Mechanism of Hepatitis C Virus (HCV)-induced Osteopontin and Its Role in Epithelial to Mesenchymal Transition of Hepatocytes*

Background: The mechanism underlying HCV-induced OPN remains elusive. Results: HCV-induced Ca2+ signaling, oxidative stress, and activation of AP-1 and Sp1 play a critical role in OPN activation. Conclusion: OPN plays important roles in EMT and cell migration induced by HCV through the activation of Akt, GSK-3β, and β-catenin. Significance: Our findings provide a novel implication of OPN in HCV-induced HCC. Osteopontin (OPN) is a secreted phosphoprotein, originally characterized in malignant-transformed epithelial cells. OPN is associated with tumor metastasis of several tumors and is overexpressed in hepatocellular carcinoma (HCC) tissue involving HCC invasion and metastasis. Importantly, OPN is significantly up-regulated in liver injury, inflammation, and hepatitis C virus (HCV)-associated HCC. However, the underlying mechanisms of OPN activation and its role in HCV-mediated liver disease pathogenesis are not known. In this study, we investigated the mechanism of OPN activation in HCV-infected cells. We demonstrate that HCV-mediated Ca2+ signaling, elevation of reactive oxygen species, and activation of cellular kinases such as p38 MAPK, JNK, PI3K, and MEK1/2 are involved in OPN activation. Incubation of HCV-infected cells with the inhibitors of AP-1 and Sp1 and site-directed mutagenesis of AP-1- and Sp1-binding sites on the OPN promoter suggest the critical role of AP-1 and Sp1 in OPN promoter activation. In addition, we show the in vivo interactions of AP-1 and Sp1 with the OPN promoter using chromatin immunoprecipitation assay. We also show the calpain-mediated processing of precursor OPN (∼75 kDa) into ∼55-, ∼42-, and ∼36-kDa forms of OPN in HCV-infected cells. Furthermore, we demonstrate the critical role of HCV-induced OPN in increased phosphorylation of Akt and GSK-3β followed by the activation of β-catenin, which can lead to EMT of hepatocytes. Taken together, these studies provide an insight into the mechanisms of OPN activation that is relevant to the metastasis of HCV-associated HCC.


Osteopontin (OPN) is a secreted phosphoprotein, originally characterized in malignant-transformed epithelial cells. OPN is associated with tumor metastasis of several tumors and is overexpressed in hepatocellular carcinoma (HCC) tissue involving HCC invasion and metastasis. Importantly, OPN is significantly up-regulated in liver injury, inflammation, and hepatitis C virus
(HCV)-associated HCC. However, the underlying mechanisms of OPN activation and its role in HCV-mediated liver disease pathogenesis are not known. In this study, we investigated the mechanism of OPN activation in HCV-infected cells. We demonstrate that HCV-mediated Ca 2؉ signaling, elevation of reactive oxygen species, and activation of cellular kinases such as p38 MAPK, JNK, PI3K, and MEK1/2 are involved in OPN activation. Incubation of HCV-infected cells with the inhibitors of AP-1 and Sp1 and site-directed mutagenesis of AP-1-and Sp1-binding sites on the OPN promoter suggest the critical role of AP-1 and Sp1 in OPN promoter activation. In addition, we show the in vivo interactions of AP-1 and Sp1 with the OPN promoter using chromatin immunoprecipitation assay. We also show the calpain-mediated processing of precursor OPN (ϳ75 kDa) into ϳ55-, ϳ42-, and ϳ36-kDa forms of OPN in HCV-infected cells. Furthermore, we demonstrate the critical role of HCV-induced OPN in increased phosphorylation of Akt and GSK-3␤ followed by the activation of ␤-catenin, which can lead to EMT of hepatocytes. Taken together, these studies provide an insight into the mechanisms of OPN activation that is relevant to the metastasis of HCV-associated HCC.
Hepatitis C virus (HCV) 2 infection leads to chronic hepatitis, which may progress to liver fibrosis, cirrhosis, and hepatocellu-lar carcinoma (HCC) (1). HCV is an enveloped single-stranded, positive sense RNA virus, which is ϳ9.6 kb in length and contains a 5Ј-untranslated region (UTR), a single open reading frame, and a 3Ј-UTR (2). The 5Ј-UTR contains an internal ribosome entry site, which directs cap-independent translation of a polyprotein precursor of ϳ3000 amino acids that is cleaved by viral proteases and host cell signal peptidases into mature structural proteins (core, E1, E2, and p7) and nonstructural proteins (NS2, NS3, NS4A, NS4B, NS5A, and NS5B) (2). The study of molecular mechanisms of HCV replication and pathogenesis has been hampered by the lack of an efficient cell culture system or a suitable small animal model. A robust and productive HCV (genotype 2a) infection system provided a major breakthrough that allows the production of infectious virus in cell culture (3,4).
HCC is one of the most common cancers, which is mainly due to an increasing incidence of hepatitis B and hepatitis C viral infections as well as alcoholic liver disease (5). HCC is a complex and heterogeneous tumor with frequent intrahepatic spread and extrahepatic metastasis (6). As with most solid malignant tumors, hepatocarcinogenesis is considered to be a multistep process involving uncontrolled cellular growth, along with modulation of the blood supply to promote tumor growth (7). HCV-related HCC has been reported to be associated with an increased recurrence after liver resection, signifying that HCV may promote tumor growth and metastasis (8). However, the underlying mechanism of HCV-induced HCC remains unknown. Oxidative stress and steatosis is suggested to play a critical role in the development of liver injury or HCC in chronic HCV infection (9,10). Our previous studies have shown the induction of oxidative stress in human hepatoma cells expressing HCV proteins or infected with HCV (11,12).
Several studies have shown that epithelial to mesenchymal transition (EMT) may be an important mechanism for HCC metastasis (13). EMT is a biological process that allows epithe-lial cells to lose their phenotypic characteristic and acquire mesenchymal cell features (13). Various molecular processes, such as activation of transcription factors, expression of specific cell surface proteins, reorganization and expression of cytoskeletal proteins, and production of ECM-degrading enzymes, are involved in initiating EMT and enabling it to be motile and invasive (13). Previously, it has been shown that HCV infection reduced hepatoma cell polarity reminiscent of EMT (14).
