Article Text

Download PDFPDF

New aspects of an anti-tumour drug: sorafenib efficiently inhibits HCV replication
  1. K Himmelsbach1,2,
  2. D Sauter1,
  3. T F Baumert1,3,
  4. L Ludwig4,
  5. H E Blum1,
  6. E Hildt1,2
  1. 1
    University of Freiburg, Department of Medicine II, Freiburg, Germany
  2. 2
    University of Kiel, Institute of Infection Medicine, Molecular Medical Virology, Kiel, Germany
  3. 3
    Institut National de la Santé et de la Recherche Médicale, U748, Strasbourg, France; Université Louis Pasteur, Strasbourg, France
  4. 4
    TU-Munich, Department of Medicine II, Munich, Germany
  1. Correspondence to Dr E Hildt, University of Kiel-UKSH, Institute of Infection Medicine, Molecular Medical Virology, Brunswiker Strasse. 4, D-24105 Kiel, Germany; hildt{at}infmed.uni-kiel.de

Abstract

Background and aims: Hepatitis C virus (HCV) infection is a major cause of chronic liver disease and is associated with significant morbidity and mortality. Since there is evidence for an interaction of NS5A with c-Raf we studied whether the c-Raf inhibitor sorafenib affects HCV replication.

Methods: HCV replicating HuH7.5 cells were treated with sorafenib and examined for HCV RNA titres by northern blotting or real time polymerase chain reaction (PCR), for core, NS3 and NS5A expression by immunostaining, and for replication by luciferase reporter assays.

Results: Here we demonstrate that in cells replicating infectious HCV particles, NS5A recruits c-Raf to the replicon complex resulting in the activation of c-Raf. Therefore, we studied the effect of inhibition of c-Raf on HCV replication using the anti-tumour drug sorafenib that is known to inhibit c-Raf with high specificity. Sorafenib efficiently blocks HCV replication and viral gene expression. In addition, in HCV-replicating cells sorafenib decreased the hyperphosphorylated form of NS5A and resulted in the formation of additional hypophosphorylated forms. Further, sorafenib caused a rapid dissociation of lipid droplets. We provide evidence that the antiviral effect of sorafenib indeed is caused by inhibition of c-Raf. By contrast, inhibition of targets downstream of c-Raf or inhibition of tyrosine kinases by sunitinib did not affect HCV replication.

Conclusion: Our data demonstrate that the well-characterised anti-tumour drug sorafenib efficiently blocks HCV replication in vitro. This novel effect of sorafenib should be further explored as an antiviral strategy for patients with chronic HCV infection.

View Full Text

Statistics from Altmetric.com

Hepatitis C virus (HCV) infection results in chronic hepatitis in more than 70% of infected individuals. At present more than 170 million people are persistently infected with HCV worldwide. Persistent HCV infection is associated with chronic inflammation of the liver (hepatitis), which can progress to liver fibrosis, cirrhosis and hepatocellular carcinoma (HCC).1 2 Since the discovery of HCV as the causative agent for non-A and non-B hepatitis in 1989,3 considerable progress has been made with respect to treatment strategies, but for a large proportion of patients, specifically those infected with HCV genotypes 1a/1b the currently available therapy with pegylated interferon α and ribavirin is ineffective. Moreover, there is no prophylactic vaccine in sight.

HCV is the sole member of the genus hepacivirus that belongs to the flaviridae family. The HCV genome is a single-stranded positive-sense RNA molecule of approx. 9600 bases in length. The viral RNA codes for a large polyprotein of approx. 3100 amino acids, which is post-translationally processed by cellular and viral proteases. The N-terminus encompasses the structural proteins core, E1 and E2, the C-terminus the p7 protein, and the non-structural (NS) proteins NS2, NS3, NS4A, NS4B, NS5A and NS5B.4 5 6 The mature NS5A protein is generated by the action of the viral NS3/NS4A serine protease. Subsequently, NS5A associates with the cytoplasmic face of the endoplasmatic reticulum (ER) via an amphipathic alpha-helix. In this association NS5A is an integral part of the replicon complex7 and required for viral morphogenesis.8 9 NS5A is a phosphoprotein that exists in a basal or in a hyperphosphorylated state (p56 and p58).6 10 11 Moreover, as an integral part of the HCV replicon complex, NS5A is able to interfere with viral proteins12 as well as with a variety of cellular proteins.13 14 15 16 17 Some of these interaction partners seem to trigger a deregulation of the host cell signal transduction, such as Grb2, PI3K, p53, or c-Raf. Overall, NS5A consists of three discrete domains I–III. Based on the crystal structure, it is assumed that domain I is involved in homodimer formation, generating a large putative RNA binding groove located at the interface of the monomers. Domains II and III are less well characterised. However, genetic mapping has shown that domain II is crucial for HCV RNA replication. With respect to c-Raf, we could recently demonstrate that the interaction with NS5A is mediated by domain II, and eventually leads to activation of c-Raf.16 In addition, recent work identified a small deletion in domain III of NS5A that disrupts the production of infectious viral particles without affecting RNA replication.8 9

