Enhanced cell survival of Hep3B cells by the hepatitis B x antigen effector, URG11, is associated with upregulation of β-catenin

Authors

  • Zhaorui Lian,

    1. Department of Pathology, Anatomy and Cell Biology, Thomas Jefferson University, Philadelphia, PA
    Search for more papers by this author
    • Zhaorui Lian and Jie Liu contributed equally to this work.

  • Jie Liu,

    1. Department of Pathology, Anatomy and Cell Biology, Thomas Jefferson University, Philadelphia, PA
    2. Department of Digestive Diseases, Xijing Hospital, Fourth Military Medical University, Xian, China
    Search for more papers by this author
    • Zhaorui Lian and Jie Liu contributed equally to this work.

  • Li Li,

    1. Department of Pathology, Anatomy and Cell Biology, Thomas Jefferson University, Philadelphia, PA
    2. Department of General Surgery, Chengzheng Hospital, Shanghai, China
    Search for more papers by this author
  • Xianxing Li,

    1. Department of General Surgery, Chengzheng Hospital, Shanghai, China
    Search for more papers by this author
  • Marcy Clayton,

    1. Department of Pathology, Anatomy and Cell Biology, Thomas Jefferson University, Philadelphia, PA
    Search for more papers by this author
  • Meng-Chao Wu,

    1. Shanghai Eastern Hospital & Institute of Hepatobiliary Surgery, Second Military Medical University, Shanghai, China
    Search for more papers by this author
  • Hong-Yang Wang,

    1. Shanghai Eastern Hospital & Institute of Hepatobiliary Surgery, Second Military Medical University, Shanghai, China
    Search for more papers by this author
  • Patrick Arbuthnot,

    1. Molecular Hepatology Research Unit, Department of Medicine, University of the Witwatersrand, Johannesburg, South Africa
    Search for more papers by this author
  • Michael Kew,

    1. Molecular Hepatology Research Unit, Department of Medicine, University of the Witwatersrand, Johannesburg, South Africa
    Search for more papers by this author
  • Daiming Fan,

    1. Department of Digestive Diseases, Xijing Hospital, Fourth Military Medical University, Xian, China
    Search for more papers by this author
  • Mark A. Feitelson

    Corresponding author
    1. Department of Pathology, Anatomy and Cell Biology, Thomas Jefferson University, Philadelphia, PA
    2. Department of Microbiology and Immunology, Kimmel Cancer Center, Thomas Jefferson University, Philadelphia, PA
    • Room 222 Alumni Hall, Department of Pathology, Anatomy and Cell Biology, Thomas Jefferson University, 1020 Locust Street, Philadelphia, PA 19107
    Search for more papers by this author
    • fax: 215-503-9982


  • Potential conflict of interest: Nothing to report.

Abstract

Intrahepatic expression of hepatitis B x antigen (HBxAg) is associated with the development of hepatocellular carcinoma (HCC), perhaps through trans-activation of selected cellular genes. When this was examined by PowerBlot analysis, upregulated levels of β-catenin and several known β-catenin effectors were observed in HBxAg-positive compared with HBxAg-negative HepG2 cells. When HBxAg was introduced into Hep3B cells, upregulated expression of wild-type β-catenin was observed. This was also observed in Hep3B cells overexpressing the HBxAg upregulated gene, URG11. Upregulated expression of URG11 and β-catenin correlated with HBxAg trans-activation function. Transient transfection assays with fragments of the β-catenin promoter showed that it was activated by both HBxAg and URG11 and inhibited by URG11-specific small inhibitory RNA. The latter also inhibited the growth of Hep3BX cells in a serum-free medium, which correlated with depressed levels of β-catenin. Activation of β-catenin effector genes was observed in cells stably expressing HBxAg or overexpressing URG11 compared with control cells transfected with the pTOPFLASH reporter plasmid. Extensive costaining between HBxAg, URG11, and β-catenin was observed in infected liver and HCC nodules, suggesting a close relationship in vivo. In conclusion, wild-type β-catenin is activated by HBxAg, in part, through the upregulated expression of the HBxAg effector URG11. URG11 stimulates the β-catenin promoter and hepatocellular growth and survival. These observations also suggest that URG11 may be a regulatory element in the β-catenin signaling pathway and may be a target for chemoprevention of HCC. (HEPATOLOGY 2006;43:415–424.)

