Hepatitis B virus X protein enhances the transcriptional activity of the androgen receptor through c-Src and glycogen synthase kinase-3β kinase pathways

Authors

  • Wan-Jen Yang,

    1. Department of Microbiology, National Taiwan University College of Medicine, Taipei, Taiwan
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Ching-Ju Chang,

    1. Department of Microbiology, National Taiwan University College of Medicine, Taipei, Taiwan
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Shiou-Hwei Yeh,

    Corresponding author
    1. Department of Microbiology, National Taiwan University College of Medicine, Taipei, Taiwan
    2. National Taiwan University Center for Genomic Medicine, National Taiwan University College of Medicine, Taipei, Taiwan
    • Department of Microbiology, National Taiwan University College of Medicine, No. 1, Jen-Ai Road, Section 1, Taipei 100, Taiwan
    Search for more papers by this author
    • fax: (886)-2-23825962

  • Wei-Hsiang Lin,

    1. Department of Microbiology, National Taiwan University College of Medicine, Taipei, Taiwan
    Search for more papers by this author
  • Sheng-Han Wang,

    1. Department of Microbiology, National Taiwan University College of Medicine, Taipei, Taiwan
    Search for more papers by this author
  • Ting-Fen Tsai,

    1. Faculty of Life Sciences and Institute of Genome Sciences, National Yang-Ming University, Taipei, Taiwan
    Search for more papers by this author
  • Ding-Shinn Chen,

    1. National Taiwan University Center for Genomic Medicine, National Taiwan University College of Medicine, Taipei, Taiwan
    2. Department of Internal Medicine, National Taiwan University College of Medicine, Taipei, Taiwan
    Search for more papers by this author
  • Pei-Jer Chen

    1. National Taiwan University Center for Genomic Medicine, National Taiwan University College of Medicine, Taipei, Taiwan
    2. Department of Internal Medicine, National Taiwan University College of Medicine, Taipei, Taiwan
    3. Graduate Institute of Clinical Medicine, National Taiwan University College of Medicine, Taipei, Taiwan
    Search for more papers by this author

  • Potential conflict of interest: Nothing to report.

Abstract

Hepatitis B virus (HBV)-related hepatocellular carcinoma (HCC) occurs predominantly in men. By enhancing the transcriptional activity of the androgen receptor (AR) gene in a ligand-dependent manner, the HBV X protein (HBx) might contribute to this disparity between sexes. To dissect the mechanisms underlying HBx-enhanced AR transactivation, we investigated the effect of HBx on two critical steps in the regulation of ligand-stimulated AR activities. One step is the dimerization of AR (through the interaction of its N-termini and C-termini), and the other step is the activation of the AR N-terminal transactivation domain (NTD). HBx increased the NTD activation of the AR through c-Src kinase. HBx also enhanced AR dimerization by inhibiting glycogen synthase kinase-3β (GSK-3β) activity, which acts as a negative regulator of the interaction between AR and the N-termini and C-termini. The HBx-enhanced AR transactivation was abolished by blocking c-Src and activating GSK-3β kinases simultaneously, suggesting that these two kinases act as major switches in the activation process. The regulatory function of both kinases has been further verified in primary hepatocytes isolated from the livers of HBx transgenic male mice. Conclusion: Our study thus identified two key kinases through which HBx enhances the AR transcriptional activity. These kinases might be potential candidates for future prevention or therapy for HBV-related HCC in men. (HEPATOLOGY 2009.)

Persistent hepatitis B virus (HBV) infection is a major risk factor of hepatocellular carcinoma (HCC), which causes more than half a million deaths each year.1 An intriguing feature of HBV-related HCC is the male predominance, with a male-to-female ratio of 5 to 11:1.2, 3 The elevated activity of the androgen axis was suggested as one major host factor for this disparity between sexes. The risk of HCC is higher in male HBV carriers with higher serum androgen concentration or androgen receptor (AR) alleles showing higher transcriptional activity.4, 5 In support, we have identified AR somatic mutations that appear only in male patients with HCC.6

Because the male predominance is more pronounced in HBV-related HCC than in hepatitis C virus–related HCC or other HCCs,2, 3 we proposed that factors specific to HBV might cooperate with the androgen-signaling pathway to accelerate hepatocarcinogenesis in men. Our previous study demonstrated that HBV X protein (HBx) enhances the transcriptional activity of the AR in an androgen concentration–dependent manner.7 The synergistic carcinogenic effect of HBx and AR was confirmed by the enhanced anchorage-independent colony-forming activity. Data from HBx transgenic mouse models, which also show a male predominance of HCC, support this concept.7–9