Osteopontin (OPN) is linked to tumor progression and metastasis in a variety of cancers, including HCC (15,16). Studies have shown the up-regulation of OPN in tumorigenesis and angiogenesis and in response to inflammation and liver injury (15,16). Recent studies have shown the correlation of serum OPN levels with hepatic inflammation and fibrosis in association with alcohol intake (17). Several viruses such as murine polyoma virus middle T antigen, HBV X protein, and HIV have been shown to induce OPN, which can lead to cell invasion and metastasis (18 -20). OPN can induce autocrine and paracrine signaling by binding to the cell surface receptors such as integrins ␣V␤1, ␣V␤3, ␣V␤5, ␣4␤1, ␣5␤1, ␣8␤1, and CD44 and transduces cell-matrix signaling directed to increased motility, invasion, and angiogenesis (15,16).
Recent studies have indicated that ϳ26% of mutations in ␤-catenin have been found in HCC associated with HCV infection (27). ␤-Catenin is a key downstream effector in the Wnt signaling pathway (28). During EMT, ␤-catenin is released from the E-cadherin and ␣-catenin complex, where it interacts with other proteins, raising the possibility that ␤-catenin signaling contributes to EMT (28). The level of cytoplasmic ␤-catenin is tightly controlled by glycogen synthase kinase 3␤ (GSK-3␤) phosphorylation, which triggers its degradation through the ubiquitin pathway (28).
In this study, we determined the mechanism of OPN activation, proteolytic processing, and secretion from HCV-infected cells. Our studies show that HCV-induced activation of OPN is mediated by altered Ca 2ϩ signaling, induction of ROS, and activation of cellular kinases. We also show the role of AP-1 and Sp1 in OPN promoter activation. Furthermore, we investigated the role of calpain in the post-translational processing of OPN. In addition, our results also show the role of Akt, GSK-3␤, and ␤-catenin in HCV-induced EMT. Collectively, these observations provide a novel role of HCV-induced OPN in EMT and migration of hepatocytes.

EXPERIMENTAL PROCEDURES
Plasmid, Antibodies, and Reagents-The infectious J6/JFH-1 cDNA (HCV genotype 2a) along with the replication-defective JFH-1/GND construct were obtained from Dr. Charles Rice (Rockefeller University, New York). The wild-type and various deletion constructs of the OPN promoter-luciferase reporter were provided by Dr. J. H. Chen (Taiwan University, Taiwan) (29). The full-length FLAG-OPN DNA (pDest490-OPN-a) was purchased from Addgene, Cambridge, MA.
Cell Lines-The human hepatoma cell line Huh7.5 (obtained from Dr. C. Rice, Rockefeller University) were grown in Dulbecco's modified Eagle' medium (DMEM) supplemented with 10% fetal calf serum, 100 units of penicillin/ml, and 100 g of streptomycin sulfate/ml. The cells were incubated at 37°C in a 5% CO 2 incubator. The study of HCV-mediated liver disease progression is difficult due to the lack of a convenient small animal model susceptible to virus infection. The cell culture system using the human hepatoma cell line Huh-7/Huh7.5 is widely used to study HCV-mediated disease pathogenesis in the HCV field.
HCV Cell Culture Infection System-The J6/JFH-1 RNA was transcribed and delivered into Huh7.5 cells by electroporation (26). The cell-free virus was propagated in Huh7.5 cell culture as described previously (4). The expressions of HCV protein in HCV-infected cells were analyzed using Western blot. The HCV cell culture supernatant was collected at appropriate time points and was used to infect naive Huh7.5 cells at a multiplicity of infection of 1 for 5-6 h at 37°C and 5% CO 2 (4,23,26). The viral titer in cell culture supernatant was expressed as focus forming unit ml Ϫ1 , which was determined by the average number of HCV-NS5A-positive foci detected at the highest dilutions as described previously (4). The cell culture supernatant collected from Huh7.5 cells expressing JFH-1/GND (replication defective virus) was used as a negative control. In most of the experiments, HCV-infected cells were serum-starved for 4 h before harvesting. Most of the experiments were performed at day 6 HCV post-infection, and we observed significant OPN induction and secretion (data not shown). For proteolytic processing of OPN, we used HCV days 7-8 post-infection because HCV has the ability to process all cleaved forms significantly at later time points (data not shown).
Hepatocyte Co-culture System-The primary human hepatocytes were obtained from Dr. Ajit Kumar (George Washington University, Washington, D. C.). Briefly, the hepatic stellate cell line (CFSC-8B) was used as a feeder cell layer, and a freshly isolated human hepatocyte suspension (Cambrex, Walkersville, MD) was seeded over the feeder cell line in a hepatocytedefined medium, as described previously (30). Primary hepatocyte cultures form spherical masses after 30 days in co-culture. The hepatocyte cultures containing spherical masses were harvested with 0.05% trypsin in hepatocyte-defined medium (supplemented with 1% FBS), reseeded in 6-well plates, and propagated in hepatocyte-defined medium. The in vitro transcribed J6/JFH-1 plasmid was transfected into primary human hepatocytes (PHH) as described previously (30). To verify if HCV particles were released in the culture supernatant of transfected PHH, conditioned media were collected and used to infect naive PHH as described previously (30). Total cellular RNA was extracted using TRIzol (Invitrogen), and HCV copy number was analyzed using quantitative RT-PCR (data not shown). For further studies, PHH or PHH infected with J6/JFH-1 HCV, at a multiplicity of infection of 1, was harvested at day 8 post-infection; cellular lysates were prepared by incubating in radioimmunoprecipitation (RIPA) buffer (50 mM Tris, pH 7.5, 150 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, 1 mM sodium orthovanadate, 1 mM sodium formate, 10 l/ml protease inhibitor mixture (Thermo Scientific, IL)) for 30 min on ice.
Humanized Mice Liver Tissue-Liver specimens from normal and human hepatocyte-engrafted MUP-uPA/SCID/bg mice infected with HCV were received from Dr. Ajit Kumar (George Washington University). Frozen samples were washed twice with cold PBS and thawed in RIPA buffer (as described above) and gently crushed with a glass rod, followed by sonication and incubation on ice for 30 min. Samples were centrifuged at 4°C (13,400 ϫ g) for 5 min, and the supernatant was collected, and the status of OPN was analyzed by Western blotting as described by us earlier (25).
