Potential conflict of interest: Nothing to report.
The hepatitis C virus (HCV) E2 protein has been shown to block apoptosis and has been suggested to facilitate persistent infection of the virus. Here, we report that the anti-apoptotic activity of E2 is mediated by activation of nuclear factor kappa B (NF-κB) that directs expression of survival gene products such as tumor necrosis factor (TNF-α) receptor-associated factor 2 (TRAF2), X-chromosome–linked inhibitor of apoptosis protein (XIAP), FLICE-like inhibitory protein (FLIP), and survivin. Increased levels of these proteins were observed in HCV-infected cells and a cell line producing HCV E2 protein. The activation of NF-κB was mediated by HCV-E2–induced expression of the molecular chaperone glucose-regulated protein 94 (GRP94). Overexpression of GRP94 alone resulted in expression of anti-apoptotic proteins and blocked apoptosis induced by tumor-necrosis–related apoptosis-inducing ligand (TRAIL). Interestingly, increased levels of GRP94 were observed in cells supporting HCV proliferation that originated from liver tissues from HCV patients. Moreover, small interfering RNA (siRNA) knock-down of GRP94 nullified the anti-apoptotic activity of HCV E2. Conclusion: These data indicate that HCV E2 blocks apoptosis induced by HCV infection and the host immune system through overproduction of GRP94, and that HCV E2 plays an important role in persistent HCV infection. (HEPATOLOGY 2008.)
Approximately 170 million people (3% of the global population) are estimated to be infected with hepatitis C virus (HCV).1 Epidemiological studies have suggested that approximately 80% of acute HCV cases develop into chronic infections,2–4 resulting in chronic hepatitis.5 This high incidence of chronic HCV infection indicates that the virus produces 1 or more proteins that actively block the anti-viral functions of the host.6 The molecular bases of the pathogenic hepatic injury and the viral persistence mechanisms are, however, largely unknown.
Some viruses ensure their survival by blocking the host anti-infective apoptotic mechanisms with a variety of viral proteins that modulate various stages of the death-signaling pathways.7–10 Therefore, HCV may express a protein(s) that blocks apoptosis in the infected cells.11–13
The transcription factor nuclear factor kappa B (NF-κB) is composed of homodimers or heterodimers of polypeptides of Rel family members.14, 15 Inactive NF-κB is restricted to the cytoplasm because of its interaction with inhibitory proteins known as IκBs, which mask the nuclear translocation signal of NF-κB. Following stimulation with, for example, the proinflammatory cytokines tumor necrosis factor alpha (TNF-α) and interleukin-1, IκBs are phosphorylated by IκB kinase (IKKα/β), which results in proteasome-dependent degradation of IκBs. The degradation of IκBs leads to the release of NF-κB and allows NF-κB to translocate into the nucleus, where it activates the transcription of target genes.16, 17 Activation of NF-κB is an immediate early step required for activation of the host immune system. Therefore, many viral proteins disrupt the innate immune responses mediated by NF-κB by nullifying signaling cascades that activate NF-κB.18
Activated NF-κB also induces expression of anti-apoptotic proteins, including X-chromosome–linked inhibitor of apoptosis protein (XIAP), inhibitors of apoptosis proteins, c-FLIP (FLICE-like inhibitory protein), TNF-α receptor–associated factor 2 (TRAF2), and Bcl-XL, which inhibit multiple stages of apoptosis.19 In this respect, activation of NF-κB was shown to facilitate some viral infections by promoting viral replication.20–22 preventing virus-induced apoptosis,23, 24 priming the host cell for infection,15, 25 and contributing to an oncogenic transformation.26, 27