The molecular mechanisms responsible for this HBx-enhanced AR activation are the next issue to be addressed. AR is a ligand-dependent transcription factor that belongs to the nuclear receptor superfamily.10, 11 Binding of androgen first induces dimerization of AR through the interaction of its N-termini and C-termini. The ligand-bound AR then translocates to the nucleus and binds to its cognate response DNA sequences to modulate the transcription of target genes through its N-terminal transactivation domain (NTD).10, 11 To investigate how HBx enhances AR transactivation, we focused on the two critical steps: the AR N–C interaction (dimerization step) and the transcriptional activation of the AR NTD (transactivation step).

Another specific aim of this study was to identify the key regulatory switches that turn on the HBx-enhanced AR transactivation. Understanding this process may help identify potential targets for targeted therapy of HBV-related HCC in men. Because effects of HBx on AR activation occur mainly in the cytosol,7 HBx might activate the AR indirectly by affecting the cytosolic signaling pathways. c-Src and glycogen synthase kinase (GSK)-3β are two cytosolic kinases regulated by HBx. By triggering the release of Ca2+ ions from the endoplasmic reticulum and mitochondria, HBx can activate c-Src kinase indirectly by activating calcium-dependent tyrosine kinase (proline-rich tyrosine kinase 2) and the focal adhesion kinase.12, 13 By inducing the association between extracellular signal–regulated protein kinase and GSK-3β, which primes GSK-3β to be phosphorylated by p90RSK at Ser9, HBx can inactivate the cytosolic GSK-3β.14, 15 These two kinases and their downstream signaling pathways are thus potential candidates as the key switches regulating HBx-enhanced AR activation.

Our current results show that HBx significantly enhances the activity of both steps, that is, the AR N–C interaction and the AR NTD activation. Further study of the role of c-Src activation and GSK-3β inactivation in regulating these two steps showed that c-Src enhances mainly the transcriptional activity of the AR NTD. However, GSK-3β inactivation contributed to the HBx-enhanced AR N–C interaction. Turning off both regulatory mechanisms simultaneously, by inhibiting c-Src and activating GSK-3β, markedly decreased the HBx-enhanced AR activation, suggesting that these two kinases are major switches that activate the HBx-enhanced AR transactivation. The results provide new insights into how HBV infection augments AR activities.

Abbreviations

AR, androgen receptor; CT-FBS, charcoal stripped fetal bovine serum; FBS, fetal bovine serum; GSK-3β, glycogen synthase kinase-3β HBV, hepatitis B virus; HBx, Hepatitis B virus X protein; HCC, heptocellular carcinoma; NTD, N-terminal transactivation domain; siRNA, small interfering RNA; tg, transgenic.

Materials and Methods

Cell Culture and Transfection.

Huh-7 and HepG2 cells were cultured and transfected with conditions and procedures as described previously.7 Porcine cytomegalovirus–HBx, pSG5-AR (AR expression plasmid), and the reporter constructs of MMTV-Luc (luciferase gene driven by MMTV promoter) or 5XARE-Luc (luciferase gene driven by the promoter containing ARE sequence in 5 repeats) were used for assaying the effect of HBx on AR transactivity. Plasmids pGal4-AD-ARN (a.a. 1∼560 of AR fused with the DNA binding domain of Gal4 at its N terminal) and pGal4-TK-Luc (luciferase reporter containing Gal4-binding domain at the promoter region) were used for assaying the effect of HBx on AR NTD (N terminal transactivation domain) activity (mammalian one-hybrid assay, Fig. 1A). The plasmids GAL4-hAR(DE) (a.a. 624-918 of AR fused with the DNA binding domain of Gal4 at its C terminal), VP16-hAR(AB) (1∼555 of AR fused with the transactivation domain of VP16 at its N terminal) and pGal4-TK-Luc were used for assaying the interaction between AR N-termini and C-termini (mammalian two-hybrid assay, Fig. 1B). Twenty hours after transfection, the cells were treated with 10 nM R1881 (Perkin-Elmer, Wellesley, MA) or with dimethylsulfoxide control, in 2% dextran-coated charcoal stripped fetal bovine serum (FBS) medium (CT-FBS). The cell lysates were harvested 20 hours after treatment for subsequent analysis.

