Hepatitis C Virus NS5A-Mediated Activation of Phos
http://www.100md.com
病菌学杂志 2005年第8期
School of Biochemistry and Microbiology and Astbury Centre for Structural Molecular Biology, University of Leeds, Leeds, United Kingdom
ABSTRACT
The hepatitis C virus (HCV) nonstructural NS5A protein has been shown to bind to and activate phosphoinositide 3-kinase (PI3K), resulting in activation of the downstream effector serine/threonine kinase Akt/protein kinase B. Here we present data pertaining to the effects of NS5A-mediated Akt activation on its downstream targets. Using a recombinant baculovirus to deliver the complete HCV polyprotein to human hepatoma cells in a tetracycline-regulable fashion, we confirm that expression of the complete HCV polyprotein also activates PI3K and Akt. We further show that this results in the inhibition of the Akt substrate Forkhead transcription factor and the stimulation of phosphorylation of a second key Akt substrate, glycogen synthase kinase-3? (GSK-3?). Phosphorylation of GSK-3? results in its inactivation; consistent with this, we show that expression of the HCV polyprotein results in the accumulation of ?-catenin. Finally, we show that levels of ?-catenin-dependent transcription are also elevated in the presence of the HCV polyprotein. Given the prevalence of ?-catenin mutations in many human tumors, especially colon and hepatocellular carcinomas, these data implicate NS5A-mediated PI3K activation as a contributory factor in the increasingly common association between HCV infection and the development of hepatocellular carcinoma.
INTRODUCTION
Hepatitis C virus (HCV) is recognized as a major public health problem; the World Health Organization estimates that as many as 170 million individuals (approximately 3% of the world population) are infected with this virus. In 85% of cases, the virus establishes a chronic infection, culminating in chronic inflammation, cirrhosis, and, increasingly, hepatocellular carcinoma (HCC). Indeed, HCV is now the leading cause of HCC and is the most common reason for liver transplantation in the West (47). Treatment is presently limited to the use of type 1 interferon in combination with the nucleoside analogue ribavirin; this therapeutic regime is successful in 40 to 80% of patients, depending on virus genotype.
The HCV genome is a single-stranded positive-sense 9.6-kb RNA molecule, comprising a single open reading frame coding for a polyprotein of 3,000 amino acids flanked by untranslated regions (UTRs). The 5' UTR contains an internal ribosome entry sequence, allowing cap-independent initiation of translation. The polyprotein is cleaved into 10 polypeptides by cellular and viral proteinases; one of these products is the nonstructural NS5A protein. NS5A has been the focus of much intensive investigation recently with regard to its potential role both in the cytopathology of HCV infection and in mediating viral immune evasion (for a review, see reference 33). NS5A has been reported to interact with a wide range of cellular proteins involved in, among others, the interferon response and cell cycle control. These interactions result in the modulation of various transcription factors, including NF-B (40, 46), STAT3 (31, 43), and AP-1 (32); furthermore, NS5A has been reported to promote anchorage-independent growth in NIH 3T3 murine fibroblasts and tumor formation in nude mice (22).
Recently, we (45) and others (26) have determined that NS5A interacts with the phosphoinositide 3-kinase (PI3K) signaling cascade. The binding of NS5A to the SH3 domain of the p85 regulatory subunit of PI3K stimulated the lipid kinase activity associated with the p110 catalytic subunit of this heterodimeric complex up to 10-fold (45). Furthermore, the activation of PI3K in cells expressing either NS5A alone or NS5A in the context of the subgenomic HCV replicon (28) resulted in increased phosphorylation and activity of the downstream kinase Akt/protein kinase B (45). The increased phosphorylation correlated with increased resistance to a variety of apoptotic stimuli, including serum starvation. Akt phosphorylates and inactivates many downstream target proteins: these include the Forkhead transcription factor (FKHR), the proapoptotic Bcl2 homologue Bad, and pro-caspase-9 (8, 11). One of the better-understood substrates of Akt is the serine/threonine kinase glycogen synthase kinase-3? (GSK-3?), which was first characterized as a negative regulator of glycogen synthesis (16) but was more recently shown to play a pivotal role in the regulation of the proto-oncogene ?-catenin. ?-Catenin has two distinct functions. Most of the protein is located at the cell membrane, where it is involved in cell-cell contact via its association with the cytoplasmic domain of E-cadherin (41). A second pool of ?-catenin is located both in the nucleus and in the cytoplasm, where it mediates Wnt signaling. In the absence of mitogenic stimulation, a multiprotein destruction complex containing GSK-3?, Axin, and the tumor suppressor adenomatous polyposis coli (APC) acts to promote the phosphorylation of serine and threonine residues in the N terminus of ?-catenin and thereby targets it for proteasome-mediated degradation via the F-box-containing protein ?TrCP (25). When Akt phosphorylates GSK-3?, it inactivates the latter, leading to the accumulation of ?-catenin, which can enter the nucleus and participate in the formation of transcriptionally active complexes with members of the Tcf/Lef family (2).
Interestingly, the stabilization of ?-catenin, as a result of mutations either in ?-catenin itself that block phosphorylation or in other components in the pathway, is associated with a range of human tumors, in particular HCC (12, 15, 24). Given that HCV infection is increasingly associated with HCC, we were interested in exploring the effects of NS5A-mediated Akt activation on the GSK-3?/?-catenin signaling pathway. We show here that NS5A, expressed either alone or in the context of the complete HCV polyprotein, mediated a reduction in the transcriptional activity of the Forkhead transcription factor, FKHR, and an increase in GSK-3? phosphorylation. These effects were dependent on the Akt pathway, as they were abolished both by a pharmacological inhibitor of PI3K and by a kinase-inactive (dominant negative) form of Akt. Furthermore, we found an increase in the overall levels of ?-catenin within cells expressing either NS5A or the complete HCV polyprotein. This accumulation correlated with increased stability of ?-catenin and resulted in a concomitant increase in the levels of ?-catenin-dependent transcription within the cells. Given these results, we propose that an increase in the activity of ?-catenin, in conjunction with an Akt-dependent decrease in apoptosis, may predispose HCV-infected hepatocytes to neoplastic transformation.
MATERIALS AND METHODS
Plasmid constructs. Vectors expressing NS5A have been described elsewhere (32, 45). The Tcf-luciferase reporter pGL3-OT, the mutant Tcf-luciferase reporter pGL3-OF (44), and the ?-catenin expression vector were provided by Bert Vogelstein (Howard Hughes Medical School, Baltimore, Md.). The hemagglutinin-tagged (HA)-GSK-3? expression vector was provided by Vivianne Ding (University of California, San Francisco). The Forkhead transcription factor luciferase reporter (pGL3-FHRE) (4) and the HA-Akt(K179M) (kinase-inactive) expression vector (13) were provided by Mike Greenberg (Harvard Medical School, Boston, Mass.).
Cell culture. The human hepatoma cell line HepG2 was maintained in minimal essential medium supplemented with 10% fetal calf serum (FCS), 2 mM L-glutamine, 100 IU of penicillin/ml, and 100 μg of streptomycin/ml plus 1% nonessential amino acids. Cos-7 (African green monkey kidney) cells were propagated in Dulbecco's modified Eagle medium supplemented as for HepG2 cells. Cells were cultured in a humidified 5% CO2 incubator at 37°C. Sf9 cells for propagation of recombinant baculoviruses were maintained in TC100 medium supplemented with 10% FCS, 100 IU of penicillin/ml, and 100 μg of streptomycin/ml. Where appropriate, cells were incubated with Ly294002 (50 μM) (Cell Signaling Technology) or cycloheximide (100 μg/ml) (Sigma).
Generation and propagation of recombinant baculoviruses. A baculovirus transfer vector carrying the replication-deficient HCV genome was generated as follows. Briefly, the previously described vector pBACH77(HV)tet (35) was restricted at the AflII site found at the beginning of the 3' UTR and at an EcoRI site immediately preceding the hepatitis delta virus ribozyme in order to remove both the ribozyme and the majority of the 3' UTR. Pfu polymerase was used to polish the resulting DNA, which was then religated, generating pBACH773'UTR. Generation of virus was as described previously (35), with the integrity of virus clones being confirmed by Southern blot analysis and immunoblotting for HCV antigens Core, E2, NS3, NS5A, and NS5B in virus-transduced HepG2 cells. The baculovirus vector for constitutive expression of the tTA tetracycline repressor-VP16 transactivator fusion protein in mammalian cells (BactTA), together with a baculovirus expressing the Escherichia coli ?-galactosidase gene under the control of the Ptet promoter (BacINDLacZtet), has been described previously (35). Tetracycline was used at a concentration of 5 μg/ml in all experiments.
Analysis of the Akt pathway by immunoblotting. HepG2 cells were seeded in 90-mm-diameter dishes precoated with rat tail collagen at a density of 2 x 106 cells per dish and 24 h later were transduced with BactTA (35) and BACH773'UTR or BacINDLacZtet for 4 h in the presence or absence of tetracycline (each virus was at a final concentration of 2.5 x 107 PFU/ml) as described previously (32, 35). The transduced cells were then transfected for 5 h with plasmids expressing HA-Akt(K179M) or HA-GSK-3? (2 μg in total) by using the Lipofectin reagent (Invitrogen) according to manufacturer's recommendations in the presence or absence of tetracycline. The total amount of DNA was kept constant by inclusion of empty vector DNA. To suppress external activation of mitogenic signaling pathways, cells were subsequently incubated in growth medium containing reduced serum (0.5%). Cells were harvested 24 h posttransfection, resuspended in 500 μl of ice-cold radioimmunoprecipitation assay lysis buffer (50 mM Tris-HCl [pH 8.0], 150 mM NaCl, 1% NP-40, 0.1% sodium dodecyl sulfate, 5 mM EDTA, with the addition of 1 mM NaOV, 1 mM NaF, 1 mM okadaic acid, and 500 nM cantharidin). After total protein quantification (BCA assay; Pierce), lysates were subjected to analysis by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and immunoblotted with antibodies directed against GSK-3?, GSK-3? phosphorylated on Ser9, Akt, Akt phosphorylated on Ser473, PDK1 phosphorylated on Ser241, or ?-catenin (all from Cell Signaling Technology) or a sheep polyclonal anti-NS5A serum (32). Blots were probed with peroxidase-conjugated secondary antibody and detected with the enhanced chemiluminescence system (Amersham Pharmacia). All experiments were repeated three times. The immunoblots shown were chosen as the clearest representations of the data.
Luciferase assays. HepG2 cells (5 x 105) were cultured in 6-well dishes precoated with rat tail collagen and transduced as described above. Transduced cells were then transfected with plasmids expressing the appropriate luciferase reporter(0.5 μg), HA-Akt(K179M) or HA-GSK-3?, or ?-catenin expression vectors (1.5 μg) with Lipofectin (Invitrogen) following the manufacturer's instructions. A Renilla luciferase reporter construct (pRLTK) was used as an internal control for transfection efficiency. The total amount of DNA was kept constant by inclusion of empty vector DNA. Cells were harvested at 24 h in 200 μl of passive lysis buffer (Promega). The quantitation of relative light units was done by using the dual luciferase Stop & Glo reagent (Promega) and a luminometer (EG&G Berthold) with a dual injector system. All assays were performed in triplicate, and each experiment was repeated a minimum of three times.
Immunofluorescence. HepG2 cells were seeded into 12-well dishes containing coverslips (105 cells per well), allowed to settle overnight, and transduced as previously described. After 24 h, the cells were fixed in 4% paraformaldehyde for 15 min, washed in phosphate-buffered saline (PBS), and permeabilized with 0.1% Triton X-100 for 10 min. After washing in PBS, the cells were incubated for 1 h with the appropriate antibody diluted in PBS containing 10% FCS. A sheep polyclonal anti-NS5A antibody (32) was used for the primary stain, and bound antibody was detected by using a donkey anti-sheep fluorescein isothiocyanate (Jackson ImmunoResearch) with a Hoechst 33342 (Molecular Probes) counterstain.
Other procedures. PI3K assays and immunoblotting analyses were performed as previously described (32, 35, 45).
