Expression of hepatitis c virus proteins inhibits interferon α signaling in the liver of transgenic...

Expression of Hepatitis C Virus Proteins Inhibits Interferon �Signaling in the Liver of Transgenic Mice

ALEX BLINDENBACHER,* FRANCOIS H. T. DUONG,* LUKAS HUNZIKER,‡

SIMONE T. D. STUTVOET,* XUEYA WANG,* LUIGI TERRACCIANO,§ DARIUS MORADPOUR,�

HUBERT E. BLUM,� TONINO ALONZI,¶ MARCO TRIPODI,¶,# NICOLA LA MONICA,** andMARKUS H. HEIM*,‡‡

*Department of Research, University Hospital Basel, Switzerland; ‡Institute for Experimental Immunology, University Hospital, Zurich,Switzerland; §Institute of Pathology, University Hospital Basel, Switzerland; �Department of Medicine II, University Hospital Freiburg, Freiburg,Germany; ¶Istituto Nazionale Malattie Infettive I.R.C.C.S., Roma, Italy; #Fondazione Istituto Pasteur Cenci-Bolognetti, Dipartimento diBiotecnologie Cellulari ed Ematologia, Universita di Roma, Italy; **Istituto di Ricerche di Biologia Molecolare P. Angeletti, Pomezia, Italy; and‡‡Department of Gastroenterology, University Hospital Basel, Switzerland

Background & Aims: Hepatitis C virus (HCV) is a majorcause of chronic liver disease, cirrhosis, and hepatocel-lular carcinoma worldwide. The majority of patientstreated with interferon alpha do not have a sustainedresponse with clearance of the virus. The molecularmechanisms underlying interferon resistance are poorlyunderstood. Interferon-induced activation of the Jak-STAT (signal transducer and activator of transcription)signal transduction pathway is essential for the induc-tion of an antiviral state. Interference of viral proteinswith the Jak-STAT pathway could be responsible forinterferon resistance in patients with chronic HCV.Methods: We have analyzed interferon-induced signaltransduction through the Jak-STAT pathway in trans-genic mice that express HCV proteins in their liver cells.STAT activation was investigated with Western blots,immunofluorescence, and electrophoretic mobility shiftassays. Virus challenge experiments with lymphocyticchoriomeningitis virus were used to demonstrate thefunctional importance of Jak-STAT inhibition. Results:STAT signaling was found to be strongly inhibited in livercells of HCV transgenic mice. The inhibition occurred inthe nucleus and blocked binding of STAT transcriptionfactors to the promoters of interferon-stimulated genes.Tyrosine phosphorylation of STAT proteins by Janus ki-nases at the interferon receptor was not inhibited. Thislack in interferon response resulted in an enhancedsusceptibility of the transgenic mice to infection with ahepatotropic strain of lymphocytic choriomeningitis vi-rus. Conclusions: Interferon-induced intracellular signal-ing is impaired in HCV transgenic mice. Interference ofHCV proteins with interferon-induced intracellular signal-ing could be an important mechanism of viral persis-tence and treatment resistance.

The majority of patients infected with hepatitis Cvirus (HCV) develop chronic liver disease that may

progress to cirrhosis and hepatocellular carcinoma.1–3

The mechanisms by which the virus evades clearance bythe host immune system, despite a polyclonal and mul-tispecific T-cell response and the presence of antibodiesagainst all viral proteins, are poorly understood.4 There isgood evidence in hepatitis B virus and lymphocyticchoriomeningitis virus (LCMV) infection that noncyto-lytic, cytokine-mediated antiviral mechanisms play a keyrole in viral clearance from liver cells.5,6 Interferons(IFNs) induce an antiviral state in hepatocytes throughthe transcriptional induction of several dozen targetgenes in the cell nucleus. After binding to their cognatereceptors on the cell membrane, IFNs activate the Jak-STAT (signal transducer and activator of transcription)signal transduction pathway.7,8 Type I IFNs (IFN� and�) bind to heterodimeric receptors consisting of IFN�receptor 1 and 2 (IFNAR1, IFNAR2). Ligand bindingresults in the activation of 2 receptor-associated proteintyrosine kinases, Jak1 and Tyk2. The activated receptorkinase complex then recruits and activates signal trans-ducer and activator of transcription 1 (STAT1), STAT2,and STAT3.9,10 STAT1 and STAT2 combine with athird transcription factor, p48 or interferon responsefactor 9, to form interferon-stimulated gene factor 3(ISGF3).7 ISGF3 binds to interferon-stimulated responseelements in the promoters of IFN-stimulated genes(ISGs). Alternatively, IFN-activated STAT1 and STAT3

Abbreviations used in this paper: CTL, cytotoxic T lymphocytes;EMSA, electrophoretic mobility shift assay; IFN, interferon; IFNAR,interferon� receptor; ISG, interferon-stimulated gene; ISGF3, inter-feron-stimulated gene factor 3; Jak, Janus kinase; LCMV, lymphocyticchoriomeningitis virus; LCMV-WE, lymphocytic choriomeningitis virus-strain WE; ORF, open reading frame; PIAS1, protein inhibitor of acti-vated signal transducer and activator of transcription 1; pfu, plaqueforming units; SOCS, suppressor of cytokine signaling; STAT, signaltransducer and activator of transcription; TNF, tumor necrosis factor.

© 2003 by the American Gastroenterological Association0016-5085/03/$30.00

doi:10.1016/S0016-5085(03)00290-7

GASTROENTEROLOGY 2003;124:1465–1475

can form homodimers or STAT1-STAT3 heterodimers.These STAT dimers bind a different class of responseelements, the gamma-activated sequence elements. Oncebound to the promoters of ISGs, STATs induce thetranscription of genes involved in the generation of anantiviral state.8,11 Signaling through the Jak-STAT path-way is of critical importance because most IFN-inducedcellular effects are lost on deletion of STAT1 in mice12,13

or in cell lines.14 Therefore, interference with Jak andSTATs could be a mechanism of viral persistence, andindeed, viruses such as adenovirus, cytomegalovirus,mumps virus, and hepatitis B virus have been shown tointerfere with Jak-STAT signaling.15–17 We hypothe-sized that interference with IFN signaling could enableHCV to escape the innate immune system. In previouswork we have found evidence for such an interferencewith Jak-STAT signaling in human osteosarcoma celllines with an inducible transgene comprising the com-plete open reading frame (ORF) of a HCV genotype 1aisolate.18,19 When viral proteins are expressed in thesecells, IFN-induced activation of ISGF3 is blocked andthe activation of the STAT1-STAT3 dimers is inhibited.Here, we report our analysis of IFN-induced Jak-STATsignaling in liver cells of mice that express HCV proteinsfrom a transgene with a liver-specific promoter. We havefound a strong inhibition of ISGF3 activation in thesemice. When infected with LCMV-strain WE (LCMV-WE), the transgenic mice could not control the viralreplication and suffered from a severe hepatitis. An in-hibition of the IFN-induced antiviral defenses in hepa-tocytes might enable HCV to establish a chronic infec-tion in the majority of patients, and could contribute toresistance to treatment with IFN�.