Figure 1.

HBx enhances both the AR N–C interaction and AR NTD activity. (A) The plasmid constructs and basic principle for measuring the AR N–C interaction using the mammalian two-hybrid assay, analyzing the interaction between GAL-hAR-C and VP16-hAR-N for their effect on the pGal4-TK-Luc reporter activity. (B) The plasmid constructs and basic principle for assaying the AR NTD activity using the mammalian one-hybrid assay, analyzing the effect of GAL-hAR-N on the pGal4-TK-Luc reporter activity. (C) Huh-7 and HepG2 cells were transfected with GAL-hAR-C, VP16-hAR-N, or pGal4-TK-Luc, with or without porcine cytomegalovirus-HBx, and the luciferase activity was measured in the presence or absence of R1881 as indicated. The cell lysates were prepared, and the luciferase activity was measured. The results shown are the average of three experiments (mean ± SD). (D) Huh-7 and HepG2 cells were transfected with GAL-hAR-N or pGal4-TK-Luc, with or without porcine cytomegalovirus-HBx, and the luciferase activity was measured in the presence or absence of R1881 as indicated. The cell lysates were prepared for assessment of luciferase activity. The results shown are the average of three experiments (mean ± SD).

For evaluating the effect of c-Src on the reporter activities, the specific kinase inhibitor PP2 (Calbiochem, San Diego, CA) was added to the cells 24 hours before cell lysate preparation.

Luciferase Reporter Assay.

The MMTV-Luc and 5XARE-Luc were used for assaying the AR-responsive transcriptional activity. The pGal4-TK-Luc was used for assaying the AR N/C interaction or AR NTD activity in mammalian two hybrid and one hybrid assay (Fig. 1A, B). The dual-luciferase reporter assay was used for analyzing the reporter activity (Promega, Madison, WI) by following the manufacturer's instructions, with pRL-CMV plasmid (the plasmid expressing renilla gene driven by pCMV promoter) (Promega, Madison, WI) co-transfected as an internal standard for normalization of transfection efficiency.1

Lentivirus-Based Knockdown of c-Src Expression.

pLKO.1-Luc, pLKO-1-GFP, and pLKO-1-c-Src constructs with si-luciferase, si-GFP, and si-c-Src driven by U6 promoter were provided by the RNAi Core Facility in the Academia Sinica (Taipei, Taiwan). The small interfering RNA (siRNA) sequence in pLKO-1-c-Src plasmid contained the sequence of 5′-GCTCGGCTCATTGAAGACAAT-3′, which specifically targets to the human c-Src (NM_198291).

The details for the preparation of lentivirus stock were described previously.16 The viral stock was used to infect the cells cultivated in 2% CT-FBS for 2 days, and then the cells were processed for the subsequent transfection experiments.

HBx Transgenic Mice.

The HBx transgenic mice established by pronucleus microinjection of C57BL/6 fertilized eggs with HBx expressed in a liver-specific transgenic vector, pAlb-In-pA-HS4, were used in the current study (with details as described previously8). All of the mice were bred and followed-up in a specific pathogen–free facility. The transgenic mice were sacrificed at different ages, and the liver tissues were collected for protein extraction and western blot analysis. The transgenic mice were also used for the hydrodynamic injection experiments and for isolating the hepatocytes for primary hepatocytes culture described as follows.

The approval for all mouse experiments was obtained from the Institutional Animal Care and Use Committee at the National Taiwan University College of Medicine, and all of the experiments were carried out according to the Committee's guidelines.

Hydrodynamic Injection Through Mouse Tail Vein.

Hydrodynamic injections for assaying the reporter activity in the liver tissues were performed by referring to the procedures reported by Huang et al.17 In brief, the plasmids including 20 μg pMMTV-luc and 2 μg pRL-CMV were diluted in a volume of 1× phosphate-buffered saline equivalent to 8% of the total body weight, and then injected into the tail vein of each mouse anesthetized by Avertin (Acros Organics, NJ) within 6-8 seconds. After 72 hours, the mice were sacrificed and their livers were processed for the subsequent protein extraction and luciferase assay.

Mouse Hepatocyte Isolation.

