector.

Transduction of woodchucks in vivo via the portal vein. Seven chronically WHV-infected woodchucks were transduced with HD-AdwIFN-α (two woodchucks: numbers 13931 and 14101), HD-AdwIFN-γ (numbers 14089, 14636, and 15727), or HD-AdGFP (numbers 14097 and 15725), respectively. In addition, one naïve woodchuck (number 15524) was transduced with HD-AdwIFN-γ. The vectors were injected directly into the portal vein by open surgery (1 × 10¹² physical particles in 1 ml of phosphate-buffered saline, corresponding to approximately 5 × 10⁹ infectious particles for HD-AdGFP; it is generally accepted that the amount of infectious particles is 100 to 1,000 times lower than the amount of physical particles). Before transduction and at weekly to monthly intervals afterwards, liver biopsies were performed by either transcutaneous needle biopsy or open surgery with a standard biopsy needle (14-gauge ABC needle; Sherwood Medical, Ballymoney, Northern Ireland) as previously described in detail (5). The biopsies were snap-frozen for virological assays and formalin fixed for immunohistochemical analysis. Blood samples were taken prior to transduction and at weekly to monthly intervals afterwards. All operations were conducted under 10% ketamine hydrochloride-2% xylazine anesthesia.

Detection of GFP by direct immunofluorescence microscopy and immunohistochemistry. Expression of GFP was detected by direct immunofluorescence microscopy of frozen sections of biopsies that were snap-frozen in liquid nitrogen. To relate the expression of GFP to the different cell types of the liver, formalin-fixed liver samples were incubated with a monoclonal antibody (MAb) to GFP (Clontech, Palo Alto, Calif.). The reaction mixture was visualized with the Power Vision system (Immunovision Technologies Co., Daly City, Calif.). Counterstaining was performed with Hämalaun.

Analysis of Mx protein induction. Frozen liver samples were ground to powder and lysed in sample buffer for sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Equal amounts of protein were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis and blotted onto nitrocellulose membrane. Mx was detected by Western blotting with a MAb raised against mouse Mx protein. This MAb recognizes Mx homologues in various species (MAb kindly provided by O. Haller, Freiburg, Germany). Bands were detected via standard peroxidase staining or via luminescence technique. The latter appeared to be the more-sensitive technique in our hands.

Detection of WHV DNA. WHV DNA in serum samples (each, 5 μl) was detected by spot blot hybridization with a full-length WHV8 genome. Serial dilutions of a WHV DNA sample with a known concentration (10⁵ to 10⁹ copies/ml) were used as standards for quantification. This method was found to be sensitive to 100 pg of WHV DNA/ml. The spots were quantified with a Cyclone storage phospho screen imager (Packard, Dreiech, Germany). WHV replication intermediates in liver biopsies were analyzed. Total liver DNA from snap-frozen liver samples was extracted with the QIAamp tissue kit (QIAGEN, Hilden, Germany) according to the manufacturer's instructions and as described previously in detail (5, 21). WHV replication intermediates were analyzed by Southern blot hybridization with a full-length WHV8 genome as a probe.

Analysis of mRNAs of woodchuck-specific IP-10 and GAPDH (glyceraldehyde-3-phosphate dehydrogenase) by Northern blotting. Total RNA was purified from snap-frozen liver biopsies with an RNeasy kit (QIAGEN) and subjected to denaturing agarose gel electrophoresis with formamide. After the RNA was transferred to a nitrocellulose sheet by vacuum blotting, the mRNAs of woodchuck-specific IFN-inducible protein 10 (IP-10) or GAPDH were detected by specific ³²P-labeled probes (partial cDNA clones), respectively. The relative signal strengths of bands on the Northern blot were quantified with a Cyclone storage phospho screen imager (Packard).

Detection of Ad DNA. To detect the persistence of the vector in liver samples, PCR amplification of the cosmid (C346) gene was performed after DNA extraction with primers C346s1 (nt 10372 to 10392; 5′GGCATCCACTAAATCCTCCC′3) and C346as1 (nt 10618 to 10598; 5′CCTTTGTTTTCTGTCTTCCC′3). PCR products were identified by gel electrophoresis.

