Hepatitis B virus X protein is essential for the activation of Wnt/β-catenin signaling in hepatoma cells

Authors


Abstract

Wnt/β-catenin signaling contributes to diverse cellular functions, such as Drosophila wing development and colon carcinogenesis. Recently, stabilizing mutations of β-catenin, a hallmark of Wnt signaling, were documented in significant numbers of primary hepatocellular carcinomas (HCC). However, whether the β-catenin mutation leads to the activation of Wnt/β-catenin signaling in hepatoma cells has not been established. We found that Wnt/β-catenin signaling could be activated by ectopic expression of Wnt-1 in some hepatoma cells, such as Hep3B and PLC/PRF/5 cells, but not in others, such as Huh7 and Chang cells. Importantly, we noted that the former were derived from hepatitis B virus (HBV)-infected livers, whereas the latter were derived from HBV-negative livers. It was then speculated that HBx, a viral regulatory protein of HBV, is involved in activating Wnt/β-catenin signaling in hepatoma cells. In agreement with this notion, ectopic expression of HBx along with Wnt-1 activated Wnt/β-catenin signaling in Huh7 cells by stabilizing cytoplasmic β-catenin. Further, we showed that such stabilization of β-catenin by HBx was achieved by suppressing glycogen synthase kinase 3 activity via the activation of Src kinase. In conclusion, the data suggest that Wnt-1 is necessary but insufficient to activate Wnt/β-catenin signaling in hepatoma cells and the enhanced stabilization of β-catenin by HBx, in addition to Wnt-1, is essential for the activation of Wnt/β-catenin signaling in hepatoma cells. (HEPATOLOGY 2004;39:1683–1693.)

Hepatocellular carcinoma (HCC), the major primary malignant tumor in the liver, is one of the human cancers clearly linked to viral infection. Chronic infection with hepatitis B virus (HBV) has been identified as the main etiological agent for HCC.1 Although the viral and environmental risk factors for HCC development have been revealed, the oncogenic pathways leading to malignant transformation of liver cells remain elusive. Activation of c-myc or cyclin D1 by DNA amplification and inactivation of tumor suppressor genes, such as p53 or Rb, have been observed in some HCCs, but at relatively low frequencies.2 Recently, however, mutations of β-catenin have been documented in significant numbers of primary HCCs.3–8 These finding raised the possibility that the activation of Wnt/β-catenin signaling contributes to hepatocarcinogenesis.

The Wnt/β-catenin signaling pathway, initially discovered by genetic analysis in the wing development of Drosophila,9 has been implicated in colon carcinogenesis.10, 11 A hallmark of Wnt signaling is the stabilization of cytoplasmic β-catenin, followed by its nuclear translocation and association with LEF/TCF transcription factors,12 which lead to the transcription of Wnt target genes.13–15 β-Catenin, first identified on the basis of its association with cadherin adhesion molecules, now is widely recognized as a key molecule of the Wnt signaling cascade. In the absence of Wnt signals, β-catenin is phosphorylated at its N-terminal serine-threonine residues by functional interactions with glycogen synthase kinase (GSK) 3β, Axin, and adenomatous polyposis coli protein, and then is degraded through the ubiquitin-dependent pathway.16 In the presence of Wnt-1, GSK3β activity is suppressed, which in turn leads to β-catenin stabilization.17 Subsequent translocation of stabilized β-catenin to the nucleus and its association with the LEF/TCF family proteins lead to transcriptional activation of target genes.18

Abnormalities in the regulation of Wnt/β-catenin signaling have been implicated in colon cancer,19–21 and this pathway was shown to be frequently activated through β-catenin accumulation as a consequence of adenomatous polyposis coli protein inactivation or β-catenin stabilizing mutation in exon 3.20–22 Recent studies have shown that mutations of β-catenin, specifically stabilizing mutations in exon 3, were detected in approximately 30% of primary HCCs.3–7 Given the high prevalence of the stabilizing mutations of β-catenin, it was speculated that Wnt/β-catenin signaling contributes to the development of HCC. To explore this notion, we examined six hepatoma cell lines for the activation of Wnt signaling by (1) measuring TCF-4-dependent transcription and (2) immunostaining β-catenin after ectopic expression of Wnt signaling molecules. Here we found that (1) Wnt-1 is necessary but insufficient for the activation of Wnt/β-catenin signaling in hepatoma cells (this finding is contrary to the fact that Wnt-1 alone is sufficient for the activation of Wnt/β-catenin signaling in cells derived from nonliver tissues12, 17, 20, 21) and (2) in addition to Wnt-1, HBx, a regulatory protein of HBV, is essential for the activation of Wnt/β-catenin signaling in hepatoma cells. The data presented here may account for the contribution of a viral factor, HBV, that has been considered as the major etiological agent for HCC development.1