The mice were anesthetized with 33% ketamine (Phoenix, St. Joseph, MD). Surgery was performed to insert a catheter into the hepatic vein. The livers were then perfused with calcium-free and magnesium-free Hank's buffer (Gibco, Carlsbad, CA) containing 0.025% collagenase type IV (Sigma-Aldrich) in Hank's buffer containing 5 mM calcium chloride. After perfusion, the livers were removed, and the cells were resuspended in hepatocyte wash medium (Gibco, Carlsbad, CA) and then filtered through a 100-μm cell strainer (BD Bioscience), counted, and tested for viability using trypan blue exclusion assay. The hepatocytes were then resuspended with William's E medium serum-free medium (Gibco, Carlsbad, CA), containing 10% FBS, 1% L-glutamine, 1% sodium pyruvate, 1% PSA (antibiotic/antimycotic solution, containing penicillin, streptomycin, & amphotericin B), 1% penicillin/streptomycin/ampicillin, and 0.1% insulin-transferrin-selenium, and plated at a density of 2 × 105 cells per well on 12-well plates. After culture overnight, the cells were processed for the transfection with the same procedures as that for culture cells.

Results

HBx Enhances Both the AR N–C Interaction and the AR NTD Activation.

We used two assays to examine the effect of HBx on the two critical steps regulating ligand-stimulated AR activation, the AR N–C interaction and the AR NTD activation. The mammalian two-hybrid assay was used to examine the AR N–C interaction, and the mammalian one-hybrid assay was used to examine the AR NTD activity; the principles of these assays are outlined in Fig. 1A, B.

Huh-7 and HepG2 cell lines were used in the analyses. In the assay of the AR N–C interaction, the reporter activity was increased significantly by the presence of HBx in a ligand-dependent manner (Fig. 1C, lane 2 versus 4). We next investigated the effect of HBx on the AR NTD activity. HBx also significantly enhanced the AR NTD activity (Fig. 1D, lanes 1 and 2 versus lanes 3 and 4). The results from both cell lines are consistent and show that HBx activates both steps of AR activation.

c-Src Activity Contributes to the AR N-Terminus Transcriptional Activation But Not to the AR N–C Interaction.

To identify the critical kinase switches controlling HBx-enhanced AR activation, we first focused on c-Src kinase. Using a PP2 inhibitor, our previous study showed the possible involvement of c-Src in HBx-induced AR transactivation.7 In our study, we adopted another approach to inhibit c-Src activity in a more specific way, namely, a lentivirus-based siRNA to knockdown the endogenous c-Src expression. The effect of this siRNA on HBx-induced AR activation was evaluated in both MMTV-Luc and 5XARE-Luc reporter constructs. Knockdown of c-Src expression significantly decreased the HBx-enhanced AR activation (Fig. 2A, lane 4 versus 8); this was consistent with the results of PP2 treatment (Fig. 2B, lane 6 versus 8). These results further support the positive regulatory role of c-Src kinase in HBx-enhanced AR transactivation.

Figure 2.

c-Src activity is involved in enhancing the AR NTD activation rather than the AR N–C interaction. In Huh-7 cells, the critical involvement of c-Src in regulating the HBx-enhanced AR transactivation was demonstrated by (A) si-c-Src and by (B) the specific inhibitor PP2. The siRNA and inhibition decreased the reporter activity of MMTV-Luc and 5XARE-Luc. The effect of inhibiting c-Src by PP2 on (C) AR N–C interaction and (D) AR NTD activity was analyzed with the mammalian two-hybrid or one-hybrid assay. The luciferase activity is the average of three experiments (mean ± SD). (E) The inhibitory effect of PP2 on AR NTD activation was verified further by transfecting the construct of the AR mutant of ARΔHBD (containing the NTD of AR) into Huh-7 cells, and the effects on MMTV-Luc reporter activity were measured. The luciferase activity is the average of three experiments (mean ± SD).

We next tried to dissect the step(s) through which c-Src activity is involved in the HBx-stimulated AR transcriptional activation. We looked for the possible involvement of c-Src in enhancing the activity of either the AR N–C interaction or the AR NTD activation. Inhibition of c-Src activity by PP2 treatment did not affect the HBx-enhanced AR N–C interaction (Fig. 2C, lane 2 versus 6; lane 6 versus lanes 7 and 8). This shows that, although HBx enhances the AR N–C interaction, c-Src activity is not involved in this regulatory step. In contrast, PP2 significantly reduced the AR NTD activity to a level similar to that in the presence or absence of HBx (Fig. 2D, lanes 3 and 4 versus lanes 7 and 8). These results suggest that c-Src activity is involved in regulating AR NTD activity.