Measurement of liver cell damage. To measure liver cell damage, sorbitol dehydrogenase (SDH) with chemicals supplied by Sigma-Aldrich Chemicals (Deisenhofen, Germany) activity was determined, e.g., after transduction, according to the manufacturer's instructions.

Expression of cytokines and GFP in vitro. The expression of biological active proteins in vitro by HD-AdwIFN-α, HD-AdwIFN-γ, and HD-AdGFP had to be demonstrated prior to in vivo experiments. The expression of GFP after transduction of 293Cre cells and WH12/6 cells with HD-AdGFP was demonstrated by direct immunofluorescence microscopy (data not shown). One preparation of the vector was determined to contain 5 × 10⁹ infectious particles by infection of 293Cre cells with HD-AdGFP corresponding to 1 ×10¹² physical particles. To test the secretion of biologically active wIFN-α or wIFN-γ by HD-AdwIFN-α or HD-AdwIFN-γ, respectively, WH12/6 cells were transduced with these vectors. Supernatants were harvested for 7 days, and the amounts of biologically active IFN were determined by the EMCV bioassay. Up to 128 U of biologically active IFN/ml in supernatants of cells transduced with HD-AdwIFN-α and up to 64 U of IFN/ml in those transduced with HD-AdwIFN-γ (Fig. 2) were measured. In supernatants of WH12/6 cells transduced with HD-AdGFP, biologically active IFN was not detectable (data not shown).

FIG. 2. Detection of biologically active IFN in supernatants of WH12/6 cells harvested up to day 7 after transduction with HD-AdwIFN-α (a) or HD-AdwIFN-γ (b) (102 physical particles per cell). IFN levels were determined by the EMCV bioassay.

Effects of expression of wIFN-α on WHV replication in primary woodchuck hepatocytes. Primary woodchuck hepatocytes of chronically infected WHV-positive woodchucks were transduced in vitro with HD-AdwIFN-α (10, 10², or 10³ physical particles per cell). The supernatants were harvested on days 1, 3, 5, 9, and 11 after transduction, and increasing amounts of biologically active IFNs were detectable by the EMCV bioassay (up to 384 U of IFN-α/ml (Fig. 3a). The expression of wIFN-α was accompanied with the induction of Mx from day 9 onwards, which was detected by Western blot analysis (data not shown). Mx, a 78-kDa protein, is selectively upregulated by IFNs like IFN-α, and is accepted as a sensitive marker of IFN activity (31). WHV replicative intermediates analyzed in parallel were reduced on day 11 after transduction in a dosage-dependent manner, indicating that wIFN-α exhibited antiviral activity (Fig. 3b). WHV replication was not altered after transduction with AdβGal as control vector. Treatment of primary woodchuck hepatocytes with wIFN-γ as a protein exhibited no effect on WHV replication. This was described previously (22).

FIG. 3. Transduction of woodchuck hepatocytes with HD-AdwIFN-α. (a) Detection of biologically active IFN in supernatants of primary hepatocytes from persistently infected woodchucks harvested up to day 11 after transduction with HD-AdwIFN-α (10, (more ...)

Transduction in vivo via the portal vein results in an efficient uptake of HD-Ad vectors in the liver. To test the transduction efficacy of HD-Ad vectors in woodchuck liver, we injected 10¹² physical particles of HD-AdGFP in a volume of 1 ml into the portal vein of one woodchuck. Liver samples that were taken 5 days after transduction were analyzed. GFP was detectable by direct immunofluorescence microscopy of frozen liver samples. Immunohistochemical staining with a MAb against GFP showed an expression of GFP in 90% of Kupffer cells, endothelial cells, and hepatocytes surrounding the central vein, whereas almost no cells near the portal vein were transduced (Fig. 4). The injection of the control vector induced no changes in liver histology; in particular, no influx of lymphocytes was seen (data not shown). GFP expression was also detected in the kidneys but not in cardiac tissue (data not shown). HD-Ad vectors are suitable for transduction of woodchuck livers, and the vector itself induced no immune response on the basis of lymphocyte influx.