Materials and MethodsAnimals

A 9.0-kilobase cDNA fragment derived from plasmidpCD(38–9.4)20 that covers the HCV ORF (aa 3 to 3010) wasused for the generation of the transgenic mice. The HCVcDNA was inserted into a A1AT minigene by homologousrecombination in bacterial cells.21 Donor plasmid p(PucII-V exA1AT) was first generated. This construct carries exons II, IV,and V and intron IV of the A1AT gene, and it contains amultiple cloning site between exon II and exon IV. The HCVcDNA, spanning from nucleotide 339 to 9416, flanked by twoPac I sites was then inserted into the Pac I site of plasmidp(PucII-V ex A1AT) such that the transgene ORF starts fromthe ATG of A1AT (in exon II), coding a fusion protein formedby 32 additional aa at the N-terminus of the HCV polyprotein.DH1 cells were used for the homologous recombination usingthe recombinant cosmid containing the entire A1AT gene asacceptor as described.22 Transgenic C57BL/6 were generatedby pronuclear microinjection in one-cell stage embryos of

inbred C57BL/6 by BRL (BRL, Fullinsdorf, Switzerland) asdescribed.23 Genotype analysis was performed by Southernblot analysis on 10 �g of total genomic DNA extracted fromtail specimens using a HCV cDNA fragment probe (spanningE2-NS3). A full description of the mice will be publishedelsewhere. The inbred C57BL/6 control mice were purchasedfrom the same provider that generated the transgenic mice(BRL). Mice were bred in a specific pathogen-free environmentunder standard conditions. For all experiments described inthis article, 7 to 14-week-old animals were used, and controlsand transgenic mice were age and sex matched. The study wasapproved by the animal care committee of the Kanton Basel.

Northern Blot

RNA was isolated from frozen tissues using Trizolreagent (Life Technologies, Gaithersburg, MD) according tothe manufacturer’s instructions. Total RNA (20�g) was usedin Northern blot analysis. HCV cDNA fragment includingregions E2-NS3 and a 28S ribosomal RNA fragment were usedas a probe and labelled by random priming.

Western Blots

For the Western blot analysis, whole cell extracts wereprepared from the middle lobe of the mouse liver. The lobe washomogenized in lysis buffer (50 mmol/L Tris pH 8, 280mmol/L NaCl, 0.5% NP40, 0.2 mmol/L EDTA, 2 mmol/LEGTA, 10% glycerol, 100 �mol/L NaVO4, 1 mmol/L PMSF,and proteinase inhibitors), incubated for 30 minutes on arotating wheel at 4°C, and centrifuged at 14,000g for 5minutes. To analyze the liver extracts for the presence of HCVcore protein, 75 �g of total protein per sample was loaded ontoa 12% SDS-PAGE gel. Membranes were probed with themurine monoclonal antibody C7-50 directed against a con-served, linear epitope in the aminoterminal region of the HCVcore protein.24 For Western blot analysis of STAT phosphor-ylation, 100 �mol/L NaVO4 was added to the lysis buffer.Tyrosine phosphorylation of STAT1 and STAT3 was testedusing the phospho-STAT1–specific antibody 9171S and thephospho-STAT3–specific antibody 9173S (New England Bio-Labs, Inc, Beverly, MA). Total amounts of STAT1 and STAT3were determined with polyclonal antibodies sc346 and sc482,respectively (both from Santa Cruz Biotechnology, Inc, SantaCruz, CA). For cytoplasmic/nuclear extracts, mouse livers werehomogenized in 200 �L of buffer A containing 20 mmol/LHepes pH 7.6, 10 mmol/L NaCl, 3 mmol/L MgCl2 , 0.1%NP-40, 10% glycerol, 0.2 mmol/L EDTA, and 1 mmol/LDTT. After centrifugation at 2000 rpm for 5 minutes, super-natant (cytoplasmic extracts) was removed and the pellet waswashed with 200 �L of buffer B containing 20 mmol/L HepespH 7.6, 0.2 mmol/L EDTA, 20% glycerol, and 1 mmol/LDTT, and centrifuged at 3000 rpm for 5 minutes. The pelletwas then resuspended in 150 �L of buffer C containing 20mmol/L Hepes pH 7.6, 400 mmol/L NaCl, 0.2% EDTA, 20%glycerol, and 1 mmol/L DTT. This nuclear extract was clearedby centrifugation and used for Western blot analysis.

1466 BLINDENBACHER ET AL. GASTROENTEROLOGY Vol. 124, No. 5

Immunostaining of Liver Sections

Cryosections were prepared at a thickness of 7 �m andfixed in 4% paraformaldehyde. They were then permeabilizedwith 0.1% Triton �-100, blocked with 3% bovine serumalbumin, and incubated overnight at 37°C with phospho-STAT1–specific antibody 9173S (New England Biolabs, Inc)or anti-core monoclonal antibody C7-50.24 After severalwashes, sections were incubated with secondary antibodieslabeled with Cy3 or FITC for 90 minutes at room temperature(Molecular Probes, Inc, Eugene, OR). Nuclei were stainedwith Hoechst stain diluted to 1:200 and sections weremounted with Fluorsave Reagent (Calbiochem, La Jolla, CA).

Electrophoretic Mobility Shift Assays

Mice were injected with the indicated doses of murineIFN� (Calbiochem) or tumor necrosis factor� (TNF�; Sigma,Buchs, Switzerland) into the tail vein. The animals wereeuthanatized 20 minutes later. Nuclear extracts and electro-phoretic mobility shift assay (EMSA) were performed asdescribed previously.25 The oligonucleotides used have thefollowing sequences: IFN-stimulated response elements, 5�-GAAAGGGAAACCGAAACTGAAGC-3�; m67, 5�-CATTTC-CCGTAAATCAT-3�; NF�B consensus site, AGTTGAGGG-GACTTTCCCAGGC-3�. The experiments were repeated severaltimes, also using murine IFN� from a different provider (Sigma).