Hepatitis B virus (HBV) carriers are at high risk for the development of hepatitis, cirrhosis, and hepatocellular carcinoma (HCC).1, 2 HCC is among the most frequent tumor types worldwide, with up to 250,000 new cases per year.3 HBV makes a genetic contribution to tumor development4 by expressing hepatitis B x antigen (HbxAg).5 HBxAg expression correlates with the intensity of liver disease,6, 7 transforms liver cells in vitro,8 while sustained high levels of HBxAg expression in transgenic mice gives rise to HCC.9, 10

HBxAg is a trans-activating protein.11 Recently, several natural effectors of HBxAg have been identified that promote hepatocellular growth and survival.12–15 One of these proteins, the upregulated gene URG11, stimulated HepG2 growth and survival in vitro and strongly accelerated tumor formation in vivo,14 suggesting that URG11 may be an oncogene or may stimulate expression of oncogenes that are important for tumor development.

The stabilization and intracellular accumulation of β-catenin, which often result from mutations in β-catenin or in β-catenin–binding proteins, contribute to the development of many tumor types.16, 17 This results in constitutive Wnt signaling, where β-catenin translocates to the nucleus and stimulates the expression of genes that promote tumorigenesis.18, 19 β-Catenin mutations are found in benign tumors of the colon20 and in small HCCs and preneoplastic liver,21, 22 suggesting they occur early in tumor development. The finding of frequent β-catenin mutations in a subset of human HCC23, 24—especially in HBV-negative tumors,25 but not in HCC nodules from ground squirrels or woodchucks chronically infected with HBV-like viruses26—implies that if activated β-catenin participates in the development of HCC, it must occur by mechanisms other than mutation. One possibility is that HBxAg upregulates expression of β-catenin. The results of the present study support this hypothesis and show that β-catenin is a target of the HBxAg effector, URG11, suggesting that URG11 promotes hepatocellular growth and survival by upregulated expression/stabilization of β-catenin.

Abbreviations

HBxAg, hepatitis B x antigen; HCC, hepatocellular carcinoma; HBV, hepatitis B virus; CAT, chloramphenicol acetyltransferase; siRNA, small inhibitory RNA; PCR, polymerase chain reaction; GSK3β, glycogen synthase kinase 3β; mRNA, messenger RNA; TCF, T cell factor.

Materials and Methods

Construction of HBxAg-Positive and Control Cultures.

The human hepatoblastoma cell line, HepG2, was transduced with a recombinant retrovirus encoding HBxAg or the bacterial chloramphenicol acetyltransferase (CAT). The resulting cultures, designated HepG2X and HepG2CAT, were selected in G418, passaged, and analyzed without selection of individual clones.27 Likewise, HBxAg was introduced and stably expressed in Hep3B and Huh7 cells.

HBxAg Mutant Expressing HepG2 and Hep3B Cell Lines.

Full-length HBx subcloned into pcDNA3 was used to make mutants lacking 10, 20, 30, or 40 amino acids from the carboxy- or amino-terminal ends, as previously described.27

Subcloning of URG11.

URG11 was previously subcloned into pcDNA3.14

PowerBlot Analysis.

To discern patterns of differentially expressed proteins in HepG2X compared with HepG2CAT cells, cell lysates were freshly prepared and shipped to the contracting company (BD Transduction Labs, BD Biosciences, San Jose, CA) for PowerBlot analysis. Cell lysates were analyzed on sodium dodecyl sulfate/polyacrylamide gels and then transferred to immobilon P membranes. The membranes were cut into strips, and each strip was blotted with a mixture of monoclonal antibodies. In the entire analysis, Western blotting was performed for 750 proteins. Images were captured electronically and were matched using PDQuest software (Biorad, Hercules, CA). The results were returned to the laboratory as images and Excel tables on CD.