Western Blot Analysis-Mock (Huh7.5) and HCV-infected cells were harvested, and cellular lysates were prepared by incu-bating in RIPA buffer for 30 min on ice. Supernatants from mock-and HCV-infected cells were concentrated using centrifugal filter units (Millipore, MA). Equal amounts of protein from lysates or cell culture supernatants were subjected to SDS-PAGE. Gels were electroblotted onto nitrocellulose membrane (Thermo Scientific, IL) in 25 mM Tris, 192 mM glycine, and 20% methanol. Membranes were incubated for 1 h in blocking buffer ((20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.1% Tween 20, 5% dry milk), probed with primary antibody for 1 h at room temperature (RT), and washed twice for 5 min with blocking buffer without milk followed by incubation with secondary antibody for 1 h at RT. After an additional washing step with blocking buffer, immunoblots were visualized using the Odyssey Infrared Imaging System (Li-Cor Biosciences, Lincoln, NE).
SYBR Green RT-PCR-Total cellular RNA was extracted using TRIzol (Invitrogen) and DNase treated with RQ1 RNasefree DNase prior to cDNA synthesis. The cDNA was reversetranscribed from 1 g of total RNA using a reverse transcription kit (Applied Biosystems). Quantitative RT-PCR was carried out using SYBR Green Master Mix (Applied Biosystems) and specific primer sets (Table 1). 18 S rRNA was used as an internal control. Amplification reactions were performed under the following conditions: 2 min at 50°C, 10 min at 95°C, 40 cycles for 10 s at 95°C, and 1 min at 60°C. Relative transcript levels were calculated using ⌬⌬C t method as specified by the manufacturer.
Quantitative Real Time RT-PCR-Total RNA was extracted from mock-and HCV-infected cells as described above. HCV RNA was quantified by real time RT-PCR using ABI PRISM 7500 sequence detector (Applied Biosystems). Amplifications were conducted in triplicate using HCV-specific primers and 6-carboxyfluorescein-and tetrachloro-6-carboxyfluoresceinlabeled probes (Applied Biosystems). The sequences for the primers and probes were designed using Primer Express software (Applied Biosystems) ( Table 1). Amplification reactions were performed in a 25-l mix using an RT-PCR core reagent kit and the template RNA. Reactions were performed in a 96-well spectrofluorometric thermal cycler under the following conditions: 2 min at 50°C, 30 min at 60°C, 10 min at 95°C, 44 cycles of 20 s at 95°C, and 1 min at 62°C. Fluorescence was monitored during every PCR cycle at the annealing step. At the termination of each PCR run, the data were analyzed by the automated system, and amplification plots were generated. To

Target genes Sense primers Antisense primers
determine the HCV RNA copy number, standards ranging from 10 1 to 10 8 copies/g were used for comparison.
Site-directed Mutagenesis-The base substitution mutations of AP-1-and Sp1-binding sites on the OPN promoter luciferase-reporter were carried out using oligonucleotide-mediated mutagenesis as described previously (26). The PCR reactions were performed with AP-1, Sp1, wild-type, and mutated primers (Table 1) according to the manufacturer's protocol (AccuPrime Pfx, Invitrogen). Briefly, the reaction buffer (5 l), consisting of 10ϫ reaction mix, 1 l (50 ng) of dsDNA template, 1.5 l (10 M) of each oligonucleotide primer, 1 l of Pfx DNA polymerase (2.5 units/l), and 40 l of double distilled H 2 O, was added to a final volume of 50 l. PCR program was as follows: initial denaturation at 95°C for 2 min; denaturation at 95°C for 15 s; annealing at 56°C for 30 s, and extension at 68°C for 5 min for 35 cycles. DpnI digestion was performed, and samples were transformed using 1 l of DpnI-treated DNA from control and sample reaction into separate 25-l aliquots of DH5␣-competent cells (Invitrogen). Clones were tested by restriction digestion, and positive clones were confirmed by DNA sequencing.
Chromatin Immunoprecipitation (ChIP) Assay-The ChIP assay was performed using Simple Chip Enzymatic Chromatin IP kit (Cell Signaling, Danvers, MA) as described previously (26). Briefly, mock-and HCV-infected cells (5 ϫ 10 7 cells) were fixed in 1% formaldehyde for 10 min to cross-link the DNA and proteins. The reaction was quenched using 125 nM glycine for 5 min, and cells were washed twice with ice-cold PBS and further suspended in ice-cold buffer A containing DTT, PMSF, and protease inhibitor mixture. The nuclei were pelleted by centrifugation at 3000 rpm for 5 min at 4°C and suspended in ice-cold buffer B ϩ DTT. The lysate was incubated with 5 l of micrococcal nuclease for 20 min at 37°C to digest DNA to a length of ϳ150 -900 bp and suspended in 1ϫ ChIP buffer followed by sonication for 30 s using Qsonica Q700. The sonicated lysates were centrifuged at 10,000 rpm for 10 min at 4°C to remove cell debris. To amplify the AP-1 site, supernatants (cross-linked chromatin preparation) were incubated with anti-c-Jun (1:50) and anti-c-Fos (1:50), and for the Sp1 site, anti-Sp1 (1:100) antibody or a normal rabbit IgG followed by an isolation procedure using protein-G magnetic beads was used. The DNA-protein interactions were reversed by heating to 65°C for 12 h. The AP-1-and Sp1-binding sites on the immunoprecipitated DNA were analyzed by quantitative RT-PCR using primers against AP-1 (c-Jun and c-Fos) and Sp1 ( Table 1). The PCR products were further visualized onto 1% agarose gel stained with 0.5 g/ml ethidium bromide.
Luciferase Assay-Mock-and HCV-infected cells were transfected with OPN promoter-luciferase reporter plasmids. At 48 h post-transfection, cells were serum-starved for 4 h, and cellular lysates were analyzed for luciferase activity using the Dual-Luciferase reporter assay kit (Promega, Madison, WI). All transfections included a Renilla expression vector as an internal control.
RNA Interference-Mock-and HCV-infected cells at day 4 were transfected with GFP siRNA (siGFP), siOPN, siCD44, and si␤3 according to the manufacturer's protocols (Santa Cruz Biotechnology). Each siRNA consists of pools of three to five target-specific 19 -25-nucleotide siRNA designed to knock down the target gene expression. For siGFP, siOPN, siCD44, and si␤3 transfections, two solutions were prepared. Solution A, containing 60 pmol of siRNA duplex, was mixed with 100 l of siRNA transfection medium. Solution B, containing 6 l of transfection reagent, was added to 100 l of siRNA transfection medium. Solutions A and B were allowed to incubate at RT for 20 min. After 20 min, solutions A and B were combined and incubated again for 20 min at RT. The combined solutions were then added to the cells and incubated for 5-7 h at 37°C and 5% CO 2 , and the transfection solution was replaced with 2 ml of complete DMEM growth media.