Abbreviations

HCC, hepatocellular carcinoma; HBV, hepatitis B virus; GSK, glycogen synthase kinase; HEK, human embryonic kidney cell line; MBP, myelin basic protein.

Materials and Methods

Cells and Transfections.

The following six human liver cell lines were used: HepG2 (American Type Culture Collection), Hep3B, Huh7 (gifts of B. Mason, Fox Chase Cancer Center), PLC/PRF/5,23 SK-Hep1, and Chang liver (American Type Culture Collection). All cells, including the human embryonic kidney cell line (HEK) 293 cells, were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and 10 μg/mL of gentamicin (Gibco/BRL). For transfections, the calcium phosphate precipitation procedure was performed as previously described24 for most experiments except one for Fig. 6. For an experiment for Fig. 6, transfections were performed by using LipofectAMINE (Life Technologies, Inc.) according to instructions from the manufacturer.

Figure 6.

Suppression of GSK3β activity by HBx-activated Src kinase. (A) Suppression of GSK3β activity by HBx plus Wnt-1. Huh7 cells were transfected with HBx and Wnt-1 expression plasmids by using LipofectAMINE, as indicated. Cells were harvested 2 days after transfection. The GSK3β kinase assays were carried out essentially as described previously,26 except myelin basic protein (MBP) was used as a substrate. After kinase reactions, immunocomplexes were separated by sodium dodecyl sulfate-polyacrylamide electrophoresis, and the upper part of gel was used to measure protein level (denoted by GSK3β) and the lower part of gel was used to measure kinase activity (denoted by MBP). (B) The data were quantified by BAS scintillation counting and normalized to the amount of GSK3β protein. The data were obtained from three independent experiments. (C) HBx activated Src kinase suppressed GSK3β activity. All experimental procedures are described in Fig. 6A. (D) Quantification was as described in Fig. 6B.

Plasmids.

The reporter plasmids (pTOPFLASH and its mutant version pFOPFLASH) and TCF-4 expression plasmids were kindly provided by B. Vogelstein and K. Kinzler (Johns Hopkins University, Baltimore, MD).20 Wnt-1, β-catenin, and β-catenin-DP, a constitutively active mutant of β-catenin, expression plasmids were kind gifts of R. Grosschedl (University of California, San Francisco).18 The HBx expression plasmid (i.e., pCMV-HBx) with C-terminal FALG tag was generated by site-directed mutagenesis. The HBV 1.2-mer replicon, which is capable of supporting viral replication, was made essentially as previously described.25 The HBx-null (HBx−) mutant of the 1.2-mer was constructed by inserting a stop codon after the first and second AUG, respectively, of the HBx open reading frame by site-directed mutagenesis.

Reporter Assays.

Reporter assays were performed as previously described.20 Cells were transfected with indicated plasmids by the calcium phosphate method. Two days after transfection, cells were harvested and assayed by using the Dual-Luciferase Reporter Assay System (Promega).

Immunoprecipitation and Kinase Assays.