Sorafenib (Nexavar, Bay 43-9006) (4-{4-[3-(4-chloro-3-trifluoromethyl-phenyl)ureido] phenoxy}pyridine-2-carboxylic acid methylamide 4-methylbenzenesulfonate) belongs to the biaryl urea class of protein kinase inhibitors. Sorafenib was identified as a potent inhibitor of c-Raf with an inhibitory concentration (IC50 values) of 6 nmol/l in vitro.18 19 Moreover, sorafenib was found to inhibit several receptor tyrosine kinases (RTKs) linked to tumour progression, including Flt-3, c-Kit, vascular endothelial growth factor receptor 2 (VEGFR2), VEGFR3, and platelet-derived growth factor receptor β (PDGFR-β). Of note is the fact that this substance already has been approved by the US Food and Drug Administration (FDA) and the European Union (EU) for the treatment of patients with renal cell carcinoma (RCC) and HCC.20 21

Based on recently developed tissue culture model systems, it is now possible to study the complete HCV life cycle in vitro, thus allowing the assessment of substances with respect to their potential effect on HCV replication.22 23 24 The present study was performed to answer the question whether modulation of c-Raf by the clinically well established small molecule inhibitor sorafenib affects HCV replication.

Materials and methods

Plasmids

Plasmids pJFH1, pJFH1/GND, pFK-lucJFH1/wt and pRafC4 have been described previously.23 25 26

Cell culture and inhibitors

The Huh7-derived cell clone Huh7.5 which is highly permissive for HCV RNA replication was used for transfection and infection assays. Cells were grown in Dulbecco’s modified Eagle’s medium (DMEM; Sigma, Deisenhofen, Germany) supplemented with 2 mmol/l l-glutamine, non-essential amino acids, 100 units of penicillin per ml, 100 μg of streptomycin per ml, and 10% fetal calf serum (DMEM complete).

HCV binding and entry were analysed as described.27 Isolation and cultivation of primary human hepatocytes was performed as described.28 29

Sorafenib was dissolved in DMSO and applied at final concentrations between 0.1 and 20 μmol/l, sunitinib was used in a final concentration of 1 nmol/l to 10 μmol/l, imatinib at a concentration of 1 nmol/l to 10 μmol/l. For inhibition of mitogen-activated protein kinase (MEK), the inhibitors PD98059 and UO126 (Calbiochem, San Diego, California, USA) were used. The incubation with the inhibitors was performed for 24 h, in case of primary human hepatocytes for 30 h. To study the effect of sorafenib on NS5A phosphorylation or on lipid droplet formation the incubation with sorafenib was performed for 2 h.

In vitro transcription and RNA transfection

In vitro transcription, electroporation of HCV RNAs, and luciferase assays were performed as described.30 All luciferase assays were done at least in triplicate.

Quantitative PCR and northern blotting

RNA isolation was performed using Trizol (Invitrogen, Karlsruhe, Germany), according to the manufacturer’s instructions. For cDNA synthesis, 2–4 μg of total RNA were treated with DNase I. First-strand synthesis was carried out using SupercriptII reverse transcriptase (Invitrogen), and semiquantitative RT-PCR was performed according to the Invitrogen protocol.

For detailed quantification of HCV RNA real time PCR was performed using a commercial assay kit (Amplicor HCV Monitor; Roche, Basel, Switzerland) according to the manufacturer’s instructions. The detection limit of the assay is at 150 genomes/ml. All experiments were performed in triplicate, one representative experiment is shown. Northern blotting was performed as described.27

Infection experiments

Cell culture medium was collected 72 h after transfection and cleared by low spin centrifugation followed by filtration through a preabsorbed 0.45 μm filter. Infection was performed as described. Infection of primary human hepatocytes was performed as described.28

Immunoprecipitation and western blot analysis

Western blotting and immunoprecipitation were performed as described.16 For detection of the NS5A protein, a rabbit-derived polyclonal serum16 was used. For detection of c-Raf, a commercial mouse monoclonal antibody was used (Transduction Laboratories, San Jose, California, USA). The antibody for detection of phospho-c-Raf (Ser338) was purchased from Upstate (Billerica, Massachusetts, USA). Specific antibodies for the detection of core and NS3 were purchased from Affinity BioReagents (Omaha, Nebraska, USA), and ViroStat (Portland, Oregon, USA), respectively. The immunocomplex assay was performed as previously described.31 The phospo-src (phospho Y416) antiserum was purchased from Cell Signalling Technology (Beverly, Massachusetts, USA).

Immunofluorescence microscopy

Immunofluorescence staining was analysed by confocal laser scanning microscopy (CLSM) using the Leica confocal microscope TCS SP2. The following antisera were used: anti-NS5A (rabbit),16 anti-HCV core (mouse; Affinity BioReagents). Bound antibodies were visualised using Cy2- and Cy3-conjugated secondary antibodies (Dianova, Hamburg, Germany). For nuclear staining, 4,6-diamidino-2-phenylindole (DAPI) (Sigma, St Louis, Missouri, USA) was used.