Patient Populations.

Paraffin blocks of paired tumorous (HCC) and nontumorous (liver) from 12 South African patients and 214 Chinese patients with HBV-associated HCC27 were used for staining. In all cases, livers with tumors had HCC, while livers without tumors had cirrhotic foci and/or hepatitis. Among Chinese carriers, there were blocks from 155 males and 59 females aged 13-75 years (mean age: 55). Among these, 23 patients had grade I HCC, 126 had grade II, 36 had grade III, and 29 had grade IV. The mean tumor diameter was 5.8 cm (range: 1-16 cm). Cirrhosis was found in 192 patients. Samples of uninfected human liver from 2 individuals were available as controls. Use of all tissues was approved by the institutional review boards at Thomas Jefferson University and collaborating institutions based on informed consent.

Immunohistochemical Staining, Northern and Western Blotting.

Staining for HBxAg and URG11 was performed using previously characterized rabbit anti-HBx and anti-URG11.14 Staining for β-catenin was performed using a mouse monoclonal β-catenin antibody (E-5; Santa Cruz Biotechnology, Santa Cruz, CA) at a 1:50 dilution. Staining was performed with each antibody on consecutive sections using the DAKO EnVision Plus System (DAKO, Carpenteria, CA). Detection was performed using the DAKO Liquid DAB+ kit (DAKO). Slides were counterstained with hematoxylin (Sigma Chemical Co., St. Louis, MO). For negative controls, an equivalent dilution of normal mouse immunoglobulin G (Vector Labs, Burlingame, CA) or rabbit preimmune serum were used in place of primary antibody. The specificity of HBxAg staining was also assessed by preabsorption of the primary antibody with an excess of immunizing antigen (X peptide) before staining and by staining uninfected livers.6, 27 For URG11 and β-catenin, uninfected liver sections were also stained with corresponding primary antibodies or normal rabbit or mouse immunoglobulin G, respectively.

Staining for HBxAg, β-catenin, and URG11 was evaluated as previously described.28 Staining intensity was scored as: weak (+1), moderate (+2), and strong (+3). The percentage of positive cells was evaluated as: up to 10% positive cells (+1), 11%-50% positive cells (+2), 51%-80% positive cells (+3), and > 81% positive cells (+4). The final score for each sample was the sum of the staining intensity plus the percentage of positive cells.

Proliferation in tumors was evaluated by staining with a mouse anti-human proliferating cell nuclear antigen (PC-10; Santa Cruz Biotechnology) (1:800 dilution). The proliferative index is the ratio of positively stained cells to the total cell count expressed as a percentage. At least 1,000 nuclei were counted for each sample.

Northern blot analysis of URG11 and β-catenin messenger RNA was performed as previously described.27 Probes for GAPDH and URG11 have been previously described as well.14 The probe for β-catenin was a fragment of β-catenin complementary DNA (encoding amino acids 350-781 and cloned into pcDNA3), which was a kind gift from Dr. Setsuo Hirohashi (National Cancer Center, Tokyo, Japan).29

Western blotting was performed with anti-HBx or anti-URG1114 and with anti–β-catenin (Santa Cruz Biotechnology) at a dilution of 1:200. For secondary antibodies, horseradish peroxidase–conjugated goat anti–rabbit immunoglobulin (1:10,000 dilution) or horseradish peroxidase–conjugated goat anti–mouse immunoglobulin (1:3,000 dilution) (Accurate, Westbury, NY) were used. The results were visualized using the ECL detection system (Amersham, Arlington Heights, IL). Mouse anti–human β-actin monoclonal antibody (Clone AC-15, Sigma) was used at 1:5,000 as an internal control.12 For the detection of activated β-catenin by Western blotting, the primary antibody used was specific for the active form of β-catenin (Upstate Cell Signaling Solution, Lake Placid, NY) used at a 1:500 dilution.

Small Inhibitory RNA Transfection and Growth Curves in Medium Containing 10% or 0% Fetal Calf Serum.