Lymphocytic Choriomeningitis Virus (LCMV)Infections

LCMV-WE was originally obtained from Dr. F. Leh-mann-Grube (Hamburg, Germany) and was propagated onL929 cells. C57BL/6 and B6 HCV transgenic mice wereinfected IV with LCMV-WE with the indicated dose. Virustiters in the liver were determined in an LCMV infectiousfocus assay as previously described.26 For the LCMV footpadswelling reaction, mice were infected with 200 plaque formingunits (pfu) LCMV-WE into the hind footpad and the swellingwas measured daily with a spring-loaded caliper. Assays forserum concentration of ALT and AST were performed at theDepartment of Clinical Chemistry, University Hospital Zu-rich, using photometric assays on a Hitachi 747 autoanalyzer(Tokyo, Japan).

ResultsHCV Transgenic Mice

To assess the effects of HCV proteins on IFN-induced intracellular signaling in liver cells in vivo, wehave generated a transgenic C57BL/6 mouse line carry-ing the entire ORF of a HCV genotype 1b isolate27 underthe control of an �1-antitrypsin promoter. The trans-genic mice were designated B6HCV. B6HCV mice arehealthy. With advanced age, more than 70% of the micedevelop fatty liver and, interestingly, a diffuse T-cellinfiltration is observed in more than half of the mice.These observations are the subject of a separate report.

Because both liver steatosis and lymphocyte infiltratescould potentially influence signaling through the Jak-STAT pathway, 7 to 14-week-old animals were used forall our experiments. In these younger mice, analysis offormalin-fixed liver sections showed no inflammatoryinfiltrates or liver steatosis (Figure 1A ). Controls wereage and sex matched. In liver cell extracts of B6HCVmice, a single full-length mRNA transcript of the trans-gene of approximately 10 kilobase can be detected byNorthern blot (Figure 1B), and strong expression ofHCV core protein was found by Western blot (Figure1C ). The core protein is fused on its N-terminus to 32 aaof �1-antitrypsin, and has a calculated molecular weightof 24 kilodalton. We could not detect additional HCVproteins by Western blot. Immunostaining of liversections showed HCV core expression in hepatocytes(Figure 1D).

Similar to findings in other transgenic mice, the ex-pression level of the transgene is not uniform. In somehepatocytes, no core protein can be detected, while othersshow strong signals. Core protein is found both in thecytoplasm and the nucleus of hepatocytes. The number ofhepatocyte nuclei that stained positive was counted inliver sections from 3 mice. The mean percentage ofnuclei containing core protein was 29% (with a standarderror of the mean of 6%). Nuclear localization of theHCV core protein is a debated issue. While some authorshave found core protein only in the cytoplasm,28 othershave found an additional nuclear localization of the pro-tein in vitro29 and in transgenic mice in vivo.30 In ourtransgenic mouse model, the high expression level of coreprotein in a subgroup of hepatocytes might allow thedetection of core protein in the nucleus. Alternatively,the fusion of the first 32 aa from �1-antitrypsin to theN-terminus of core protein might have changed thesubcellular distribution of core.

IFN-Induced Jak-STAT Signaling

In the first set of experiments, B6HCV and con-trol C57BL/6 mice were injected with murine IFN�(mIFN�) into the tail vein (Figure 2A ). The animalswere euthanatized 20 minutes later. Nuclear extractswere prepared from liver cells and tested for ISGF3induction by EMSA using an IFN-stimulated responseelement oligonucleotide probe. A total of 12 controlmice and 16 B6HCV mice were used in 4 independentexperiments with 5–10 mice each. In C57BL/6 mice, astrong ISGF3 shift was detected after treatment with1000 IU mIFN�/g body weight, whereas the same dosewas ineffective in B6HCV mice (Figure 2A, lanes 6, 7,and 8, respectively). This inhibition of STAT signalingcould be overcome by higher doses of mIFN�. Theintravenous injection of 4000 IU mIFN�/g body weight

May 2003 HCV INTERFERENCE WITH INTERFERON SIGNALING 1467

Figure 1. Expression of HCV core protein in livercells of B6HCV mice. (A) Formalin-fixed liver sec-tions of C57BL/6 and of B6HCV mice were stainedwith hematoxylin-eosin according to standard pro-cedures. Neither inflammatory infiltrates nor liversteatosis is present in these 14-week-old animals.The yellow size bars indicate 100�m. (B) Liver-specific expression of the A1AT-HCV transgene.Northern blot analysis was performed using RNAextracted from 1-month-old transgenic mouse tis-sues probed with a HCV-specific probe. Twenty �gof total RNA were loaded per lane. A probe for the28S ribosomal RNA was used as an internal con-trol. Sg, salivary glands; M, muscle; Sk, skin; H,heart; K, kidney; Sp, spleen; L, liver of transgenicmice; C, control (liver of wild-type mouse). (C) Livercell extract from a C57BL/6 (lane 1) and a B6HCVmouse (lane 2) were analyzed for the presence ofHCV core protein by Western blot analysis. Theasterisk (*) indicates the core protein. The posi-tions of the molecular weight markers are indicatedto the left. (D) HCV core protein can be detected byimmunostaining in liver cells of B6HCV mice. Coreprotein is present both in the cytoplasm and in thenucleus (arrows). The size bars indicate 20�m.


induces a strong STAT activation as shown by EMSAwith the m67 probe (Figure 2B). The inhibition of STATsignaling is not due to nonspecific toxic effects of theviral proteins because signaling of TNF� through NF�Bwas not inhibited (Figure 2C ).

The absence of ISGF3 and STAT1 and STAT3 ho-modimer gel shift complexes could result from a lack ofactivation of STATs through tyrosine phosphorylation atthe IFN� receptor kinase complex. The phosphorylationof STATs on a single tyrosine residue carboxy-terminal oftheir SH2 domain is required for dimerization and nu-clear translocation.31 In HCV protein containing livercells of B6HCV mice, this decisive STAT activationevent could be impaired through a down-regulation ofIFNAR1, IFNAR2, Jak1, Tyk2, or the STATs, orthrough the induction of suppressor of cytokine signal-ing 1 or 3 (SOCS1 or SOCS3).32 However, tyrosinephosphorylation of STAT1 and STAT3 after injection of1000 IU mIFN�/g body weight is not impaired in theliver of B6HCV mice compared with C57BL/6 controlmice (Figure 3A ). This observation is in accordance withour previous findings in inducible cell lines in vitro.18