Glucose-regulated protein 94 (GRP94) is the endoplasmic reticulum (ER)-resident member of the heat-shock-protein 90 (Hsp90) family.28–30 Hsp90 and GRP94 interact with their counterparts (client proteins) and protect them from ubiquitin-dependent proteasomal degradation.31–33 Although the GRP94 protein is expressed constitutively in all cell types, its expression is up-regulated under various stress conditions including low glucose levels, low extracellular pH, expression of mutated proteins, and viral infections.34–36 Heat-shock proteins have a cytoprotective function and modulate apoptosis directly or indirectly. Previous studies have shown that expression of GRP94 is increased in tumor cells, including hepatocellular carcinoma, colorectal carcinoma, and lung cancer cells,37–39 and that GRP94 has an anti-apoptotic effect on some tumor cells.40 Moreover, increased levels of GRP94 were observed when a chronic hepatitis B virus (HBV) infection progressed to cirrhosis and hepatocellular carcinoma (HCC).41–43 Inhibitors of Hsp90 and GRP94 (such as geldanamycin and its less toxic derivative 17-AAG) have been investigated for efficacy in cancer treatment. Geldanamycin and 17-AAG inhibit Hsp90 and GRP94 by binding competitively to its N-terminal adenosine triphosphate–binding site. Inhibition of adenosine triphosphatase activity results in misfolding and degradation of client proteins through the ubiquitin–proteasome pathway. 17-AAG is undergoing phase II clinical trials44 for use in metastatic melanoma, prostate cancer, and multiple myeloma.45
We have previously shown that HCV E2 inhibits the apoptosis of host cells by preventing activation of caspases.46 Here, we report that the anti-apoptotic activity of E2 is mediated by activation of NF-κB, which directs expression of survival gene products such as TRAF2, XIAP, FLIP, and survivin. Increased levels of these proteins were observed in both HCV-infected cells and an HCV-E2–expressing cell line. The activation of NF-κB was mediated by HCV-E2–induced increased expression of the molecular chaperone GRP94. Overexpression of GRP94 alone resulted in expression of the anti-apoptotic proteins and blocked apoptosis induced by tumor necrosis-related apoptosis-inducing ligand (TRAIL). Interestingly, increased levels of GRP94 were observed in cells supporting HCV proliferation that originated from liver tissues of HCV patients. Moreover, small interfering RNA (siRNA) knock-down of GRP94 reduced the anti-apoptotic activity of HCV E2. These data indicate that HCV E2 blocks apoptosis induced by HCV infection and the host immune system through overproduction of GRP94, and that HCV E2 plays an important role in persistent HCV infection.
17-AAG, less toxic derivative of geldanamycin; ER, endoplasmic reticulum; FLIP, FLICE-like inhibitory protein; GRP94, glucose-regulated protein 94; HBV, hepatitis B virus; HCC, hepatocellular carcinoma; HCV, hepatitis C virus; Hsp90, heat-shock-protein 90; NF-κB, nuclear factor kappa B; PARP, poly (adenosine diphosphate-ribose) polymerase; siRNA, small interfering RNA; TNF-α, tumor necrosis factor alpha; TRAF2, tumor necrosis factor receptor-associated factor; TRAIL, tumor-necrosis–related apoptosis-inducing ligand; 2XIAP, X-chromosome–linked inhibitor of apoptosis protein;
Materials and Methods
Establishment of a Cell Line Expressing GRP94.
A plasmid pCDNA 3.1-GRP94 expressing GRP9447 was provided by Dr. Tohru Mizushima (Kumamoto University, Japan). A control plasmid pCDNA 3.1 (10 μg) and the plasmid pCDNA 3.1-GRP94 (10 μg) were transfected into Huh-7 cells by electroporation. From 48 hours posttransfection, cells were maintained in Dulbecco's modified Eagle's medium containing G418 (600 μg /mL; Calbiochem). After 3 weeks of selection, G418-resistant cell colonies were pooled and cultivated for further analyses.
Antibodies and Chemicals.
The monoclonal antibody H52 against E2 was a gift from Dr. Dubuisson at the University of Lille. Actin antibody was purchased from ICN. Antibodies against GRP94, Hsp90, survivin, and IKKα/β were purchased from Santa Cruz. Antibodies against TRAF2 and XIAP were purchased from BD, and 17-AAG was obtained from Sigma Aldrich Co.
Cell Culture and Transient Transfection.
Huh-7 cells and 293T cells were grown at 37°C in Dulbecco's modified Eagle's medium (Gibco) supplemented with antibiotics (penicillin 100 U/mL, streptomycin 10 g/mL) and 10% fetal bovine serum (Hyclone) in the presence of 6.0% CO2. FK-E1(UAA) and FK-E2(UAA) replicons, and control (vector), and GRP94 (GRP94-overexpressing) cell lines were grown under the same conditions but with the addition of the antibiotic G418 (600 g/mL; Calbiochem). The control and spE2 (HCV-E2–expressing cell line)46 were grown under the same conditions but with the addition of the antibiotic hygromycin B (300 g/mL; Calbiochem). 293T cells were electroporated as previously described.46
Knockdown of GRP94 Using siRNA.