RESULTS
Expression of the complete HCV polyprotein stimulates both PI3K and Akt activity in HepG2 cells. We have previously shown that NS5A expression, either alone or as part of the HCV subgenomic replicon, induced activation of both PI3K and Akt in both Cos-7 cells and the human hepatoma cell line Huh-7 (45). Activation of PI3K was mediated via a direct protein-protein interaction between NS5A and the SH3 domain of the PI3K p85 regulatory subunit. Given that the subgenomic replicon expressed only the nonstructural NS3 through NS5B proteins (constituting the C-terminal two-thirds of the polyprotein), we considered it important to verify that these effects of NS5A were also manifest in the context of the complete polyprotein. Although full-length replicons expressing the complete HCV polyprotein are now available, for the reasons outlined below we decided to utilize a novel baculovirus delivery system developed in our laboratory to express the HCV polyprotein in hepatocyte-derived cell lines. This recombinant virus contains the complete HCV polyprotein coding sequence flanked by the 5' UTR but lacking the 3' UTR. A tetracycline-regulable promoter (Ptet) (23) is positioned upstream of the 5' UTR such that the initiation of transcription coincides with the start of the 5' UTR (Fig. 1a). The lack of the 3' UTR means that this virus (BacH773'UTR) is replication defective (34). We have previously shown that these baculovirus vectors work most efficiently in the hepatoblastoma-derived cell line HepG2 (35), so we conducted subsequent experiments in this cell line. The use of this tetracycline-regulable system provided two key advantages over the use of the replicon system for these experiments. First, experiments could be performed on homogeneous cell populations not subjected to long-term selection (with the potential for selection of cells with defects in signaling pathways). Second, the tight regulation allowed for robust internal controls, as signaling events could be compared in samples derived from a single population of transduced cells incubated in the presence or absence of tetracycline prior to analysis. In addition, as discussed below, Huh-7 cells were not suitable for some of the studies described here due to inherent defects in the regulation of GSK-3? phosphorylation (15). As shown in Fig. 1b and c, when cells were cotransduced with BacH773'UTR and a baculovirus expressing the tTA tetracycline repressor-VP16 transcriptional activator fusion protein (BactTA) (35), expression of the HCV polyprotein in HepG2 cells was tetracycline regulable. In the presence of tetracycline, expression was tightly repressed, whereas in the absence of the antibiotic, the HCV polyprotein was expressed and efficiently proteolytically processed, as judged by immunoblotting for either the E2 envelope glycoprotein (Fig. 1b, left panel) or NS5A (Fig. 1b, right panel). This result is consistent with our previous observations with a LacZ-expressing baculovirus (BacINDLacZtet [35]), in which we showed that the tetracycline-responsive promoter is tightly regulated in HepG2 cells: in the presence of tetracycline, levels of ?-galactosidase were barely detectable above background, whereas in the absence of tetracycline they were increased 1,000-fold (35). As shown in Fig. 1c, the BacMAM delivery system was highly efficient in HepG2 cells; in the absence of tetracycline, essentially all of the cells in the population expressed the HCV polyprotein, as judged by indirect immunofluorescent detection of NS5A.
We previously showed that the expression of either the HCV subgenomic replicon or NS5A alone resulted in a 10-fold increase in the activity of PI3K in vivo (45). As shown in Fig. 2a, the incubation of cells cotransduced with BacH773'UTR and BactTA in the absence of tetracycline resulted in a threefold increase in PI3K activity compared to cells incubated in the presence of tetracycline. Levels of p85 in immunoprecipitates were verified by immunoblotting (Fig. 2a, panel ii), and there were no significant differences, demonstrating that the increased PI3K activity was not due to elevated levels of PI3K expression. Panel iii of Fig. 2a confirms appropriate expression of NS5A. This result confirms and extends our previous data (45), demonstrating that the expression of the complete HCV polyprotein also stimulates PI3K activity. Our previous study had also shown that NS5A-mediated activation of PI3K resulted in subsequent activation of the protein kinase Akt; we therefore investigated the activation status of Akt in these baculovirus-transduced cells. As shown in panel i of Fig. 2b, levels of Ser473-phosphorylated Akt were elevated in cells incubated in the absence of tetracycline (Fig. 2b, compare lanes 1 and 2). To confirm that this was mediated through the PI3K pathway, cells were also treated with 50 μM Ly294002, a specific PI3K inhibitor, for 24 h prior to harvest. This treatment significantly reduced both basal and HCV-induced activation of Akt (Fig. 2b, panel ii, lanes 3 and 4). No activation of either PI3K or Akt was seen in cells cotransduced with BactTA and a control baculovirus (BacINDLacZtet) expressing ?-galactosidase (data not shown). As expected, there were no differences observed in the total levels of Akt expression (Fig. 2b, panel ii) or the levels of phosphorylated PDK1, the kinase that phosphorylates Akt (Fig. 2b, panel iii). PDK1 has previously been shown to be constitutively active (1). Appropriate expression of NS5A in lysates of cells incubated without tetracycline confirmed the controlled expression of the complete HCV polyprotein (Fig. 2b, panel iv). These data indicate that Akt is activated in a PI3K-dependent manner in cells expressing the whole HCV genome.
Expression of the complete HCV polyprotein modulates Akt-mediated downstream events. Activated Akt phosphorylates and inactivates a number of downstream targets, including FKHR. Once phosphorylated, FKHR binds to 14-3-3 proteins and is sequestered in the cytoplasm. In order to assess the downstream effects of HCV-induced Akt phosphorylation, we therefore analyzed the activity of FKHR by using a luciferase reporter construct responsive to this transcription factor. HepG2 cells were transduced with BacH773'UTR and BactTA and subsequently transfected with an FKHR-luciferase construct (Fig. 3a). In the presence of serum, levels of FKHR-responsive luciferase were low, whereas in the absence of serum there was a fivefold increase in luciferase levels, consistent with the role of FKHR in the induction of apoptosis (Fig. 3a, compare the black bars). Expression of the HCV genome following the removal of tetracycline resulted in a 45% reduction in FKHR activity in serum-starved cells (Fig. 3a, compare black and white bars), whereas no HCV-mediated inhibition of FKHR activity was seen in cells grown in the presence of serum, presumably due to constitutive activation of PI3K and Akt activity by serum components (e.g., growth factors or insulin). There was no significant inhibition of FKHR activity in control cells transduced with BacINDLacZtet (LacZ) following removal of tetracycline (Fig. 3a, compare black and white bars). The HCV-mediated inhibition was abolished by transfection with a plasmid expressing an HA-tagged kinase-inactive (dominant negative) mutant of Akt (K179M). To confirm the role of NS5A-mediated activation of PI3K, we performed a similar study in Cos-7 cells transiently transfected with plasmids expressing NS5A and HA-Akt(K179M) (Fig. 3b). As in HepG2 cells, levels of FKHR activity were elevated in Cos-7 cells incubated in the absence of serum (Fig. 3b, black bars) compared with cells grown in 10% serum. Transfection with an NS5A plasmid (Fig. 3b, white bars) resulted in a 60% reduction in FKHR activity in serum-starved cells; this inhibition was abolished by cotransfection with a plasmid expressing HA-Akt(K179M) (Fig. 3b, grey bars). These data are consistent with the activation of Akt by NS5A, resulting in increased Akt-dependent phosphorylation and inactivation of FKHR.
A second key substrate of Akt is the serine/threonine kinase GSK-3?. Our initial experiments to determine whether the phosphorylation status of GSK-3? was influenced by NS5A-mediated PI3K activation revealed a limitation of the replicon system in the context of this study. Using an antibody specific for GSK-3? phosphorylated on Ser9, we were unable to observe elevated levels of phosphorylated GSK-3? in Huh-7 cells expressing the HCV subgenomic replicon or NS5A alone in comparison to control cells expressing neomycin phosphotransferase (45) (data not shown). Subsequent analysis of the literature revealed that Huh-7 cells, along with some other HCC cells lines, such as Hep3B and Mahlavu cells, contain very high constitutive levels of phosphorylated GSK-3? (15), consistent with an inherent defect in GSK-3? regulation. In contrast, HepG2 cells exhibited a very low level of GSK-3? phosphorylation, comparable to normal primary hepatocytes. We therefore transduced HepG2 cells with the recombinant baculoviruses described above and analyzed GSK-3? phosphorylation. Figure 3c, panel i shows that levels of GSK-3? phosphorylation were significantly increased in cells expressing the HCV polyprotein in the absence of tetracycline compared to cells incubated in the presence of tetracycline (Fig. 3c, compare lanes 1 and 2). We confirmed that this result was due to stimulation of the PI3K/Akt signaling pathway by demonstrating that phosphorylation of GSK-3? was inhibited either by Ly294002 or by transfection of an HA-Akt(K179M) plasmid (Fig. 3c, panel i, lanes 3 to 6). Again, overall levels of GSK-3? were unaffected (Fig. 3c, panel ii), and panels iii and iv of Fig. 3c confirm appropriate expression of Akt and NS5A.
Expression of the complete HCV polyprotein results in the accumulation of ?-catenin. A key downstream target of GSK-3? is the proto-oncogene ?-catenin. Normally, ?-catenin is sequestered in the cytoplasm as part of a complex with Axin and APC. One pool of GSK-3? is also associated with the Axin complex and, when active, mediates serine/threonine phosphorylation of residues at the N terminus of ?-catenin, thus providing a recognition signal for ?TrCP, the receptor component of the Skp1/Cullin/F-box protein (SCF)-?TrCP E3 ubiquitin ligase complex, targeting the protein for proteasomal degradation (25). Increased phosphorylation and concomitant inhibition of GSK-3? would therefore be predicted to prevent ?-catenin phosphorylation, resulting in decreased ubiquitin-mediated proteolysis and increased levels of ?-catenin. To determine whether HCV polyprotein-mediated stimulation of Akt and GSK-3? phosphorylation was able to effect a similar stabilization, steady-state levels of ?-catenin were examined in HepG2 cells transduced with the recombinant baculoviruses described above. HepG2 cells express two ?-catenin alleles, one wild-type form and a mutated form lacking amino acids 25 to 140 (5, 14), which cannot be phosphorylated and thus accumulates in the nucleus, where it is constitutively active. Levels of endogenous wild-type ?-catenin in HepG2 cells are therefore low (5), and to facilitate the analysis we transfected them with a ?-catenin expression plasmid (37). It should be pointed out that the endogenous truncated ?-catenin in HepG2 cells cannot be detected with the monoclonal antibody used in this study, as it binds to an epitope near the N terminus and thus did not interfere with the interpretation of the data. Panel i of Fig. 4a shows that in cells transduced with BacH773'UTR and BactTA, levels of ?-catenin were increased in the absence of tetracycline (Fig. 4a, compare lanes 1 and 2). Again we confirmed that this increase was mediated through the PI3K/Akt pathway by demonstrating that the incubation of cells with Ly294002, or transfection with an HA-Akt(K179M) plasmid, blocked the increase in ?-catenin levels (Fig. 4a, lanes 3 to 6). Panels ii and iii of Fig. 4a confirm appropriate expression of Akt and NS5A. To confirm that baculovirus transduction and/or treatment with tetracycline did not affect ?-catenin levels, we also examined lysates from mock-transduced cells or cells transduced with BacINDLacZtet and BactTA. Figure 4b, panel i shows that there was no difference in the levels of ?-catenin in any of these lysates; panel ii confirms appropriate expression of ?-galactosidase. We conclude from these experiments that the expression of the HCV polyprotein results in increased accumulation of ?-catenin.
HCV-mediated accumulation of ?-catenin is due to increased protein stability. Although it was likely that the increased steady-state levels of ?-catenin observed in cells expressing the complete HCV polyprotein were due to a reduction in proteasomal degradation, we wished to confirm this by analyzing the stability of ?-catenin. To examine this, we transduced HepG2 cells as described above and treated them with cycloheximide to block protein synthesis. Samples were taken at intervals after treatment, and levels of ?-catenin were analyzed by immunoblotting. As shown in Fig. 4c, panel i, in the presence of tetracycline, levels of ?-catenin declined over an 8-h time course, consistent with proteasomal degradation. However, in the absence of tetracycline, ?-catenin levels were initially higher than in the presence of tetracycline (as seen in Fig. 4a) and did not decline as rapidly over the course of the experiment. Figure 4c, panel ii shows that levels of a control protein, the endoplasmic reticulum resident chaperone calreticulin, remained constant both in the presence and in the absence of tetracycline; while panel iii shows that NS5A was in fact somewhat more unstable than ?-catenin, very little NS5A was present at 8 h after cycloheximide treatment. To confirm that this stabilization of ?-catenin was mediated via the PI3K/Akt pathway, cells were transfected with an HA-Akt(K179M) plasmid (Fig. 4c, panel iv). As expected, in these cells ?-catenin was rapidly degraded, and the expression of the HCV polyprotein had no effect on ?-catenin stability.