Viral proteins could also impair the nuclear translo-cation of phosphorylated STATs. Such a block in nuclearimport in the presence of a normal STAT1 tyrosinephosphorylation at the receptor-kinase complex would

result in an accumulation of phospho-STAT1 in thecytosol and a relative lack phospho-STAT1 in the nu-cleus. However, in both controls and B6HCV mice, mostphosphorylated STAT1 molecules were found in nuclearextracts and not in cytoplasmic extracts (Figure 3B). Weconclude that nuclear translocation is not significantlyimpaired in liver cells of HCV transgenic mice. Alter-natively, viral proteins could stimulate the dephosphor-ylation of STATs in the nucleus through the up-regula-tion of one or more as yet unidentified phosphatases. Toaddress this issue with a more quantitative method thanWestern blot detection of phospho-STAT1 in cytoplas-mic and nuclear extracts, we stained cryosections fromlivers of 5 C57BL/6 and 4 B6HCV mice, respectively,with a phosphospecific STAT1 antibody (Figure 3C ) andcounted the number of cells that showed nuclear local-ization of phosphorylated STAT1. All mice have beeninjected intravenously with 1000 IU mIFN�/g bodyweight and euthanatized 20 minutes later. Approxi-mately 2000 cells were counted for each mouse. Theaverage percentage of phospho-STAT1–positive nucleiwas 48.0% in C57BL/6 mice (standard deviation, 8.3;standard error of the mean, 3.7), but only 31.3% inB6HCV mice (standard deviation, 7.9; standard error ofthe mean, 2.3) (Figure 3C ). The absence of detectableSTAT gel shifts in nuclear extracts from B6HCV mice

Figure 2. Activation of STATs after IV cytokine injections in B6HCV and C57BL/6 mice. (A) Nuclear extracts from liver cells were tested by EMSAwith the interferon-stimulated response elements oligonucleotide. An ISGF3 shift (arrow) is readily detected in C57BL/6 mice injected with 1000IU mIFN�/g body weight, but not in B6HCV mice. C57BL/6 and B6HCV mice were injected with PBS, 100 IU mIFN�/g body weight, or 1000 IUmIFN�/g body weight as indicated. (B) EMSA with the m67 oligonucleotide. The positions of gel shift complexes containing STAT3 homodimers,STAT1-STAT3 heterodimers, and STAT1 homodimers are indicated. No induction of STAT gel shift complexes can be detected in both C57BL/6and B6HCV mice after injection with 250 IU mIFN�/g body weight (lanes 1 and 2, respectively). Injection of 1000 IU mIFN�/g body weight inducesSTAT dimers in a C57BL/6 mouse (lane 3) but not in a B6HCV mouse (lane 4). In both the C57BL/6 (lane 5) and the B6HCV mouse (lane 6),STAT gel shifts are induced by 4000 IU mIFN�/g body weight. (C) TNF�-induced NF�B activation is not impaired in B6HCV mice. EMSA with NF�Bconsensus oligonucleotide. Injection of 5 ng TNF�/g body weight results in NF�B shifts (arrow) in both a C57BL/6 mouse (lane 1) and in twoB6HCV mice (lanes 2 and 3). PBS controls are shown in lane 4 (C57BL/6) and lane 5 (B6HCV).


could be explained by this reduction of nuclear phospho-STAT1 because the sensitivity and detection limit of theEMSA with liver cell nuclear extracts is not known.Alternatively, the DNA binding of phosphorylatedSTAT1 complexes could be inhibited by viral proteins or

by endogenous inhibitors such as protein inhibitor ofactivated STAT1 (PIAS1) which would be up-regulatedin the liver cells of B6HCV mice. By Western blotanalysis with PIAS1 antibodies we could not detectincreased expression of PIAS1 in liver cells of B6HCV

Figure 3. Tyrosine phosphorylation and nuclear translocation ofSTATs. (A) Tyrosine phosphorylation of STATs is not impaired inB6HCV mice compared with C57BL/6 mice. Mice were injected IV withPBS or with 1000 IU mIFN�/g body weight and euthanized 20 minuteslater. The arrows indicate STAT1 and STAT3, respectively. The posi-tion of the 116 kilodalton and 85 kilodalton size markers is shown. (B)Nuclear import of phospho-STAT1 is not inhibited in B6HCV mice.Cytoplasmic as well as nuclear extracts were prepared from livers ofcontrol mice and HCV transgenic mice 20 minutes after injection withPBS or 1000 IU mIFN�/g body weight as indicated. Tyrosine phos-phorylated STAT1 was detected by Western blot analysis with phos-pho-STAT1–specific antibodies. (C) Phospho-STAT1 nuclear transloca-tion is impaired in B6HCV mice. The scale bar represents 50 �m.Hepatocytes with phospho-STAT1 nuclear localization were countedunder a fluorescent microscope. Five C57BL/6 mice and 4 B6HCVmice were analyzed. For each animal, 10 areas were randomly chosenfor counting. About 2000 cells were counted per animal. The histo-gram shows the average percentage of cell nuclei that stain positivewith the phospho-STAT1–specific antibody 9171S in C57BL/6 andB6HCV mice, respectively. The error bars represent the 95% confi-dence interval for the mean values.

Figure 4. Increased susceptibility of B6HCV mice to LCMV-WE infec-tion. (A) B6HCV (red circles) and C57BL/6 (blue rectangles) micewere injected IV with the indicated doses of LCMV-WE. Serum ALTvalues (U/L) were determined at day 0, 4, 8, and 12. Three mice eachwere used per group and per LCMV-WE dose. The rectangles andcircles represent mean values; the bars show the standard error ofthe mean. (B) All C57BL/6 and B6HCV mice were euthanatized 12days after IV injection of 2 � 105 pfu of LCMV. Formalin-fixed liversections were stained with H&E according to standard procedures.Representative samples are shown. Whereas C57BL/6 mice show nosigns of hepatitis, a moderately severe hepatitis with slight lobulardisarray, diffuse intrasinusoidal mononuclear infiltrates, and apopto-tic hepatocytes (arrow) are seen in B6HCV mice. The yellow size barsindicate 100 �m. (C) Virus titers were measured 12 days afterinfection with the indicated doses. Results from B6HCV mice areshown in red, controls in blue. Two mice each were used per groupand LCMV-WE dose. The bars are standard error of the means. Theassay has a limit of detection at 4.2 log 10 pfu (shown as yellow bar).After injection of 2 � 103 pfu in both groups and after injection of 2 �104 pfu in controls, the values were below the limit of detection.


mice (data not shown). However, yet unknown inhibitorsof STAT DNA binding could be involved in the ob-served inhibition of STAT gel shifts.