Hep3B cells (1 × 105 cells) stably expressing HBxAg or URG11 were seeded overnight, in triplicate, into 60-mm dishes. Cells were transiently transfected with small inhibitory RNAs (siRNAs) against URG11 or irrelevant siRNAs using oligofectamine (Invitrogen, Carlsbad, CA). For transfection, 20 μL of a 20-μmol/L stock of each of 2 URG11-specific siRNAs (residues 420-438, CAGACGGAUUGCUGUACUU; and residues 1385-1403, ACACAGACUUUACCUACAA) or 2 irrelevant siRNAs (which have sequences different from any cellular gene) were used (Dharmacon, Lafayette, CO). Cells were incubated for 4 hours at 37°C and then grown in complete or serum-free medium. The number of viable cells was determined daily for up to 4 days by trypan blue staining and the modified tetrazolium salt assay (Cell Titer 96 Non-radioactive Promega). Parallel cultures were analyzed by flow cytometry at the Kimmel Cancer Center of Thomas Jefferson University.

HBxAg Trans-Activation Assay.

The levels of HBxAg trans-activation were determined exactly as previously described.27

β-Catenin Promoter and Activity Assays.

To test for β-catenin promoter activity, 3 fragments of the β-catenin promoter cloned into pSEAP-basic (a gift from Dr. Frans van Roy, University of Ghent)30 were used for transient transfections into Hep3B cells. The fragments were designated as FRAG 1 (−103 to +41), FRAG 2 (−298 to +139), and FRAG3 (the 6-kb fragment adjacent to the 5′ end of the gene). Promoter activity was evaluated by measuring secreted human placental alkaline phosphatase at 48 hours after transfection using a commercially available Phospha-Light chemiluminescent report gene assay.31 To measure β-catenin activity, cells were transiently transfected with T cell factor (TCF) reporter plasmids pTOPFLASH (which is β-catenin responsive) or pFOPFLASH (a mutant that is β-catenin unresponsive) (Upstate Technology, Charlottesville, VA). Luciferase activity was determined after 48 hours.

β-Catenin Exon 3 Sequencing.

Genomic DNA from microdissected tumor tissue from 50-μm sections was extracted using the DNeasy tissue kit (QIAGEN, Valencia, CA). β-Catenin exon 3 was amplified by PCR in all samples. The primers were 5′-ATTTGATGGAGTTGGACATGG-3′ (sense; primer 1) and 5′- TGTTCTTGAGTGAAGGACTGA-3′ (antisense; primer 2). For samples that did not give a sharp band following PCR amplification, seminested PCR was performed using the inner primer 5′-TCTTCCTCAGGATTGCCTT-3′ (antisense; primer 3) with primer 1 above. PCR amplification was performed using the Advantage 2 Polymerase Kit (BD Biosciences/Clontech, Palo Alto, CA) according to the manufacturer's instructions. The PCR conditions were as follows: 95°C for 1 minute, 95°C for 20 seconds, 68°C for 1.5 minutes for 30 cycles, and 68°C for 2 minutes. PCR products were purified using the MiniElute PCR purification kit (QIAGEN), then sequenced at the DNA sequencing facility at the Kimmel Cancer Center of Thomas Jefferson University. Each sample was sequenced in both directions, and 2 separate PCR products were examined for each sample in which a mutation was found.

Proteasome Inhibition.

Hep3BX, Hep3B-URG11, and Hep3B-pcDNA3 cells were treated with the proteasome inhibitor, MG132 (Sigma) at a final concentration of 20 μmol/L for 6 hours prior to analysis for β-catenin levels in cell lysates.

Statistics.

The growth of Hep3BX cells treated with siRNAs, activation of the β-catenin promoter fragments, β-catenin activity, and experiments involving flow cytometry were evaluated using the Student t test. Statistical analysis and graphic presentation were performed using SPSS version 10.0 software (SPSS Inc., Chicago, IL). Spearman rank correlation coefficient analysis was used to assess the correlation between ordinal variables. The χ2 test was used for testing relationships between categorical variables as appropriate. A P value of less than .05 was considered significant.

Results

Upregulated Expression of β-Catenin in HepG2X Cells.