FIG. 4. Expression of GFP in the liver of a chronically WHV-infected woodchuck on day 5 after transduction with 109 expressing particles of HD-AdGFP by immunohistochemical staining with a MAb against GFP (arrow points to positively stained cells). Magnification, (more ...)

Quantification of cytokine expression in vivo. After we demonstrated that biologically active cytokines were expressed in vitro and that woodchuck livers can be transduced by HD-Ad vectors in vivo, five chronically WHV-infected woodchucks were transduced with HD-AdwIFN-α or HD-AdwIFN-γ (two and three animals, respectively) by injection of the vectors via the portal vein. During 3 weeks after transduction, the production of wIFN-α and wIFN-γ in the sera of these animals was determined by the virus protection assay. After transduction with HD-AdwIFN-α, up to 512 U of biologically active wIFN/ml was measured in sera by week 2, and after transduction with HD-AdwIFN-γ 16 U of wIFN/ml was measured by week 1 (Fig. 5). This amount of IFN detected after transduction with HD-AdwIFN-γ is comparable to the amount we observed during acute WHV infection in weeks 2 to 6 after inoculation (M. Fiedler, unpublished data). Transduction of woodchucks with HD-AdGFP induced no detectable production of IFN (data not shown).

FIG. 5. Biologically active IFN in sera of woodchucks in the first 3 weeks after transduction of the livers of WHV carriers with 1012 physical particles of HD-AdwIFN-α or HD-AdwIFN-γ. IFN levels were determined by the EMCV bioassay.

Effect of IFN-α and IFN-γ expression in the liver on WHV replication. The influence of transduction with HD-AdwIFN-α on WHV replication was monitored in liver and serum samples over a period of 41 to 47 weeks. A pronounced reduction in WHV replicative intermediates in the liver was observed between weeks 4 and 11 in woodchuck number 13931 (Fig. 6) and between weeks 1 and 7 in woodchuck number 14101 (Fig. 7), respectively. Reduction of replication was accompanied by a reduction of WHV DNA in serum for about 1 log unit: from 8 × 10⁸ to 8 × 10⁷ copies of WHV DNA/ml (woodchuck number 13931) (Fig. 6) and from 2.7 × 10⁸ to 5 × 10⁷ copies/ml (woodchuck number 14101) (Fig. 7). The reduction of WHV DNA levels in sera was transient in woodchuck number 13931; in woodchuck number 14101, WHV DNA fluctuated during followup at levels lower than those seen before transduction. HD-Ad vector DNA was detectable by PCR in all liver biopsies after transduction up to the end of the observation time (data not shown). To demonstrate the successful expression of wIFN-α in the liver, Mx induction was measured. Mx induction was detectable in liver biopsies 1 or 2 weeks after transduction by Western blot analysis and persisted up to the end of the observation period. The amount of Mx expressed, however, declined over time in woodchuck number 13931. An induction of large amounts of Mx was observed up to week 11, respectively (Fig. 6d). In woodchuck number 14101, Mx induction remained at a constant level between weeks 2 and 42 (Fig. 7d). The band detected in week 42 seems stronger than those seen before. This is probably due to a change of the detection technique from staining with peroxidase to detection via luminescence. We demonstrated an HD-Ad vector-mediated delivery of the wIFN-α gene: Mx induction as a marker for production of wIFN-α persisted in the liver for at least 47 weeks. However, only a transient inhibition of WHV replication was observed.

FIG. 6. Transduction of woodchuck number 13931 with 1012 physical particles of HD-AdwIFN-α at week 0. (a) Course of WHV replicative intermediates in livers by Southern blot analysis. RC, relaxed-circular DNA; SS, single-stranded DNA; DS, double-stranded (more ...)

FIG. 7. Transduction of woodchuck number 14101 with 1012 physical particles of HD-AdwIFN-α at week 0. (a) Course of WHV replicative intermediates in livers by Southern blot analysis. RC, relaxed-circular DNA; SS, single-stranded DNA; DS, double-stranded (more ...)