Impaired antiviral response in vivo. We nexttested the consequences of impaired IFN signaling forhost resistance to a viral infection. Susceptibility toLCMV has been shown to be greatly increased in micethat lack IFNAR1.33 The hepatotropic strain LCMV-WEis a noncytopathic RNA virus that causes a dose-depen-dent immunpathology. Liver damage is mediated by acytotoxic T-cell response.34 We hypothesized that theinhibition of IFN signaling by HCV proteins in livercells of B6HCV transgenic mice would increase the viralreplication, and consequently result in an increased im-munopathologic liver inflammation. Therefore, trans-genic and control mice were injected IV with variousdoses of the hepatotropic LCMV-WE. LCMV-WE in-fects hepatocytes, Kupffer, and sinusoidal endothelialcells.35 In support of our hypothesis, B6HCV mice suf-fered a more severe hepatitis (Figure 4).

After injection of 2 � 104 pfu LCMV or 2 � 105 pfuLCMV C57BL/6 controls had only a slight increase ofserum amino transferase levels at day 8 after infectionwith a return to normal values at day 12 (Figure 4A ).The liver histology of C57BL/6 mice showed a minimalmononuclear infiltrate in liver sinusoids at day 8, and anormal liver at day 12 (Figure 4B). On the contrary,B6HCV mice had very high amino transferase values,and liver histology showed a more severe hepatitis withdiffuse hydropic swelling and ballooning of hepatocytes,diffuse intrasinusoidal mononuclear infiltrates, and apo-ptotic hepatocytes (Figure 4).

A likely explanation for the observed difference be-tween controls and HCV transgenic mice is our findingof an increased LCMV replication in livers of B6HCVmice (Figure 4C ). When measured 12 days after intra-venous infection of 2 � 105 pfu LCMV, the difference inviral replication was up to 100-fold. This increased viralreplication is probably a consequence of the inhibition ofthe IFN-induced antiviral defense caused by the inter-ference of HCV proteins with the Jak-STAT pathway.Alternatively, the transgene insertion could have inci-dentally resulted in an ineffective cytotoxic T lympho-cytes (CTL) response against LCMV, causing an insuffi-cient CTL-mediated control of LCMV replication.However, local infection with LCMV resulted in a nor-mal CTL-mediated immunopathology in HCV trans-genic mice, as shown by footpad swelling after injectionof 200 pfu LCMV-WE in the foot of B6HCV and controlmice. Both groups showed a similar increase of 150%–170% 8 days after infection (data not shown). Furtherevidence for an unimpaired immune response in B6HCV

mice was obtained with vesicular stomatitis virus chal-lenge experiments. Infection of mice with vesicular sto-matitis virus, a neurotropic rhabdovirus with a highmortality rate in mice lacking the IFN� receptor,33 wascontrolled by an equally efficient immune response inboth C57BL/6 and B6HCV mice. Of 3 B6HCV and 3control mice infected with 2 � 106 pfu vesicular stoma-titis virus, all survived (data not shown).

In an observational study with 12 B6HCV and 12control mice infected with various doses of LCMV-WE,6 of 12 B6HCV mice and 2 of 12 controls died between10 and 12 days after infection. The death of the controlsis in the range of the CTL-mediated immunopathologyusually observed in wild-type mice (around 10–15%mortality). In the case of the B6HCV mice, we believethat the liver disease contributed to the excess mortality.

Interference of viral proteins with nuclear local-ization of STATs. Our observation of a nuclear localiza-tion of the HCV core protein in about 30% of thehepatocytes (Figure 1C ) and the 35% reduction of phos-pho-STAT1 localized in the nucleus after IFN� treat-ment in B6HCV mice as compared with C57BL/6 mice(Figure 3C ) suggested that the 2 phenomena might beinterdependent. Therefore, we performed double stainingexperiments of liver sections from C57BL/6 and B6HCVmice after mIFN� treatment. Interestingly, most cellsthat were strongly stained with antibodies to HCV coreprotein had no detectable phospho-STAT1 in the nucleus(Figure 5B, D, and F ).

Discussion

To establish a persistent infection in humans,HCV must, at least to a degree, evade the antiviralcellular defense mechanisms induced by IFNs. The verysame defense strategies could also contribute to the ob-served resistance to IFN-based treatments in the majorityof patients with chronic hepatitis C. Over the past severalyears, several groups have reported that HCV proteinsNS5A or E2 can interfere with the IFN-induced antiviraleffector protein PKR.36–38 These results have been ob-tained by biochemical methods using purified proteins orby analyses of transfected mammalian or yeast cells, andthe significance of these findings for chronic hepatitis Cin patients remains to be determined. As an alternativeto the inhibition of antiviral effector systems, the viruscould already interfere with IFN-induced signaling andthereby prevent the induction of these systems. BecauseJak-STAT signaling is required for IFN induction of anantiviral state, such an inhibition could be a very efficientdefense strategy for HCV. We report here that HCVproteins expressed from a transgene in mouse liver cells


indeed inhibit signaling through the Jak-STAT path-way.

Because HCV cannot be propagated in cultured cells,and small animal models for chronic hepatitis C are stillnot available, we have generated a transgenic mouse thatconstitutively expresses the entire ORF of a genotype 1bHCV isolate. These mice provide an opportunity tostudy interference of HCV proteins with IFN-inducedsignaling or with IFN-induced effector systems in livercells in vivo, but they have some properties that shouldbe kept in mind. First, the expression level of the HCVtransgene in hepatocytes is heterogeneous. We do nothave an explanation for the patchy staining of hepato-cytes, but similar staining patterns have been reported inother transgenic mouse lines.39,40 We detect strong sig-nals with core-specific antibodies in about 30% of hepa-tocytes. However, it is likely that viral proteins are

expressed in more than 30% of hepatocytes, but at alower level not detectable by immunofluorescence. Sec-ond, the core protein contains 32 additional amino acidsfrom the N-terminus of �1-antitrypsin. It is unlikelythat the observed inhibition of Jak-STAT signaling iscaused by novel properties specific to the fusion proteinbecause we have observed an inhibition of Jak-STATsignaling in cell lines that express the core protein in itsregular form.18 Third, we have little information on theother viral proteins. We cannot detect them by Westernblot analysis; therefore, we cannot know if they arecorrectly cleaved from the HCV polyprotein precursorand present in undetectably low concentrations, or ifthey are completely absent. This is an important caveat,because even a low amount of viral proteins could stillhave a major impact on cellular processes such as signal-ing. In fact, it is known that HCV protein expression

Figure 5. Double immunostain-ing for HCV core protein andphospho-STAT1 in C57BL/6 andB6HCV after injection of mIFN�.(A) Hoechst stain of C57BL/6mouse liver section. (B) Hoechststain of B6HCV mouse liver sec-tion. (C) phospho-STAT1 stain-ing, C57BL/6, (D) phospho-STAT1 staining, B6HCV. (E) HCVcore protein staining, C57BL/6,(F ) HCV core protein staining,B6HCV. As expected, no HCVcore protein could be detected inC57BL/6 mice. In B6HCV mice,approximately 30% of hepato-cytes show HCV core protein nu-clear localization and its pres-ence in the nucleus is negativelycorrelated with the presence ofactivated STAT1. The white ar-rows show nuclei of hepato-cytes that stain positive for HCVcore but negative for phospho-STAT1. The blue arrows showhepatocytes with nuclear stain-ing for phospho-STAT1 but with-out nuclear staining for HCVcore. The yellow size bar indi-cates 20�m.


levels in human liver samples from patients with chronichepatitis C are generally low, often not detectable byWestern blot analysis.