To confirm that this effect on ?-catenin stability was due to NS5A-mediated stimulation of the PI3K/Akt pathway, we repeated this experiment in Cos-7 cells transiently transfected with an NS5A expression plasmid. Figure 4d, panel i shows that in control cells, degradation of ?-catenin occurred at a rate similar to that in HepG2 cells. In cells expressing NS5A, levels of ?-catenin remained constant throughout the time course (Fig. 4d, panel ii). Transfection with an HA-Akt(K179M) plasmid was able to partially restore the degradation of ?-catenin, consistent with the role of the PI3K/Akt pathway in regulating ?-catenin stability.
Expression of the complete HCV polyprotein results in elevated ?-catenin-dependent transcriptional activity. Nuclear accumulation of ?-catenin allows association with Tcf-Lef transcription factors and stimulates the expression of a series of cellular genes that are involved in cell proliferation (2). To test whether stabilization of ?-catenin by NS5A led to increased nuclear transcriptional activity of ?-catenin, HepG2 cells were transduced with recombinant baculoviruses as described above and subsequently transfected with luciferase reporters carrying three copies of either wild-type Tcf-binding sites (pGL3-OT) or mutated binding sites that were nonresponsive to Tcf (pGL3-OF) (42). As shown in Fig. 5a, levels of luciferase activity from the pGL3-OT reporter were approximately twofold higher in the absence of tetracycline than in cells incubated in the presence of tetracycline, consistent with HCV-mediated stimulation of wild-type ?-catenin-dependent transcription. No such increase was seen in cells transduced with BacINDLacZtet and BactTA when they were incubated in the absence of tetracycline. The relatively high basal activity of the reporter in HepG2 cells is consistent with the presence of the truncated, constitutively active allele of ?-catenin (5, 14). We confirmed that this stimulation was mediated via the PI3K/Akt pathway by transfecting cells with an HA-Akt(K179M) plasmid. As expected, this resulted in a dramatic reduction in ?-catenin-mediated reporter gene expression, consistent with the involvement of the PI3K/Akt pathway. Although the expression of HA-Akt(K179M) reduced overall levels of ?-catenin activity, an HCV polyprotein-mediated increase in reporter activity was still apparent. This is consistent with the fact that Akt functions as a multimer (8), and the presence of an Akt(K179M) monomer within the multimer does not completely abolish wild-type Akt function. When cells were transfected with a ?-catenin expression plasmid to increase overall levels of the wild-type protein, luciferase levels were elevated as expected; again, the expression of the HCV polyprotein stimulated the reporter twofold, whereas BacINDLacZtet had no effect. To confirm that the activation of ?-catenin transcriptional activity was due to Akt activation, we transfected cells with expression plasmids for both ?-catenin and HA-Akt(K179M). Under these conditions, the activation of ?-catenin was abolished. Lastly, we transfected cells with a plasmid expressing HA-tagged GSK-3?. Although this had no effect on basal ?-catenin activity, it blocked the HCV polyprotein-mediated enhancement, providing further evidence for the role of PI3K/Akt in the accumulation of ?-catenin; in the presence of high levels of GSK-3?, there would be an increase in the amount of unphosphorylated (active) GSK-3? and a concomitant decrease in the levels of ?-catenin. No significant responses were seen with the mutant (pGL3-OF) reporter (data not shown).
To confirm the involvement of NS5A, we repeated these experiments in transiently transfected Cos-7 cells. As shown in Fig. 5b, luciferase levels were increased threefold following cotransfection with an NS5A expression vector. Transfection of an HA-Akt(K179M) plasmid was less effective at reducing basal ?-catenin-responsive transcription in these cells. This result could be due to higher levels of Akt in Cos-7 cells than in HepG2 cells (data not shown); however, HA-Akt(K179M) did block NS5A-mediated augmentation of ?-catenin activity. Transfection with a ?-catenin expression plasmid increased overall luciferase levels while maintaining the NS5A-mediated augmentation. As observed in HepG2 cells expressing the full-length polyprotein (Fig. 5a), the transfection of cells with expression plasmids for both ?-catenin and HA-Akt(K179M) abolished the activation of ?-catenin. We further showed that the effect of HA-Akt(K179M) was titratable, as the level of ?-catenin transcription was inversely proportional to the amount of HA-Akt(K179M) expression plasmid transfected. Lastly, transfection of an HA-GSK-3? plasmid reduced both basal ?-catenin-responsive transcription and NS5A-mediated augmentation. We conclude from these data that the increased stability of ?-catenin in cells expressing either NS5A or the complete HCV polyprotein results in elevated levels of ?-catenin-responsive transcription.
DISCUSSION
We have previously shown that the HCV NS5A protein activates PI3K by binding to the SH3 domain of the PI3K p85 regulatory subunit, and this subsequently results in the activation of the downstream effector kinase Akt (45). In this paper, we extend these results to analyze some of the downstream events resulting from the activation of PI3K and Akt. We demonstrate that the expression of NS5A, either alone or in the context of the full-length HCV polyprotein, results in the increased phosphorylation of Akt and the inactivation of the Akt substrates FKHR and GSK-3?. Furthermore, GSK-3? inactivation results in the stabilization of the proto-oncogene ?-catenin and thereby stimulates ?-catenin-dependent transcription (Fig. 6). Interestingly, a recent paper (9) identified the yeast Akt homologue as one of eight kinases in the yeast "kinome" able to directly phosphorylate NS5A; recombinant human Akt was also able to phosphorylate NS5A in vitro, suggesting that there may be further interactions between NS5A and the Akt pathway. However, the site(s) of Akt phosphorylation on NS5A and the functional significance remain to be determined.
GSK-3? and ?-catenin represent a point of convergence of signaling from both PI3K/Akt and the Wnt pathway. The latter was originally described for Drosophila, where it plays a key role in embryogenesis; binding of the secreted Wnt protein to its receptor (Frizzled) deactivates a complex consisting of GSK-3?, APC, and Axin, blocking ?-catenin phosphorylation. Interestingly, different pools of GSK-3? exist within cells; as well as that associated with the Axin complex, there are other pools within both the cytoplasm and the nucleus. These pools of GSK-3? phosphorylate a variety of substrates and are involved in a number of processes, such as glycogen metabolism, vesicle transport, and translation (for a review, see reference 16). Components of the Wnt pathway, such as Dishevelled (Dvl), have been shown to be required to link activated Akt to GSK-3? associated with the Axin complex (19). It will be of interest to determine whether such factors are also required for the inactivation of GSK-3? by Akt in HCV-expressing cells. In addition, it will be of interest to investigate the effects of HCV on other substrates and processes regulated by GSK-3?.
Over recent years, the Wnt/?-catenin signaling pathway has been identified as a common target for perturbation by viruses. With the notable exception of human immunodeficiency virus type 1 (HIV-1), these are all tumorigenic DNA viruses that establish chronic infections, consistent with the fact that ?-catenin-regulated genes positively regulate cell cycle and division (e.g., c-Myc, cyclin D1). Interestingly, the mechanisms by which these viruses target ?-catenin are all distinct. The -herpesvirus Epstein-Barr virus (EBV) encodes a plasma membrane-localized latent membrane protein 2A (LMP2A) that activates PI3K and Akt, resulting in GSK-3? inactivation and ?-catenin stabilization (38). The EBV LMP1 has independently been shown to activate PI3K/Akt, although it has not been determined whether this contributes to ?-catenin stabilization in EBV-infected cells. In contrast, the latency-associated nuclear antigen (LANA) of another -herpesvirus, Kaposi's sarcoma-associated herpesvirus, binds to GSK-3? and sequesters it in the nucleus, preventing ?-catenin phosphorylation (17, 18). Hepatitis B virus X protein achieves ?-catenin stabilization by suppressing GSK-3? activity in an Src-kinase-dependent manner (6). The large T antigen of the human neurotropic polyomavirus JCV interacts directly with ?-catenin, stabilizing it and promoting its nuclear accumulation (20). Lastly, the Vpu protein of HIV-1 binds to ?TrCP and blocks the ubiquitinylation and proteasomal degradation of ?-catenin (and other SCF-?TrCP substrates, such as IB and ATF4) (3). A number of other viral proteins have been shown to stimulate PI3K/Akt signaling; these include human papillomavirus type 16 E5, polyoma middle T antigen, and HIV-1 Nef (reviewed in reference 10). However, whether these proteins also upregulate downstream events involving GSK-3? or ?-catenin remains to be evaluated.
Our data showed that the HCV polyprotein or NS5A alone activated ?-catenin transcriptional activity two- to threefold (Fig. 5). Although this is a modest increase, it is comparable to the magnitude of ?-catenin activation observed in the context of other viral proteins, e.g., EBV LMP2A, Kaposi's sarcoma-associated herpesvirus LANA, and JCV large T antigen. The majority of these observations were made in transformed cell lines; in this regard, the involvement of ?-catenin in oncogenesis suggests that the basal levels of ?-catenin may already be elevated in these cells. It is conceivable that in the context of a natural infection of primary, untransformed cells, the magnitude of ?-catenin activation might be much greater for all of these viral proteins, including NS5A.
In addition to disruption by oncogenic viruses, the genetic dysregulation of the Wnt signaling pathway has been shown to play a pivotal role in carcinogenesis in a number of human tumor types, in particular colon carcinoma, in which 80 to 90% of cases have mutations either in APC or in ?-catenin itself (42), and HCC, in which 20 to 30% of cases are associated with mutations in Axin-1 or ?-catenin (14). Activating mutations in ?-catenin are predominantly located at sites of serine/threonine phosphorylation near the N terminus, although some larger deletions have been identified. Interestingly, in the case of HCC, stabilization of ?-catenin alone is insufficient to transform hepatocytes. The targets for ?-catenin-mediated transcriptional activation include many genes that control cell growth and cell cycle, such as c-Myc and cyclin D1, as well as genes involved in invasion, such as matrilysin (30), providing a rationale for the involvement of ?-catenin in tumorigenesis.
HCV infection is increasingly associated with the development of HCC (49). The fact that HCC does not occur in all cases of HCV, together with the extended time period between initial infection and the development of tumors (as long as 30 or 40 years), suggests that HCV is not directly tumorigenic. It is more likely that HCV infection predisposes patients to the development of tumors, such that a further event will precipitate the neoplastic transformation of infected hepatocytes. Given the constant exposure of hepatocytes to xenobiotics, many of which may be potentially carcinogenic, such a second event is likely to be a relatively frequent occurrence. In one study (27), activating mutations in ?-catenin were found in 41% of HCC cases associated with HCV infection. All of these tumors showed increased nuclear accumulation of ?-catenin; however, additional tumors without ?-catenin mutations also showed this accumulation, suggesting that a distinct mechanism might operate in these cases. Although the development of HCC is certainly influenced by a combination of viral, host, and environmental factors, the data presented here suggest that NS5A-mediated stabilization of ?-catenin may be a contributing factor in the association between HCV infection and HCC. This hypothesis is strengthened by previous results from our laboratory (45) and others (7, 21, 36) showing that NS5A mediates an increase in the resistance of cells to proapoptotic stimuli by a variety of mechanisms. As well as direct effects on the PI3K pathway, NS5A has been reported to bind to and inhibit both p53 (29) and Bax (7), thereby blocking the apoptotic response. We propose, therefore, that NS5A, as well as acting in some as-yet-undefined fashion during the process of viral RNA replication, can be regarded as a viral "survival factor," promoting the persistence of HCV. In the short term, it will be important to more fully understand the biochemical mechanism by which NS5A activates PI3K, leading to ?-catenin stabilization; however, in the long term, our data suggest that the development of chemotherapeutics targeted to NS5A might be beneficial in treating chronic HCV infection by potentiating the elimination of virus-infected cells.
ACKNOWLEDGMENTS
We thank Bert Vogelstein, Vivianne Ding, and Mike Greenberg for the kind gifts of plasmid reagents.
This work was supported by grants from the British Medical Research Council (G9801522) and the Wellcome Trust (0671250). A.S. was supported by a Biotechnology and Biological Sciences Research Council Ph.D. studentship.
A.S. and A.M. contributed equally to this work.
Present address: MRC Protein Phosphorylation Unit, School of Life Sciences, University of Dundee, Dundee DD1 5EH, United Kingdom.
REFERENCES
Alessi, D. R., M. Deak, A. Casamayor, F. B. Caudwell, N. Morrice, D. G. Norman, P. Gaffney, C. B. Reese, C. N. MacDougall, D. Harbison, A. Ashworth, and M. Bownes. 1997. 3-Phosphoinositide-dependent protein kinase-1 (PDK1): structural and functional homology with the Drosophila DSTPK61 kinase. Curr. Biol. 7:776-789.