Glutamyl oxaloacetic transaminase quantification and MTT assay

Glutamyl oxaloacetic transaminase (GOT) was quantified using the reflotron system according to the manufacturer’s instructions (Roche). The MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) assay was performed as described.32

Alkaline phosphatase treatment

HCV replicating cells were lysed by 0.1% TritonX-100 on ice for 10 min. Phosphatase treatment was performed for 1 h at 37°C using 0.5 U calf intestinal phosphatase (CIP) (NEB) per μl lysate in 1× Buffer 3 (NEB).

Two-dimensional separation

Isoelectric focusing and subsequent sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) were performed as described.31

Results

c-Raf is part of the replicon complex

In a recent report, it was shown that in a subgenomic replicon system NS5A (HCV genotype 1b) can bind to c-Raf.16 As shown in fig 1a,b, we confirmed this interaction via co-immunoprecipation experiments of lysates derived from infectious HCV-replicating cells. Given the fact that NS5A is an integral part of the replicon complex in cells replicating infectious HCV, we wondered whether – mediated by a specific interaction with this factor – c-Raf might also be associated with this complex. To test this hypothesis, cellular lysates from HCV-replicating Huh7.5 cells were precipitated using a c-Raf-specific antiserum. Unrelated antisera (anti-hexokinase and anti-HBs) served as controls. The precipitate was analysed for the presence of HCV-specific RNA by RT-PCR. As shown in fig 1c, HCV-specific RNA was indeed specifically co-precipitated by the c-Raf-specific antiserum, but not by the control sera, demonstrating that c-Raf is part of the replicon complex.

Figure 1

c-Raf is part of the replicon complex. Hepatitis C virus (HCV)-replicating cells were lysed for 96 h after elctroporation and subjected to immunoprecipitation by the indicated antisera, followed by further analysis. (a) Western blotting using a NS5A-specific rabbit-derived antiserum for detection. Precipitation using a mouse-derived HBsAg-specific antiserum (anti-HBs) served as negative controls. Precipitation using a NS5A-specific serum served as positive control. c-Raf was precipitated by a mouse-derived antiserum. To enable the detection of the hypo- and the hyperphosphorylated NS5A forms the samples were loaded on a 8% gel. For detection a peroxidase-conjugated rabbit-derived NS5A-specific antiserum was used. (b) Western blotting using a c-Raf-specific mouse-derived antiserum for detection. Precipitation using a rabbit-derived HBsAg-specific antiserum and Huh7.5 cells derived lysate (---) served as negative controls. Precipitation using a c-Raf-specific rabbit-derived serum served as positive control. NS5A was precipitated by a rabbit-derived antiserum and loaded in duplicate. (c) For precipitation a c-Raf-specific antiserum was used. Precipitation with an unrelated antiserum (anti-hexokinase), and a sample which had not been treated with antiserum served as negative controls. The immunoprecipitates were analysed by reverse transcription polymerase chain reaction (RT-PCR). To prove the specificity of the RT-PCR a negative control with water instead of template, and a positive control with a c-Raf-specific expression plasmid as template, were also carried out. (d) Western blot analysis of lysates from HCV-replicating cells (lanes 1, 3, 5), and control-transfected cells (XXXXX (GND)) (lanes 2, 4, 6), using anti-phospho-c-Raf (S338) (Upstate), and anti-Raf-1 (BDTransduction Labs, San Jose, California, USA) in parallel. Immunostaining (anti-phospho-Raf-1 (S338)/anti-Raf-1) was quantified. The bar graphs show the results of three independent experiments.

Since c-Raf phosphorylation/activity is often modulated by interaction of this kinase with other proteins, we wondered whether – mediated by the specific interaction with NS5A within the replicon complex – this might also be the case in HCV-replicating cells. As shown in fig 1d, the analysis of c-Raf phosphorylation by western blotting using a phospho-Raf-specific serum indeed revealed an increased amount of phospho-Raf in HCV-replicating cells compared to uninfected controls. Immunocomplex assays consistently demonstrated an induction of c-Raf activity in these cells (data not shown).

Taken together, these data indicate that in HCV-replicating cells (1) c-Raf is associated with the replicon complex, and (2) c-Raf is activated.

Inhibition of HCV replication by sorafenib

In view of our finding that c-Raf is associated with the replicon complex and activated in HCV-replicating cells, we studied the inhibition of c-Raf as a novel antiviral strategy to control HCV replication. Given the potential therapeutic implications, we used the small molecule inhibitor sorafenib (Bay 43-9006; Nexavar), a well-characterised inhibitor of c-Raf that has already been clinically approved by the FDA for treatment of renal cell carcinoma and HCC. To test its potential effect on HCV replication, Huh7.5 cells were transfected with the reporter virus RNA JFH-1/luc and incubated with increasing concentrations of sorafenib (5–20 μmol/l). As shown in fig 2a, the analysis of luciferase activity revealed a concentration-dependent inhibition of HCV replication by sorafenib with an IC50 of 7.2 μmol/l.