The GSK3β kinase assays were carried out essentially as described previously,26 except that myelin basic protein (MBP) was used as a substrate. For kinase assays, cells were lysed in a buffer containing 10 mM Tris (pH 7.4), 50 mM NaCl, 3 mM ethylenediaminetetraacetic acid, 6 mM EGTA, 1% Triton X-100, 1 mM benzamidine, 20 mM sodium pyrophosphate, 2 mM sodium vanadate, 10 mM KH2PO4, 10 mM MgCl2, 50 mM β-glycerophosphate (pH 7.3), 0.2 mM polymethylsulfonyl fluoride (PMSF). Cell lysates were microcentrifuged for 15 minutes at 4°C. The GSK3β monoclonal antibody (Transduction Laboratory) was added and incubated for 1 hour at 4°C. Then, protein-G Sepharose (Sigma, St. Louis, MO) was added, and the mixture was incubated at 4°C for an additional 1 hour. These mixtures were washed with lysis buffer four times and once with kinase buffer (20 mM β-glycerophosphate, pH 7.3, 5 mM MgCl2, 1 mM EGTA, 10% glycerol, 1 mM sodium vanadate, 0.2 mM PMSF). Immunocomplexes were incubated for 20 minutes at 37°C in 30 μL kinase buffer containing 2 μg MBP, 20 μCi [γ-32P] adenosine triphosphate, and 250 μM unlabeled adenosine triphosphate. Reactions were stopped by boiling with sodium dodecyl sulfate sample buffer, and proteins were separated by 12% sodium dodecyl sulfate-polyacrylamide electrophoresis.

Results

Wnt-1 Is Necessary But Insufficient for Activating Wnt/β-Catenin Signaling in Hepatoma Cells.

Although activation mutations of β-catenin have been described in hepatoma cells and primary HCCs, earlier studies have not established whether Wnt/β-catenin signaling in these hepatoma cells is functionally activated by β-catenin mutations.3–6 To address this question, we examined TCF-4-dependent transcriptional activation by luciferase reporter gene assay. The reporter plasmid contained multimerized TCF-4 binding sites (pTOPFLASH and its mutant version pFOPFLASH) linked to the luciferase gene.20 Six hepatoma cell lines were transfected by various combinations of plasmids that express Wnt-1, TCF-4, and β-catenin. Unexpectedly, we found that some hepatoma cells were refractory to the activation of Wnt/β-catenin signaling on the addition of Wnt-1.

First, in Huh 7 and SK-Hep1 cells, the activity of the reporter gene was increased significantly only when cells were cotransfected by TCF-4 and β-catenin expression plasmids, which served as the positive control for the activation of Wnt signaling (Fig. 1A,B). Ectopic expression of Wnt-1 alone failed to activate TCF-4-dependent transcription. Possibly, the amount of endogenous β-catenin or the TCF-4 protein level was not sufficient to respond to the transient reporter assay. To examine this possibility, we cotransfected either the β-catenin or TCF-4 expression plasmid along with the Wnt-1 expression plasmid. Nonetheless, transfection of Wnt-1 plus either TCF-4 or β-catenin did not activate TCF-4-dependent transcription (Fig. 1A,B). We interpreted this data to indicate that ectopic expression of Wnt-1 failed to stabilize β-catenin in Huh7 and SK-Hep1 cells. An essentially identical result was obtained with Chang liver cells (data not shown).

Figure 1.

TCF-4 dependent transcriptional activation in response to Wnt signaling molecules in hepatoma cell lines. Hepatoma cells (A–D) were transfected with 1 μg of each indicated expression plasmid, along with 1 μg of pTOPFLASH or pFOPFLASH reporter plasmids. The amount of DNA in each transfection was kept constant by adding an appropriate amount of the pcDNA1/Amp vector. pTOPFLASH (closed bar) or pFOPFLASH (open bar) reporter genes contained either multimerized wild-type or mutant TCF-4 binding sites positioned 5′ of the luciferase reporter gene, respectively.20, 21 Fold activation was quantified relative to the level of luciferase activity from cells transfected with pTOPFLASH only. Data represent the mean of three or more independent transfections. Error bars represent standard errors of the mean.

Second, in Hep3B and PLC/PRF/5 cells, significant transcriptional activation was not observed on transfection of Wnt-1, β-catenin or TCF-4 alone or cotransfection of Wnt-1 plus TCF-4 (Fig. 1C,D). By contrast, transfection of Wnt-1 plus β-catenin expression plasmids led to significant induction of TCF-4-dependent transcription (Fig. 1C,D). These data suggested that ectopic expression of Wnt-1 resulted in the accumulation of β-catenin in the cytoplasm and subsequent nuclear localization and induction of TCF-4-dependent transcription in Hep3B and PLC/PRF/5 cells.

Third, in HepG2 cells, apparently, Wnt signaling was already activated in parental cells (data not shown). This result is consistent with a previous report in which the stabilizing exon 3 mutation of β-catenin in HepG2 cells was described.3 Collectively, disparate responses among hepatoma cells to Wnt signaling molecules allowed us to divide these hepatoma cells into three groups: (1) Huh7, SK-Hep-1, and Chang cells; (2) Hep3B and PLC/PRF/5 cells; and (3) HepG2 cells.