Barker, N., P. J. Morin, and H. Clevers. 2000. The Yin-Yang of TCF/beta-catenin signaling. Adv. Cancer Res. 77:1-24.
Besnard-Guerin, C., N. Belaidouni, I. Lassot, E. Segeral, A. Jobart, C. Marchal, and R. Benarous. 2004. HIV-1 Vpu sequesters beta-transducin repeat-containing protein (betaTrCP) in the cytoplasm and provokes the accumulation of beta-catenin and other SCFbetaTrCP substrates. J. Biol. Chem. 279:788-795.
Brunet, A., A. Bonni, M. J. Zigmond, M. Z. Lin, P. Juo, L. S. Hu, M. J. Anderson, K. C. Arden, J. Blenis, and M. E. Greenberg. 1999. Akt promotes cell survival by phosphorylating and inhibiting a Forkhead transcription factor. Cell 96:857-868.
Carruba, G., M. Cervello, M. D. Miceli, R. Farruggio, M. Notarbartolo, L. Virruso, L. Giannitrapani, R. Gambino, G. Montalto, and L. Castagnetta. 1999. Truncated form of beta-catenin and reduced expression of wild-type catenins feature HepG2 human liver cancer cells. Ann. N. Y. Acad. Sci. 886:212-216.
Cha, M. Y., C. M. Kim, Y. M. Park, and W. S. Ryu. 2004. Hepatitis B virus X protein is essential for the activation of Wnt/beta-catenin signaling in hepatoma cells. Hepatology 39:1683-1693.
Chung, Y. L., M. L. Sheu, and S. H. Yen. 2003. Hepatitis C virus NS5A as a potential viral Bcl-2 homologue interacts with Bax and inhibits apoptosis in hepatocellular carcinoma. Int. J. Cancer 107:65-73.
Coffer, P. J., J. Jin, and J. R. Woodgett. 1998. Protein kinase B (c-Akt): a multifunctional mediator of phosphatidylinositol 3-kinase activation. Biochem. J. 335:1-13.
Coito, C., D. L. Diamond, P. Neddermann, M. J. Korth, and M. G. Katze. 2004. High-throughput screening of the yeast kinome: identification of human serine/threonine protein kinases that phosphorylate the hepatitis C virus NS5A protein. J. Virol. 78:3502-3513.
Cooray, S. 2004. The pivotal role of phosphatidylinositol 3-kinase-Akt signal transduction in virus survival. J. Gen. Virol. 85:1065-1076.
Cory, S., and J. M. Adams. 2002. The Bcl2 family: regulators of the cellular life-or-death switch. Nat. Rev. Cancer 2:647-656.
Cui, J., X. D. Zhou, Y. K. Liu, Z. Y. Tang, and M. Romeih. 2003. Wnt signaling in hepatocellular carcinoma: analysis of mutation and expression of beta-catenin, T-cell factor-4 and glycogen synthase kinase 3-beta genes. J. Gastroenterol. Hepatol. 18:280-287.
Datta, S. R., H. Dudek, X. Tao, S. Masters, H. Fu, Y. Gotoh, and M. E. Greenberg. 1997. Akt phosphorylation of BAD couples survival signals to the cell-intrinsic death machinery. Cell 91:231-241.
de La Coste, A., B. Romagnolo, P. Billuart, C. A. Renard, M. A. Buendia, O. Soubrane, M. Fabre, J. Chelly, C. Beldjord, A. Kahn, and C. Perret. 1998. Somatic mutations of the beta-catenin gene are frequent in mouse and human hepatocellular carcinomas. Proc. Natl. Acad. Sci. USA 95:8847-8851.
Desbois-Mouthon, C., M. J. B. V. Eggelpoel, E. Beurel, M. Boissan, R. Delelo, A. Cadoret, and J. Capeau. 2002. Dysregulation of glycogen synthase kinase-3 beta signaling in hepatocellular carcinoma cells. Hepatology 36:1528-1536.
Doble, B. W., and J. R. Woodgett. 2003. GSK-3: tricks of the trade for a multi-tasking kinase. J. Cell Sci. 116:1175-1186.
Fujimuro, M., and S. D. Hayward. 2003. The latency-associated nuclear antigen of Kaposi's sarcoma-associated herpesvirus manipulates the activity of glycogen synthase kinase-3?. J. Virol. 77:8019-8030.
Fujimuro, M., F. Y. Wu, C. ApRhys, H. Kajumbula, D. B. Young, G. S. Hayward, and S. D. Hayward. 2003. A novel viral mechanism for dysregulation of beta-catenin in Kaposi's sarcoma-associated herpesvirus latency. Nat. Med. 9:300-306.
Fukumoto, S., C. M. Hsieh, K. Maemura, M. D. Layne, S. F. Yet, K. H. Lee, T. Matsui, A. Rosenzweig, W. G. Taylor, J. S. Rubin, M. A. Perrella, and M. E. Lee. 2001. Akt participation in the Wnt signaling pathway through Dishevelled. J. Biol. Chem. 276:17479-17483.
Gan, D. D., and K. Khalili. 2004. Interaction between JCV large T-antigen and beta-catenin. Oncogene 23:483-490.
Ghosh, A. K., M. Majumder, R. Steele, K. Meyer, R. Ray, and R. B. Ray. 2000. Hepatitis C virus NS5A protein protects against TNF-alpha mediated apoptotic cell death. Virus Res. 67:173-178.
Ghosh, A. K., R. Steele, K. Meyer, R. Ray, and R. B. Ray. 1999. Hepatitis C virus NS5A protein modulates cell cycle regulatory genes and promotes cell growth. J. Gen. Virol. 80:1179-1183.
Gossen, M., and H. Bujard. 1992. Tight control of gene expression in mammalian cells by tetracycline-responsive promoters. Proc. Natl. Acad. Sci. USA 89:5547-5551.
Gotoh, J., M. Obata, M. Yoshie, S. Kasai, and K. Ogawa. 2003. Cyclin D1 over-expression correlates with beta-catenin activation, but not with H-ras mutations, and phosphorylation of Akt, GSK3 beta and ERK1/2 in mouse hepatic carcinogenesis. Carcinogenesis 24:435-442.
Hart, M., J. P. Concordet, I. Lassot, I. Albert, R. del los Santos, H. Durand, C. Perret, B. Rubinfeld, F. Margottin, R. Benarous, and P. Polakis. 1999. The F-box protein beta-TrCP associates with phosphorylated beta-catenin and regulates its activity in the cell. Curr. Biol. 9:207-210.
He, Y. P., H. H. Nakao, S. L. Tan, P. J. Polyak, P. Neddermann, S. Vijaysri, B. L. Jacobs, and M. G. Katze. 2002. Subversion of cell signaling pathways by hepatitis C virus nonstructural 5A protein via interaction with Grb2 and P85 phosphatidylinositol 3-kinase. J. Virol. 76:9207-9217.
Huang, H., H. Fujii, A. Sankila, B. M. Mahler-Araujo, M. Matsuda, G. Cathomas, and H. Ohgaki. 1999. Beta-catenin mutations are frequent in human hepatocellular carcinomas associated with hepatitis C virus infection. Am. J. Pathol. 155:1795-1801.
Krieger, N., V. Lohmann, and R. Bartenschlager. 2001. Enhancement of hepatitis C virus RNA replication by cell culture-adaptive mutations. J. Virol. 75:4614-4624.
Lan, K. H., M. L. Sheu, S. J. Hwang, S. H. Yen, S. Y. Chen, J. C. Wu, Y. J. Wang, N. Kato, M. Omata, F. Y. Chang, and S. D. Lee. 2002. HCV NS5A interacts with p53 and inhibits p53-mediated apoptosis. Oncogene 21:4801-4811.
Levy, L., C. Neuveut, C. A. Renard, P. Charneau, S. Branchereau, F. Gauthier, J. T. Van Nhieu, D. Cherqui, A. F. Petit-Bertron, D. Mathieu, and M. A. Buendia. 2002. Transcriptional activation of interleukin-8 by beta-catenin-Tcf4. J. Biol. Chem. 277:42386-42393.
Macdonald, A., K. Crowder, A. Street, C. McCormick, and M. Harris. 2004. The hepatitis C virus NS5A protein binds to members of the Src family of tyrosine kinases and regulates kinase activity. J. Gen. Virol. 85:721-729.
Macdonald, A., K. Crowder, A. Street, C. McCormick, K. Saksela, and M. Harris. 2003. The hepatitis C virus NS5A protein inhibits activating protein-1 function by perturbing Ras-ERK pathway signalling. J. Biol. Chem. 278:17775-17784.
Macdonald, A., and M. Harris. 2004. Hepatitis C virus NS5A: tales of a promiscuous protein. J. Gen. Virol. 85:2485-2502.
McCormick, C. J., L. Challinor, A. Macdonald, D. J. Rowlands, and M. Harris. 2004. Introduction of replication-competent hepatitis C virus transcripts using a tetracycline-regulable baculovirus delivery system. J. Gen. Virol. 85:429-439.
McCormick, C. J., D. J. Rowlands, and M. Harris. 2002. Efficient delivery and regulable expression of hepatitis C virus full length and mini-genome constructs in hepatocyte-derived cell lines using baculovirus vectors. J. Gen. Virol. 83:383-394.
Miyasaka, Y., N. Enomoto, M. Kurosaki, N. Sakamoto, N. Kanazawa, T. Kohashi, E. Ueda, S. Maekawa, H. Watanabe, N. Izumi, C. Sato, and M. Watanabe. 2003. Hepatitis C virus nonstructural protein 5A inhibits tumor necrosis factor-alpha-mediated apoptosis in Huh7 cells. J. Infect. Dis. 188:1537-1544.
Morin, P. J., A. B. Sparks, V. Korinek, N. Barker, H. Clevers, B. Vogelstein, and K. W. Kinzler. 1997. Activation of beta-catenin-Tcf signaling in colon cancer by mutations in beta-catenin or APC. Science 275:1787-1790.
Morrison, J. A., A. J. Klingelhutz, and N. Raab-Traub. 2003. Epstein-Barr virus latent membrane protein 2A activates ?-catenin signaling in epithelial cells. J. Virol. 77:12276-12284.
Owsianka, A., R. F. Clayton, L. D. Loomis-Price, J. A. McKeating, and A. H. Patel. 2001. Functional analysis of hepatitis C virus E2 glycoproteins and virus-like particles reveals structural dissimilarities between different forms of E2. J. Gen. Virol. 82:1877-1883.
Park, K. J., S. H. Choi, S. Y. Lee, S. B. Hwang, and M. M. C. Lai. 2002. Nonstructural 5A protein of hepatitis C virus modulates tumor necrosis factor alpha-stimulated nuclear factor kappa B activation. J. Biol. Chem. 277:13122-13128.
Polakis, P. 2000. Wnt signaling and cancer. Genes Dev. 14:1837-1851.
Polakis, P., M. Hart, and B. Rubinfeld. 1999. Defects in the regulation of beta-catenin in colorectal cancer. Adv. Exp. Med. Biol. 470:23-32.
Sarcar, B., A. K. Ghosh, R. Steele, R. Ray, and R. B. Ray. 2004. Hepatitis C virus NS5A mediated STAT3 activation requires co-operation of Jak1 kinase. Virology 322:51-60.
Shih, I. M., J. Yu, T. C. He, B. Vogelstein, and K. W. Kinzler. 2000. The beta-catenin binding domain of adenomatous polyposis coli is sufficient for tumor suppression. Cancer Res. 60:1671-1676.
Street, A., A. Macdonald, K. Crowder, and M. Harris. 2004. The hepatitis C virus NS5A protein activates a phosphoinositide 3-kinase dependent survival signalling cascade. J. Biol. Chem. 279:12232-12241.
Waris, G., A. Livolsi, V. Imbert, J. F. Peyron, and A. Siddiqui. 2003. Hepatitis C virus NS5A and subgenomic replicon activate NF-kappaB via tyrosine phosphorylation of IkappaBalpha and its degradation by calpain protease. J. Biol. Chem. 278:40778-40787.
Willems, M., H. J. Metselaar, H. W. Tilanus, S. W. Schalm, and R. A. de Man. 2002. Liver transplantation and hepatitis C. Transplant Int. 15:61-72.
Yanagi, M., R. H. Purcell, S. U. Emerson, and J. Bukh. 1997. Transcripts from a single full-length cDNA clone of hepatitis C virus are infectious when directly transfected into the liver of a chimpanzee. Proc. Natl. Acad. Sci. USA 94:8738-8743.