To verify the inhibitory effect of PP2 on the AR NTD activation, we used an ARΔHBD construct (AR expression plasmid containing the AR NTD and the DNA-binding domain covering amino acids 1-670) to analyze the effect of PP2 treatment on its cognate MMTV-Luc reporter construct. HBx also enhanced the reporter activity in this construct, and the activity was repressed by PP2 treatment (Fig. 2E, lanes 5 and 6 versus lanes 7 and 8). This confirmed the role of c-Src activity in regulating the HBx-enhanced AR NTD activity.

c-Src Enhances the AR NTD Activity by Increasing the Phosphorylation of AR Ser81 and Increasing the AR Protein Level.

The next question was how the c-Src activity is involved in regulating the AR NTD activity. Because c-Src can activate various downstream kinases and signaling pathways, we proposed that c-Src can affect the AR NTD activity by regulating the phosphorylation of AR. HBx enhanced the phosphorylation level of AR Ser81, and this phosphorylation was down-regulated by PP2 treatment (Fig. 3A, lane 6 versus 5) or by knockdown of the expression of c-Src (Fig. 3B, lane 8 versus 4). This suggests that c-Src contributes to the HBx-enhanced AR phosphorylation at least through the residue of Ser81.

Figure 3.

c-Src enhances the AR NTD activity by increasing the phosphorylation of AR Ser81 and by increasing the AR protein level. The HBx-enhanced phosphorylation of AR Ser81 was suppressed by inhibiting c-Src activity with either (A) PP2 or (B) lentivirus-based siRNA against c-Src in Huh-7 cells. The cell lysates were harvested for western blot analysis using the antibodies indicated. (C) HBx-enhanced AR transactivation was decreased by the ARS81A construct but not by the ARY534F construct (upper panel). The luciferase activities are the average of three experiments (mean ± SD). The cell lysates were harvested for western blot analysis using the antibodies indicated (lower panel).

To test further the functional effect of this enhanced Ser81 phosphorylation on HBx-enhanced AR transactivation, we mutated this residue to Ala and evaluated its influence on 5XARE-Luc reporter activity. The HBx-induced AR activity was decreased to approximately 50% compared with that of wild-type AR (Fig. 3C, lane 4 versus 8), suggesting that the elevated phosphorylation at Ser81 contributes to the higher HBx-enhanced AR activity. Because Guo et al.18 recently showed that c-Src kinase can activate the transcriptional activity of the AR in androgen-refractory prostate cancers by increasing the phosphorylation of the AR at Tyr534,18 we included the ARY534F mutant in our analysis. Mutation at this residue did not affect the HBx-enhanced AR transcriptional activity (Fig. 3C, lane 4 versus 12), suggesting that c-Src–mediated Try534 phosphorylation of the AR does not contribute to HBx-enhanced AR activation.

Cell lysates processed for western blot analysis produced an intriguing finding. The amount of protein was significantly lower in the ARS81A than in the wild-type AR or the ARY534F (Fig. 3C, lower panel, lanes 6 and 8 versus lanes 2 and 4, and lanes 10 and 12). These results raise the possibility that the c-Src–mediated AR activation occurs through increasing AR phosphorylation at Ser81 and that this might be associated with the increased accumulation of the AR in a ligand-dependent manner.

GSK-3β Activity Is Involved in Regulating the HBx-Enhanced AR N–C Interaction But Not the AR NTD Activation.

We next investigated the involvement of GSK-3β in the HBx-enhanced AR activation because this is inactivated by HBx.14, 15 We first examined whether GSK-3β activity is regulated in the process of HBx-enhanced AR activation. Phosphorylation of the N-terminus of GSK-3β on Ser9 can occupy its phospho-priming binding site at the C-terminus and thus inhibit the enzyme activity, which is one major mechanism that regulates GSK-3β activity.19 We therefore evaluated the activity of GSK-3β by western blot analysis hybridization with anti-PS9-GSK-3β antibody. The phosphorylation on Ser9 of GSK-3β increased, which can be attributed mainly to the presence of HBx (Fig. 4A, lanes 3 and 4 versus lanes 1 and 2), indicating that GSK-3β activity is decreased in the process. These results confirm previous findings of the inhibitory effect of HBx on GSK-3β activity.14, 15 We next transfected a constitutively active GSK-3β(S9A) construct, in which the phospho-Ser9, the phospho-priming binding site of GSK-3β, was absent. We investigated whether such inactivation contributes to the HBx-enhanced AR activation. A reporter assay showed that GSK-3β(S9A) had approximately 50% of the HBx-enhanced AR activation (Fig. 4B, lane 5 versus 6), suggesting its involvement in the activation process.