Cell Migration Assay-Mock-and HCV-infected cells were plated in 6-well culture plates, incubated overnight, and then followed by scratching using a sterile pipette tip across the monolayer. Cell debris was removed by washing with PBS, and the cells were cultured in serum-free media. Images were captured after wounding at 0 and 48 h post-wounding. The percent migration distance was calculated according to the following formula: percent migration distance ϭ percent wound width at time 0 Ϫ percent wound width at 48 h.
Statistical Analysis-Error bars show the standard deviations of the means of data from three individual trials. Two-tailed unpaired t tests were used to compare experimental conditions to those of the respective controls. The significance level was set at p value of 0.05.

HCV Induces OPN Expression and Secretion-To determine
whether HCV infection induces OPN protein expression, cellular lysates from mock-infected (Huh7.5) and HCV-infected Huh7.5 cells were subjected to Western blot analysis using anti-OPN antibody. We observed increased expression of OPN in HCV-infected Huh7.5 cells compared with mock-infected cells (Fig. 1A). To investigate if HCV infection induces OPN secretion, cell culture supernatants were collected from mock-and HCV-infected cells and subjected to Western blot analysis using anti-OPN antibody. The results show increased OPN secretion by HCV-infected cells compared with mock cells (Fig. 1B).
To demonstrate if the induction of OPN was due to increased expression of OPN mRNA, total cellular RNA was extracted from mock-and HCV-infected cells, and OPN mRNA expression was quantified by RT-PCR. The results show significantly higher OPN mRNA expression (ϳ6-fold) in HCV-infected cells compared with mock cells (Fig. 1C). To determine the level of HCV infection, total cellular RNA was extracted from the above cells and subjected to quantitative RT-PCR. The results showed a significant increase in HCV RNA copy number in HCV-infected cells (Fig. 1D). Taken together, these results suggest that HCV infection induces OPN expression as well as secretion.
HCV Activates OPN via Ca 2ϩ Signaling and Elevation of ROS-Our previous studies have shown that HCV-induced Ca 2ϩ signaling and elevation of ROS play a key role in the cell signaling cascade (12,23,25,26). To determine whether HCV-induced Ca 2ϩ signaling and induction of ROS activate OPN, mock-and HCV-infected cells were incubated with nontoxic doses of antioxidant (NAC), an inhibitor of mitochondrial Ca 2ϩ uptake (Ru360), intracellular Ca 2ϩ chelator (BAPTA-AM), and extracellular Ca 2ϩ chelator (EGTA). Cellular lysates as well as cell culture supernatants were immunoblotted with anti-OPN antibody. The results show an increased OPN expression and secretion in HCV-infected cells, which was reduced in the presence of NAC, Ru360, and BAPTA-AM (Fig. 2, A and B) but not with EGTA, which serves as negative control.
To demonstrate if HCV-induced Ca 2ϩ signaling and induction of ROS also activate endogenous OPN mRNA, mock-and HCV-infected cells were treated with the above inhibitors, and OPN mRNA was analyzed by quantitative RT-PCR. We observed increased OPN mRNA expression (ϳ3-fold) in HCVinfected cells compared with mock cells, which was reduced in the presence of NAC, Ru360, and BAPTA-AM but not with EGTA (Fig. 2C).
To evaluate if HCV-induced Ca 2ϩ signaling and induction of ROS stimulate OPN promoter induction, mock-and HCV-infected cells were transfected with the OPN promoter (Ϫ500/ 20)-luciferase reporter followed by treatment with the above inhibitors and subjected to Dual-Luciferase assay. The results show increased OPN promoter (Ϫ500/20) luciferase activity (ϳ2.7-fold) in HCV-infected cells compared with mock cells, which was reduced in the presence of NAC, Ru360, and BAPTA-AM (Fig. 2D). In contrast, we did not observe any effect of EGTA (extracellular Ca 2ϩ chelator) (Fig. 2D). No significant cytotoxicity was observed in mock-and HCV-infected cells treated with these inhibitors as shown by us earlier (12). Taken together, these results suggest that HCV-mediated altered Ca 2ϩ signaling in the ER and elevation of ROS production in the mitochondria are critical for OPN induction as well as secretion.
In addition, our ongoing studies suggest that HCV core, E1/E2, and NS3/4A induced OPN activation in hepatoma cells (data not shown). These proteins are also known to induce ROS in human hepatoma cells (31)(32)(33).
HCV Infection Induces OPN via Activation of Cellular Kinases-Previously, we have shown that HCV-induced Ca 2ϩ signaling and elevation of ROS activate several cellular kinases (23,25). To investigate the role of HCV-activated cellular kinases on OPN induction and secretion, mock-and HCV-infected cells were incubated with inhibitors of p38 MAPK (SB203580), JNK (SP600125), PI3-kinase (LY294002), MEK1/2 (U0126), and PKC (Go6976). Cellular lysates and cell culture supernatants were subjected to Western blot analysis using anti-OPN antibody. Increased expression and secretion of OPN were observed in HCV-infected cells, which were abrogated in the presence of above kinase inhibitors; however, no inhibition was observed with PKC inhibitor (Go6976), which serves as a control (Fig. 3, A and B).
To determine the role of HCV-activated cellular kinases on endogenous OPN mRNA expression, total cellular RNA was extracted from mock-and HCV-infected cells treated with the above inhibitors, and the OPN mRNA level was analyzed by SYBR Green RT-PCR. We observed increased OPN mRNA expression (ϳ3-fold) in HCV-infected cells compared with mock cells, which was reduced in the presence of above kinase inhibitors except PKC inhibitor (Go6976) (Fig. 3C).
To further strengthen the role of HCV-induced cellular kinases on OPN promoter activation, mock-and HCV-infected cells were transfected with the wild-type OPN promoter (Ϫ500/20)-luciferase reporter, followed by treatment with the above kinase inhibitors, and subjected to Dual-Luciferase assay. The results show increased OPN promoter luciferase activity (ϳ3-fold) in HCV-infected cells compared with mock cells, which was abrogated in the presence of the above kinase inhibitors (Fig. 3D). In contrast, we did not observe any effect by PKC inhibitor (Go6976) (Fig. 3D). Collectively, these results demonstrate that HCV infection stimulates the induction of OPN via activation of cellular kinases. No significant cytotoxicity was observed in mock-and HCV-infected cells treated with these inhibitors as shown by us recently (26).