Zoulim, F., M. Chevallier, M. Maynard, and C. Trepo. 2003. Clinical consequences of hepatitis C virus infection. Rev. Med. Virol. 13:57-68.(Andrew Street, Andrew Mac)
ABSTRACT
The hepatitis C virus (HCV) nonstructural NS5A protein has been shown to bind to and activate phosphoinositide 3-kinase (PI3K), resulting in activation of the downstream effector serine/threonine kinase Akt/protein kinase B. Here we present data pertaining to the effects of NS5A-mediated Akt activation on its downstream targets. Using a recombinant baculovirus to deliver the complete HCV polyprotein to human hepatoma cells in a tetracycline-regulable fashion, we confirm that expression of the complete HCV polyprotein also activates PI3K and Akt. We further show that this results in the inhibition of the Akt substrate Forkhead transcription factor and the stimulation of phosphorylation of a second key Akt substrate, glycogen synthase kinase-3? (GSK-3?). Phosphorylation of GSK-3? results in its inactivation; consistent with this, we show that expression of the HCV polyprotein results in the accumulation of ?-catenin. Finally, we show that levels of ?-catenin-dependent transcription are also elevated in the presence of the HCV polyprotein. Given the prevalence of ?-catenin mutations in many human tumors, especially colon and hepatocellular carcinomas, these data implicate NS5A-mediated PI3K activation as a contributory factor in the increasingly common association between HCV infection and the development of hepatocellular carcinoma.
INTRODUCTION
Hepatitis C virus (HCV) is recognized as a major public health problem; the World Health Organization estimates that as many as 170 million individuals (approximately 3% of the world population) are infected with this virus. In 85% of cases, the virus establishes a chronic infection, culminating in chronic inflammation, cirrhosis, and, increasingly, hepatocellular carcinoma (HCC). Indeed, HCV is now the leading cause of HCC and is the most common reason for liver transplantation in the West (47). Treatment is presently limited to the use of type 1 interferon in combination with the nucleoside analogue ribavirin; this therapeutic regime is successful in 40 to 80% of patients, depending on virus genotype.
The HCV genome is a single-stranded positive-sense 9.6-kb RNA molecule, comprising a single open reading frame coding for a polyprotein of 3,000 amino acids flanked by untranslated regions (UTRs). The 5' UTR contains an internal ribosome entry sequence, allowing cap-independent initiation of translation. The polyprotein is cleaved into 10 polypeptides by cellular and viral proteinases; one of these products is the nonstructural NS5A protein. NS5A has been the focus of much intensive investigation recently with regard to its potential role both in the cytopathology of HCV infection and in mediating viral immune evasion (for a review, see reference 33). NS5A has been reported to interact with a wide range of cellular proteins involved in, among others, the interferon response and cell cycle control. These interactions result in the modulation of various transcription factors, including NF-B (40, 46), STAT3 (31, 43), and AP-1 (32); furthermore, NS5A has been reported to promote anchorage-independent growth in NIH 3T3 murine fibroblasts and tumor formation in nude mice (22).
Recently, we (45) and others (26) have determined that NS5A interacts with the phosphoinositide 3-kinase (PI3K) signaling cascade. The binding of NS5A to the SH3 domain of the p85 regulatory subunit of PI3K stimulated the lipid kinase activity associated with the p110 catalytic subunit of this heterodimeric complex up to 10-fold (45). Furthermore, the activation of PI3K in cells expressing either NS5A alone or NS5A in the context of the subgenomic HCV replicon (28) resulted in increased phosphorylation and activity of the downstream kinase Akt/protein kinase B (45). The increased phosphorylation correlated with increased resistance to a variety of apoptotic stimuli, including serum starvation. Akt phosphorylates and inactivates many downstream target proteins: these include the Forkhead transcription factor (FKHR), the proapoptotic Bcl2 homologue Bad, and pro-caspase-9 (8, 11). One of the better-understood substrates of Akt is the serine/threonine kinase glycogen synthase kinase-3? (GSK-3?), which was first characterized as a negative regulator of glycogen synthesis (16) but was more recently shown to play a pivotal role in the regulation of the proto-oncogene ?-catenin. ?-Catenin has two distinct functions. Most of the protein is located at the cell membrane, where it is involved in cell-cell contact via its association with the cytoplasmic domain of E-cadherin (41). A second pool of ?-catenin is located both in the nucleus and in the cytoplasm, where it mediates Wnt signaling. In the absence of mitogenic stimulation, a multiprotein destruction complex containing GSK-3?, Axin, and the tumor suppressor adenomatous polyposis coli (APC) acts to promote the phosphorylation of serine and threonine residues in the N terminus of ?-catenin and thereby targets it for proteasome-mediated degradation via the F-box-containing protein ?TrCP (25). When Akt phosphorylates GSK-3?, it inactivates the latter, leading to the accumulation of ?-catenin, which can enter the nucleus and participate in the formation of transcriptionally active complexes with members of the Tcf/Lef family (2).
Interestingly, the stabilization of ?-catenin, as a result of mutations either in ?-catenin itself that block phosphorylation or in other components in the pathway, is associated with a range of human tumors, in particular HCC (12, 15, 24). Given that HCV infection is increasingly associated with HCC, we were interested in exploring the effects of NS5A-mediated Akt activation on the GSK-3?/?-catenin signaling pathway. We show here that NS5A, expressed either alone or in the context of the complete HCV polyprotein, mediated a reduction in the transcriptional activity of the Forkhead transcription factor, FKHR, and an increase in GSK-3? phosphorylation. These effects were dependent on the Akt pathway, as they were abolished both by a pharmacological inhibitor of PI3K and by a kinase-inactive (dominant negative) form of Akt. Furthermore, we found an increase in the overall levels of ?-catenin within cells expressing either NS5A or the complete HCV polyprotein. This accumulation correlated with increased stability of ?-catenin and resulted in a concomitant increase in the levels of ?-catenin-dependent transcription within the cells. Given these results, we propose that an increase in the activity of ?-catenin, in conjunction with an Akt-dependent decrease in apoptosis, may predispose HCV-infected hepatocytes to neoplastic transformation.
MATERIALS AND METHODS
Plasmid constructs. Vectors expressing NS5A have been described elsewhere (32, 45). The Tcf-luciferase reporter pGL3-OT, the mutant Tcf-luciferase reporter pGL3-OF (44), and the ?-catenin expression vector were provided by Bert Vogelstein (Howard Hughes Medical School, Baltimore, Md.). The hemagglutinin-tagged (HA)-GSK-3? expression vector was provided by Vivianne Ding (University of California, San Francisco). The Forkhead transcription factor luciferase reporter (pGL3-FHRE) (4) and the HA-Akt(K179M) (kinase-inactive) expression vector (13) were provided by Mike Greenberg (Harvard Medical School, Boston, Mass.).
Cell culture. The human hepatoma cell line HepG2 was maintained in minimal essential medium supplemented with 10% fetal calf serum (FCS), 2 mM L-glutamine, 100 IU of penicillin/ml, and 100 μg of streptomycin/ml plus 1% nonessential amino acids. Cos-7 (African green monkey kidney) cells were propagated in Dulbecco's modified Eagle medium supplemented as for HepG2 cells. Cells were cultured in a humidified 5% CO2 incubator at 37°C. Sf9 cells for propagation of recombinant baculoviruses were maintained in TC100 medium supplemented with 10% FCS, 100 IU of penicillin/ml, and 100 μg of streptomycin/ml. Where appropriate, cells were incubated with Ly294002 (50 μM) (Cell Signaling Technology) or cycloheximide (100 μg/ml) (Sigma).
Generation and propagation of recombinant baculoviruses. A baculovirus transfer vector carrying the replication-deficient HCV genome was generated as follows. Briefly, the previously described vector pBACH77(HV)tet (35) was restricted at the AflII site found at the beginning of the 3' UTR and at an EcoRI site immediately preceding the hepatitis delta virus ribozyme in order to remove both the ribozyme and the majority of the 3' UTR. Pfu polymerase was used to polish the resulting DNA, which was then religated, generating pBACH773'UTR. Generation of virus was as described previously (35), with the integrity of virus clones being confirmed by Southern blot analysis and immunoblotting for HCV antigens Core, E2, NS3, NS5A, and NS5B in virus-transduced HepG2 cells. The baculovirus vector for constitutive expression of the tTA tetracycline repressor-VP16 transactivator fusion protein in mammalian cells (BactTA), together with a baculovirus expressing the Escherichia coli ?-galactosidase gene under the control of the Ptet promoter (BacINDLacZtet), has been described previously (35). Tetracycline was used at a concentration of 5 μg/ml in all experiments.
Analysis of the Akt pathway by immunoblotting. HepG2 cells were seeded in 90-mm-diameter dishes precoated with rat tail collagen at a density of 2 x 106 cells per dish and 24 h later were transduced with BactTA (35) and BACH773'UTR or BacINDLacZtet for 4 h in the presence or absence of tetracycline (each virus was at a final concentration of 2.5 x 107 PFU/ml) as described previously (32, 35). The transduced cells were then transfected for 5 h with plasmids expressing HA-Akt(K179M) or HA-GSK-3? (2 μg in total) by using the Lipofectin reagent (Invitrogen) according to manufacturer's recommendations in the presence or absence of tetracycline. The total amount of DNA was kept constant by inclusion of empty vector DNA. To suppress external activation of mitogenic signaling pathways, cells were subsequently incubated in growth medium containing reduced serum (0.5%). Cells were harvested 24 h posttransfection, resuspended in 500 μl of ice-cold radioimmunoprecipitation assay lysis buffer (50 mM Tris-HCl [pH 8.0], 150 mM NaCl, 1% NP-40, 0.1% sodium dodecyl sulfate, 5 mM EDTA, with the addition of 1 mM NaOV, 1 mM NaF, 1 mM okadaic acid, and 500 nM cantharidin). After total protein quantification (BCA assay; Pierce), lysates were subjected to analysis by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and immunoblotted with antibodies directed against GSK-3?, GSK-3? phosphorylated on Ser9, Akt, Akt phosphorylated on Ser473, PDK1 phosphorylated on Ser241, or ?-catenin (all from Cell Signaling Technology) or a sheep polyclonal anti-NS5A serum (32). Blots were probed with peroxidase-conjugated secondary antibody and detected with the enhanced chemiluminescence system (Amersham Pharmacia). All experiments were repeated three times. The immunoblots shown were chosen as the clearest representations of the data.
Luciferase assays. HepG2 cells (5 x 105) were cultured in 6-well dishes precoated with rat tail collagen and transduced as described above. Transduced cells were then transfected with plasmids expressing the appropriate luciferase reporter(0.5 μg), HA-Akt(K179M) or HA-GSK-3?, or ?-catenin expression vectors (1.5 μg) with Lipofectin (Invitrogen) following the manufacturer's instructions. A Renilla luciferase reporter construct (pRLTK) was used as an internal control for transfection efficiency. The total amount of DNA was kept constant by inclusion of empty vector DNA. Cells were harvested at 24 h in 200 μl of passive lysis buffer (Promega). The quantitation of relative light units was done by using the dual luciferase Stop & Glo reagent (Promega) and a luminometer (EG&G Berthold) with a dual injector system. All assays were performed in triplicate, and each experiment was repeated a minimum of three times.
Immunofluorescence. HepG2 cells were seeded into 12-well dishes containing coverslips (105 cells per well), allowed to settle overnight, and transduced as previously described. After 24 h, the cells were fixed in 4% paraformaldehyde for 15 min, washed in phosphate-buffered saline (PBS), and permeabilized with 0.1% Triton X-100 for 10 min. After washing in PBS, the cells were incubated for 1 h with the appropriate antibody diluted in PBS containing 10% FCS. A sheep polyclonal anti-NS5A antibody (32) was used for the primary stain, and bound antibody was detected by using a donkey anti-sheep fluorescein isothiocyanate (Jackson ImmunoResearch) with a Hoechst 33342 (Molecular Probes) counterstain.
Other procedures. PI3K assays and immunoblotting analyses were performed as previously described (32, 35, 45).