After transduction with HD-AdwIFN-γ, a delayed reduction of WHV replicative intermediates was observed in the liver of one of three animals between weeks 7 and 18 after transduction. This reduction was accompanied with a decline of WHV DNA in the periphery from 5 × 10⁹ in week 1 to 5.8 × 10⁸ in week 16 in woodchuck number 14089 (Fig. 8). The level of WHV DNA remained reduced up to week 43. In animal number 15727, a reduction in the number of replicative intermediates in the liver at week 4 after transduction was detected; however, WHV DNA in sera did not decline but increased between weeks 11 and 16 (Fig. 9). With the third HD-AdwIFN-γ-transduced woodchuck (number 14636), no reduction of replicative intermediates in the liver was observed; however, a late reduction of WHV DNA in serum was measured between weeks 26 and 40, from 4.5 × 10⁸ at week 14 to 6.3 × 10⁷ at week 35, for example (Fig. 10). With all three animals, the vector was detectable throughout the whole observation period after transduction by PCR (23, 43, and 60 weeks, respectively). As expected, Mx was not induced (Fig. 8, 9, and 10). The changes in viremia seen with these three animals were delayed or not in parallel in liver and serum samples. These effects are probably due to a natural fluctuation of viremia observed in chronically WHV-infected woodchucks (see also results obtained with control animals).

FIG. 8. Transduction of woodchuck number 14089 with 1012 physical particles of HD-AdwIFN-γ at week 0. (a) Course of WHV replicative intermediates in livers by Southern blot analysis. RC, relaxed circular DNA; SS, single-stranded DNA; DS, double-stranded (more ...)

FIG. 9. Transduction of woodchuck number 15727 with 1012 physical particles of HD-AdwIFN-γ at week 0. (a) Course of WHV replicative intermediates in livers by Southern blot analysis. RC, relaxed circular DNA; SS, single-stranded DNA; DS, double-stranded (more ...)

FIG. 10. Transduction of woodchuck number 14636 with 1012 physical particles of HD-AdwIFN-γ at week 0. (a) Course of WHV replicative intermediates in livers by Southern blot analysis. RC, relaxed circular DNA; SS, single-stranded DNA; DS, double-stranded (more ...)

To demonstrate the expression of wIFN-γ in vivo after transduction with HD-AdwIFN-γ, we transduced a naïve woodchuck with the same vector and determined if IFN-γ-inducible genes like IP-10 were present in the liver (23). We were able to demonstrate an early and strong induction of woodchuck-specific IP-10 mRNA 1 week after transduction, which decreased clearly in week 2 (see Fig. 13). These results give clear evidence for the expression of biologically active wIFN-γ after transduction with HD-AdwIFN-γ.

FIG. 13. Transduction of a naïve woodchuck (number 15524) with 1012 physical particles of HD-AdwIFN-γ. Detection of mRNA of woodchuck-specific IP-10 and GAPDH by Northern blotting.

Transduction of the livers of two animals with HD-AdGFP as controls had no significant influence on WHV replication markers in livers or sera (woodchuck numbers 15725 and 13933) (Fig. 11 and Fig. 12). Woodchuck number 15725 presented with a fluctuation of WHV DNA levels in serum (e.g., 5 × 10⁹ in week 0 and 9.8 × 10⁸ in week 4), which was not accompanied by a reduction of replicative intermediates in the liver. A fluctuation of viremia of less than 1 log step in serum is a phenomenon which could be observed regularly in chronically WHV-infected woodchucks. A parallel reduction of WHV DNA in serum of about 1 log step and of replicative intermediates in the liver in the same time interval is necessary to define a therapeutic effect. Mx was not induced after transduction with HD-AdGFP.

FIG. 11. Transduction of woodchuck number 15725 with 1012 physical particles of HD-AdGFP at week 0. (a) Course of WHV replicative intermediates in livers by Southern blot analysis. RC, relaxed circular DNA; SS, single-stranded DNA; DS, double-stranded DNA. (b) (more ...)

FIG. 12. Transduction of woodchuck number 13933 with 1012 physical particles of HD-AdGFP at week 0. (a) Course of WHV replicative intermediates in livers by Southern blot analysis. RC, relaxed circular DNA; SS, single-stranded DNA; DS, double-stranded DNA. (b) (more ...)