Figure 2

Inhibition of hepatitis C virus (HCV) replication by sorafenib. (a) Analysis of viral replication by luciferase reporter gene assay. Forty-eight hours after electroporation, HCV JFH-1 luc-replicating Huh7.5 cells were treated for 24 h with the indicated concentrations of sorafenib. Treatment with an equivalent amount of DMSO (0.1%) served as a control. (b) Analysis of HCV replication by reverse transcription polymerase chain reaction (RT-PCR) using a 5′UTR-specific primer pair. Forty-eight hours after electroporation, HCV JFH-1 wt-replicating Huh7.5 cells were incubated for 24 h with the indicated concentrations of sorafenib. Treatment with an equivalent amount of DMSO (0.1%) served as a control. Amplification of actin-specific sequences was performed to ensure that comparable amounts of RNA were analysed in all experiments. (c) HCV JFH-1infected cells were incubated for 24 h with the indicated concentrations of sorafenib. Treatment with an equivalent amount of dimethyl sulfoxide (DMSO) (0.1%) served as a control. RNA was isolated and analysed by northern blot using an NS5A-specific probe. Equal loading was controlled using an actin-specific probe. (d) Western blot analysis of cellular lysate derived from HCV-JFH1-replicating Huh7.5 cells that had been grown for 24 h in the absence or presence of the indicated concentrations of sorafenib using a NS5A- or an HCV core-specific antiserum. Equal loading was controlled by re-probing of the membrane with an actin-specific serum. Lysates were separated on a 12% gel to enable detection of NS5A and core on the same membrane. (e) MTT assay of Huh7.5 cells that were grown for 24 h in the presence of the indicated amounts of sorafenib. The assay was performed in triplicate. Incubation with 200 μmol/l PDTC served as positive control. (f) HCV-replicating Huh7.5 cells were incubated for 24 h with the indicated concentrations of sorafenib. After 24 h the medium was removed, the cells were washed and incubated again for 48 h with the indicated concentrations of sorafenib. The number of genomes (genomes/ml) in the supernatant was quantified by real time PCR. The detection limit (150 genomes/ml) is indicated by the dotted line. (g) Reinfection of Huh 7.5 cells with the dialysed supernatants derived from cells treated with the indicated amounts of sorafenib. Twelve hours after infection cells were washed and grown for 72 h. The number of genomes (genomes/ml) in the supernatant was quantified by real time PCR. MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide.

To further investigate the effect of sorafenib on HCV replication, RNA was isolated from sorafenib-treated and untreated HCV-replicating Huh7.5 cells, followed by quantification of the HCV RNA by RT-PCR and northern blotting. As shown in fig 2b,c, increasing concentrations of sorafenib resulted in a dose-dependent disappearance of HCV-specific RNA.

To analyse the effect of sorafenib on the synthesis of HCV proteins, western blot analyses of lysates from HCV-replicating cells treated with sorafenib were performed using HCV-, core- and NS5A-specific antisera. As shown in fig 2d, increasing concentrations of sorafenib indeed reduced the amount of HCV-specific proteins below the western blot detection level, demonstrating a strong inhibitory effect of sorafenib on the synthesis of HCV proteins. To investigate whether sorafenib exerts a general inhibitory effect on viral replication adenovirus replication and poliovirus were studied under identical conditions. In both cases no effect on virus replication was observed (data not shown).

To investigate the specificity of the observed effects the cytotoxicity of sorafenib was studied by MTT assay. HuH7.5 cells were grown for 24 h in the presence of the indicated concentrations of sorafenib. The MTT assay revealed that for concentrations up to 12.5 μmol/l no significant cytotoxicity was observed (fig 2e). Comparable results were obtained by analysis of GOT activity in the cell culture supernatant or in the serum of sorafenib-treated mice (data not shown). These data argue against a cytotoxic effect of the applied sorafenib concentrations.

To demonstrate that the effect of sorafenib does not depend on a general inhibition of cell proliferation and to quantify directly the effect of sorafenib on the amount of secreted HCV particles, confluent HCV replicating cells were treated with sorafenib. The quantification of genome copies by real time PCR demonstrates that sorafenib strongly inhibits HCV replication (fig 2f). Moreover, Huh7.5 cells were infected with dialysed supernatant derived from untreated or sorafenib-treated HCV-replicating cells. Medium was changed after 12 h. The amount of de novo synthesised genomes was quantified 84 h p.i. by real time PCR. The real time PCR demonstrated that in case of the cells infected with the supernatant derived from the untreated cells productive de novo synthesis was observed, while in case of the cells infected with supernatant derived from sorafenib-treated cells no productive infection was established (fig 2g).