To further demonstrate the role of AP-1 and Sp1 on OPN promoter activation, mock-and HCV-infected cells were transfected with wild-type OPN promoter-luciferase reporter (Ϫ500/20) and treated with nontoxic doses of the inhibitors of AP-1 (tanshinone IIA) and Sp1 (mithramycin A), and Dual-Luciferase assay was performed. The results showed ϳ3-fold increased luciferase activity by HCV that was reduced in the presence of the above inhibitors, suggesting that AP-1 and Sp1 are critical for OPN promoter activation (Fig. 4C). However, Stat3 inhibitor V (Stattic), used as negative control, did not show any effect on OPN promoter activation.
To confirm the role of AP-1-and Sp1-binding sequences in OPN promoter activity, wild-type OPN promoter-luciferase reporter and those with mutated AP-1-and Sp1-binding sites were transfected into mock-and HCV-infected cells and subjected to Dual-Luciferase assay. The results showed ϳ2.6and ϳ3-fold increased luciferase activity of wild-type AP-1 and Sp1 OPN promoter in HCV-infected cells compared with mock cells (Fig. 4D). In contrast, the OPN promoter containing mutated AP-1-or Sp1-binding sites showed reduced luciferase activity in HCV-infected cells (Fig. 4D). These results confirm the role of HCV-induced AP-1 and Sp1 in OPN promoter activation. We observed no significant cytotoxicity using tanshinone IIA and mithramycin A inhibitors as shown by us recently (26). The mutated bases of AP-1-and Sp1-binding sites on OPN promoter are shown in red in Fig. 4E.
In Vivo Interaction of HCV-induced AP-1 and Sp1 on OPN Promoter-To demonstrate if AP-1 and Sp1 interact with the OPN promoter in vivo in HCV-infected cells, chromatin immunoprecipitation (ChIP) assay was performed using c-Jun, c-Fos, and Sp1 antibodies as per the manufacturer's instruction (SimpleChIP Enzymatic Chromatin IP kit, Cell Signaling Tech-nology). The DNA quantitative RT-PCR analysis showed that c-Jun-, c-Fos-, and Sp1-specific antibodies immunoprecipitated chromatin from HCV-infected cells (Fig. 5A). However, immunoprecipitation with nonspecific antibody (normal rabbit IgG) did not amplify the DNA fragments. The PCR amplification of input chromatin before immunoprecipitation was served as a positive control. The amplified DNA fragments were further confirmed by agarose gel electrophoresis (Fig. 5B). These results indicate that AP-1 and Sp1 form a protein-DNA transcriptional regulatory complex by binding to the OPN promoter in HCV-infected cells.
HCV Induces Proteolytic Processing of OPN-To determine whether HCV infection induces proteolytic cleavage of precursor OPN, cellular lysates from mock-and HCV-infected cells at day 8 post-infection were subjected to Western blot analysis using anti-OPN antibody. We observed increased expression of the precursor form of OPN (ϳ75 kDa) followed by its cleavage into various forms of OPN (ϳ55, ϳ42, and ϳ36 kDa) in HCVinfected cells compared with mock-infected cells (Fig. 6A, lane  2). Furthermore, cellular RNA from above mock-and HCVinfected cells were amplified using OPN-specific primers by semi-quantitative RT-PCR. We observed a single OPN band that corresponds to full-length OPN mRNA, but no splice variants were observed (Fig. 6B). To determine whether the activation of OPN was specific to HCV replication, Huh7.5 cells were electroporated with J6/JFH-1(wild-type) and JFH-1/GND (replication defective mutant) RNA. We observed the expression of OPN precursor and cleaved forms in J6/JFH-1RNA-electroporated Huh7.5 cells. In contrast, we did not observe any expression of OPN in JFH-1/GND RNA-electroporated cells (Fig. 6C).
To further demonstrate whether proteolytic processing of OPN is specific to HCV, mock-and HCV-infected Huh7.5 cells were transiently transfected with full-length FLAG-OPN expression vector. The immunoblot results show the expression of precursor (full-length) OPN (ϳ75 kDa) and one cleaved form of OPN (ϳ36 kDa) in HCV-infected cells using anti-FLAG antibody (Fig. 6D, lane 4), but we did not observe the cleavage of FLAG-OPN in Huh7.5 cells (Fig. 6D, lane 2). However, we observed precursor and three cleaved products (55,42, and 36 kDa) of OPN in the above lysates (Fig. 6D,  lanes 3 and 4) when incubated using anti-OPN antibody (Fig.  6E, lane 2). These results suggest that the cleavage site of ϳ36-kDa OPN might be close to FLAG sequences at the C terminus, and anti-FLAG was able to detect only precursor and 36-kDa cleaved product. However, anti-OPN antibody was able to detect both exogenous and endogenous OPN (Fig. 6E)

Mechanism of OPN Activation by HCV
role in the proteolysis of OPN, Huh7.5 cells were treated with H 2 O 2 (500 M) followed by Western blot analysis using anti-OPN antibody. The results show the increased expression of precursor OPN (ϳ75 kDa) in H 2 O 2 -treated mock cells compared with untreated cells (Fig. 6F, lane 2). However, we did not observe any cleavage of OPN in Huh7.5 cells treated with H 2 O 2 (Fig. 6F, lane 2). Collectively, these results suggest that the OPN processing occurs in HCV-infected cells.
To verify the OPN cleavage in primary human hepatocytes (HPP), cellular lysates from PHH-and HCV-infected PHH were subjected to Western blot analysis using anti-OPN antibody. Similar cleaved products of OPN (ϳ55, ϳ42, and ϳ36 kDa) were observed (Fig. 6G, lane 2) as described above. To further strengthen our results, equal amounts of cellular lysates from normal and HCV-infected human hepatocyte-engrafted MUP-uPA/SCID/bg mice were subjected to Western blot analysis using anti-human OPN. We observed increased expression of precursor OPN (ϳ75 kDa) in HCV-infected samples followed by its cleavage into 55 and 42 kDa (Fig. 6H, lane 2). However, we could not detect the 36-kDa cleaved form of OPN in HCV-infected human hepatocyte-engrafted MUP-uPA/SCID/bg mouse liver tissue samples (Fig. 6H, lane 2). These results suggest that at later time points HCV has the ability to cleave OPN into various forms.