Side effects of HD-Ad vector-mediated gene delivery in woodchucks. The transduction of livers of woodchucks with the different vectors induced no obvious side effects. The levels of SDH as a marker for liver cell damage fluctuated in all animals, but did not increase significantly after transduction with HD-Ad vectors. Geometric mean concentrations per liter were as follows (ranges are in parentheses): for woodchuck number 13931, 37 U (6 to 57 U); woodchuck number 14101, 70 U (25 to 161 U); woodchuck number 14089, 24 U (7 to 96 U); woodchuck number 15727, 21 U (13 to 22 U); and woodchuck number 14636, 44 U (17 to 153 U). The fluctuation of the concentrations is comparable to that seen normally with chronic carriers (data not shown). In naïve woodchucks, SDH levels tend to be lower (mean, 29 U/liter; range, 6 to 70 U/liter), whereas the concentrations rise up to about 400 to 800 U/liter in acute infection (data not shown).

Transduction of primary hepatocytes of chronically WHV-infected woodchucks in vitro with HD-AdwIFN-α resulted in the secretion of IFN-α, the induction of Mx expression, and a reduction of WHV replication. These promising in vitro data encouraged us to investigate the effects of HD-AdwIFN-α on WHV replication in chronically WHV-infected woodchucks. We transduced livers of woodchucks with the wIFN-α gene in vivo in chronic WHV infection and demonstrated for the first time the delivery of cytokine genes by HD-Ad vectors into the liver during chronic hepadnaviral infection. HD-Ad vector DNA persisted at least for 1 year in the livers of these animals. After transduction, wIFN-α became detectable in the sera of the animals, Mx expression was induced in the liver, and WHV replication was reduced. A decrease of viral particles determined by WHV DNA in sera of about one log and of WHV replicative intermediates in livers was induced; however, this was a transient effect up to weeks 7 and 11, respectively. The reduction of replication was associated with an induction of Mx in the livers of these animals. Mx was detectable as early as 1 or 2 weeks after transduction, and its expression persisted up to the end of the observation period for nearly 1 year; however, the amount of Mx declined over time in one of two animals. A close association between intrahepatic IFN-α expression and Mx upregulation has been demonstrated previously by immunostaining procedures in human livers (18). The half-life of Mx is about 2 days (33). Therefore, we can assume that the expression of Mx can be correlated with an ongoing expression of wIFN-α after transduction with HD-AdwIFN-α. Thus, wIFN-α was expressed at lower levels at the end of the observation period. The IFN-α level was still high enough to induce Mx but too low to mediate antiviral effects. The reduction of IFN-α expression could possibly be due to a downregulation of the promoter. The role of Mx as mediator of the antiviral activity of IFN-α in chronic HBV infection has been the subject of controversial discussions in recent years. Gordien at al. demonstrated the inhibition of HBV replication by Mx in Huh7 cells at a posttranscriptional level (6). Rang et al., however, were able to suppress HBV replication by IFN-α in Mx-deficient HEp2 cells, indicating that Mx is dispensable for the antiviral activity of IFN-α (29). Our results are compatible with both theories; therefore, the correlation between Mx expression and reduction of hepadnavirus replication remains an open question.