The functional importance of impaired IFN-inducedsignaling through the Jak-STAT pathway is demon-strated in the virus challenge experiments with LCMV-WE. Here we show for the first time that expression ofHCV proteins in hepatocytes has an influence on thecourse of a viral infection in an animal. B6HCV mice areclearly more susceptible to LCMV-induced hepatitis thanthe control mice. The most likely explanation for themore severe hepatitis and the increased mortality inB6HCV mice is the increased viral replication in livercells and thereby the increased CTL-mediated lysis ofinfected hepatocytes. We believe that hepatocytes thatexpress HCV proteins do not mount an effective antiviraldefense against LCMV because the endogenous IFN thatbinds to its receptors on the surface of hepatocytes cannotsignal to the nucleus.

In a recent report, expression of HCV core protein inliver cells after infection with a replication-defectiverecombinant adenovirus coding for core was found toinduce normal immune responses against adenovirus.41

In the same report, HCV core transgenic mice wereshown to mount a normal CTL response against adeno-virus. The authors concluded that core protein has noimmunomodulatory effect on virus-induced cellular im-munity. Experiments performed by another group usingcore–E1-E2 transgenic mice infected with replication-defective adenovirus also found an intact cellular andhumoral immune response against adenovirus.42 The au-thors concluded that HCV core and envelope proteins donot inhibit the hepatic antiviral mechanism in general.In agreement with these reports, we also find a normalimmune response against LCMV in our HCV transgenicmice. B6HCV mice have a normal footpad-swelling assayafter local injection of 200 pfu LCMV-WE in the foot aswell as an unimpaired immune response against theneurotropic rhabdovirus vesicular stomatitis virus. How-ever, when challenged with the replication-competenthepatotropic LCMV-WE virus, B6HCV mice show aprofound defect in the control of viral replication becauseof the defect of the IFN-stimulated antiviral intracellulardefense. Because the immune response is intact, B6HCVmice suffer a stronger and often fulminant hepatitis. Inpatients with chronic hepatitis C, liver histology usuallydoes not show a similar degree of necroinflammatorychanges. Indeed, for the host response against HCV,clearance of the virus without destruction of liver cellshas been suggested to be an important mechanism.43 Itdepends on IFN�–induced noncytopathic clearancethrough the induction of intracellular antiviral defense

mechanisms in hepatocytes.5,6 An inhibition of IFN-induced intracellular signaling by viral proteins wouldprevent the induction of an antiviral state and allow anactive viral replication despite the presence of a strongcellular immune response.

Our data provide evidence that the inhibition of STATsignaling occurs at the level of DNA binding of theactivated STAT transcription factors. We found an intacttyrosine phosphorylation of STAT1 and STAT3 (Figure3A ), indicating a functional IFN receptor and a normalactivity of the receptor-associated kinases Jak1 and Tyk2.For this reason, an up-regulation of the well-knownpathway inhibitors SOCS1 and SOCS3 is unlikely, andindeed, we found no such induction of SOCS1 or SOCS3(data not shown). Next, nuclear import of STATs couldbe inhibited by viral proteins. A block in nuclear importin the presence of normal STAT activation at the receptorkinase complex could lead to an accumulation of phos-phorylated STAT1 in the cytoplasm. However, we couldnot detect such an accumulation in cytoplasmic extractsof B6HCV mice. In the nucleus, STAT dephosphoryla-tion is an important mechanism for the termination ofSTAT signaling. Viral proteins could induce nuclearphosphatases. The phospho-STAT1 immunostaining ex-periments (Figure 3C ) provide evidence for increaseddephosphorylation of STATs because the number of nu-clei with a positive phospho-STAT1 signal is reduced.Finally, STAT signaling can be blocked at the level ofDNA binding. So far, the only protein known to inter-fere with DNA binding of activated STATs is PIAS1,and we could not detect an up-regulation of PIAS1 byWestern blot analysis (data not shown). However, thestrong inhibition of DNA binding found in the gel shiftassays (Figure 2) indicates that a direct or indirect inter-ference of viral proteins with DNA binding is an impor-tant mechanism of HCV-induced inhibition of Jak-STAT signaling.

As outlined previously, we cannot unequivocally iden-tify the specific viral protein(s) that is (are) responsiblefor the observed inhibition of the Jak-STAT pathway.Our immunofluorescence results suggest a role of coreprotein in the inhibition of the Jak-STAT pathway in ourtransgenic mice. We found core protein expression in thenucleus of about 30% of the hepatocytes, and, interest-ingly, most of those very same hepatocytes did not havethe active tyrosine-phosphorylated form of STAT1 in thenucleus. The core protein itself could stimulate a rapiddephosphorylation of STATs in the nucleus. Alterna-tively, high expression levels of core might just be anindicator of an above-average expression of the entireviral transgene. In this case, the inhibition of Jak-STATsignaling could be caused by any other viral protein(s).


Whatever the mechanism, hepatocytes with detectablecore protein in the nucleus show a lack of phosphorylatedSTAT1 in the nucleus.

The same mechanism of interference with the Jak-STAT pathway could take place in chronic viral hepati-tis. IFN would then be actively inducing an antiviralstate only in hepatocytes that are not infected by HCV orthat have very low cellular viral loads with concomitantlow expression levels of viral proteins. It is a matter ofongoing debate what percentage of hepatocytes is actu-ally infected by HCV in chronic hepatitis C. Publisheddata indicate that viral replication and gene expressiontakes place in less than 20% of hepatocytes.44 In view ofthese data, one may speculate that IFN treatment inpatients with chronic hepatitis C actually induces anantiviral state in uninfected hepatocytes only, while thepatient’s own HCV-specific CD8-positive cytotoxic Tlymphocytes would have to clear the virus by attackinginfected hepatocytes. On the other hand, high doses ofIFN can overcome the block of IFN�-induced Jak-STATsignaling in our mouse model (Figure 2B). Interestingly,a similar observation has been made in patients infectedwith HCV. It has been shown that treatment with higherIFN� doses results in a higher response rate in patientswith chronic hepatitis C.45 Apparently, higher dosescould overcome the resistance against IFN� treatment inabout 15% of patients.45 However, adverse effects ofIFN� limit the dose that can be administered to about100 IU/g body weight 3 times per week. In this context,a more promising strategy to overcome viral interferencewith IFN signaling would be to search for compoundsthat counteract the effects of viral proteins on Jak-STATsignaling.