To corroborate these findings, we next examined cellular localization of β-catenin in hepatoma cells by immunostaining. Consistent with a previous report,3 nuclear staining of β-catenin was observed in parental HepG2 cells (Fig. 2A), although the pattern of β-catenin staining in the nucleus appeared as dense speckles, possibly associated with the nucleoli in HepG2 cells (Fig. 2A). In contrast, in other hepatoma cells, β-catenin was localized predominantly to the plasma membrane (Fig. 2A). Not surprisingly, when TCF-4 was coexpressed, ectopically expressed β-catenin was localized to the nucleus of hepatoma cells, regardless of Wnt-1 expression (Fig. 2B), consistent with previous reports with other cells.18, 27 It should be noted, however, that the nuclear localization of overexpressed β-catenin induced by TCF-4 does not bear on the mechanism by which endogenous β-catenin translocates to the nucleus on Wnt signaling. In addition, when TCF-4 and β-catenin plasmids were cotransfected into HepG2 cells, rather dense but homogenous staining of β-catenin was detected, suggesting that ectopically expressed β-catenin had been translocated predominantly to the nucleus (Fig. 2B). However, in almost all hepatoma cells except HepG2 cells, when the β-catenin expression plasmid alone was transfected, no nuclear staining of β-catenin was observed (Fig. 2C). Likewise, when the Wnt-1 expression plasmid was transfected, no nuclear staining of β-catenin was observed in the hepatoma cells, suggesting that Wnt-1 alone is insufficient to stabilize the endogenous β-catenin in transient transfections (Fig. 2D). Significantly, when Wnt-1 and β-catenin were coexpressed, nuclear staining of β-catenin was detected in Hep3B and PLC/PRF/5 cells but not in Huh7 cells, Chang liver cells, or SK-Hep1 cells (Fig. 2E; data not shown). More importantly, nuclear staining of β-catenin was detected in hepatoma cells in which Wnt signaling was shown to be activated by the reporter assay (Fig. 1). The results of immunostaining of β-catenin, in conjunction with data from the reporter assay, strongly suggest that Wnt-1 expression stabilizes ectopically expressed β-catenin in Hep3B or PLC/PRF/5 cells, but not in Huh7, Chang liver, and SK-Hep1 cells. Collectively, our data led us to conclude that Wnt-1 is necessary but insufficient for activating Wnt/β-catenin signaling in some hepatoma cells. One interpretation of these data is that one or more Wnt/β-catenin signaling molecules upstream of β-catenin are negatively regulated in Huh7 cells, Chang liver, and SK-Hep1 cells, but not in Hep3B cells and PLC/PRF/5 cells.

Figure 2.

Cellular localization of β-catenin in hepatoma cells after ectopic expression of Wnt signaling molecules. The four indicated hepatoma cell lines were transfected transiently with the indicated expression plasmids (A–E). Two days after transfection, cells were fixed, were stained with the anti-β-catenin antibody (Transduction Laboratories), and were visualized by confocal microscopy.

Huh7 Cells Are Refractory to the Activation of Wnt/β-Catenin Signaling on the Addition of Wnt-1.

To substantiate our observation that some hepatoma cells were refractory to Wnt/β-catenin signaling, we compared Huh7 cells to the well-characterized HEK293 cell line28 on the addition of Wnt-1 with respect to stabilization of cytosolic β-catenin molecules, a hallmark of Wnt/β-catenin signaling. As a source of Wnt-1 protein, we used conditioned media prepared from Huh7 cells transfected with the Wnt-1 expression plasmid. After the addition of the conditioned media containing Wnt-1, the level of cytosolic β-catenin was monitored by Western blot analysis (Fig. 3A). Consistent with the previous report,28 in HEK293 cells, the amount of β-catenin gradually increased until 6 hours after the addition of Wnt-1 media and then slowly decreased and returned to its initial level, reflecting transient activation of Wnt/β-catenin signaling on the addition of Wnt-1 media. By contrast, in Huh7 cells, stabilization of cytosolic β-catenin was not found on the addition of Wnt-1 media, whereas the β-catenin in whole cells was readily detectable and remained unchanged (Fig. 3A; data not shown). The TCF-4 reporter assay performed in parallel indicated that significant transcriptional activation of luciferase activity was observed in HEK293 cells by Wnt-1 transfection, but not in Huh7 cells (Fig. 3B), confirming that Huh7 cells are refractory to the activation of Wnt/β-catenin signaling.