Gene transfer of wIFN-γ in vivo by transduction with an HD-Ad vector also resulted in the production of biologically active IFN. The induction of IFN-inducible IP-10 in a naïve woodchuck transduced with HD-AdwIFN-γ gave clear evidence for in vivo expression of wIFN-γ. Despite our demonstration, in principle, of the response of liver cells to the production of wIFN-γ after transduction, we found no reduction in WHV replicative intermediates in the liver or viral particles in sera of these animals. In week 2 after transduction of the naïve woodchuck, the expression was already strongly reduced (Fig. 13). We can only speculate that this downregulation could be a physiological reaction to avoid the toxic effects of IFN-γ in vivo. The amount of IFN detectable in sera after transduction of chronically WHV-infected woodchucks by the EMCV bioassay was low, however, compared to that seen in sera of woodchucks in acute WHV infection (Fiedler, unpublished). The difference in the amount of IFN measured in serum after transduction with HD-AdwIFN-α can be caused by the different activities of these two cytokines measured in cytopathic effect inhibition assays. We observed previously that wIFN-γ did not reduce WHV replication in primary hepatocytes of persistently infected woodchucks, whereas it suppressed WHV replication in in vitro-infected primary hepatocytes in naïve animals (22). This lack of effect on hepadnavirus replication in vitro may be due to the absence of immunomodulatory cells. On the other hand, not only direct antiviral effects of IFN-γ but also indirect mechanisms can induce a reduction of replication in vivo. In chronically WHV-infected woodchucks, however, wIFN-γ was not able to mediate a significant reduction in WHV replication. The delayed fluctuation of viremia observed in this study can most probably be attributed to the normal fluctuation of viremia seen in chronically WHV-infected woodchucks. We suppose that the hepatocytes of chronically WHV-infected woodchucks may be functionally altered in their response to wIFN-γ. In vivo inflammatory cytokines, including IFN-γ, were shown to be elevated in the livers of WHV carriers compared to levels in livers of naïve animals (12, 26). The continuous exposure of the hepatocytes to IFN-γ during chronic WHV infection may induce an adaptation to the presence of IFN-γ. This hypothesis is supported by results showing that hepatocytes in WHV carriers expressed higher levels of MHC-I heavy-chain mRNAs than did those in naïve animals (22, 26). In contrast, a transfer of HBV-specific cytotoxic T lymphocytes in the HBV transgenic mouse model reduces HBV replication mainly via secretion of IFN-γ (10). Although HBV transgenic mice show liver-specific HBV-gene expression and replication, they are not exposed to inflammatory cytokines before immune transfer. Hepatocytes of HBV transgenic mice are, therefore, naïve to IFN-γ and are not adapted to long-term elevations in IFN-γ levels.

After transduction, expression of GFP or cytokines was predominantly found in liver cells around the central vein, as shown with an HD-AdGFP-transduced control animal. It is difficult to assess the influence of the local production of cytokines on liver cells in areas with a high percentage of transduced cells versus the influence on cells in other regions. This depends on the distribution and the half-life of the cytokines in the liver. We can only speculate if the irregular transduction of liver cells had any influence on the therapeutic effect of wIFN-γ or wIFN-α.

Local expression of IFN-α in chronic hepadnavirus infection did not exhibit a significant and long-lasting therapeutic effect on WHV replication. The way that wIFN-α was administered in our study seemed to have no advantage over subcutaneous administration of IFN-α. Combination therapy with lamivudine and/or serial injections of the vector would probably be more effective. Transduction with HD-AdwIFN-α, however, could prevent the development of HCC. The risk of HCC is 100 times higher in HBV-infected patients than in uninfected patients (2). Epidemiological observations of humans (15) and preliminary data from an HCC transgenic mouse model (25) seem to justify this gene therapy approach. Long-term administration of IFN-α was associated with a decrease or a delay of oncogenesis in these mice (25). By transduction of chronically WHV-infected woodchucks with HD-AdwIFN-α, HCC development may be postponed.

In conclusion, we have established a genetic approach for expressing cytokines in the liver in chronic hepadnavirus infection. Prerequisites for successful realization for this and future studies are the already-described lack of a vector-induced immune response, accompanied with the long-term expression of the gene products of the HD-Ad vectors that we also observed with the woodchuck model. In future experiments, the procedures will be optimized: repeated injections of the vector and a combination of cytokine gene delivery with antiviral drugs are under consideration. Mechanisms which could lead to a silencing of expression of the gene product have to be investigated. An optimized long-term expression of IFN-α by transduction with HD-AdwIFN-α may reduce replication and prevent the development of HCC in chronic hepadnavirus infections.

This work was supported by the Dr. Mildred Scheel Stiftung für Krebsforschung (grant 10-2067-Ro 1). Technician Anja Busse was paid by this grant.

We thank O. Haller (Department of Virology, Institute of Medical Mikrobiology and Hygiene, University of Freiburg, Freiburg, Germany) for providing the antibody against Mx, L. Vollbracht (Institute of Clinical Chemistry, University Clinic Essen, Essen, Germany) for performing the SDH assays, and S. Viazov (Institute of Virology, University Clinic Essen) for critical reading of the manuscript. The excellent technical assistance of A. Busse and S. Zimmermann is gratefully appreciated.