To investigate whether, in addition to its inhibitory effect on virus replication, sorafenib might also affect virus binding to the cell surface and/or viral entry into the cell, hepatocyte binding and entry assays were performed. These analyses revealed that sorafenib neither impairs binding of HCV to the hepatocyte nor its entry into the hepatocyte (data not shown).

Taken together, our results indicate that sorafenib is a potent inhibitor of HCV replication and protein synthesis without affecting virus binding and/or entry into hepatocytes.

Sorafenib abolishes HCV propagation in HCV-infected primary human hepatocytes

To exclude the possibility that the sorafenib effect on HCV replication is restricted to Huh7.5 cells and to test the compound’s inhibitory effect under more physiological conditions, primary human hepatocytes were used. For this purpose, primary human hepatocytes were infected with HCV and subsequently incubated with different concentrations of sorafenib for 30 h or 72 h. As shown in fig 3, analysis of HCV replication by western blotting using a NS5A-specific antiserum showed that sorafenib decreases HCV replication in infected cells below the detection level. The same results were obtained for HCV RNA quantified by RT-PCR (data not shown). Overall, these data demonstrate that HCV replication is also efficiently blocked by sorafenib in HCV infected primary human hepatocytes.

Figure 3

Inhibition of hepatitis C virus (HCV) replication by sorafenib in vivo. (a)Western blot analysis of cellular lysate from HCV-JFH1-infected primary human hepatocytes (PHH) using a NS5A-specific antiserum. Forty-eight hours after the infection, PHH were incubated for 30 h with the indicated concentrations of sorafenib. Lysate from untreated Huh7.5 cells served as a positive control, lysate from uninfected PHH as a negative control. Equal loading was controlled by re-probing of the membrane with an actin-specific antiserum. Samples were loaded on a 12% gel. (b) Serum samples of HCV positive patients were analysed by real time polymerase chain reaction (PCR) for quantification of HCV genomes. The white bar represents the samples before sorafenib treatment, the black bar the samples taken after 3 months of sorafenib treatment. In the case of patient 1 the viral load under sorafenib therapy was below the detection level of 150 genomes/ml (dotted line).

To study the potential clinical relevance of these cell culture-based data we analysed sera from HCV positive male HCC patients before the sorafenib therapy was started and about 3 months after start of the sorafenib therapy. The patients received a monotherapy with 2× 400 mg sorafenib/day. In total, three patients were found for this analysis. The real time PCR revealed for all three patients a significant decrease in the viral load under sorafenib therapy. (patient 1: 6.8×10e5 genomes/ml before therapy, less than 150 genomes/ml (detection limit) under therapy; patient 2: 1.5×10e6 genomes/ml before therapy, 9 ×10e2 genomes/ml under therapy; patient 3: 1.3×10e6 genomes/ml, c before therapy, 5.3×10e3 under therapy), (fig 3b).

Taken together these data indicate that sorafenib could have a potential relevance for the treatment of chronic HCV infection.

The sorafenib effect on HCV replication is mediated by a direct inhibition of c-Raf

Sorafenib is a potent and highly specific inhibitor of c-Raf, but at higher concentrations also affects other targets, ie, the tyrosine kinases c-Kit, VEGFR2, VEGFR3, and PDGFR-β. To investigate whether the effect of sorafenib on HCV replication is due to an inhibition of these tyrosine kinases HCV replicating cells were incubated with increasing concentrations of the tyrosine kinase inhibitor sunitinib. Analysis of HCV replication by western blotting or real time PCR revealed that inhibition of tyrosine kinases does not affect HCV replication. Functionality of the applied sunitinib concentrations was shown by decrease of phosphorylated-src (pY416-src) (fig 4a). To quantify directly the effect of sunitinib on the amount of secreted HCV particles, confluent HCV replicating cells were treated with sunitinib. The number of genome copies was quantified by real time PCR. The real time PCR demonstrates that sunitinib does not affect HCV replication (fig 4b). Moreover, Huh7.5 cells were infected with dialysed supernatant derived from untreated or sunitinib-treated HCV-replicating cells. Medium was changed after 12 h. The amount of de novo synthesised RNA copies was quantified 84 h p.i. by real time PCR. The real time PCR demonstrated that there is no significant difference in the amount of viral genomes in the supernatant between the cells infected with the supernatant derived from the untreated cells or sunitinib-treated cells (fig 4c). This confirms that the sunitinib-dependent inhibition of tyrosine kinases does not affect HCV replication.