Figure 3.

Huh7 cells are refractory to the activation of Wnt/β-catenin signaling upon Wnt-1. (A) Accumulation of cytosolic β-catenin in HEK293 cells but not in Huh7 cells. Cells were incubated with Wnt-1-conditioned media for the indicated times (h). The conditioned media containing Wnt-1 was prepared as previously described.28 Cytosolic fractions were prepared essentially as previously described28 and were immunoblotted for β-catenin and tubulin, which served as the loading control. (B) Ectopic expression of Wnt-1 led to TCF-4-dependent transcription in HEK293 cells, but not in Huh7 cells. Cells were transfected with Wnt-1, as indicated. The TCF-4-dependent reporter assay is described in Fig. 1.

HBx Plays an Essential Role in Activating Wnt/β-Catenin Signaling in Hepatoma Cells.

Intriguingly, it was noted that cells that had been activated by ectopic expression of Wnt-1 (Hep3B and PLC/PRF/5 cells) were derived from HBV-infected liver,23, 29 whereas others (Huh7, Chang liver, and SK-Hep1 cells) were derived from HBV-negative liver. This important difference immediately raised the possibility that HBx, a viral regulatory protein presumably expressed only in these HBV-infected hepatoma cells, likely was responsible for these disparate responses to Wnt-1. One expectation of this notion was that ectopic expression of HBx confers to Huh7 cells the ability to respond to Wnt-1 and led to the activation of Wnt/β-catenin signaling. To substantiate the proposed role of HBx in Wnt signaling, we measured the luciferase level after cotransfection of Huh7 cells with Wnt-1 plus HBx expression plasmids. Indeed, we found that as the amount of transfected HBx plasmid increased, luciferase activity increased up to fivefold in a dose-dependent manner (Fig. 4A). In other words, in agreement with the hypothesis, ectopic expression of HBx along with Wnt-1 led to the activation of Wnt signaling in Huh7 cells. To corroborate the findings of the reporter assay, we cotransfected Wnt-1 plus HBx expression plasmids and immunostained for β-catenin (Fig. 4B). Indeed, nuclear translocation of β-catenin was observed, when HBx plus Wnt-1 were coexpressed. Immunostaining of HBx performed in parallel revealed that 48% (61 of 126) of HBx-stained cells exhibited nuclear staining of β-catenin, indicating that β-catenin was nuclear translocated in significant fraction of HBx-expressing cells. The data led us to conclude that HBx plays an essential role in the activation of Wnt/β-catenin signaling in Huh7 cells. This conclusion was strengthened by the observation that HBx is, in fact, functionally expressed in PLC/PRF/5 cells derived from HBV-infected liver.30

Figure 4.

Evidence that HBx is essential for the activation of Wnt signaling in Huh7 cells. (A) HBx-activated TCF-4-dependent transcription in Huh7 cells. Cells were transfected with increasing amounts of HBx expression plasmid (0.05 μg, 0.1 μg, 0.15 μg, 0.2 μg), along with 1 μg of each indicated expression plasmid and either the pTOPFLASH or pFOPFLASH reporter plasmids. Transfection and luciferase assay protocols are described in Fig. 1. Cells cotransfected with TCF-4 and β-catenin served as the positive control. (B) Nuclear localization of β-catenin in Huh7 cells after ectopic expression of HBx, along with Wnt-1 and β-catenin. Cells were transfected with various expression plasmids and immunostained as in Fig. 2. β-Catenin staining, DAPI staining, and their superimposed images are shown.

HBx Induces the Stabilization and Subsequent Nuclear Translocation of β-Catenin by Activating Src Kinase.