In summary, we have found a functionally importantinhibition of IFN�-induced signaling through the Jak-STAT pathway in liver cells of HCV transgenic mice invivo. Interference with this intracellular signaling path-way reduces the amount of phosphorylated STAT1 in thenucleus of cells expressing viral proteins, and results in aloss of binding of STAT transcription factor complexes totheir cognate response elements in the promoters ofISGs. This mechanism could contribute to escape fromthe antiviral effects of endogenous IFNs and, thereby, toHCV persistence in infected individuals. Moreover, itmight contribute to the poor response to IFN-basedtherapies in the majority of HCV-infected patients.Strategies aimed at restoring IFN signaling in infectedliver cells might allow for more successfully treatingpatients in the future.

References1. European Association for the Study of the Liver. Consensus

statement. EASL International Consensus Conference on Hepa-

titis C. Paris, February 26–28, 1999. European Association forthe Study of the Liver. J Hepatol 1999;30:956–961.

2. Alter MJ, Margolis HS, Krawczynski K, Judson FN, Mares A,Alexander WJ, Hu PY, Miller JK, Gerber MA, Sampliner RE, MeeksEL, Beach MJ. The natural history of community-acquired hepati-tis C in the United States. The Sentinel Counties Chronic non-A,non-B Hepatitis Study Team. N Engl J Med 1992;327:1899–1905.

3. El-Serag HB, Mason AC. Risk factors for the rising rates of primaryliver cancer in the United States. Arch Intern Med 2000;160:3227–3230.

4. Cerny A, Chisari FV. Pathogenesis of chronic hepatitis C: immu-nological features of hepatic injury and viral persistence. Hepa-tology 1999;30:595–601.

5. Guidotti LG, Rochford R, Chung J, Shapiro M, Purcell R, ChisariFV. Viral clearance without destruction of infected cells duringacute HBV infection. Science 1999;284:825–829.

6. Guidotti LG, Borrow P, Brown A, McClary H, Koch R, Chisari FV.Noncytopathic clearance of lymphocytic choriomeningitis virusfrom the hepatocyte. J Exp Med 1999;189:1555–1564.

7. Darnell JE Jr, Kerr IM, Stark GR. Jak-STAT pathways and transcrip-tional activation in response to IFNs and other extracellular sig-naling proteins. Science 1994;264:1415–1421.

8. Heim MH. Intracellular signalling and antiviral effects of interfer-ons. Dig Liver Dis 2000;32:257–263.

9. Schindler C, Shuai K, Prezioso VR, Darnell JE Jr. Interferon-dependent tyrosine phosphorylation of a latent cytoplasmic tran-scription factor. Science 1992;257:809–813.

10. Silvennoinen O, Schindler C, Schlessinger J, Levy DE. Ras-inde-pendent growth factor signaling by transcription factor tyrosinephosphorylation. Science 1993;261:1736–1739.

11. Darnell JE Jr. STATs and gene regulation. Science 1997;277:1630–1635.

12. Meraz MA, White JM, Sheehan KCF, Bach EA, Rodig SJ, Dighe AS,Kaplan DH, Riley JK, Greenlund AC, Campbell D, Carvermoore K,Dubois RN, Clark R, Aguet M, Schreiber RD. Targeted disruptionof the Stat1 gene in mice reveals unexpected physiologic speci-ficity in the Jak-Stat signaling pathway. Cell 1996;84:431–442.

13. Durbin JE, Hackenmiller R, Simon MC, Levy DE. Targeted disrup-tion of the mouse Stat1 results in compromised innate immunityto viral disease. Cell 1996;84:443–450.

14. Horvath CM, Darnell JE Jr. The antiviral state induced by alphainterferon and gamma interferon requires transcriptionally activeStat1 protein. J Virol 1996;70:647–650.

15. Look DC, Roswit WT, Frick AG, Gris-Alevy Y, Dickhaus DM, WalterMJ, Holtzman MJ. Direct suppression of Stat1 function duringadenoviral infection. Immunity 1998;9:871–880.

16. Foster GR, Ackrill AM, Goldin RD, Kerr IM, Thomas HC, Stark GR.Expression of the terminal protein region of hepatitis B virusinhibits cellular responses to interferons alpha and gamma anddouble-stranded RNA. Proc Natl Acad Sci U S A 1991;88:2888–2892.

17. Miller DM, Rahill BM, Boss JM, Lairmore MD, Durbin JE, WaldmanJW, Sedmak DD. Human cytomegalovirus inhibits major histo-compatibility complex class II expression by disruption of theJak/Stat pathway. J Exp Med 1998;187:675–683.

18. Heim MH, Moradpour D, Blum HE. Expression of hepatitis C virusproteins inhibits signal transduction through the jak-STAT path-way. J Virol 1999;73:8469–8475.

19. Moradpour D, Kary P, Rice CM, Blum HE. Continuous human celllines inducibly expressing hepatitis C virus structural and non-structural proteins. Hepatology 1998;28:192–201.

20. Tomei L, Failla C, Santolini E, De Francesco R, La Monica N. NS3is a serine protease required for processing of hepatitis C viruspolyprotein. J Virol 1993;67:4017–4026.

21. Tripodi M, Perfumo S, Ali R, Amicone L, Abbott C, Cortese R.Generation of small mutation in large genomic fragments by


homologous recombination: description of the technique andexamples of its use. Nucleic Acids Res 1990;18:6247–6251.

22. Amicone L, Galimi MA, Spagnoli FM, Tommasini C, De Luca V,Tripodi M. Temporal and tissue-specific expression of the METORF driven by the complete transcriptional unit of human A1ATgene in transgenic mice. Gene 1995;162:323–328.

23. Hogan B, Beddington R, Costantini F, Lacy E. Manipulating themouse embryo. A laboratory manual. 2nd ed. Cold Spring Harbor,NY, CSHL Press, 1994.

24. Moradpour D, Englert C, Wakita T, Wands JR. Characterization ofcell lines allowing tightly regulated expression of hepatitis C viruscore protein. Virology 1996;222:51–63.

25. Heim MH, Gamboni G, Beglinger C, Gyr K. Specific activation ofAP-1 but not Stat3 in regenerating liver in mice. Eur J Clin Invest1997;27:948–955.