Duplex siRNAs targeted to GRP94 and TRAF2, and control siRNAs were purchased from Bioneer Inc. (Korea). The siRNA sequences targeting GRP94 (GRP94 siRNA1 and GRP94 siRNA2) were 5′-UGA UGU GGA UGG UAC AGU A dTdT-3′ and 5′-UAC UGU ACC AUC CAC AUC A dTdT-3′. The control siRNA sequence was 5′-CCU ACG CCA CCA AUU UCG UdTdT-3′. The siRNA sequence targeting TRAF2 (TRAF2 siRNA) was 5′-CAA CCA GAA GGU GAC CUU A dTdT-3. To transfect siRNA into Huh-7 cells, 100 nM siRNA mixed with 3 L Lipofectamine 2000 (Invitrogen) was added to each well of a 6-well plate. The cells were harvested 72 hours after the first transfection.
Luciferase assays were performed using a luciferase assay kit (Promega), according to the manufacturer's instructions. The luciferase activities in the cells were normalized by the protein concentrations, as determined by the Bradford assay.
Western Blot Analysis.
Proteins were resolved by 8% or 13.5% sodium dodecyl sulfate polyacrylmide gel electrophoresis and transferred to nitrocellulose membranes (Amersham). The membranes were blocked overnight with 5% skimmed milk in Tris-buffered saline (TBS) [20 mM Tris-HCl (pH 7.5), 150 mM NaCl, 0.5% Tween-20] and then incubated with monoclonal antibodies against actin (1:5000), GRP94 (1:1000), Hsp90 (1:1000), TRAF2 (1:500), XIAP (1:1000), survivin (1:500), glyceraldehyde-3-phosphate dehydrogenase (1:5000), and IKKα/β for 5 hours. A monoclonal anti-human actin antibody was used as a control. Horseradish-peroxidase–conjugated anti-mouse, anti-goat, or anti-rabbit immunoglobulin Gs were used as secondary antibodies (1:5000), and bands were visualized by enhanced chemiluminescence according to the manufacturer's instructions (Amersham).
Human tissue specimens were supplied from the Liver Cancer Specimen Bank supported by National Research Resource Bank Program of the Korea Science and Engineering Foundation in the Ministry of Science and Technology. The consents to use the tissue specimens for research purposes were obtained from patients, and the utilization of the specimens for this research was authorized by the Institutional Review Board of College of Medicine, Yonsei University.
For fluorescence microscopic analysis, frozen liver tissues originated from patients with tumors were fixed in 4% paraformaldehyde overnight at 4°C and embedded in optimal cutting temperature (OCT) (Fisher Scientific) for sectioning. The sections (12 μm) were stained with hematoxylin-eosin. For immunostaining, sections were fixed in 4% paraformaldehyde and the antigenic epitopes were exposed by treating with 10 mM citrate buffer and heating in a microwave oven. Sections were incubated in blocking solution (3% bovine serum albumin, 5% horse serum, and 0.5% Tween-20 in phosphate-buffered saline) at room temperature for 4 hours, followed by an additional incubation with monoclonal antibody against E2 (H52) (1:100 dilution) and goat polyclonal antibody against GRP94 (1:100 dilution; Santa Cruz Biotechnology). Specific binding was detected with Alexa 488–labeled anti-mouse and Alexa 594 immunoglobulin Gdash;labeled anti-goat ummunoglobulin G (Molecular Probes). Fluorescence microscopy was performed as described previously.46 A primary rabbit antibody against GRP94 and a mouse antibody against E2 were used to detect co-localization of the 2 proteins.
HCV E2 Augments Expression of Survival Genes Related With NF-κB Activation.
We have previously shown that the HCV E2 protein inhibits apoptosis triggered through the mitochondrial pathway.46 To determine the molecular basis of the HCV E2 anti-apoptotic activity, we monitored the levels of proteins known to be involved in cell survival and anti-apoptosis processes by western blot assays in HCV-E2–expressing cells (spE2, a derivative of the hepatocellular carcinoma cell line Huh-7). To minimize the artificial effects of individual colonies expressing E2, we used a pool of E2-gene–containing colonies generated by permanent cell line selection methods. Of the proteins tested, the levels of the anti-apoptotic proteins XIAP, TRAF2, and FLIP were increased in the spE2 cells (Fig. 1A). We measured the levels of phosphorylated IKKα/β, which is a measure of NF-κB activation, with an antibody against phosphorylated IKKα/β; this is because the anti-apoptotic proteins XIAP, TRAF2, and FLIP are under the control of NF-κB. The level of phosphorylated IKKα/β was increased in the spE2 cells (Fig. 1A).