The necessity of HBx led us to speculate that HBx complements the essential function needed for the activation of Wnt/β-catenin signaling that lacks in hepatoma cells, but not in others such as HEK293 cell. To explore the notion, cells were transfected as indicated and treated with a proteosome inhibitor, MG132 (Fig. 5A). After the MG132 treatment, it was evident that β-catenin staining was enhanced in the cytoplasm with distinctive perinuclear staining; however, β-catenin was excluded from the nucleus, suggesting that stabilization of β-catenin by MG132 was insufficient for nuclear translocation. By contrast, even in the absence of HBx, Wnt-1 transfection led to nuclear translocation of β-catenin in cells treated with MG-132, suggesting that the implicated function lacked in hepatoma cell could be complemented experimentally either by HBx or MG132 treatment.

Figure 5.

Evidence that HBx stabilizes β-catenin via the activation of Src kinase. (A, B) Nuclear localization of β-catenin was suppressed by Src-DN. Immunostaining of cells for β-catenin was performed as described in Fig. 2. Cells were transfected as indicated. MG132, a proteosome inhibitor; Src-DN, a dominant-negative mutant of Src kinase. (C) HBx expression was not suppressed by Src-DN. Western blot analysis of HBx was performed with anti-FLAG antibody (Sigma). (D) The nuclear accumulation of β-catenin in Wnt-1 plus HBx-transfected cells was blocked by Src-DN. Biochemical fractionation of the transfected cells was performed essentially as described previously.44 Coomassie staining of the nuclear fraction served as a loading control. Quantification of the band was carried out with a calibrated GS-800 densitometer (Bio-Rad).

In defining HBx activities, recent studies have pointed out the central role of Src kinase activation.31, 32 Thus, we asked whether HBx contributes to Wnt/β-catenin signaling via the activation of Src kinase. We examined the effect of HBx activation of Src kinase on the nuclear localization of β-catenin by cotransfecting Wnt-1, HBx, and Src-DN, an Src dominant-negative mutant.33 Importantly, ectopic transfection of the Src dominant-negative plasmid diminished the nuclear staining of β-catenin in Wnt-1 plus HBx-transfected cells (Fig. 5B). The data indicated that HBx, via the activation of Src kinase, led to the nuclear localization of β-catenin in Huh7 cells. However, the Src-DN did not block the nuclear translocation of β-catenin induced by Wnt-1 in the MG132-treated cells. This finding, together with the above, indicated that nuclear translocation of β-catenin induced by Wnt-1 in the MG132-treated cells was not Src kinase-dependent, whereas that by Wnt-1 plus HBx was Src kinase-dependent (Fig. 5A). Further, the above observation implicated that HBx-activated Src kinase lies upstream of ubiquitin-mediated β-catenin degradation, which can be inhibited by MG132. A possibility that HBx expression was suppressed by the Src-DN was excluded, because a comparable amount of HBx was detected by Western blot analysis when Src-DN was cotransfected (Fig. 5C).

To substantiate these findings in the context of Wnt/β-catenin signaling, we examined whether the accumulation of β-catenin in the nucleus could be induced by Wnt-1 plus HBx and if so, whether the nuclear translocation of β-catenin is Src dependent. Western blot analysis after biochemical fractionation of the transfected cells indicated that (1) the level of β-catenin was found elevated by approximately threefold in the nucleus of cells transfected by Wnt-1 plus HBx, but not by Wnt-1 or HBx alone (Fig. 5D, lanes 2 and 3 versus lane 4) and (2) the enhanced nuclear accumulation of β-catenin was blocked by Src-DN (Fig. 5D, lane 5). This biochemical analysis, in conjunction with the data shown above (Fig. 5A,B), led us to conclude that HBx led to stabilization and subsequent nuclear localization of β-catenin via the activation of Src kinase. Moreover, this result is remarkably consistent with the findings in Fig. 1 and 2 showing that Wnt-1 is insufficient to activate Wnt/β-catenin signaling in Huh7 cells. Considering lower transfection efficiency (approximately 5%–10%; data not shown), the magnitude of enhanced accumulation of β-catenin by threefold was considered significant.

HBx Activation of Src Kinase Suppressed GSK3β Activity.