Figure 4

Sorafenib-dependent inhibition of hepatitis C virus (HCV) replication is mediated by c-Raf. (a) Western blot analysis of lysates from HCV-JFH1-replicating Huh7.5 cells that had been grown for 24 h in the presence of the indicated concentrations of the tyrosine kinase inhibitor sunitinib. Equal loading was controlled by re-probing of the membrane with an actin-specific antiserum. Lysates were separated on a 12% gel to enable detection of NS5A and core on the same membrane. To demonstrate functionality of sunitinib cells were treated with the indicated amounts of sunitinib for 4 h followed by a stimulation with EGF (25 ng/ml) for 15 min. For detection of phosphorylated-src a pY416-src-specific serum was used. Equal loading was controlled by re-probing of the membrane with an actin-specific antiserum. (b) HCV-replicating Huh7.5 cells were incubated for 24 h with the indicated concentrations of sunitinib. After 24 h the medium was removed, the cells were washed and incubated again for 48 h with the indicated concentrations of sunitinib. The number of genomes (genomes/ml) in the supernatant was quantified by real time polymerase chain reaction (PCR). (c) Reinfection of Huh 7.5 cells with the dialysed supernatants derived from cells treated with the indicated amounts of sunitinib. Twelve hours after infection cells were washed and grown for 72 h. The number of genomes (genomes/ml) in the supernatant was quantified by real time PCR. (d) Western blot analysis of lysates from HCV-JFH1-replicating Huh7.5 cells that had been grown for 24 h in the presence of the indicated concentrations of sorafenib. Equal loading was controlled by re-probing of the membrane with an actin-specific antiserum. Phosphorylation (activation) of c-Raf was analysed using anti-phospho-c-Raf (S338) (Upstate). Lysates were separated on a 12% gel to enable detection of NS5A and core on the same membrane. (e) Western blot analysis of lysate from HCV-JFH1wt or HCV-JFH-1GND-transfected Huh7.5 cells using a phospho-Erk-specific antiserum. Seventy-two hours after transfection, cells were incubated with the indicated concentrations of sorafenib for 24 h. Lysate from untreated Huh7.5 cells served as a positive control, lysate from uninfected cells as a negative control. Equal loading was controlled by re-probing of the membrane with an actin-specific antiserum. HCV replication was demonstrated by probing with a NS5A-specific antiserum. Samples were loaded on a 12% gel. (f) Western blot analysis of lysates from HCV-JFH1-replicating Huh7.5 cells that had been grown for 24 h in the presence of the indicated concentrations of the MEK-specific inhibitors UO126 or PD98059 using a NS5A-specific antiserum. Equal loading was controlled by re-probing of the membrane with an actin-specific antiserum. Samples were loaded on a 12% gel. (g) Analysis of viral replication by luciferase reporter gene assay. Forty-eight hours after electroporation, HCV JFH-1 luc-replicating Huh7.5 cells were transfected with pRafC4 which encodes a transdominant negative Raf mutant. Transfection with the empty vector pCDNA3.1 served as a negative control. (h) Western blot analysis of lysate from HCV-JFH1-replicating Huh7.5 cells using a NS5A-specific antiserum. Forty-eight hours after electroporation, HCV JFH-1-replicating cells were transfected with pRafC4. Transfection with the empty vector pCDNA3.1 served as a control. Samples were loaded on a 12% gel.

To demonstrate that under the chosen conditions sorafenib indeed inhibits c-Raf the phosphorylation of c-Raf was analysed using the phospho-Ser 338-specific antiserum. The western blot shows that increasing concentrations of sorafenib result in a decrease of Ser-338-phosphorylated c-Raf, indicating that c-Raf is indeed inhibited and as observed before a reduction of NS5A and core (fig 4d). In accordance to the inhibitory effect of sorafenib on c-Raf it was found that the activation of c-Raf targets was decreased. Western blotting using a phospho-Erk-specific antiserum revealed a dose-dependent decrease of phospho-Erk, reflecting a strong inhibition of c-Raf (fig 4e) Again, the inhibition of c-Raf correlated with the inhibition of HCV replication as demonstrated by the decreasing amounts of NS5A. Based on these data, an IC50 of 8.1 μmol/l for the sorafenib-dependent inhibition of c-Raf was determined. To further exclude the possibility that the observed effect of sorafenib on HCV replication might not be due to its direct inhibitory effect on c-Raf, but the result of an indirect inhibition of targets downstream from c-Raf, HCV-replicating cells were treated with two different MEK inhibitors (UO126 and PD98059) at a concentration corresponding to their 10-fold IC50. Interestingly, there was no inhibitory effect on HCV replication as demonstrated by western blotting (fig 4f).

To further demonstrate that the inhibitory effect of sorafenib on HCV replication is due to the compound’s specific interference with c-Raf, a transdominant-negative mutant of c-Raf (RafC4,26) was used to specifically block c-Raf activity. As shown in fig 4g,h, both analyses of HCV replication by luciferase assay in JFH-1luc-transfected cells, and of viral proteins in JFH-1 infected cells by western blotting, demonstrated that the specific inhibition of c-Raf by sorafenib indeed resulted in a strong reduction of HCV replication (Note that due to the limited transfection efficiency in JFH-1 cells, only about 30% of the HCV replicating cells express the transdominant negative Raf mutant.)

Taken together, these data indicate that the sorafenib-dependent inhibition of c-Raf is at least one essential factor that constitutes to the observed impairment of HCV replication by sorafenib.