26. Battegay M, Cooper S, Althage A, Banziger J, Hengartner H,Zinkernagel RM. Quantification of lymphocytic choriomeningitisvirus with an immunological focus assay in 24- or 96-well plates.J Virol Methods 1991;33:191–198 [published errata appears inJ Virol Methods 1991;35:115 and 1992;38:263].

27. Takamizawa A, Mori C, Fuke I, Manabe S, Murakami S, Fujita J,Onishi E, Andoh T, Yoshida I, Okayama H. Structure and organi-zation of the hepatitis C virus genome isolated from humancarriers. J Virol 1991;65:1105–1113.

28. Barba G, Harper F, Harada T, Kohara M, Goulinet S, Matsuura Y,Eder G, Schaff Z, Chapman MJ, Miyamura T, Brechot C. HepatitisC virus core protein shows a cytoplasmic localization and asso-ciates to cellular lipid storage droplets. Proc Natl Acad Sci U S A1997;94:1200–1205.

29. Yasui K, Wakita T, Tsukiyama-Kohara K, Funahashi SI, IchikawaM, Kajita T, Moradpour D, Wands JR, Kohara M. The native formand maturation process of hepatitis C virus core protein. J Virol1998;72:6048–6055.

30. Moriya K, Fujie H, Shintani Y, Yotsuyanagi H, Tsutsumi T, Ishi-bashi K, Matsuura Y, Kimura S, Miyamura T, Koike K. The coreprotein of hepatitis C virus induces hepatocellular carcinoma intransgenic mice. Nat Med 1998;4:1065–1067.

31. Shuai K, Stark GR, Kerr IM, Darnell JE Jr. A single phosphoty-rosine residue of Stat91 required for gene activation by interfer-on-gamma. Science 1993;261:1744–1746.

32. Wang Q, Miyakawa Y, Fox N, Kaushansky K. Interferon-alphadirectly represses megakaryopoiesis by inhibiting thrombopoi-etin-induced signaling through induction of SOCS-1. Blood 2000;96:2093–2099.

33. Muller U, Steinhoff U, Reis LF, Hemmi S, Pavlovic J, ZinkernagelRM, Aguet M. Functional role of type I and type II interferons inantiviral defense. Science 1994;264:1918–1921.

34. Zinkernagel RM, Haenseler E, Leist T, Cerny A, Hengartner H,Althage A. T cell-mediated hepatitis in mice infected with lympho-cytic choriomeningitis virus. Liver cell destruction by H-2 class I-restricted virus-specific cytotoxic T cells as a physiological corre-late of the 51Cr-release assay? J Exp Med 1986;164:1075–1092.

35. Lohler J, Gossmann J, Kratzberg T, Lehmann-Grube F. Murinehepatitis caused by lymphocytic choriomeningitis virus. I. Thehepatic lesions. Lab Invest 1994;70:263–278.

36. Gale MJ Jr, Korth MJ, Tang NM, Tan S-L, Hopkins DA, Dever TE,

Polyak SJ, Gretch DR, Katze MG. Evidence that hepatitis C virusresistance to interferon is mediated through repression of thePKR protein kinase by the nonstructural 5A protein. Virology1997;230:217–227.

37. Polyak SJ, Paschal DM, McArdle S, Gale MJ Jr, Moradpour D,Gretch DR. Characterization of the effects of hepatitis C virusnonstructural 5A protein expression in human cell linesand on interferon-sensitive virus replication. Hepatology 1999;29:1262–1271.

38. Taylor DR, Shi ST, Romano PR, Barber GN, Lai MM. Inhibition ofthe interferon- inducible protein kinase PKR by HCV E2 protein.Science 1999;285:107–110.

39. Kawamura T, Furusaka A, Koziel MJ, Chung RT, Wang TC,Schmidt EV, Liang TJ. Transgenic expression of hepatitis C virusstructural proteins in the mouse. Hepatology 1997;25:1014–1021.

40. Honda A, Arai Y, Hirota N, Sato T, Ikegaki J, Koizumi T, Hatano M,Kohara M, Moriyama T, Imawari M, Shimotohno K, Tokuhisa T.Hepatitis C virus structural proteins induce liver cell injury intransgenic mice. J Med Virol 1999;59:281–289.

41. Liu ZX, Nishida H, He JW, Lai MM, Feng N, Dennert G. HepatitisC virus genotype 1b core protein does not exert immunomodula-tory effects on virus-induced cellular immunity. J Virol 2002;76:990–997.

42. Sun J, Bodola F, Fan X, Irshad H, Soong L, Lemon SM, Chan T-S.Hepatitis C virus core and envelope proteins do not suppress thehost’s ability to clear a hepatic viral infection. J Virol 2001;75:11992–11998.

43. Thimme R, Oldach D, Chang K-M, Steiger C, Ray SC, Chisari FV.Determinants of viral clearance and persistence during acutehepatitis C virus infection. J Exp Med 2001;194:1395–1406.

44. Lau JY, Krawczynski K, Negro F, Gonzalez-Peralta RP. In situdetection of hepatitis C virus-a critical appraisal. J Hepatol 1996;24:43–51.

45. Poynard T, Leroy V, Cohard M, Thevenot T, Mathurin P, Opolon P,Zarski JP. Meta-analysis of interferon randomized trials in thetreatment of viral hepatitis C: effects of dose and duration.Hepatology 1996;24:778–789.

Address reprint requests to: Markus H. Heim, M.D., Departments ofResearch and Gastroenterology, University Hospital Basel, Hebelstrasse20, CH-4031, Basel, Switzerland. e-mail: [email protected]; fax:(41) 61 265 23 50.

Supported by Swiss National Science Foundation grants 3231-054973 and 3100-052626, Swiss Cancer League grant 922-09-1999,Roche Research Foundation grant 90-1999, the Astra-Fonds, and theStiftung fur Gastroenterologische Forschung. D.M. and H.E.B. are sup-ported by grant Mo 799/1-2 from the Deutsche Forschungsgemein-schaft. M.T. is supported by the AIRC (Associazione Italiana Ricerca sulCancro), by progetto strategico 2000 “Epatite C” Ministero della Sanitaand by MURST (Ministero della Ricerca Scientifica e Tecnologica), Italy.

The authors thank Rolf Zinkernagel, George Thomas, FrancescoNegro, and Gennaro Ciliberto for stimulating discussions and readingof the manuscript.

A.B., F.H.T.D., L.H., and S.T.D.S. contributed equally to this work.


Expression of hepatitis c virus proteins inhibits interferon α signaling in the liver of transgenic...

Documents

Transcript of Expression of hepatitis c virus proteins inhibits interferon α signaling in the liver of transgenic...