Next, we sought to determine the mechanism by which HBx contributes to the stablilization of β-catenin. In fact, GSK3β activity has been shown to determine the stability of cytosolic β-catenin, and we thus examined the activity of GSK3β, whose activity is suppressed on the binding of Wnt-1 ligand to its receptor.28 In particular, we investigated whether ectopic expression of HBx could inhibit GSK3β in the presence of Wnt-1 by monitoring the GSK3β kinase activity (Fig. 6A). To achieve higher transfection efficiency, cells was transfected by liposome. Two days after transfection of Huh7 cells, GSK3β was immunoprecipitated with the GSK3β antibody and assayed in vitro for kinase activity by using MBP as a substrate as described previously.34 As shown in Fig. 6A,B, Wnt-1 or HBx alone marginally suppressed the kinase activity of GSK3β. However, when cells were cotransfected with Wnt-1 plus HBx, GSK3β activity decreased by 40%; it should be noted that given that not all cells were transfected (approximately 50%–60% transfection efficiency; data not shown), this magnitude of suppression was considered to be near complete, and further, this magnitude of reduction in GSK3β activity was shown to be sufficient to induce Wnt/β-catenin signaling in other contexts.35 Thus, it seemed that HBx works with Wnt-1 synergistically in suppressing GSK3β activity. Moreover, the data are in good agreement with the observation independently made by the reporter assay (Fig. 4) and immunostaining (Fig. 5). In addition, we examined the effect of HBx activation of Src kinase on the inhibition of GSK3β kinase activity by cotransfecting Src-DN, Wnt-1, and HBx. As shown in Fig. 6C,D, when cells were cotransfected with Src-DN, the GSK3β activity was restored to a level similar to the level of the mock-transfected cell. The data suggested that the HBx activated Src kinase is mainly responsible for suppressing the GSK3β activity.

HBx Expressed From the HBV Replicon Recapitulates the HBx-Mediated Activation of Wnt/β-Catenin Signaling.

One caveat to the transfection studies concerns the fact that the expression level of the HBx protein driven by a strong cytomegalovirus promoter is believed to be far greater than the physiological level expressed during chronic HBV infection.36 To address this concern, we sought to examine the effect of modest level of HBx expressed from an HBV replicon on Wnt signaling. An HBV replicon, a greater-than-genome-length construct (i.e., 1.2 mer), was used that is capable of recapitulating viral genome replication in hepatoma cells (data not shown). Huh7 cells were transfected with the HBV 1.2-mer construct, and the luciferase expression was assayed. As shown in Fig. 7A, transfection of the wild-type HBV 1.2-mer (HBx+) led to significantly increased TCF-4-dependent transcription in a dose-dependent manner, whereas transfection of the HBx-null mutant (HBx) of HBV 1.2 mer did not. The reporter assay results further supported the notion that HBx, but not other HBV proteins, is responsible in responding to Wnt-1 in HBV-positive hepatoma cells, as shown in Fig. 1C,D. However, the observation that Src-DN blocked reporter gene activity suggested that the activation of Wnt signaling by the HBV 1.2 mer was mediated via the activation of Src kinase. Based on these data, we concluded that HBx, whether expressed from a replicating HBV genome (Fig. 7A) or from a strong heterologous cytomegalovirus promoter (Fig. 4A), confers to hepatoma cells the ability to respond to Wnt-1 and leads to the activation of Wnt signaling via the activation of Src kinase. However, a concern that Src-DN generally suppresses transcription from host RNA polymerase II promoters also was eliminated by an experiment with MHC class II promoter (data not shown).

Figure 7.

The HBx expressed from the hepatitis B virus (HBV) replicon-activated Wnt/β-catenin signaling. (A) TCF-4-dependent reporter assay by HBx expressed from its endogenous promoter in the context of the viral genome replication. Huh7 cells were transfected with increasing amounts of the 1.2-mer replicon construct, a wild-type (HBx+) or an HBx-null mutant (HBx), along with 1 μg of Wnt-1 or Src-DN. The 1.2-mer HBV replicon, a greater-than-genome-length construct, was used to express HBx in the context of viral genome replication. (B) TCF-4-dependent transcription was not suppressed by Src-DN. Huh7 cells were transfected with 1 μg of β-catenin-DP or Src-DN. β-catenin-DP, a constitutively active mutant.

To address whether the inhibitory effect of Src-DN on Wnt signaling observed in Fig. 7A is specific to the HBx activated Wnt signaling, we examined the effect of Src-DN on the Wnt signaling that is activated by a constitutively active β-catenin. The result shown in Fig. 7B indicated that the elevated luciferase activity induced by β-catenin-DP, a constitutively active β-catenin, was not suppressed by cotransfected Src-DN, suggesting that the effect of Src-DN on the activated Wnt signaling is specific to the HBx activated Wnt signaling.