Similarly, increased levels of the anti-apoptotic proteins XIAP, TRAF2, and survivin, which are under the control of NF-κB, were observed in Huh-7 cells containing the HCV replicon FK-E2(UAA) expressing HCV nonstructural proteins (NS3, NS4A, NS4B, NS5A, and NS5B), which are required for replication of HCV RNA, and structural proteins (core, E1, and E2)46 (Fig. 1B). Conversely, the levels of these anti-apoptotic proteins in Huh-7 cells containing the HCV replicon FK-E1(UAA) that expresses HCV nonstructural proteins and the structural proteins core and E1, but not E2,46 were the same as those in control Huh-7 cells without an HCV replicon (Fig. 1B). These data indicate that the levels of the anti-apoptotic proteins XIAP, TRAF2, FLIP, and survivin are strongly increased with HCV E2, regardless of the presence of other HCV proteins.
We examined the protein level of IκBα in control Huh-7 cells (Fig. 1C) and in the Huh-7 cells containing replicons FK-E1(UAA) and FK-E2(UAA) (Fig. 1C) by western blot assays; this is because the phosphorylation-dependent degradation of IκBαs is an essential step for NF-κB activation after TNF-α treatment. Similar levels of IκBα proteins were observed in the control Huh-7 cells and in the Huh-7 cells containing FK-E1(UAA) before TNF-α treatment (Fig. 1C). Within 5 minutes of TNF-α treatment, IκBα proteins were not detected in the control Huh-7 cells (Fig. 1C). In Huh-7 cells containing the replicon FK-E1(UAA) that expresses NS3–5B, core, and E1 proteins, the degradation of IκBα was delayed by up to 10 minutes after treatment with TNF-α (Fig. 1C). This indicates that 1 of the HCV proteins encoded by the FK-E1(UAA) replicon blocks degradation of IκBα. The molecular basis of this effect remains unknown. Importantly, the amount of IκBα in untreated cells was lowest in the cells containing the FK-E2(UAA) replicon and expressing HCV NS3–5B, core, E1, and E2 (Fig. 1C). Moreover, the residual IκBα proteins were not detected in the FK-E2(UAA)–containing cells within 5 minutes of TNF-α treatment (Fig. 1C). This indicates that NF-κB may be partially activated in E2-expressing cells, because of the reduced basal level of IκBα, but that it is fully activated soon after treatment with TNF-α, even in the presence of other HCV proteins. Taken together, these data indicate that HCV E2 may induce expression of anti-apoptotic proteins through activation of NF-κB. It should be noted that some viruses are known to constitutively activate NF-κB to prevent host cell apoptosis, leading to persistent infection.8, 48, 49
GRP94 Is Overexpressed in Cells Expressing HCV E2 Protein.
We attempted to determine the mechanism by which HCV E2 activates NF-κB. We focused on 1 report that suggested that E2 activates the GRP94 promoter.50 GRP94 is a molecular chaperone in the lumen of the endoplasmic reticulum (ER). Its abnormal expression correlates with carcinogenesis, progression, and prognosis of HCC triggered by HBV infection.41 Therefore, we investigated the expression levels and subcellular localization of E2 and GRP94 by western blot analysis (Fig. 3C; Fig. 4C) and immunocytochemical analysis with spE2 (Fig. 2A) and FK-E2(UAA) cells (Fig. 2B). To minimize the artificial effects of individual colonies overexpressing HCV E2, we used a pool of E2-overexpressing cells (spE2) that had been selected during generation of a permanent cell line. Higher levels of GRP94 were observed in HCV-E2–expressing cells compared with nonexpressing cells, as reported previously.50 Both E2 and GRP94 were found to be localized in the ER. Interestingly, the GRP94 proteins seemed to be co-localized with E2, as indicated by yellow regions in Fig. 2A/B. The intensities of red signals revealed that GRP94 protein levels were increased in the spE2 cells (Fig. 2A) and Huh-7 cells containing the FK-E2(UAA) replicon expressing high levels of E2 (Fig. 2B).
GRP94 Activates NF-κB.
The relationship between increased expression of GRP94 and activation of NF-κB was investigated because both of these biological events are triggered by HCV E2. We established a cell line that overexpressed GRP94 (Fig. 3A). To minimize the artificial effects of individual colonies overexpressing GRP94, we used a pool of GRP94-overexpressing cells (GRP94) selected out in the process of a permanent cell line generation. Expression levels of the NF-κB–related gene products TRAF2, XIAP, and survivin were increased in GRP94-overexpressing cells (Fig. 3A), as seen in HCV-E2–overexpressing cells (Fig. 1A, B). Moreover, the expression levels of IKKα/β, which is also under the control of NF-κB, were markedly increased in GRP94-overexpressing cells. Conversely, levels of Hsp90, a GRP94-related protein, and actin (a negative control) were not changed in GRP94-overexpressing cells. These results indicate that GRP94 affects NF-κB–related gene expression.