Figure 4.

GSK-3β activity is involved in regulating the HBx-enhanced AR N–C interaction but not the AR NTD activation. (A) The presence of HBx can enhance the phosphorylation of GSK-3β (at Ser9), indicating its inhibitory effect on GSK-3β activity. The cell lysates were harvested for western blot analysis using the antibodies indicated. (B) The 5XARE-Luc reporter assay in HepG2 cells showed that the HBx-enhanced AR activation was reduced partly by transfecting the constitutively active form of GSK-3β, GSK-3β(S9A). The effect of the constitutively active form of GSK-3β(S9A) on the (C) AR N–C interaction and (D) AR NTD activation was analyzed using a mammalian two-hybrid and a one-hybrid assay. The results are the average of three experiments (mean ± SD).

We tried again to dissect the role of GSK-3β inactivation in regulating the two critical steps of ligand-stimulated AR activation, the AR N–C interaction and AR NTD activation. The mammalian two-hybrid assay indicated that GSK-3β(S9A) significantly down-regulates the HBx-enhanced AR N–C interaction (Fig. 4C, lane 5 versus 6). In contrast, GSK-3β(S9A) did not affect the HBx-enhanced AR NTD activity (Fig. 4D, lane 5 versus 6), suggesting that this step of AR activation is not regulated by the inactivation of GSK-3β. We speculate that inactivation of GSK-3β also contributes to the HBx-enhanced AR activation, which contributes mainly by increasing the AR N–C interaction.

c-Src and GSK-3β Play a Major Regulatory Role in Switching on the HBx-Enhanced AR Activation In Vitro and In Vivo.

To evaluate the contribution of c-Src activation and GSK-3β inactivation in the process of HBx-enhanced AR activation, we turned off each kinase switch either individually or in combination and evaluated the resulting effect on the 5XARE-Luc reporter activity. First, in the cell culture system, each kinase switch was turned off individually by treating with PP2 (for blocking the c-Src activity) and by transfecting with GSK-3β(S9A) (for activating the GSK-3β activity). The reporter activity was down-regulated by turning off each switch, although it remained higher than that of the ligand-stimulated AR (in the absence of HBx) (Fig. 5A, lanes 8 and 9 versus lane 2). Notably, turning off both switches in combination markedly down-regulated the reporter activity to a level lower than that of the ligand-stimulated AR (in the absence of HBx) (Fig. 5A, lane 10 versus 2). These results support the idea that the two kinases are the major switches that cooperatively turn on the HBx-enhanced AR activation.

Figure 5.

c-Src and GSK-3β play major regulatory roles as switches to turn on the HBx-enhanced AR activation both in vitro and in vivo. (A) In HepG2 cells, the c-Src and GSK-3β switches were turned off by PP2 and by transfecting with the constitutively active form of GSK-3β(S9A), either individually or in combination. The HBx-enhanced AR activation was decreased partly by turning off each switch and was decreased dramatically when both switches were turned off in combination. (B) The transactivation of AR in liver tissues between HBx transgenic mice (n = 6) and wild-type mice (n = 8) at the age of 2.5 months were compared by measuring the 5XARE-Luc activity. (C) The lysates from liver tissues at different ages of HBx transgenic mice were processed for western blot analysis hybridization with the antibodies indicated. As a control, the liver tissue from a wild-type C57BL/6 mouse (aged 2 months) was included in the analysis. (D) The effect of turning off each switch on the 5XARE-Luc activity was measured in primary hepatocytes isolated from HBx transgenic mice at ages of 2, 3, and 5 months. As a control, the liver tissues from wild-type mice at ages of 2 and 5 months were included in the reporter activity. The results are the average of three experiments (mean ± SD).