Discussion

Although the literature on the stabilizing mutation of β-catenin in HCC is growing,3–7 it has not been established whether such β-catenin mutation leads to the functional activation of Wnt/β-catenin signaling in hepatoma cells. During our effort to investigate the functional integrity of Wnt/β-catenin signaling in hepatoma cells, we unexpectedly found that HBV-negative hepatoma cells are refractory to Wnt/β-catenin signaling, whereas HBV-positive hepatoma cells normally responded to Wnt-1 (Figs. 1 and 2). We then speculated that HBx, a viral regulatory protein, may contribute to Wnt/β-catenin signaling. Importantly, we demonstrated that Wnt/β-catenin signaling could be activated by Wnt-1 only if HBx is provided in trans, that contributes to the stabilization of β-catenin by suppressing GSK3 activity (Figs. 4–6).

Inability of the HBV-negative hepatoma cells to respond to Wnt-1 in the reporter assay led us to compare Huh7 cells to HEK293 cell line, which is well characterized with respect to their responses to Wnt-1.28 The observation that adding the Wnt-1 molecule failed to stabilize cytoplasmic β-catenin in Huh7 cells suggested that a pathway upstream of β-catenin in Wnt/β-catenin signaling is suppressed or negatively regulated in these hepatoma cells (Fig. 3). The finding that HBx confers to Huh7 cells the ability to respond to Wnt-1 was taken to imply that HBx overrides the negative regulation imposed on hepatoma cells, thereby leading to the activation of Wnt/β-catenin signaling in Huh7 cells (Fig. 4). An observation that MG132 treatment can replace HBx further suggested that HBx contributes to the activation of Wnt/β-catenin signaling by stabilizing β-catenin (Fig. 5A). This result is interpreted to indicate that the extent of which the β-catenin is stabilized on Wnt-1 binding to its receptor is insufficient for the activation of Wnt/β-catenin signaling in hepatoma cells; thus, enhanced stabilization provided by HBx or MG132 treatment is needed (Fig. 5A). However, Wnt-1 is sufficient to activate Wnt/β-catenin signaling in the HBV-positive hepatoma cells, such as Hep3B and PLC/PRF/5 cells, presumably because of the presence of functional HBx in these cells.

Recent studies have suggested that most, if not all, of HBx activities may be ascribed to the activation of Src kinase.37 The observation that HBx-activated Src kinase is involved in the activation of the Wnt/β-catenin signaling pathway (Fig. 5) led us to consider other consequences of Src kinase activation in this pathway. Relevantly, a study has shown that tyrosine phosphorylation of β-catenin by Src kinase results in an increase in its free cytosolic pool and correlates directly with increased cell migration and loss of epithelial morphological features.38 Other groups have shown that phosphorylation of β-catenin by Src kinase leads to nuclear localization of phosphorylated β-catenin and disruption of cadherin-mediated adhesion.39 Moreover, HBx has been shown to induce adherens junction disruption in a Src kinase-dependent manner, resulting in an invasive phenotype.40 The work presented here clearly indicates that the stabilization of β-catenin by the HBx-activated Src kinase largely contribute to the activation of Wnt/β-catenin signaling, although the phosphorylation of β-catenin by Src kinase remains to be seen. Collectively, the activated Src kinase could lead to the stabilization of cytosolic β-catenin via two distinct mechanisms: (1) by suppressing GSK3β, which is a negative regulator of the Wnt/β-catenin signaling (Fig. 6), and (2) by stimulating tyrosine phosphorylation and subsequent nuclear localization of β-catenin.38, 39

HBx, a regulatory protein of HBV, has gained considerable attention owing to its presumptive roles in hepatocarcinogenesis. Although HBx has pleiotropic activities that may be related to oncogenesis, its contribution to tumor formation has been less clear.1, 41, 42 Given the importance of Src kinase activation in defining HBx activities,31, 32, 43 it was speculated that HBx activation of Src kinase could have a profound effect on the development of HCC. The data presented here provide a novel mechanism by which HBx could contribute to hepatocarcinogenesis via the activation of Wnt/β-catenin signaling.

Ancillary