Subcellular localizations of GRP94 and IKKα/β were monitored by immunocytochemical analysis (Fig. 3B). GRP94 proteins were localized to the ER in cells expressing high and normal levels of GRP94 and IKKα/β (Fig. 3B). Interestingly, higher levels of IKKα/β were observed in cells expressing higher levels of GRP94 (Fig. 3B), which is consistent with the western blot results (Fig. 3A). Interestingly, the GRP94 proteins were partially co-localized with the cytoplasmic IKKα/β proteins in GRP94-overexpressing cells.
The role of GRP94 in NF-κB activation was further investigated by an RNA-interference method using siRNAs against GRP94. siRNAs against GRP94 and a control siRNA were transfected into control and spE2 cells, and the levels of NF-κB–related gene products were monitored by western blot analysis (Fig. 3C). The siRNAs against GRP94 (GRP94 siRNA-1 and GRP94 siRNA-2) partially reduced the levels of GRP94 in normal cells (Fig. 3C) and in GRP94-overexpressing spE2 cells (Fig. 3C) compared with control siRNA. Treatment with siRNA against GRP94 reduced the basal levels of GRP94, XIAP, TRAF2, surviving, and IKKα/β in normal cells (Fig. 3A), and those of proteins overshot by production of HCV E2 (Fig. 3A). By contrast, the level of the negative control actin was not affected by the presence of the GRP94-specific siRNAs (Fig. 3C). There were no reductions in the expression levels of E2 and Hsp90 with siRNA treatment (Fig. 3C). These data indicate that increased expression of NF-κB–related genes is mediated by the overproduction of GRP94, which in turn is induced by HCV E2. In other words, GRP94 is a mediator of anti-apoptotic activity of HCV E2. Curiously, increased expression of GRP78 was consistently observed in the cells treated with siRNAs against GRP94 (Fig. 3C). We speculate that the increase in GRP78, another molecular chaperone in the ER lumen, may compensate for the reduced levels of the molecular chaperone GRP94 by an unknown mechanism.
The activation of NF-κB by overproduction of GRP94 was investigated by measuring the activity of a reporter gene (firefly luciferase) under the control of the NF-κB response element (Fig. 4A). Overproduction of GRP94 increases NF-κB–dependent luciferase activity by approximately 3-fold to 5-fold compared with the control vector (Fig. 4A). An increase in NF-κB reporter activity of approximately 10-fold was observed with TNF-α treatment (Fig. 4A). These data indicate that overproduction of GRP94 activates NF-κB.
We tried to determine the stage of the signaling cascade at which NF-κB is activated by GRP94. First, the role of IKKα/β was investigated because it plays a key role in activation of NF-κB by cytokines such as TNF-α. The effects of BAY11-7082, a specific inhibitor of IKKα/β, on the expression of NF-κB–related gene products were monitored using GRP94-overexpressing (Fig. 4B) and control cell lines (Fig. 4B). The levels of XIAP, TRAF2, survivin, and IKKα/β were markedly reduced in GRP94-expressing and control cells treated with BAY11-7082 (Fig. 4B). These data indicate that IKKα/β activation is needed for GRP94-induced activation of NF-κB. Curiously, the levels of GRP94 protein were reduced by BAY11-7082 treatment (Fig. 4B). This indicates that expression of GRP94 also may be influenced by NF-κB. If this is the case, GRP94 production and NF-κB activation form a positive feedback loop; however, further experiments are needed to confirm this hypothesis.
We also investigated the role of TRAF2, a TNF-α-receptor–associated protein essential for TNF-α signaling upstream of IKKα/β, in HCV E2 activation of NF-κB. The effects of siRNA against TRAF2 on expression of NF-κB–related gene products were monitored by western blot analysis. As expected, the siRNA against TRAF2 blocked expression of TRAF2 (Fig. 4C). However, there was no reduction in the levels of NF-κB–related proteins (XIAP, survivin, IKKα/β) after treatment with siRNAs against TRAF2 (Fig. 4C). These data indicate that activation of NF-κB by HCV E2 is independent of TRAF2, TNF-α receptor, and related proteins because TRAF2 mediates the NF-κB activation through TNFR1, TNFR2, and CD40.
GRP94 Prevents Apoptosis Induced by TRAIL.