RESULTS
Expression of the complete HCV polyprotein stimulates both PI3K and Akt activity in HepG2 cells. We have previously shown that NS5A expression, either alone or as part of the HCV subgenomic replicon, induced activation of both PI3K and Akt in both Cos-7 cells and the human hepatoma cell line Huh-7 (45). Activation of PI3K was mediated via a direct protein-protein interaction between NS5A and the SH3 domain of the PI3K p85 regulatory subunit. Given that the subgenomic replicon expressed only the nonstructural NS3 through NS5B proteins (constituting the C-terminal two-thirds of the polyprotein), we considered it important to verify that these effects of NS5A were also manifest in the context of the complete polyprotein. Although full-length replicons expressing the complete HCV polyprotein are now available, for the reasons outlined below we decided to utilize a novel baculovirus delivery system developed in our laboratory to express the HCV polyprotein in hepatocyte-derived cell lines. This recombinant virus contains the complete HCV polyprotein coding sequence flanked by the 5' UTR but lacking the 3' UTR. A tetracycline-regulable promoter (Ptet) (23) is positioned upstream of the 5' UTR such that the initiation of transcription coincides with the start of the 5' UTR (Fig. 1a). The lack of the 3' UTR means that this virus (BacH773'UTR) is replication defective (34). We have previously shown that these baculovirus vectors work most efficiently in the hepatoblastoma-derived cell line HepG2 (35), so we conducted subsequent experiments in this cell line. The use of this tetracycline-regulable system provided two key advantages over the use of the replicon system for these experiments. First, experiments could be performed on homogeneous cell populations not subjected to long-term selection (with the potential for selection of cells with defects in signaling pathways). Second, the tight regulation allowed for robust internal controls, as signaling events could be compared in samples derived from a single population of transduced cells incubated in the presence or absence of tetracycline prior to analysis. In addition, as discussed below, Huh-7 cells were not suitable for some of the studies described here due to inherent defects in the regulation of GSK-3? phosphorylation (15). As shown in Fig. 1b and c, when cells were cotransduced with BacH773'UTR and a baculovirus expressing the tTA tetracycline repressor-VP16 transcriptional activator fusion protein (BactTA) (35), expression of the HCV polyprotein in HepG2 cells was tetracycline regulable. In the presence of tetracycline, expression was tightly repressed, whereas in the absence of the antibiotic, the HCV polyprotein was expressed and efficiently proteolytically processed, as judged by immunoblotting for either the E2 envelope glycoprotein (Fig. 1b, left panel) or NS5A (Fig. 1b, right panel). This result is consistent with our previous observations with a LacZ-expressing baculovirus (BacINDLacZtet [35]), in which we showed that the tetracycline-responsive promoter is tightly regulated in HepG2 cells: in the presence of tetracycline, levels of ?-galactosidase were barely detectable above background, whereas in the absence of tetracycline they were increased 1,000-fold (35). As shown in Fig. 1c, the BacMAM delivery system was highly efficient in HepG2 cells; in the absence of tetracycline, essentially all of the cells in the population expressed the HCV polyprotein, as judged by indirect immunofluorescent detection of NS5A.
We previously showed that the expression of either the HCV subgenomic replicon or NS5A alone resulted in a 10-fold increase in the activity of PI3K in vivo (45). As shown in Fig. 2a, the incubation of cells cotransduced with BacH773'UTR and BactTA in the absence of tetracycline resulted in a threefold increase in PI3K activity compared to cells incubated in the presence of tetracycline. Levels of p85 in immunoprecipitates were verified by immunoblotting (Fig. 2a, panel ii), and there were no significant differences, demonstrating that the increased PI3K activity was not due to elevated levels of PI3K expression. Panel iii of Fig. 2a confirms appropriate expression of NS5A. This result confirms and extends our previous data (45), demonstrating that the expression of the complete HCV polyprotein also stimulates PI3K activity. Our previous study had also shown that NS5A-mediated activation of PI3K resulted in subsequent activation of the protein kinase Akt; we therefore investigated the activation status of Akt in these baculovirus-transduced cells. As shown in panel i of Fig. 2b, levels of Ser473-phosphorylated Akt were elevated in cells incubated in the absence of tetracycline (Fig. 2b, compare lanes 1 and 2). To confirm that this was mediated through the PI3K pathway, cells were also treated with 50 μM Ly294002, a specific PI3K inhibitor, for 24 h prior to harvest. This treatment significantly reduced both basal and HCV-induced activation of Akt (Fig. 2b, panel ii, lanes 3 and 4). No activation of either PI3K or Akt was seen in cells cotransduced with BactTA and a control baculovirus (BacINDLacZtet) expressing ?-galactosidase (data not shown). As expected, there were no differences observed in the total levels of Akt expression (Fig. 2b, panel ii) or the levels of phosphorylated PDK1, the kinase that phosphorylates Akt (Fig. 2b, panel iii). PDK1 has previously been shown to be constitutively active (1). Appropriate expression of NS5A in lysates of cells incubated without tetracycline confirmed the controlled expression of the complete HCV polyprotein (Fig. 2b, panel iv). These data indicate that Akt is activated in a PI3K-dependent manner in cells expressing the whole HCV genome.
Expression of the complete HCV polyprotein modulates Akt-mediated downstream events. Activated Akt phosphorylates and inactivates a number of downstream targets, including FKHR. Once phosphorylated, FKHR binds to 14-3-3 proteins and is sequestered in the cytoplasm. In order to assess the downstream effects of HCV-induced Akt phosphorylation, we therefore analyzed the activity of FKHR by using a luciferase reporter construct responsive to this transcription factor. HepG2 cells were transduced with BacH773'UTR and BactTA and subsequently transfected with an FKHR-luciferase construct (Fig. 3a). In the presence of serum, levels of FKHR-responsive luciferase were low, whereas in the absence of serum there was a fivefold increase in luciferase levels, consistent with the role of FKHR in the induction of apoptosis (Fig. 3a, compare the black bars). Expression of the HCV genome following the removal of tetracycline resulted in a 45% reduction in FKHR activity in serum-starved cells (Fig. 3a, compare black and white bars), whereas no HCV-mediated inhibition of FKHR activity was seen in cells grown in the presence of serum, presumably due to constitutive activation of PI3K and Akt activity by serum components (e.g., growth factors or insulin). There was no significant inhibition of FKHR activity in control cells transduced with BacINDLacZtet (LacZ) following removal of tetracycline (Fig. 3a, compare black and white bars). The HCV-mediated inhibition was abolished by transfection with a plasmid expressing an HA-tagged kinase-inactive (dominant negative) mutant of Akt (K179M). To confirm the role of NS5A-mediated activation of PI3K, we performed a similar study in Cos-7 cells transiently transfected with plasmids expressing NS5A and HA-Akt(K179M) (Fig. 3b). As in HepG2 cells, levels of FKHR activity were elevated in Cos-7 cells incubated in the absence of serum (Fig. 3b, black bars) compared with cells grown in 10% serum. Transfection with an NS5A plasmid (Fig. 3b, white bars) resulted in a 60% reduction in FKHR activity in serum-starved cells; this inhibition was abolished by cotransfection with a plasmid expressing HA-Akt(K179M) (Fig. 3b, grey bars). These data are consistent with the activation of Akt by NS5A, resulting in increased Akt-dependent phosphorylation and inactivation of FKHR.
A second key substrate of Akt is the serine/threonine kinase GSK-3?. Our initial experiments to determine whether the phosphorylation status of GSK-3? was influenced by NS5A-mediated PI3K activation revealed a limitation of the replicon system in the context of this study. Using an antibody specific for GSK-3? phosphorylated on Ser9, we were unable to observe elevated levels of phosphorylated GSK-3? in Huh-7 cells expressing the HCV subgenomic replicon or NS5A alone in comparison to control cells expressing neomycin phosphotransferase (45) (data not shown). Subsequent analysis of the literature revealed that Huh-7 cells, along with some other HCC cells lines, such as Hep3B and Mahlavu cells, contain very high constitutive levels of phosphorylated GSK-3? (15), consistent with an inherent defect in GSK-3? regulation. In contrast, HepG2 cells exhibited a very low level of GSK-3? phosphorylation, comparable to normal primary hepatocytes. We therefore transduced HepG2 cells with the recombinant baculoviruses described above and analyzed GSK-3? phosphorylation. Figure 3c, panel i shows that levels of GSK-3? phosphorylation were significantly increased in cells expressing the HCV polyprotein in the absence of tetracycline compared to cells incubated in the presence of tetracycline (Fig. 3c, compare lanes 1 and 2). We confirmed that this result was due to stimulation of the PI3K/Akt signaling pathway by demonstrating that phosphorylation of GSK-3? was inhibited either by Ly294002 or by transfection of an HA-Akt(K179M) plasmid (Fig. 3c, panel i, lanes 3 to 6). Again, overall levels of GSK-3? were unaffected (Fig. 3c, panel ii), and panels iii and iv of Fig. 3c confirm appropriate expression of Akt and NS5A.
Expression of the complete HCV polyprotein results in the accumulation of ?-catenin. A key downstream target of GSK-3? is the proto-oncogene ?-catenin. Normally, ?-catenin is sequestered in the cytoplasm as part of a complex with Axin and APC. One pool of GSK-3? is also associated with the Axin complex and, when active, mediates serine/threonine phosphorylation of residues at the N terminus of ?-catenin, thus providing a recognition signal for ?TrCP, the receptor component of the Skp1/Cullin/F-box protein (SCF)-?TrCP E3 ubiquitin ligase complex, targeting the protein for proteasomal degradation (25). Increased phosphorylation and concomitant inhibition of GSK-3? would therefore be predicted to prevent ?-catenin phosphorylation, resulting in decreased ubiquitin-mediated proteolysis and increased levels of ?-catenin. To determine whether HCV polyprotein-mediated stimulation of Akt and GSK-3? phosphorylation was able to effect a similar stabilization, steady-state levels of ?-catenin were examined in HepG2 cells transduced with the recombinant baculoviruses described above. HepG2 cells express two ?-catenin alleles, one wild-type form and a mutated form lacking amino acids 25 to 140 (5, 14), which cannot be phosphorylated and thus accumulates in the nucleus, where it is constitutively active. Levels of endogenous wild-type ?-catenin in HepG2 cells are therefore low (5), and to facilitate the analysis we transfected them with a ?-catenin expression plasmid (37). It should be pointed out that the endogenous truncated ?-catenin in HepG2 cells cannot be detected with the monoclonal antibody used in this study, as it binds to an epitope near the N terminus and thus did not interfere with the interpretation of the data. Panel i of Fig. 4a shows that in cells transduced with BacH773'UTR and BactTA, levels of ?-catenin were increased in the absence of tetracycline (Fig. 4a, compare lanes 1 and 2). Again we confirmed that this increase was mediated through the PI3K/Akt pathway by demonstrating that the incubation of cells with Ly294002, or transfection with an HA-Akt(K179M) plasmid, blocked the increase in ?-catenin levels (Fig. 4a, lanes 3 to 6). Panels ii and iii of Fig. 4a confirm appropriate expression of Akt and NS5A. To confirm that baculovirus transduction and/or treatment with tetracycline did not affect ?-catenin levels, we also examined lysates from mock-transduced cells or cells transduced with BacINDLacZtet and BactTA. Figure 4b, panel i shows that there was no difference in the levels of ?-catenin in any of these lysates; panel ii confirms appropriate expression of ?-galactosidase. We conclude from these experiments that the expression of the HCV polyprotein results in increased accumulation of ?-catenin.
HCV-mediated accumulation of ?-catenin is due to increased protein stability. Although it was likely that the increased steady-state levels of ?-catenin observed in cells expressing the complete HCV polyprotein were due to a reduction in proteasomal degradation, we wished to confirm this by analyzing the stability of ?-catenin. To examine this, we transduced HepG2 cells as described above and treated them with cycloheximide to block protein synthesis. Samples were taken at intervals after treatment, and levels of ?-catenin were analyzed by immunoblotting. As shown in Fig. 4c, panel i, in the presence of tetracycline, levels of ?-catenin declined over an 8-h time course, consistent with proteasomal degradation. However, in the absence of tetracycline, ?-catenin levels were initially higher than in the presence of tetracycline (as seen in Fig. 4a) and did not decline as rapidly over the course of the experiment. Figure 4c, panel ii shows that levels of a control protein, the endoplasmic reticulum resident chaperone calreticulin, remained constant both in the presence and in the absence of tetracycline; while panel iii shows that NS5A was in fact somewhat more unstable than ?-catenin, very little NS5A was present at 8 h after cycloheximide treatment. To confirm that this stabilization of ?-catenin was mediated via the PI3K/Akt pathway, cells were transfected with an HA-Akt(K179M) plasmid (Fig. 4c, panel iv). As expected, in these cells ?-catenin was rapidly degraded, and the expression of the HCV polyprotein had no effect on ?-catenin stability.