Because our conclusions so far were based on studies of cell cultures, we next tried to evaluate whether such regulation also occurs in hepatocytes in vivo. We used the HBx transgenic (tg) mouse model, in which more than 95% of male mice exhibit HCC at the age of 14 to approximately 18 months.8 We used hydrodynamic injection of 5XARE-Luc reporter construct into the tail vein and measured the reporter activity in lysates of the livers of wild-type and HBx tg male mice. The transcriptional activity of the AR was significantly higher in HBx tg male mice than in the wild-type mice (Fig. 5B). We next examined the activation status of c-Src and GSK-3β in liver tissues collected from male HBx tg mice of different ages, by hybridization with antibodies against phosphorylation-specific c-Src and GSK-3β. Compared with the wild-type mice, c-Src activation and GSK-3β inactivation occurred in the livers of HBx tg male mice during carcinogenesis (Fig. 5C, lane 1 versus lanes 2-7). This HBx tg mouse model reproduced the association between higher AR activity and the switching on of c-Src activation and GSK-3β inactivation.

To examine further the contribution of c-Src activation and GSK-3β inactivation in regulating the increased AR activity identified in HBx tg male mice, we isolated the hepatocytes from HBx tg male mice at the ages of 2, 3, and 5 months for primary culture. The hepatocytes were transfected with the 5XARE-Luc reporter in combination with transfection by GSK-3β(S9A) or were treated with PP2 as indicated in Fig. 5D. The reporter activity was significantly higher in hepatocytes isolated from HBx tg mice than in cells isolated from wild-type mice (Fig. 5D, lanes 3, 7, and 11 versus lanes 1 and 2), which is consistent with the results of the hydrodynamic injection experiment. The reporter activity was down-regulated by turning off each switch individually and even more dramatically by turning off both switches in combination (Fig. 5D, lanes 4, 5, and 6 versus lane 3; lanes 8, 9, and 10 versus lane 7; and lanes 12, 13, and 14 versus lane 11). This trend was similar to that identified in the cell culture system. The results from these primary hepatocytes support the critical involvement of these two switches regulated by c-Src and GSK-3β in controlling HBx-enhanced AR activation in vivo in HBx tg mice.

Discussion

We have shown previously that HBx can enhance the transcriptional activity of the AR in a ligand-dependent manner.7 Our current study explored the underlying molecular mechanisms by focusing on the two steps involved in regulating the ligand-stimulated AR activation: AR NTD activation and AR N–C interaction. Interestingly, we found that HBx can enhance both steps through different mechanisms that are regulated by two switch kinases: the c-Src kinase, which regulates the AR NTD activity, and GSK-3β kinase, which regulates the AR N–C interaction (This is shown schematically in Fig. 6).

Figure 6.

A proposed model illustrating the potential role of c-Src and GSK-3β kinases as key switches in regulating the HBx-induced AR activation. HBx can enhance the AR N–C interaction by inactivating GSK-3β and enhance the AR NTD activity by activating c-Src activity. These two steps are critical for the ligand-stimulated AR transcriptional activity. HSP, heat shock protein.

Our previous study, in which we treated the cells with the c-Src inhibitor PP2, showed that c-Src kinase participates in HBx-enhanced AR activation.7 Our current study using an siRNA approach to knock down the c-Src expression confirmed our previous data. We found the c-Src contributes mainly to the regulation of the AR NTD activation instead of the AR N–C interaction. By mutating the residue of Ser81 to Ala, we found that the HBx-enhanced phosphorylation at the AR Ser81 contributes to the increased level of AR protein in a ligand-dependent manner, which may lead to the higher transcriptional activity of the AR.

The kinase responsible for phosphorylating the AR at Ser81 has been identified recently as Cdk1, which functions mainly in the nucleus.20 Kesler et al.21 reported that compartmentalization of the AR determines its phosphorylation status. The androgen-stimulated phosphorylation at Ser81, Ser256, and Ser308 was significantly lower when the AR was forced to localize in the cytoplasm.21 This implies that phosphorylation of the AR at Ser81 occurs preferentially in the nucleus by action of its putative kinase Cdk1. We found that, in the presence of ligand, ARS81A existed in a less-phosphorylated form of AR protein compared with the wild-type AR; this effect was even more evident in the presence of HBx (Fig. 3C). We propose that HBx, through c-Src, mediates an increase in the phosphorylation at the AR Ser81, which regulates its retention in the nucleus and leads to higher transcriptional activity for its target promoters. We are currently investigating this possibility. It is still unclear how c-Src influences the downstream signaling pathway or pathways targeted by Cdk1, and how this leads to increased phosphorylation of the AR Ser81.