When lysates from HBxAg-positive and HBxAg-negative cells were subjected to PowerBlot analysis, β-catenin expression was elevated by 4.5 ± 0.5–fold in HepG2X compared with HepG2CAT cells compared with endogenous β-actin (Fig. 1). In addition, c-myc was upregulated 4.7 ± 0.7–fold, cyclin D1 was upregulated by 6.6 ± 1.1–fold, and cyclo-oxygenase-2 was upregulated by 5.9 ± 0.6–fold in HepG2X compared with HepG2CAT cells (Fig. 1). This suggests that HBxAg expression is associated with elevated levels of β-catenin and several β-catenin effectors that are known to be upregulated in HCC.32–35

Figure 1.

PowerBlot analysis of (A) HepG2X cells and (B) HepG2CAT cells. The results show the expression of β-catenin, cyclin D1, cyclo-oxygenase-2, and c-myc. The results from 1 of 3 experiments are shown. COX-2, cyclo-oxygenase-2.

Expression of HBxAg and β-Catenin In Vivo.

To test whether the expression of β-catenin is up-regulated in vivo, tumorous and adjacent nontumorous liver samples from 214 Chinese and 12 South African patients were stained for HBxAg, and consecutive slides were stained for β-catenin. In tumor samples from Chinese patients, HBxAg was detected in 144 (67%) and β-catenin in 171 (80%) (Table 1). Costaining was observed between HBxAg (Fig. 2A) and β-catenin (Fig. 2B) in 133 cases (62%) (P < .01) (Table 1), although HBxAg staining was predominantly cytoplasmic, while β-catenin staining was membranous, cytoplasmic, and nuclear. Staining with preimmune rabbit serum did not result in any signal (Fig. 2C). HBxAg staining was also blocked by preincubating anti-HBx with the HBx synthetic peptides used for immunization (data not shown). Nuclear β-catenin was observed in 48 (33%) of HBxAg-positive tumors, but in only 7 (9%) of HBxAg-negative tumors (P < .01), implying that HBxAg expression in tumors is associated with the activation of β-catenin in vivo.36

Table 1. Summary of Results for URG11, β-Catenin, and HBxAg Staining
PopulationNo. of PatientsNo. PositiveCostainingCorrelation CoefficientP Values
Xβ-CatURG11X+X+URG11+X+X+URG11+X+X+URG11+
   β-CatURG11β-Catβ-CatURG11β-Catβ-CatURG11β-Cat
  1. Abbreviations: X, HBxAg; β-Cat, β-catenin.

South African12            
 Tumorous 81088680.6650.6650.671<.05<.05<.05
 Nontumorous 1010111010100.7720.8790.878<.01<.01<.01
Chinese214            
 Tumorous 1441711431331271420.6640.7970.727<.01<.01<.01
 Nontumorous 1741711801641641650.6320.6290.656<.01<.01<.01
Figure 2.

Immunohistochemical staining with (A,E) anti-HBx, (B,F) anti–β-catenin, (C,G) preimmune serum, or (D,H) anti-URG11 of consecutive sections of (A-D) tumor and (E-H) liver with cirrhosis from an HBV carrier. Inserts in the upper left-hand corner of selected panels are higher magnification images of each corresponding panel. (I) Staining of uninfected liver with anti–β-catenin (left panel), anti-URG11 (middle panel) or anti-HBx (right panel). (J) Proliferating cell nuclear antigen staining of HCC samples from β-catenin–positive (left) and β-catenin–negative (right) patients. HBx, hepatitis B x; HCC, hepatocellular carcinoma; PCNA, proliferating cell nuclear antigen.

In nontumorous samples from Chinese patients, HBxAg was detected in 174 (81%) and β-catenin in 171 (80%) (Table 1). Again, HBxAg staining (Fig. 2E) correlated with that of β-catenin (Fig. 2F) in 164 cases (P < .01). Staining with preimmune rabbit serum did not result in any signal (Fig. 2G).