Previously, full-length OPN has been shown to be cleaved into various forms by thrombin and matrix metalloproteases (15,36,37). To investigate if protease plays an important role in HCV-induced OPN cleavage, mock-and HCV-infected cells were treated with nontoxic doses of various inhibitors against calpain II inhibitor (ALLM), calpain I inhibitor (ALLN), caspase-3 inhibitor I, and proteosome inhibitor (MG-132), and cellular lysates were immunoblotted using anti-OPN antibody. The results show that the cleavage of OPN in HCV-infected cells was significantly blocked by ALLM and ALLN, leading to stabilization of the precursor form of OPN (Fig. 6I, lanes 3 and  4). In contrast, we did not observe any effect of caspase-3 and MG-132 inhibitors in the abrogation of OPN cleavage (Fig. 6I,  lanes 5 and 6). In addition, we also used MMP-2/MMP-9 inhibitor V (Calbiochem) and did not observe the stabilization of precursor OPN (data not shown). Moreover, no cytotoxicity was observed in mock-and HCV-infected cells treated with these inhibitors as shown by us (38). Collectively, these results indicate the involvement of calpain proteases in the post-translational processing of HCV-induced OPN.
HCV Induces EMT-To determine whether HCV induces EMT, cellular lysates from mock-(Huh7.5 cells and PHH) and HCV-infected cells were subjected to Western blot analysis using antibodies against E-cadherin (epithelial marker) and N-cadherin (mesenchymal marker). The results show decreased expression of E-cadherin and increased expression of N-cadherin in HCVinfected Huh7.5 cells as well as PHH compared with mock cells (Fig. 7, A and B). FIGURE 5. HCV induces in vivo interaction of AP-1 and Sp1 on OPN promoter. A, cross-linked chromatin preparations from mock-(Huh7.5) and HCV-infected cells were immunoprecipitated using anti-c-Jun, c-Fos, and Sp1 antibody or normal rabbit IgG. The AP-1-and Sp1-binding sites on the immunoprecipitated DNA were determined by quantitative RT-PCR using the primers as per the manufacturer's instruction. Amplification of input chromatin prior to immunoprecipitation served as positive controls for chromatin extraction and PCR amplification. Chromatin immunoprecipitation using a nonspecific antibody (normal rabbit IgG) served as negative controls. The data represents the means Ϯ S.D. of three independent experiments performed in triplicate. * denotes p Ͻ 0.05 compared with Huh7.5 cells. B, equal amounts of the PCR products were visualized onto 1% agarose gel.
To demonstrate if HCV-activated OPN induces EMT, HCVinfected cells were transfected with siGFP (control siRNA), siOPN, siCD44, and si␤3. Cellular lysates were subjected to Western blot analysis using anti-OPN, anti-CD44, and anti-␤3 antibody. The results show significant knockdown of OPN, CD44, and ␤3 expression in HCV-infected cells transfected with the above siRNA (Fig. 7C). The above cellular lysates were immunoblotted with anti-E-cadherin and anti-N-cadherin. The results show decreased expression of E-cadherin in HCVinfected cells compared with mock-infected cells (Fig. 7D, lane  2), which reappeared in HCV-infected cells transfected with siOPN, siCD44, and si␤3 (lanes 4 -6) but not with siGFP (lane  1 and 2). B, HCV induces OPN mRNA expression. Total cellular RNA was extracted using TRIzol (Invitrogen) from the above cells (A) followed by cDNA synthesis. OPN mRNA was amplified using OPN primers by semi-quantitative RT-PCR. Equal amounts of amplified products were loaded onto 1% agarose gel. OPN gene expression was compared by 18 S rRNA. C, J6/JFH-1 activates OPN. The plasmid pFL-J6/JFH1 encoding the HCV J6/JFH-1 genome was linearized with XbaI for in vitro transcription using the AmpliScribe T 7 transcription kit (Epicenter Technologies). Huh7.5 cells were electroporated with J6/JFH-1 (wild type) and JFH-1/GND (replication defective mutant) RNA (10 -12 g)/60-mm plate. At day 8 post-infection, equal amounts of cellular lysates were subjected to Western blot analysis using anti-OPN antibody. Tubulin was used as a protein loading control. NS3 represents the level of HCV replication. D, mock-and HCV-infected cells were transfected with FLAG-tagged OPN-a plasmid (pDest490-OPN-a) using Lipofectamine 2000 as per the manufacturer's instruction. At 48 h posttransfection, cellular lysates were subjected to Western blot analysis using anti-FLAG antibody. E, equal amounts of cellular lysates from FLAG-OPN-transfected cells (B) were immunoblotted with anti-OPN antibody. F, mock cells were treated with H 2 O 2 (500 M) for 20 min, and cellular lysates were immunoblotted using anti-OPN antibody. G, equal amounts of cellular lysates from PHH and HCV-infected PHH were immunoblotted using anti-OPN antibody. H, equal amounts of cellular lysates from normal and HCV-infected human hepatocyte-engrafted MUP-uPA/SCID/bg mice were subjected to Western blot analysis using antihuman OPN. I, HCV-infected Huh7.5 cells were treated with various inhibitors, ALLM (100 M), ALLN (100 M), and MG-132 (20 M), for 6 h and caspase-3 (50 M) for 1 h in serum-free media. Cellular lysates were subjected to Western blot analysis using anti-OPN antibody. Immunoblot with anti-actin and anti-tubulin was used as protein loading control. HCV NS3 represents HCV infection. DECEMBER 27, 2013 • VOLUME 288 • NUMBER 52

JOURNAL OF BIOLOGICAL CHEMISTRY 37003
3). However, the expression of N-cadherin was increased in HCV-infected cells compared with mock-infected cells (Fig.  7D, lane 2), which was abrogated in cells transfected with siOPN, siCD44, and si␤3 (lanes 4 -6) but not with siGFP (lane 3). Collectively, these results suggest that HCV has the ability to induce EMT via OPN, including cell surface receptors CD44 and ␣V␤3.
To determine the level of HCV infection, HCV-infected PHH and Huh7.5 cells were immunostained using anti-HCV NS5A antibody. The results showed about Ͼ95% of the cells were infected with HCV as observed by immunofluorescence microscopy (Fig. 7, E and F).