A recent publication reported that c-Src activity contributes to AR transactivation by stimulating specific AR phosphorylation at Y534, which is critical for the carcinogenesis of androgen-refractory prostate cancers.18 Phosphorylation at Y534 occurs preferentially in the androgen-independent carcinogenic process, in which some growth factors are activated abnormally. This abnormal activation contributes to the activation of c-Src, which leads to AR Y534 phosphorylation and in turn accelerates the entrance of the AR into the nucleus in a ligand-independent manner. However, this mechanism does not seem to contribute to HBx-enhanced AR activation because the ARY534F had no detectable effect on HBx-enhanced AR activation. In addition, when investigating whether c-Src is activated in the HBx-enhanced AR activation process, we noted a mild elevation of c-Src activity in our transient transfection assay (data not shown). This activation was further supported by the HBx tg mouse HCC model, which showed increased phosphorylation of c-Src-Y416 in the liver tissues from male mice during carcinogenesis (Fig. 5C). c-Src activity is elevated in a significant proportion of primary HCC tissues,22, 23 supporting the idea that this regulatory mechanism occurs in human hepatocarcinogenesis.

In the current study, the HBx-enhanced AR N–C interaction has been further verified by measuring the fluorescence resonance energy transfer activity between fluorophores attached to the NTD and ligand-binding domain of the AR using the assay developed by Schaufele et al.24 (data not shown). This event is not regulated by c-Src activity but instead is mediated by inactivation of GSK-3β. Several lines of evidence from human and mouse studies show that phospho-GSK-3β (at Ser9) is overexpressed in HCC tissues compared with normal liver tissues, suggesting that GSK-3β is generally inactivated in the hepatocarcinogenic process.25, 26 We found evidence of inactivation of GSK-3β in both the cell culture system and in the HBx tg male mouse liver tissues, which supports the role of inactivation of the enzyme in the HBx-mediated carcinogenic process in vivo. We have further tried to test whether the effect of GSK-3β on the AR N–C interaction is mediated through the accumulation of β-catenin, which occurs when GSK-3β is inactivated.14, 15 We have knocked down the expression of β-catenin but found no detectable effect on the resulting AR N–C interaction (data not shown). Therefore, some other unidentified mechanisms that underlie the regulation remain to be addressed.

The results of our current study support the critical role of c-Src activation and GSK-3β inactivation in HBx-enhanced AR transcriptional activation, although the possibility that some other regulatory mechanisms exists besides these two cannot be excluded. We propose that these two kinases are potential targets for prevention and therapy of HBV-related male HCC. Some current drugs have been designed for targeted therapy by inhibiting c-Src activity.27 However, a drug that activates GSK-3β is not available yet. One potential approach for this purpose is to activate GSK-3β by inactivating the upstream kinases that regulate the level of GSK-3β in the cell, such as extracellular signal–regulated protein kinase or phosphoinositide 3-kinase/Akt; these two kinases are known to inactivate GSK-3β under different pathological conditions.28, 29 Our preliminary results demonstrate the feasibility of this approach of inhibiting phosphoinositide 3-kinase/Akt and c-Src at the same time to almost totally abolish the HBx-enhanced AR activation (data not shown). The efficacy of such combination therapy in prevention or therapy of HBV-related male HCC can be tested in the HBx tg mouse HCC model.

Actually, in addition to these two kinase switches, the AR protein itself certainly can also be considered as one potential candidate for the targeted therapy of male HCCs. Especially we noted that HCV can also activate the AR pathway, through its activation of signal transducer and activator of transcription 3.30 Ma et al.31 well demonstrated the effective suppression of HCC by targeting AR in the DEN (N′-N′-diethylnitrosamine)-induced HCC mouse model.31 However, in this animal model, knockout of AR in liver can only delay but not block the tumor occurrence, suggesting that some other factors besides AR contribute to the carcinogenic process. Intriguingly, estrogen pathway was identified as a critical factor responsible for the gender disparity in this model. It can inhibit the release of interleukin-6 from the Kuppfer cells, which is critical for tumor formation.32 The protective role of estrogen pathway in hepatocarcinogenesis has further been supported by our recent finding that microRNA-18a–mediated suppression of ERα (estrogen receptor alpha) can promote the proliferation of hepatoma cells.16 Therefore, in addition to AR, some other factors such as the estrogen receptor, also needs to be included for consideration of the effective treatment strategy.

Ancillary