HBxAg was detected in 8 (67%) HCC samples from South African patients. All HBxAg-positive tumors exhibited cytoplasmic and nuclear β-catenin (P < .05). In nontumorous livers, cytoplasmic HBxAg was accompanied by strong, membranous β-catenin staining in 10 of 12 cases (P < .01) (Table 1). Hence, there was costaining between HBxAg and β-catenin in both liver and tumor compartments, with activation of β-catenin seen predominantly in tumorous tissue in samples analyzed from 2 different populations.

In uninfected liver, β-catenin staining showed a weak, exclusively membranous distribution (Fig. 2I, left panel), while HBxAg staining was undetectable (Fig. 2I, right panel) as previously described.7, 13–16, 30

To determine whether upregulated β-catenin in tumors correlated with proliferation, consecutive tumor sections from both groups of patients were stained for proliferating cell nuclear antigen. Proliferating cell nuclear antigen staining was detected in 83 ± 10.8% of β-catenin–positive HCC, but only in 50 ± 12.3% of β-catenin–negative HCC (P < .01), suggesting that activated β-catenin is associated with accelerated tumor growth. An example of the staining differences is presented in Fig. 2J.

Given that the cytoplasmic/nuclear localization of β-catenin in HCC may be associated with mutations in the glycogen synthase kinase 3β (GSK3β) binding site within exon 3 of the β-catenin gene,24 PCR amplication of exon 3 from the paraffin blocks of each tumor, followed by DNA sequencing, was performed. The results showed wild-type β-catenin exon 3 sequences in all tumors, except in patient 4 from South Africa, who had a GAT → GCT mutation within codon 43. The finding that the GSK3β site mutations occur in approximately 20% of HCCs,25 especially among HBV-negative HCCs, and that mutations in other molecules of the β-catenin signaling pathway are rare,37 suggests that HBxAg activates wild-type β-catenin.

Association of Activated β-Catenin With URG11.

HBxAg upregulates genes that promote cell growth, survival, and/or tumorigenesis.12–15 To examine whether any of these activate β-catenin, each was stably transfected into Hep3B cells, which encodes only wild-type β-catenin. In contrast, HepG2 makes wild-type and an activated mutant of β-catenin.38 Hep3B cultures were then stained for β-catenin and analyzed for β-catenin messenger RNA (mRNA) and protein. β-Catenin was upregulated in Hep3BX cells or in Hep3B cells overexpressing URG11 compared with control cells (Fig. 3A -C). This was verified by Western (Fig. 3D) and Northern (Fig. 3E) blotting. Hence, URG11 is associated with increased steady-state levels of β-catenin mRNA and protein.

Figure 3.

Detection of β-catenin in Hep3B cells. Staining for β-catenin in Hep3B cells transfected with (A) pcDNA3-HBx, (B) pcDNA3-URG11, or (C) pcDNA3. (D) Western blot analysis of β-catenin in Hep3B cells and (E) corresponding Northern blot analysis of RNA extracted from Hep3B cells stably transfected with pcDNA3 (Con), pcDNA3-HBx (HBx), or pcDNA3-URG11 (URG11). (D) Protein signals were normalized to β-actin, whereas (E) mRNA ratios were normalized to GAPDH mRNA. Con, control; HBx, hepatitis B x; G3PDH, glyceraldehyde 3-phosphate dehydrogenase.

To determine whether this relationship exists in vivo, the two were compared by staining of HCC and nontumorous liver sections. URG11 staining was observed in 143 (67%) of Chinese tumor samples, and costaining with β-catenin was observed in 142 (66%) (P < .01) (Table 1). In adjacent nontumorous liver, URG11 staining was observed in 180 (84%) cases, and costaining with β-catenin was observed in 165 (77%) (P < .01) (Table 1). Examples of URG11 staining in tumorous and nontumorous livers are presented in Fig. 2D and 2H, respectively. Similar costaining was also seen in South African samples (Table 1). URG11 staining in normal, uninfected livers was spotty, cytoplasmic, and weak (Fig. 2I, middle panel) compared with tumorous (Fig. 2D) and infected (Fig. 2H) livers. Hence, there appears to be a link between URG11 and elevated wild-type β-catenin in two different populations with a high incidence of HBV-associated HCC.