The role of GRP94 in inhibition of apoptosis during HCV infection was investigated using an siRNA against GRP94 and a control siRNA, the GRP94-overexpressing, and the E2-expressing cell lines (Fig. 5). We used the strong apoptosis inducer TRAIL to mimic the physiological conditions of a virus infection that induces TRAIL expression.51 Cleavage of poly (adenosine diphosphate-ribose) polymerase (PARP), which is a substrate of activated caspase-3/7, was used to monitor the process of apoptosis. After treatment with TRAIL, cleavage of PARP was observed in control cells irrespective of the presence of control siRNA or GRP94 siRNA (Fig. 5). Conversely, cleavage of PARP was blocked in the GRP94-overexpressing (Fig. 5) and in the HCV-E2–expressing cells (Fig. 5). This indicates that GRP94 overexpression can inhibit TRAIL-induced apoptosis. Importantly, these anti-apoptotic activities of GRP94 and HCV E2 were abrogated by treatment with siRNA against GRP94 (Fig. 5). This effect is consistent with inhibition of NF-κB–related gene expression by siRNAs against GRP94 (Fig. 3C). These data indicate that the anti-apoptotic activity of HCV E2 is mediated by overexpression of GRP94. It should be noted that basal expression of GRP94 was not sufficient to block apoptosis by TRAIL (Fig. 5). This may indicate that basal levels of anti-apoptotic proteins such as XIAP, FLIP, and survivin in normal cells are too low to block apoptotic signals generated by TRAIL even with basal levels of GRP94 (Fig. 1A, B).
GRP94 Is Needed for Blockage of Apoptosis Induced by TRAIL in HCV-Infected Cells.
To confirm that the anti-apoptotic activity of E2 is induced by GRP94 and NF-κB–related gene products in HCV-infected cells, the levels of GRP94, XIAP, TRAF2, FLIP, and Bcl-xL were monitored using an HCV-infection system. We used the infectious JFH1 strain of HCV52 and a JFH1 derivative known as JFH5a-Rluc with a reporter gene (Renilla luciferase) in the NS5A of HCV (Fig. 6A) that is suitable for quantification of HCV infection.53 Three days after infection, Renilla luciferase activity was greatly increased in the Huh-7 cells infected with JFH5a-Rluc, indicating that the infection and replication of HCV occurred properly (data not shown). Western blot analysis of GRP94 and NF-κB–related proteins showed that the levels of cell-survival–related gene products (XIAP, FLIP, and Bcl-xL) were greatly increased in the HCV-infected cells (Fig. 6B). Considering that HCV E2 induces GRP94 production and that overexpressed GRP94 activates NF-κB, the induction of these proteins is attributed to the overproduction of GRP94 and NF-κB activation, as indicated by the increased levels of GRP94 and phosphorylated IκBα, and the reduced level of IκBα (Fig. 6B). Consistent with this hypothesis, immunocytochemical analysis showed greatly increased levels of GRP94 in the Huh-7 cells infected with JFH5a-Rluc (Fig. 6C). Note that higher levels of GRP94 are observed in the cells infected with HCV compared with the uninfected cells (Fig. 6C).
The apoptosis-blocking effects of GRP94 in HCV-infected cells were observed with TRAIL, which induces apoptosis of cancer cells and virus-infected cells51 with or without treatment of an siRNA against GRP94 (Fig. 6D). The cleavage of PARP in mock-infected cells was observed in TRAIL-treated cells, even in the presence of basal levels of GRP94 (Fig. 6D). Marked changes in PARP cleavage efficiency were observed in HCV-infected cells. There were almost no reductions in the levels of PARP cleavage in HCV-infected cells, even after TRAIL treatment (Fig. 6D). This indicates that 1 of the HCV proteins blocks TRAIL-induced apoptosis in HCV-infected cells. The blockade of apoptosis in HCV-infected cells was nullified by knocking down of GRP94, as shown by monitoring the PARP cleavage levels in GRP94-specific siRNA-treated HCV-infected cells (Fig. 6D).
The role of GRP94 in HCV infection was monitored by measuring the infectivity of the HCV strain JFH 5a-Rluc in the presence of TRAIL. Infectivity of HCV was reduced by about 20% with TRAIL treatment (Fig. 6D). Infectivity of HCV was reduced by approximately 10% with GRP94-specific siRNA treatment (Fig. 6D). The most marked reduction in HCV infectivity (almost 70%) was observed in cells treated with TRAIL and the siRNA against GRP94 (Fig. 6D). These data indicate that GRP94 may play a key role in HCV infection and establishment of chronic hepatitis.
Increased Level of GRP94 Protein Is Observed in the HCV-Infected Liver Cells of HCV Patients.