Sorafenib affects the post-translational modification of NS5A

Several previous studies suggest that NS5A phosphorylation might affect HCV replication. We therefore hypothesised that the inhibitory effect of sorafenib might correlate with a modulation of the NS5A phosphorylation pattern. HCV-infected cells were incubated for 2 h with sorafenib before lysis. The lysate was analysed by one-dimensional or two-dimensional gel electrophoresis followed by western blotting. As shown in fig 5a, after one-dimensional separation (SDS-PAGE), the total amount of upper band (designated hyperphosphorylated NS5A) decreased in sorafenib-treated cells compared to controls. To confirm that the hyperphosphorylated form is affected alkaline phosphatase treatment of lysate derived from sorafenib treated cells was performed. The hyperphosphorylated form is much more susceptible to alkaline phosphatase-dependent dephosphorylation than the hypophosphorylated form.33 34 The alkaline phosphatase treatment demonstrates that hyperphosphorylated form disappears, while the hypophosphorylated form was not significantly affected. Lysate from untreated cells served as control (fig 5b). For a more detailed analysis, two-dimensional separations of lysates from sorafenib-treated and untreated HCV-replicating cells were performed. As shown in fig 5c, these studies demonstrate that the hyperphosphorylated NS5A variant disappeared in sorafenib-treated cells. Further, these analyses demonstrated that sorafenib resulted in the formation of additional forms with low pI (pH 3.8–4.2) most likely representing additional hypophosphorylated forms of NS5A, not present in untreated HCV-infected cells (fig 5c). These data demonstrate that sorafenib affects the post-translational modification of NS5A.

Figure 5

Sorafenib affects the phosphorylation pattern of NS5A. (a) One-dimensional SDS-PAGE of lysate from HCV-JFH1-replicating Huh7.5 cells followed by western blot analysis using a NS5A-specific antiserum. Seventy-two hours after infection, the cells had been incubated for 2 h with increasing concentrations of sorafenib. DMSO treatment served as a control. To separate hypo- and hyperphosphorylated forms an 8% gel was used. (b)Western blot analysis using a NS5A-specific antiserum of cellular lysate derived from untreated HCV positive cells (lane 1) or sorafenib treated (lanes 2 and 3). In the case of lane 3 the lysate was treated with alkaline phosphatase (lane 3). (c) Two-dimensional separation of lysate from HCV-JFH1-infected cells. Seventy-two hours after infection cells were treated for 2 h with 20 μmol/l sorafenib or with DMSO. Cells were lysed in two-dimensional lysis buffer (6 mol/l urea, 2% CHAPS, phosphatase inhibitor cocktail 2 (Roche, Penzberg, Germany), and IPG buffer (pH 3–10; GE Healthcare, Uppsala, Sweden) for 10 min. In the first dimension, lysate was subjected to isoelectric focussing over a pH range from 3 to 6. SDS-PAGE on an 8% gel was run as the second dimension. The gel was analysed by western blotting using a NS5A-specific antiserum. Two different exposures are shown. The changes in the hyperphosphorylated form of NS5A are indicated by the green arrows, the changes in the hypophosphorylated form by the red arrows. DMSO, dimethyl sulfoxide; HCV, hepatitis C virus; SDS-PAGE, sodium dodecyl sulfate polyacrylamide gel electrophoresis.

Inhibition of c-Raf results in the resolution of lipid droplets

Lipid droplets were identified as central intracellular compartments for HCV replication and morphogenesis.35 Thus, we aimed at studying the question whether sorafenib affects the subcellular localisation of viral structural and non-structural proteins. For this purpose, confocal immunofluorescence microscopy of untreated and sorafenib-treated HCV-infected cells was performed. In untreated cells, using core- or NS5A-specific antisera, specific structures representing lipid droplets could be identified (fig 6). By contrast, in cells treated with sorafenib for 2 h, these structures completely disappeared (fig 6). These data suggest that sorafenib affects the integrity of a structure that is central for HCV replication and/or morphogenesis.

Figure 6

Inhibition of c-Raf affects subcellular localisation of HCV proteins. Confocal laser scanning immunofluorescence microscopy (200- and 630-fold magnification) of formaldehyde-fixed HCV-infected Huh7.5 cells which had been grown 72 h after infection for 2 h in the presence of 15 μmol/l sorafenib (lower panel). DMSO-treated cells served as a control (upper panel). Immunofluorescence staining was performed using a rabbit-derived polyclonal NS5A-specific serum (red) and a mouse-derived core-specific serum (green). Cy2- and Cy3-conjugated secondary antibodies were used for detection. DAPI was used for visualisation of the nuclei. DAPI, 4,6-diamidino-2-phenylindole; DMSO, dimethyl sulfoxide; HCV, hepatitis C virus.