β-Catenin is a Target of HBxAg trans-Activation.

To test whether β-catenin is trans-activated by HBxAg, HepG2 and Hep3B cells were stably transfected with different truncation mutants of HBxAg (Fig. 4A). Expression of the recombinant HBxAg polypeptides was verified by Western blotting (Fig. 4B). The trans-activation activity of each mutant was then assessed by reporter gene assay (Fig. 4C). When Northern blot analysis was performed for URG11 and β-catenin mRNAs (Fig. 4D), there was a direct correlation between HBxAg trans-activation function and the mRNA levels of URG11 and β-catenin. Although performed in HepG2 cells, similar results were obtained in Hep3B cells (data not shown). Hence, URG11 and β-catenin are targets for HBxAg trans-activation, which is consistent with their coexpression in vivo.

Figure 4.

Characterization of HBxAg mutants. (A) Diagram of each HBx mutant indicating the first and last amino acids at the ends of the coding region (line) and their corresponding lane in panels B-D. (B) Western blots of HepG2 cells transfected with each of the HBxAg constructs in panel A. (C) Corresponding trans-activation activities of HBxAg polypeptides expressed in panel B. (D) Northern blot analysis of URG11 mRNA or β-catenin mRNA in the cultures expressing different HBxAg polypeptides (as in panel B). Northern blots were normalized to GAPDH mRNA in each lane. The ratios of each mRNA relative to vector transfected cells are listed below each blot. HBxAg, hepatitis B x antigen; HBx, hepatitis B x; mRNA, messenger RNA; G3PDH, glyceraldehyde 3-phosphate dehydrogenase.

To test whether β-catenin is a transcriptional target of HBxAg, Hep3B cells were transiently cotransfected with the pSEAP-FRAG1, 2 or 3 reporter gene constructs containing different portions of the β-catenin promoter plus pcDNA3, pcDNA3-HBx, or pcDNA3-URG11. The results show that HBxAg and URG11 stimulate the β-catenin promoter FRAG 2 and FRAG 3 (Fig. 5A). When URG11-specific siRNA was added, it blocked activation of the β-catenin promoter by HBxAg or URG11. Hence, stimulation of the β-catenin promoter by HBxAg was URG11-dependent. This inhibition was specific, because irrelevant siRNA did not alter the ability of HBxAg or URG11 to stimulate the β-catenin promoter.

Figure 5.

β-Catenin gene and pathway activation. (A) Activity of the β-catenin promoter in Hep3B cells transiently cotransfected with pSEAP-FRAG1, 2 or 3, and with pcDNA3-HBx, pcDNA3-URG11, or pcDNA3 with or without siRNAs as indicated. Results are shown for 48 hours posttransfection and represent the average of 2 experiments, each done in triplicate. (B) Activity of the β-catenin reactive promoter pTOPFLASH (white bars) and control promoter pFOPFLASH (gray bars) in Huh7 cells transiently transfected with pcDNA3, pcDNA3-HBx, or pcDNA3-URG11 as indicated. The results represent the average of 3 independent experiments. (C) Levels of β-catenin in Hep3B-pcDNA3, Hep3BX, and Hep3B-URG11 cells in the absence (lanes 1-3) or presence (lanes 4-6) of the proteasome inhibitor MG132 as indicated. Ratios were normalized to β-actin. (D) Levels of activated β-catenin in Hep3B control cells (Con), Hep3BX cells (HBx), and Hep3B-URG11 cells (URG11) via Western blotting. Corresponding levels of GSK3β were also assayed (middle panel) and normalized to β-actin. The relative ratios, after normalization, are shown below each panel. HBx, hepatitis B x; siRNA, small inhibitory RNA; Con, control; GSK3β, glycogen synthase kinase 3β.

HBxAg and URG11 Stimulate β-Catenin Signaling.

Hep3B and Huh7 cells were transiently transfected with the β-catenin responsive TCF reporter plasmid to assess the activity of β-catenin signaling by HBxAg and URG11. The results show that HBxAg and URG11 significantly stimulate the transcriptional activity of a β-catenin reporter plasmid (Fig. 5B), suggesting elevated levels and nuclear translocation of β-catenin is also associated with elevated β-catenin signaling activity.