The pathological relevance of GRP94 overproduction in HCV-infected hepatocyte was investigated using liver tissue samples from HCV patients that were obtained from the Liver Cancer Bank in Korea (http://www.liverca.com/). Clinical background data are summarized in Tables 1 and 2. Overall morphologies of normal liver, HCV-infected liver, and HCV-positive cancerous liver tissues were observed by hematoxylin-eosin staining (Fig. 7A). Abnormal morphologies of liver tissues were observed in HCV-infected and cancerous liver tissues (Fig. 7A). HCV was confirmed in liver tissues from a chronic HCV patient and from an HCC patient chronically infected with HCV, determined by western blot analysis with an antibody against HCV E2 (Fig. 7B). Increased levels of GRP94 were detected in these patient tissues (Fig. 7B). Expression levels of the anti-apoptotic protein XIAP, which is under the control of NF-κB, were greatly increased in the liver tissues from HCV patients and from HCC patients infected with HCV compared with levels in normal liver tissue (Fig. 7B). The relationship between HCV E2 and production of GRP94 was further analyzed by immunohistochemical analysis of liver tissues (Fig. 7C). As expected, no HCV E2 protein was observed in the normal liver tissue (Fig. 7C), and a low level of GRP94 was detected in the tissue (Fig. 7C, panels a, b, g, h, s, and t). Some cells in the liver tissues of the HCV patient (Fig. 7C, panels c, d, i, j, u, and v) and the HCV-positive HCC patient (Fig. 7C, panels e, f, k, l, w, and x) revealed HCV E2 protein. Increased levels of GRP94 proteins were also observed in some cells of the liver tissues of the HCV patient (Fig. 7C) and the HCV-positive HCC patient (Fig. 7C). Importantly, the increased levels of GRP94 coincided with presence of HCV E2 in the patients' liver cells (Fig. 7C). These results indicate that GRP94 expression is increased in HCV-infected cells of the patients and that the GRP94 activity seen in cell-culture systems (induction of survival gene products through activation of NF-κB) is likely to occur in HCV patients.
Table 1. Description of Normal Liver
Pathologic Status of Adjacent Liver
Viral Marker (HCV)
Within normal limit
Within normal limit
Table 2. Clinical Background of HCC Patient With HCV Infection
Pathologic Status of Adjacent Liver
6.4 × 6.5
6.5 × 5.5
5.5 × 4.5
Many viruses encode proteins that suppress or delay apoptosis of host cells long enough for the viruses to replicate or to establish persistent infection.7, 15, 24, 51 The HCV envelope protein E2 has been shown to have anti-apoptotic activity and to augment colony formation by HCV replicons.46 Here, we report the molecular basis of the anti-apoptotic activity of HCV E2. Several lines of evidence indicate that the anti-apoptotic activity of HCV E2 is mediated by overproduction of the molecular chaperone GRP94. First, HCV E2 induces overproduction of GRP94 (Fig. 2), as previously described by Liberman et al.50 Second, both HCV-E2–expressing cells (spE2) and GRP94-overexpressing cells induced expression of NF-κB–related anti-apoptotic gene products such as XIAP, FLIP, and survivin (Fig, 1A, B, and 3A). Third, the expression of NF-κB–related gene products in HCV-E2–expressing cells was reduced by the knockdown of GRP94 with siRNAs (Fig. 3C). Fourth, GRP94 was induced in cells infected with the HCV strain JFH 5a-Rluc (Fig. 6B, C). Moreover, marked induction of anti-apoptotic proteins such as XIAP, FLIP, and Bcl-xL, which are under the control of NF-κB, was observed in the HCV-infected cells (Fig. 6B). The induction of GRP94 by HCV infection was also seen in the liver tissues of HCV patients (Fig. 7C). Finally, TRAIL-induced apoptosis was inhibited in E2-expressing cells and GRP94-overexpressing cells (Fig. 5). Importantly, the anti-apoptotic activity of E2 was nullified by treatment of E2-overexpressing cells or HCV-infected cells with an siRNA against GRP94 (Figs. 5, 6D). Taken together, these data indicate that HCV E2 facilitates persistent infection of HCV by inducing GRP94.