To confirm that this effect on ?-catenin stability was due to NS5A-mediated stimulation of the PI3K/Akt pathway, we repeated this experiment in Cos-7 cells transiently transfected with an NS5A expression plasmid. Figure 4d, panel i shows that in control cells, degradation of ?-catenin occurred at a rate similar to that in HepG2 cells. In cells expressing NS5A, levels of ?-catenin remained constant throughout the time course (Fig. 4d, panel ii). Transfection with an HA-Akt(K179M) plasmid was able to partially restore the degradation of ?-catenin, consistent with the role of the PI3K/Akt pathway in regulating ?-catenin stability.
Expression of the complete HCV polyprotein results in elevated ?-catenin-dependent transcriptional activity. Nuclear accumulation of ?-catenin allows association with Tcf-Lef transcription factors and stimulates the expression of a series of cellular genes that are involved in cell proliferation (2). To test whether stabilization of ?-catenin by NS5A led to increased nuclear transcriptional activity of ?-catenin, HepG2 cells were transduced with recombinant baculoviruses as described above and subsequently transfected with luciferase reporters carrying three copies of either wild-type Tcf-binding sites (pGL3-OT) or mutated binding sites that were nonresponsive to Tcf (pGL3-OF) (42). As shown in Fig. 5a, levels of luciferase activity from the pGL3-OT reporter were approximately twofold higher in the absence of tetracycline than in cells incubated in the presence of tetracycline, consistent with HCV-mediated stimulation of wild-type ?-catenin-dependent transcription. No such increase was seen in cells transduced with BacINDLacZtet and BactTA when they were incubated in the absence of tetracycline. The relatively high basal activity of the reporter in HepG2 cells is consistent with the presence of the truncated, constitutively active allele of ?-catenin (5, 14). We confirmed that this stimulation was mediated via the PI3K/Akt pathway by transfecting cells with an HA-Akt(K179M) plasmid. As expected, this resulted in a dramatic reduction in ?-catenin-mediated reporter gene expression, consistent with the involvement of the PI3K/Akt pathway. Although the expression of HA-Akt(K179M) reduced overall levels of ?-catenin activity, an HCV polyprotein-mediated increase in reporter activity was still apparent. This is consistent with the fact that Akt functions as a multimer (8), and the presence of an Akt(K179M) monomer within the multimer does not completely abolish wild-type Akt function. When cells were transfected with a ?-catenin expression plasmid to increase overall levels of the wild-type protein, luciferase levels were elevated as expected; again, the expression of the HCV polyprotein stimulated the reporter twofold, whereas BacINDLacZtet had no effect. To confirm that the activation of ?-catenin transcriptional activity was due to Akt activation, we transfected cells with expression plasmids for both ?-catenin and HA-Akt(K179M). Under these conditions, the activation of ?-catenin was abolished. Lastly, we transfected cells with a plasmid expressing HA-tagged GSK-3?. Although this had no effect on basal ?-catenin activity, it blocked the HCV polyprotein-mediated enhancement, providing further evidence for the role of PI3K/Akt in the accumulation of ?-catenin; in the presence of high levels of GSK-3?, there would be an increase in the amount of unphosphorylated (active) GSK-3? and a concomitant decrease in the levels of ?-catenin. No significant responses were seen with the mutant (pGL3-OF) reporter (data not shown).
To confirm the involvement of NS5A, we repeated these experiments in transiently transfected Cos-7 cells. As shown in Fig. 5b, luciferase levels were increased threefold following cotransfection with an NS5A expression vector. Transfection of an HA-Akt(K179M) plasmid was less effective at reducing basal ?-catenin-responsive transcription in these cells. This result could be due to higher levels of Akt in Cos-7 cells than in HepG2 cells (data not shown); however, HA-Akt(K179M) did block NS5A-mediated augmentation of ?-catenin activity. Transfection with a ?-catenin expression plasmid increased overall luciferase levels while maintaining the NS5A-mediated augmentation. As observed in HepG2 cells expressing the full-length polyprotein (Fig. 5a), the transfection of cells with expression plasmids for both ?-catenin and HA-Akt(K179M) abolished the activation of ?-catenin. We further showed that the effect of HA-Akt(K179M) was titratable, as the level of ?-catenin transcription was inversely proportional to the amount of HA-Akt(K179M) expression plasmid transfected. Lastly, transfection of an HA-GSK-3? plasmid reduced both basal ?-catenin-responsive transcription and NS5A-mediated augmentation. We conclude from these data that the increased stability of ?-catenin in cells expressing either NS5A or the complete HCV polyprotein results in elevated levels of ?-catenin-responsive transcription.
DISCUSSION
We have previously shown that the HCV NS5A protein activates PI3K by binding to the SH3 domain of the PI3K p85 regulatory subunit, and this subsequently results in the activation of the downstream effector kinase Akt (45). In this paper, we extend these results to analyze some of the downstream events resulting from the activation of PI3K and Akt. We demonstrate that the expression of NS5A, either alone or in the context of the full-length HCV polyprotein, results in the increased phosphorylation of Akt and the inactivation of the Akt substrates FKHR and GSK-3?. Furthermore, GSK-3? inactivation results in the stabilization of the proto-oncogene ?-catenin and thereby stimulates ?-catenin-dependent transcription (Fig. 6). Interestingly, a recent paper (9) identified the yeast Akt homologue as one of eight kinases in the yeast "kinome" able to directly phosphorylate NS5A; recombinant human Akt was also able to phosphorylate NS5A in vitro, suggesting that there may be further interactions between NS5A and the Akt pathway. However, the site(s) of Akt phosphorylation on NS5A and the functional significance remain to be determined.
GSK-3? and ?-catenin represent a point of convergence of signaling from both PI3K/Akt and the Wnt pathway. The latter was originally described for Drosophila, where it plays a key role in embryogenesis; binding of the secreted Wnt protein to its receptor (Frizzled) deactivates a complex consisting of GSK-3?, APC, and Axin, blocking ?-catenin phosphorylation. Interestingly, different pools of GSK-3? exist within cells; as well as that associated with the Axin complex, there are other pools within both the cytoplasm and the nucleus. These pools of GSK-3? phosphorylate a variety of substrates and are involved in a number of processes, such as glycogen metabolism, vesicle transport, and translation (for a review, see reference 16). Components of the Wnt pathway, such as Dishevelled (Dvl), have been shown to be required to link activated Akt to GSK-3? associated with the Axin complex (19). It will be of interest to determine whether such factors are also required for the inactivation of GSK-3? by Akt in HCV-expressing cells. In addition, it will be of interest to investigate the effects of HCV on other substrates and processes regulated by GSK-3?.
Over recent years, the Wnt/?-catenin signaling pathway has been identified as a common target for perturbation by viruses. With the notable exception of human immunodeficiency virus type 1 (HIV-1), these are all tumorigenic DNA viruses that establish chronic infections, consistent with the fact that ?-catenin-regulated genes positively regulate cell cycle and division (e.g., c-Myc, cyclin D1). Interestingly, the mechanisms by which these viruses target ?-catenin are all distinct. The -herpesvirus Epstein-Barr virus (EBV) encodes a plasma membrane-localized latent membrane protein 2A (LMP2A) that activates PI3K and Akt, resulting in GSK-3? inactivation and ?-catenin stabilization (38). The EBV LMP1 has independently been shown to activate PI3K/Akt, although it has not been determined whether this contributes to ?-catenin stabilization in EBV-infected cells. In contrast, the latency-associated nuclear antigen (LANA) of another -herpesvirus, Kaposi's sarcoma-associated herpesvirus, binds to GSK-3? and sequesters it in the nucleus, preventing ?-catenin phosphorylation (17, 18). Hepatitis B virus X protein achieves ?-catenin stabilization by suppressing GSK-3? activity in an Src-kinase-dependent manner (6). The large T antigen of the human neurotropic polyomavirus JCV interacts directly with ?-catenin, stabilizing it and promoting its nuclear accumulation (20). Lastly, the Vpu protein of HIV-1 binds to ?TrCP and blocks the ubiquitinylation and proteasomal degradation of ?-catenin (and other SCF-?TrCP substrates, such as IB and ATF4) (3). A number of other viral proteins have been shown to stimulate PI3K/Akt signaling; these include human papillomavirus type 16 E5, polyoma middle T antigen, and HIV-1 Nef (reviewed in reference 10). However, whether these proteins also upregulate downstream events involving GSK-3? or ?-catenin remains to be evaluated.
Our data showed that the HCV polyprotein or NS5A alone activated ?-catenin transcriptional activity two- to threefold (Fig. 5). Although this is a modest increase, it is comparable to the magnitude of ?-catenin activation observed in the context of other viral proteins, e.g., EBV LMP2A, Kaposi's sarcoma-associated herpesvirus LANA, and JCV large T antigen. The majority of these observations were made in transformed cell lines; in this regard, the involvement of ?-catenin in oncogenesis suggests that the basal levels of ?-catenin may already be elevated in these cells. It is conceivable that in the context of a natural infection of primary, untransformed cells, the magnitude of ?-catenin activation might be much greater for all of these viral proteins, including NS5A.
In addition to disruption by oncogenic viruses, the genetic dysregulation of the Wnt signaling pathway has been shown to play a pivotal role in carcinogenesis in a number of human tumor types, in particular colon carcinoma, in which 80 to 90% of cases have mutations either in APC or in ?-catenin itself (42), and HCC, in which 20 to 30% of cases are associated with mutations in Axin-1 or ?-catenin (14). Activating mutations in ?-catenin are predominantly located at sites of serine/threonine phosphorylation near the N terminus, although some larger deletions have been identified. Interestingly, in the case of HCC, stabilization of ?-catenin alone is insufficient to transform hepatocytes. The targets for ?-catenin-mediated transcriptional activation include many genes that control cell growth and cell cycle, such as c-Myc and cyclin D1, as well as genes involved in invasion, such as matrilysin (30), providing a rationale for the involvement of ?-catenin in tumorigenesis.
HCV infection is increasingly associated with the development of HCC (49). The fact that HCC does not occur in all cases of HCV, together with the extended time period between initial infection and the development of tumors (as long as 30 or 40 years), suggests that HCV is not directly tumorigenic. It is more likely that HCV infection predisposes patients to the development of tumors, such that a further event will precipitate the neoplastic transformation of infected hepatocytes. Given the constant exposure of hepatocytes to xenobiotics, many of which may be potentially carcinogenic, such a second event is likely to be a relatively frequent occurrence. In one study (27), activating mutations in ?-catenin were found in 41% of HCC cases associated with HCV infection. All of these tumors showed increased nuclear accumulation of ?-catenin; however, additional tumors without ?-catenin mutations also showed this accumulation, suggesting that a distinct mechanism might operate in these cases. Although the development of HCC is certainly influenced by a combination of viral, host, and environmental factors, the data presented here suggest that NS5A-mediated stabilization of ?-catenin may be a contributing factor in the association between HCV infection and HCC. This hypothesis is strengthened by previous results from our laboratory (45) and others (7, 21, 36) showing that NS5A mediates an increase in the resistance of cells to proapoptotic stimuli by a variety of mechanisms. As well as direct effects on the PI3K pathway, NS5A has been reported to bind to and inhibit both p53 (29) and Bax (7), thereby blocking the apoptotic response. We propose, therefore, that NS5A, as well as acting in some as-yet-undefined fashion during the process of viral RNA replication, can be regarded as a viral "survival factor," promoting the persistence of HCV. In the short term, it will be important to more fully understand the biochemical mechanism by which NS5A activates PI3K, leading to ?-catenin stabilization; however, in the long term, our data suggest that the development of chemotherapeutics targeted to NS5A might be beneficial in treating chronic HCV infection by potentiating the elimination of virus-infected cells.
ACKNOWLEDGMENTS
We thank Bert Vogelstein, Vivianne Ding, and Mike Greenberg for the kind gifts of plasmid reagents.
This work was supported by grants from the British Medical Research Council (G9801522) and the Wellcome Trust (0671250). A.S. was supported by a Biotechnology and Biological Sciences Research Council Ph.D. studentship.
A.S. and A.M. contributed equally to this work.
Present address: MRC Protein Phosphorylation Unit, School of Life Sciences, University of Dundee, Dundee DD1 5EH, United Kingdom.
REFERENCES
Alessi, D. R., M. Deak, A. Casamayor, F. B. Caudwell, N. Morrice, D. G. Norman, P. Gaffney, C. B. Reese, C. N. MacDougall, D. Harbison, A. Ashworth, and M. Bownes. 1997. 3-Phosphoinositide-dependent protein kinase-1 (PDK1): structural and functional homology with the Drosophila DSTPK61 kinase. Curr. Biol. 7:776-789.
Barker, N., P. J. Morin, and H. Clevers. 2000. The Yin-Yang of TCF/beta-catenin signaling. Adv. Cancer Res. 77:1-24.