Discussion

The currently available antiviral strategies to treat patients with chronic hepatitis C are effective in only approx. half of the patients and are limited by severe side effects that often necessitate discontinuation of treatment.36 37 Therefore, novel antiviral concepts are being developed and explored. In this context we explored the efficiency of sorafenib in the HCV replication model system.

This study shows that in HCV-replicating cells, c-Raf is recruited to the replicon complex via NS5A, resulting in the activation of c-Raf. Based on these findings we wondered whether inhibition of c-Raf might affect HCV replication. To study this question, sorafenib, a well-characterised small molecule inhibitor of c-Raf, was chosen. Sorafenib has already been approved by the FDA and the EU for the treatment of RCC and HCC.20 Here, we demonstrate in the HCV replicon/infection model that sorafenib strongly inhibits HCV replication as demonstrated by the disappearance of viral RNA and of viral proteins below the detection limit. These effects were observed at concentrations of about 10–15 μmol/l. These levels are observed in the plasma from patients treated with sorafenib 400 mg twice daily18 21 At this dose sorafenib is well-tolerated with some grade 3 or 4 toxicities, however. The most common side effects of sorafenib are skin manifestations including alopoecia, hand–foot syndrome and erythema, presumably caused by the interference of sorafenib with VEGFR signalling.

Sorafenib is a potent inhibitor of c-Raf, but in addition, further targets can be affected by sorafenib, ie, VEGFR tyrosine kinase signalling. Inhibition of tyrosine kinases by sunitinib, however, did not affect HCV replication. Inhibition of c-Raf can result in a decreased proliferation. HCV preferentially replicates in proliferating cells. Since treatment of confluent HCV replicating cells with sorafenib causes inhibition of HCV replication as well (data not shown), it can be excluded that a simple inhibitory effect on cell proliferation is causative for the observed repression of HCV replication. This is in accordance with a recent report describing the robust HCV replication in resting cells.40 To demonstrate that the sorafenib effects are mediated by c-Raf, titration experiments using different concentrations of sorafenib were performed, and ie, the effect on c-Raf S338 phosphorylation was analysed. These experiments demonstrated comparable IC50 values for the sorafenib-dependent inhibition of c-Raf and of HCV replication (7–8 μmol/l) and are in accordance to data obtained for PLC/PRF/5 hepatoma cells (Alexander cell line).38 Furthermore, while inhibition of targets downstream from c-Raf had no effect on HCV replication, inhibition of c-Raf by co-expression of a transdominant negative c-Raf mutant resulted in a suppression of HCV replication. This suggests that the observed inhibitory effect of sorafenib on HCV replication is specifically c-Raf-mediated. The inhibition of viral replication by interference with a central signal transducer has also been described for the replication of influenza virus and MEK, or hepatitis B virus and c-Raf or Erk.39 40

NS5A is a phosphoprotein existing in two different phosphorylated forms the hypo- and the hyperphosphorylated state.6 10 11 41 Given the interaction between NS5A and c-Raf, we analysed whether sorafenib affects the state of phosphorylation of NS5A. In general, short-term treatment of HCV-replicating cells with sorafenib demonstrated a decrease of the hyperphosphorylated form of NS5A. A more detailed analysis by two-dimensional separation, however, revealed that sorafenib on the one hand decreased the hyperphosphorylated form, but on the other hand resulted in the formation of additional forms, most likely representing hypophosphorylated variants. These results suggest that different kinases are involved in the phosphorylation of NS5A. In contrast to previous reports indicating that a decrease of NS5A hyperphosphorylation by inhibition of CK1a is associated with an increased HCV replication,11 41 42 43 our results suggest that the sorafenib-dependent decrease of NS5A hyperphosphorylation and the formation of additional hypophosphorylated forms is associated with a reduced HCV replication. This discrepancy might be due to the fact that CK1a and c-Raf are directly or indirectly involved in the phosphorylation of different sites of NS5A that have different functions with respect to HCV RNA replication and viral morphogenesis. Viral morphogenesis indeed seems to be affected by sorafenib because it induces the rapid and complete destruction of vesicular structures that might represent lipid droplets. This is consistent with recent reports demonstrating that NS5A is not only part of the replicon complex and thereby involved in RNA replication, but also plays an essential role in viral morphogenesis, by mediating the transfer of de novo synthesised genomes from the replicon complex to the site of virus assembly in lipid droplets.8 9 35 Although the immunofluorescence microscopy demonstrates that core proteins are still present, it cannot be excluded that the disappearance of these vesicular structures is due to the impaired de novo synthesis of core protein under these conditions.

From our data, we conclude that sorafenib-dependent specific inhibition of c-Raf presents a novel strategy to inhibit HCV replication which should be further evaluated with respect to its clinic potential for the treatment of patients with chronic HCV infection.

Acknowledgments

We thank Dr M Mayr, Munich, for her help in identifying sera from sorafenib-treated patients.

REFERENCES

View Abstract

Footnotes

  • Funding This work was supported by a grant from the DFG-excellence cluster “Inflammation at Interfaces” to EH.

  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.