The increased level of GRP94 results in production of anti-apoptotic proteins such as XIAP, FLIP, and survivin (Fig. 1A and 1B). It is most likely that these survival-related genes are induced by activation of NF-κB because these genes are under the transcriptional control of NF-κB. Moreover, expression of these proteins requires the activation of IKKα/β (Fig. 4B) and the reduction of IκBα level (Fig. 1C), both of which are prerequisite events for NF-κB activation. The molecular basis of GRP94-induced NF-κB activation is not fully understood. Nevertheless, it seems to be clear that IKKα/β activation plays a key role in the GRP94-mediated activation of NF-κB because an IKK inhibitor (BAY11-7082) blocked NF-κB activation by GRP94 (Fig. 4B). It should be noted that the IKKα/β proteins were redistributed in the GRP94-overexpressing cells and were partially co-localized with GRP94 (Fig. 3B). Moreover, IKKα/β proteins were co-immunoprecipitated with GRP94 (data not shown). This may indicate that GRP94 forms a complex with IKKα/β and activates IKKα/β directly or indirectly through an unidentified protein. It remains unclear how a protein in the ER lumen (GRP94) can activate IKKα/β, which is on the cytosolic side of the ER membrane. In this respect, it should be noted that GRP94 is known to exist at multiple loci (cell surface, transmembrane, and ER lumen).54 The GRP94 in the membrane may participate in activation of IKKα/β. TRAF2, which plays a key role in TNF-α receptor–mediated NF-κB activation55 does not seem to be involved in the NF-κB activation mediated by GRP94 because an siRNA against TRAF2 does not hamper the NF-κB activation by GRP94 (Fig. 4C). The molecular detail of the mechanism by which GRP94 activates IKKα/β remains to be determined.
Two other HCV proteins are known to activate NF-κB. Core56–58 and NS5A59, 60 were shown to activate NF-κB through distinct pathways. However, the activities of these proteins were not sufficiently strong to trigger induction of survival genes such as XIAP and survivin, as shown in cells containing the FK-E1(UAA) replicon, which produces core and NS5A but not E2 (Fig. 1B). Conversely, these genes were induced in cells containing the FK-E2(UAA) replicon, which produces core, NS5A, and E2 (Fig. 1B). This indicates that HCV E2 is the key protein involved in induction of survival genes, even though core and NS5A also may contribute to activation of NF-κB.
HCV E2 has been reported to trigger transcription of 2 molecular chaperones, GRP94 and GRP78, in the lumen of ER, although the molecular detail of the mechanism involved has not been determined.50 Our research was focused on GRP94 rather than GRP78, which also has been shown to have cytoprotective function.40, 61 because GRP78 has been reported not to activate NF-κB.62 We found that HCV E2 and GRP94 are partially co-localized in the ER (Fig. 2). Moreover, HCV E2 and GRP94 form a complex, as shown by co-immunoprecipitation analysis (data not shown). Nevertheless, we do not know whether the GRP94–HCV-E2 complex is involved in the transcriptional activation of GRP78 and GRP94. One plausible hypothesis for the induction of these molecular chaperons is that “ER stress” generated by the production of HCV E263 may trigger the production of GRP78 and GRP94 similarly to the induction of these proteins by the large surface protein of HBV.64 Investigations into the mechanism of GRP94 induction by HCV E2 are in progress.
Overproduction of GRP94 by HCV may be related to the pathological progress of HCV patients. It is likely that the anti-apoptotic activity of E2 contributes to persistent HCV infections, resulting in chronic hepatitis, by blocking innate and acquired immunity. Moreover, the anti-apoptotic activity also may contribute to generation of HCC. In this respect, it is noteworthy that transgenic mice expressing HCV core-E1-E2 proteins produced larger HCC, compared with control mice or transgenic mice expressing HCV core alone, when they were treated with diethylnitrosamine.65 Moreover, the HCV transgenic mice liver producing core-E1-E2 underwent significantly less apoptosis than transgenic tissue expressing core only in a Fas-induced apoptosis analysis system.65 It should be noted that very high levels of GRP94 proteins were observed in the tumor tissues of the HCV patient (Fig. 7C) and that a large portion of GRP94 proteins are redistributed to the edges of cells (Fig. 7C). Interestingly, a similar GRP94 distribution pattern has been reported in other tumor cells.42 Moreover, the increase in GRP94 level was shown to be correlated with the disease progression of HBV infection66 although unrelated to HCV, HBV shows pathological progress similar to that of HCV. The levels of GRP94 were greatest in HCC patients and lowest in patients with chronic HBV infection, with intermediate levels reported in patients with cirrhosis.66 This may indicate that the pathological progress of hepatotropic viruses (HBV and HCV) is mediated by GRP94, although it is not clear which HBV protein induces production of GRP94.
Investigations into the function of GRP94 will further our understanding of the pathogenic process after HCV infection and aid in the development of anti-HCV drugs that activate defense mechanisms against viral infections.