Besnard-Guerin, C., N. Belaidouni, I. Lassot, E. Segeral, A. Jobart, C. Marchal, and R. Benarous. 2004. HIV-1 Vpu sequesters beta-transducin repeat-containing protein (betaTrCP) in the cytoplasm and provokes the accumulation of beta-catenin and other SCFbetaTrCP substrates. J. Biol. Chem. 279:788-795.
Brunet, A., A. Bonni, M. J. Zigmond, M. Z. Lin, P. Juo, L. S. Hu, M. J. Anderson, K. C. Arden, J. Blenis, and M. E. Greenberg. 1999. Akt promotes cell survival by phosphorylating and inhibiting a Forkhead transcription factor. Cell 96:857-868.
Carruba, G., M. Cervello, M. D. Miceli, R. Farruggio, M. Notarbartolo, L. Virruso, L. Giannitrapani, R. Gambino, G. Montalto, and L. Castagnetta. 1999. Truncated form of beta-catenin and reduced expression of wild-type catenins feature HepG2 human liver cancer cells. Ann. N. Y. Acad. Sci. 886:212-216.
Cha, M. Y., C. M. Kim, Y. M. Park, and W. S. Ryu. 2004. Hepatitis B virus X protein is essential for the activation of Wnt/beta-catenin signaling in hepatoma cells. Hepatology 39:1683-1693.
Chung, Y. L., M. L. Sheu, and S. H. Yen. 2003. Hepatitis C virus NS5A as a potential viral Bcl-2 homologue interacts with Bax and inhibits apoptosis in hepatocellular carcinoma. Int. J. Cancer 107:65-73.
Coffer, P. J., J. Jin, and J. R. Woodgett. 1998. Protein kinase B (c-Akt): a multifunctional mediator of phosphatidylinositol 3-kinase activation. Biochem. J. 335:1-13.
Coito, C., D. L. Diamond, P. Neddermann, M. J. Korth, and M. G. Katze. 2004. High-throughput screening of the yeast kinome: identification of human serine/threonine protein kinases that phosphorylate the hepatitis C virus NS5A protein. J. Virol. 78:3502-3513.
Cooray, S. 2004. The pivotal role of phosphatidylinositol 3-kinase-Akt signal transduction in virus survival. J. Gen. Virol. 85:1065-1076.
Cory, S., and J. M. Adams. 2002. The Bcl2 family: regulators of the cellular life-or-death switch. Nat. Rev. Cancer 2:647-656.
Cui, J., X. D. Zhou, Y. K. Liu, Z. Y. Tang, and M. Romeih. 2003. Wnt signaling in hepatocellular carcinoma: analysis of mutation and expression of beta-catenin, T-cell factor-4 and glycogen synthase kinase 3-beta genes. J. Gastroenterol. Hepatol. 18:280-287.
Datta, S. R., H. Dudek, X. Tao, S. Masters, H. Fu, Y. Gotoh, and M. E. Greenberg. 1997. Akt phosphorylation of BAD couples survival signals to the cell-intrinsic death machinery. Cell 91:231-241.
de La Coste, A., B. Romagnolo, P. Billuart, C. A. Renard, M. A. Buendia, O. Soubrane, M. Fabre, J. Chelly, C. Beldjord, A. Kahn, and C. Perret. 1998. Somatic mutations of the beta-catenin gene are frequent in mouse and human hepatocellular carcinomas. Proc. Natl. Acad. Sci. USA 95:8847-8851.
Desbois-Mouthon, C., M. J. B. V. Eggelpoel, E. Beurel, M. Boissan, R. Delelo, A. Cadoret, and J. Capeau. 2002. Dysregulation of glycogen synthase kinase-3 beta signaling in hepatocellular carcinoma cells. Hepatology 36:1528-1536.
Doble, B. W., and J. R. Woodgett. 2003. GSK-3: tricks of the trade for a multi-tasking kinase. J. Cell Sci. 116:1175-1186.
Fujimuro, M., and S. D. Hayward. 2003. The latency-associated nuclear antigen of Kaposi's sarcoma-associated herpesvirus manipulates the activity of glycogen synthase kinase-3?. J. Virol. 77:8019-8030.
Fujimuro, M., F. Y. Wu, C. ApRhys, H. Kajumbula, D. B. Young, G. S. Hayward, and S. D. Hayward. 2003. A novel viral mechanism for dysregulation of beta-catenin in Kaposi's sarcoma-associated herpesvirus latency. Nat. Med. 9:300-306.
Fukumoto, S., C. M. Hsieh, K. Maemura, M. D. Layne, S. F. Yet, K. H. Lee, T. Matsui, A. Rosenzweig, W. G. Taylor, J. S. Rubin, M. A. Perrella, and M. E. Lee. 2001. Akt participation in the Wnt signaling pathway through Dishevelled. J. Biol. Chem. 276:17479-17483.
Gan, D. D., and K. Khalili. 2004. Interaction between JCV large T-antigen and beta-catenin. Oncogene 23:483-490.
Ghosh, A. K., M. Majumder, R. Steele, K. Meyer, R. Ray, and R. B. Ray. 2000. Hepatitis C virus NS5A protein protects against TNF-alpha mediated apoptotic cell death. Virus Res. 67:173-178.
Ghosh, A. K., R. Steele, K. Meyer, R. Ray, and R. B. Ray. 1999. Hepatitis C virus NS5A protein modulates cell cycle regulatory genes and promotes cell growth. J. Gen. Virol. 80:1179-1183.
Gossen, M., and H. Bujard. 1992. Tight control of gene expression in mammalian cells by tetracycline-responsive promoters. Proc. Natl. Acad. Sci. USA 89:5547-5551.
Gotoh, J., M. Obata, M. Yoshie, S. Kasai, and K. Ogawa. 2003. Cyclin D1 over-expression correlates with beta-catenin activation, but not with H-ras mutations, and phosphorylation of Akt, GSK3 beta and ERK1/2 in mouse hepatic carcinogenesis. Carcinogenesis 24:435-442.
Hart, M., J. P. Concordet, I. Lassot, I. Albert, R. del los Santos, H. Durand, C. Perret, B. Rubinfeld, F. Margottin, R. Benarous, and P. Polakis. 1999. The F-box protein beta-TrCP associates with phosphorylated beta-catenin and regulates its activity in the cell. Curr. Biol. 9:207-210.
He, Y. P., H. H. Nakao, S. L. Tan, P. J. Polyak, P. Neddermann, S. Vijaysri, B. L. Jacobs, and M. G. Katze. 2002. Subversion of cell signaling pathways by hepatitis C virus nonstructural 5A protein via interaction with Grb2 and P85 phosphatidylinositol 3-kinase. J. Virol. 76:9207-9217.
Huang, H., H. Fujii, A. Sankila, B. M. Mahler-Araujo, M. Matsuda, G. Cathomas, and H. Ohgaki. 1999. Beta-catenin mutations are frequent in human hepatocellular carcinomas associated with hepatitis C virus infection. Am. J. Pathol. 155:1795-1801.
Krieger, N., V. Lohmann, and R. Bartenschlager. 2001. Enhancement of hepatitis C virus RNA replication by cell culture-adaptive mutations. J. Virol. 75:4614-4624.
Lan, K. H., M. L. Sheu, S. J. Hwang, S. H. Yen, S. Y. Chen, J. C. Wu, Y. J. Wang, N. Kato, M. Omata, F. Y. Chang, and S. D. Lee. 2002. HCV NS5A interacts with p53 and inhibits p53-mediated apoptosis. Oncogene 21:4801-4811.
Levy, L., C. Neuveut, C. A. Renard, P. Charneau, S. Branchereau, F. Gauthier, J. T. Van Nhieu, D. Cherqui, A. F. Petit-Bertron, D. Mathieu, and M. A. Buendia. 2002. Transcriptional activation of interleukin-8 by beta-catenin-Tcf4. J. Biol. Chem. 277:42386-42393.
Macdonald, A., K. Crowder, A. Street, C. McCormick, and M. Harris. 2004. The hepatitis C virus NS5A protein binds to members of the Src family of tyrosine kinases and regulates kinase activity. J. Gen. Virol. 85:721-729.
Macdonald, A., K. Crowder, A. Street, C. McCormick, K. Saksela, and M. Harris. 2003. The hepatitis C virus NS5A protein inhibits activating protein-1 function by perturbing Ras-ERK pathway signalling. J. Biol. Chem. 278:17775-17784.
Macdonald, A., and M. Harris. 2004. Hepatitis C virus NS5A: tales of a promiscuous protein. J. Gen. Virol. 85:2485-2502.
McCormick, C. J., L. Challinor, A. Macdonald, D. J. Rowlands, and M. Harris. 2004. Introduction of replication-competent hepatitis C virus transcripts using a tetracycline-regulable baculovirus delivery system. J. Gen. Virol. 85:429-439.
McCormick, C. J., D. J. Rowlands, and M. Harris. 2002. Efficient delivery and regulable expression of hepatitis C virus full length and mini-genome constructs in hepatocyte-derived cell lines using baculovirus vectors. J. Gen. Virol. 83:383-394.
Miyasaka, Y., N. Enomoto, M. Kurosaki, N. Sakamoto, N. Kanazawa, T. Kohashi, E. Ueda, S. Maekawa, H. Watanabe, N. Izumi, C. Sato, and M. Watanabe. 2003. Hepatitis C virus nonstructural protein 5A inhibits tumor necrosis factor-alpha-mediated apoptosis in Huh7 cells. J. Infect. Dis. 188:1537-1544.
Morin, P. J., A. B. Sparks, V. Korinek, N. Barker, H. Clevers, B. Vogelstein, and K. W. Kinzler. 1997. Activation of beta-catenin-Tcf signaling in colon cancer by mutations in beta-catenin or APC. Science 275:1787-1790.
Morrison, J. A., A. J. Klingelhutz, and N. Raab-Traub. 2003. Epstein-Barr virus latent membrane protein 2A activates ?-catenin signaling in epithelial cells. J. Virol. 77:12276-12284.
Owsianka, A., R. F. Clayton, L. D. Loomis-Price, J. A. McKeating, and A. H. Patel. 2001. Functional analysis of hepatitis C virus E2 glycoproteins and virus-like particles reveals structural dissimilarities between different forms of E2. J. Gen. Virol. 82:1877-1883.
Park, K. J., S. H. Choi, S. Y. Lee, S. B. Hwang, and M. M. C. Lai. 2002. Nonstructural 5A protein of hepatitis C virus modulates tumor necrosis factor alpha-stimulated nuclear factor kappa B activation. J. Biol. Chem. 277:13122-13128.
Polakis, P. 2000. Wnt signaling and cancer. Genes Dev. 14:1837-1851.
Polakis, P., M. Hart, and B. Rubinfeld. 1999. Defects in the regulation of beta-catenin in colorectal cancer. Adv. Exp. Med. Biol. 470:23-32.
Sarcar, B., A. K. Ghosh, R. Steele, R. Ray, and R. B. Ray. 2004. Hepatitis C virus NS5A mediated STAT3 activation requires co-operation of Jak1 kinase. Virology 322:51-60.
Shih, I. M., J. Yu, T. C. He, B. Vogelstein, and K. W. Kinzler. 2000. The beta-catenin binding domain of adenomatous polyposis coli is sufficient for tumor suppression. Cancer Res. 60:1671-1676.
Street, A., A. Macdonald, K. Crowder, and M. Harris. 2004. The hepatitis C virus NS5A protein activates a phosphoinositide 3-kinase dependent survival signalling cascade. J. Biol. Chem. 279:12232-12241.
Waris, G., A. Livolsi, V. Imbert, J. F. Peyron, and A. Siddiqui. 2003. Hepatitis C virus NS5A and subgenomic replicon activate NF-kappaB via tyrosine phosphorylation of IkappaBalpha and its degradation by calpain protease. J. Biol. Chem. 278:40778-40787.
Willems, M., H. J. Metselaar, H. W. Tilanus, S. W. Schalm, and R. A. de Man. 2002. Liver transplantation and hepatitis C. Transplant Int. 15:61-72.
Yanagi, M., R. H. Purcell, S. U. Emerson, and J. Bukh. 1997. Transcripts from a single full-length cDNA clone of hepatitis C virus are infectious when directly transfected into the liver of a chimpanzee. Proc. Natl. Acad. Sci. USA 94:8738-8743.
Zoulim, F., M. Chevallier, M. Maynard, and C. Trepo. 2003. Clinical consequences of hepatitis C virus infection. Rev. Med. Virol. 13:57-68.(Andrew Street, Andrew Mac)