OPN Induces Hepatoma Cells Migration-To determine whether HCV-induced OPN plays a role in human hepatoma cell migration, HCV-infected cells (from the same pool of Fig.  7F) were transfected with siGFP, siOPN, siCD44, and si␤3 and were subjected to wound healing assay. The results show increased cell migration in HCV-infected cells at 48 h compared with mock cells, which was reduced in HCV-infected cells transfected with siOPN, siCD44, and si␤3 but not with siGFP (control siRNA) (Fig. 8A). The increased migration depth at 48 h was recorded as ϳ85% in HCV-infected cells that was reduced to ϳ39, 67, and 54% in HCV-infected cells transfected with siOPN, siCD44, and si␤3, respectively (Fig. 8B), suggesting the critical role of HCV-induced OPN in wound healing.
OPN-induced ␤-Catenin-mediated Signaling Cascade in HCV-infected Cells-␤-Catenin is known to be a downstream effector of Wnt signaling cascades and plays a critical role in EMT progression (28). Similarly, OPN and cell surface receptors, ␣V␤3 and CD44, have also been known to be critical for EMT in various cancers cells (15,16,38). To determine whether HCV-activated OPN induces EMT via Akt, GSK-3␤, and ␤-catenin phosphorylation, the cellular lysates from siRNA-transfected cells (Fig. 7C) were immunoblotted using anti-Akt (Ser 473 ), GSK-3␤ (Ser 9 ), and anti-␤-catenin (Ser 33/37 / Thr 41 ). The results show increased phosphorylation of Akt, GSK-3␤, and ␤-catenin in HCV-infected cells compared with mock cells, which was reduced in HCV-infected cells transfected with siOPN, siCD44, and si␤3 but not with siGFP FIGURE 7. HCV induces EMT. A and B, mock-(Huh7.5 and PHH) and HCV-infected Huh7.5 cells and PHH were harvested, and equal amounts of cellular lysates were immunoblotted using anti-E-cadherin and anti-N-cadherin. C, knockdown of OPN, CD44, and ␤3 gene expression. HCV-infected cells were transfected with siGFP, siOPN, siCD44, and si␤3 as described under "Experimental Procedures." At 72 h post-transfection, equal amounts of cellular lysates were immunoblotted using anti-OPN, anti-CD44, and anti-␤3. D, equal amounts of cellular lysates from above transfected cells (C) were immunoblotted using anti-E-cadherin and anti-N-cadherin. Anti-tubulin was used a protein loading control. E and F, HCV-infected PHH and Huh7.5 cells at day 4 post-infection were immunostained using HCV NS5A antibody (magenta). DAPI was used as a nuclear stain in blue. Scale bar, 10 M. (Fig. 9). In contrast, we did not observe any significant change in the total expression of GSK-3␤ and ␤-catenin (Fig.  9). Taken together, these results suggest that HCV-induced OPN is critical for Akt-, GSK-3␤-, and ␤-catenin-mediated EMT.

DISCUSSION
HCC represents an extremely poor prognostic cancer that remains one of the most common and aggressive human malignancies worldwide (5,6). Current therapies are inefficient, mainly due to usually late diagnosis and high recurrence rate within the remaining cirrhotic liver after surgical resection. It is a multistep process involving different genetic alterations that ultimately lead to malignant transformation of hepatocytes (39). HBV and HCV chronic infections account for 75% of HCCs, whereas nonviral etiologies such as alcohol and genetic or metabolic disorders represent less than 25% of cases. HCVinduced chronic liver injury can lead to progressive fibrosis, cirrhosis, and eventually HCC (40). HCV-associated HCC has been reported to promote tumor growth and metastasis due to an increased recurrence after liver resection (8). In addition, expression of HCV structural proteins in transgenic mouse liver may contribute to the development of hepatic steatosis, fibrosis, and HCC (41,42). However, the underlying mechanisms responsible for the metastatic spread of HCV-induced HCC are not fully elucidated.
Previous studies have demonstrated that OPN is involved in tumor metastasis and has been detected not only in numerous cancers but also in the plasma of HCV-related HCC patients, suggesting a correlation between high levels of OPN expression and malignant invasion (16,43). Recently, ␣-fetoprotein has been shown to be a valid serological marker for HCV-related HCC, but ␣-fetoprotein alone is not sufficient unless ultrasound is available (43,44). OPN is mainly expressed in transformed malignant epithelial cells and has been suggested as an additional biomarker for HCC screening due to inadequacy of ␣-fetoprotein (43,44). However, the role of OPN in HCV-induced HCC remains unclear.
In this study, we determined the molecular mechanism of OPN induction and secretion by HCV infection. Our results  DECEMBER 27, 2013 • VOLUME 288 • NUMBER 52

Mechanism of OPN Activation by HCV
suggest that HCV infection induces OPN via altered Ca 2ϩ homeostasis in the ER and elevation of ROS in the mitochondria (Fig. 2). Previously, Ca 2ϩ -mediated mitochondrial dysfunction has been suggested to play an important role in HCVmediated liver disease pathogenesis (45). Interestingly, the HCV core has also been shown to target mitochondria and to increase Ca 2ϩ -dependent ROS production in mouse liver (46). Moreover, an increased efflux of Ca 2ϩ from the ER can leads to migration and metastatic signaling cascade of several cancer cells through the Ca 2ϩ -dependent activation of S100A4, a member of the S100 family EF-hand calcium-binding protein (47). In cancer cells, mild oxidative stress activates the cell signaling cascade such as proliferation, migration, and invasion, but high oxidative stress can induce cell death (48). Our results are consistent with the previous studies that ROS up-regulated OPN and antioxidants prevented this effect in the mouse model (49).
In our findings, we observed that the induction of OPN in HCV-infected cells is mediated by the activation of cellular kinases such as p38 MAPK, JNK, PI3K, and MEK1/2 (Fig. 3). Interestingly, several kinases such as MAPK (JNK and p38), ERK, and PI3K are known to be involved in transcriptional regulation of metastasis-related genes leading to EMT, cell migration, and tumor invasion (50). It is well established that HCV activates several oncogenic transcription factors via phosphorylation by these cellular kinases (23)(24)(25). The OPN promoter is known to be regulated by the binding of several transcription factors (16,29,34). Our data suggest that AP-1 and Sp1 are responsible for OPN gene expression, which is consistent with transactivation of the OPN promoter by the human T-cell leukemia virus type 1-encoded Tax protein via AP-1 (51). In addition, we also observed ϳ3-fold luciferase activity in the wild-type OPN promoter and ϳ2-fold in the OPN deletion construct Ϫ267/20 (Fig. 4B), indicating the roles of the AP-2-, GATA-1-, and TCF-1-binding site on OPN promoter activation, which are our future plans of study.
Previously, it has been reported that full-length OPN is composed of ϳ314 amino acid residues, which may exist as functionally important cleaved products as well as the occasional splice variants (16,37,43). In HCV-infected hepatoma cells, we could observe only the expression of the full-length OPN mRNA (Fig. 6B), indicating that HCV infection does not induce the formation of OPN splice variants. In contrast, recent studies have shown the formation of two splice variants of OPN in HCC (43). This could be due to different cell types and the source of HCC tissue samples used in those studies. We observed the induction of full-length OPN polypeptide (ϳ75 kDa), which is cleaved into ϳ55, ϳ42, and ϳ36 kDa. The most intriguing finding of our studies is the involvement of calcium-activated calpain proteases in HCV-induced proteolytic processing of OPN. Calpain is activated by Ca 2ϩ binding to its catalytic domain, which stimulates its protease activity (52). In HCV-infected cells, calpain inhibitors (ALLM and ALLN) completely blocked the proteolytic processing of OPN (Fig. 6I), suggesting the role of Ca 2ϩ -mediated calpain proteases in the cleavage of OPN. This observation is supported by previous studies where enhanced calpain activity promoted cell motility and invasion in various cancers, which was confirmed by in vivo studies using siRNA against m-calpain (53,54).
OPN is known to bind with widely expressed receptors, ␣V␤3 and CD44, to induce the signaling cascade responsible for cell adhesion, migration, and tumor progression (15,16,43). Our data suggest that HCV infection can induce secretion of OPN. These findings are supported by the previous reports where interaction of secreted OPN to integrins and CD44 is more efficient than native OPN and are associated with cell migration, invasion, and metastasis of a number of malignant tumors, including HCC (15,16,36,37,43). Recently, the enhanced level of OPN was observed in the plasma and liver of transgenic mice expressing OPN in hepatocytes, suggesting that extrahepatic manifestation could occur in other organs due to Th1 immune reaction provoked by circulation of OPN (55).
␤-Catenin is a proto-oncogene, a key downstream target of GSK-3␤. Aberrant activation of Wnt/␤-catenin signaling results in enhanced cell growth and malignant cellular transformation. The hallmark of canonical Wnt signaling activation is the stabilization and nuclear translocation of ␤-catenin. In HCV-infected cells, we observed increased phosphorylation of GSK-3␤, which can lead to the activation of ␤-catenin followed by increased transcriptional activity critical for EMT progression as described in recent studies (56). Our results also suggest that the activation of ␤-catenin is mediated by HCV-induced OPN via ␣V␤3 and CD44 (Fig. 9). Our findings are consistent with previous reports where authors have concluded that GSK-3␤ and ␤-catenin activity can be modulated by virus-encoded proteins such as the latent membrane protein 2A of Epstein-Barr virus and hepatitis B virus X protein (57,58). Recently, the HCV core and NS5A have been shown to inactivate GSK3-␤ activity and to subsequently increase nuclear FIGURE 9. Role of HCV-induced OPN in cell signaling cascade. Mock-and HCV-infected cells were transfected with siGFP, siOPN, siCD44, and si␤3 as described under "Experimental Procedures." Equal amounts of cellular lysates were subjected to Western blot analysis using anti-p-Akt, anti-p-GSK-3␤, antip-␤-catenin, anti-GSK-3␤, and anti-␤-catenin antibodies. Actin serves as a protein loading control.
accumulation of ␤-catenin in human hepatoma cells (59,60). Furthermore, our findings suggest the critical involvement of HCV-induced OPN on the phosphorylation of Akt via ␣V␤3 and CD44 (Fig. 9). These results are consistent with the previous studies showing OPN-mediated phosphorylation of Akt via ␣V␤3 and CD44 (15,43). It has been shown that Akt-mediated phosphorylation of ␤-catenin leads to increased transcriptional activity in the nucleus (61). Previously, we and others have reported that HCV can stimulate the phosphorylation of Akt, a downstream target of OPN, followed by phosphorylation of GSK-3␤, and results in ␤-catenin-mediated EMT and cell migration (23,43,61,62).
The epithelial marker, E-cadherin, is primarily expressed in epithelia at the sites of cell-cell contacts, whereas in most cancers of epithelial origin, E-cadherin-mediated cell-cell adhesion is lost concomitantly with progression toward tumor malignancy (63). Loss of E-cadherin promotes the progression from a benign tumor to carcinoma. Our results demonstrate the loss of E-cadherin and gain of N-cadherin expression in HCV-infected cells (Fig. 7), suggesting that HCV infection has the ability to induce EMT of hepatoma cells, which are consistent with HCV infection and HCV glycoproteins (E1 and E2) involved in hepatoma cell migration through the activation of EMT markers Twist and Snail (56,64). In our observations, OPN binding with integrin ␣V␤3 and CD44 increased induction of N-cadherin leading to EMT and cell migration in HCV-infected hepatoma cells (Figs. 7D and 8) (15,43). Moreover, studies have also shown that HCV core protein expression induces EMT in the cholangiocarcinoma cell line, but in cell culture or in transgenic mice it led to the development of steatosis, a risk factor that contributes to HCC (65,66).
In summary, our studies demonstrate the mechanism of HCV-induced OPN activation, which involves altered Ca 2ϩ homeostasis, elevation of ROS, and activation of various cellular kinases, followed by activation of transcription factors AP-1 and Sp1. We also showed that OPN proteolytic processing in HCV-infected cells is dependent on calpain proteases. Furthermore, we investigated that HCV infection induces EMT and cell migration via OPN through ␣V␤3 and CD44 receptors. In addition, we also demonstrated that HCV infection induces OPN-dependent increased phosphorylation of Akt, GSK-3␤, and ␤-catenin, which can lead to tumor progression and EMT of human hepatoma cells (Fig. 10). Our results provide novel insight into the mechanisms of HCV-induced OPN activation leading to hepatoma cell migration and HCC.