Proteasome Activity and Phosphorylation of β-Catenin.

To assess the contribution of the proteasome to β-catenin degradation, Hep3BX and Hep3B-URG11 cultures were treated with or without the protease inhibitor, MG132, and the cell lysates assayed for β-catenin by Western blotting. In the absence of MG132, β-catenin levels in Hep3BX and Hep3B-URG11 cells were five- to sixfold higher compared with Hep3B-vector control cells (Fig. 5C). When MG132 was added, β-catenin accumulated in control cells, but little further accumulation was observed in Hep3BX or Hep3B-URG11 cells (P < .005), suggesting that most of the β-catenin stabilization in Hep3B-URG11 and Hep3BX cells was due to proteasome inhibition.

To determine whether HBxAg, through URG11, results in activation of β-catenin, cell lysates were prepared from Hep3BX, Hep3B-URG11 and control cells, and then analyzed for activated β-catenin by Western blotting. The results show elevated levels of activated β-catenin in Hep3BX cells (6.3 ± 0.5) and Hep3B-URG11 cells (6.7 ± 0.76) compared with control cells (Fig. 5D). This was accompanied by significantly diminished levels of GSK3β in Hep3BX (0.21 ± 0.04) but not in Hep3B-URG11 cells (Fig. 5D), suggesting that one way in which HBxAg (but not URG11) activates β-catenin is through downregulated expression of GSK3β. When URG11-specific siRNA was cotransfected into Hep3BX cells, levels of activated β-catenin were near baseline (Fig. 6C), suggesting that HBxAg activation of β-catenin is URG11-dependent.

Figure 6.

Growth of Hep3BX cells in culture with (A) 10% serum or (B) 0% serum following transient transfection with URG11-specific siRNA (closed square) or control siRNA (open diamond). Each growth curve is the average of 3 experiments. (C) β-Catenin levels in Hep3BX or Hep3B-URG11 cells transiently transfected with irrelevant siRNA or URG11-specific siRNA as indicated. Western blotting was performed on cell lysates collected 48 hours after transfection. β-Actin was used for normalization. MTT assay, modified tetrazolium salt assay; siRNA, small inhibitory RNA.

URG11 Upregulated Expression of β-Catenin Promotes Growth in Serum-Free Medium.

To determine whether HBxAg-stimulated hepatocellular growth and survival is URG11-dependent, Hep3BX cells were transiently transfected with URG11-specific or irrelevant siRNA, and growth was determined in 10% or 0% serum. There was little difference in the growth of Hep3BX cells in 10% serum whether or not cells were treated with URG11 or irrelevant siRNAs (Fig. 6A). In 0% serum, Hep3BX growth was inhibited with URG11-specific but not irrelevant siRNA (Fig. 6B), suggesting that the ability of HBxAg to stimulate growth is partially dependent on the upregulated expression of URG11. When flow cytometry was performed on identically treated cells grown in 0% serum, 26% of URG11 siRNA transfected cells had less than 2N DNA content, while 19.8% of cells transfected with control siRNAs did (P < .02) (data not shown). Introduction of URG11-specific siRNA into Hep3BX cells also resulted in decreased levels of β-catenin while irrelevant siRNA did not (Fig. 6C). These results suggest a linkage between upregulated URG11, upregulated β-catenin, and enhanced hepatocellular growth. Similar results were obtained with Huh7-HBx and Huh7-URG11 cells (data not shown).

Discussion

HBxAg alters the expression of genes that promote growth, inhibit apoptosis, and/or stimulate tumorigenesis.12–15 This study suggests that one of these upregulated proteins is β-catenin, since it accumulates in HBxAg positive tissue culture cells (Figs. 1 and 3) and co-stains with upregulated β-catenin in HCC and surrounding nontumorous liver in HBV carriers (Fig. 2, Table 1). That this up-regulation contributes to tumor growth is suggested by the observation that proliferating cell nuclear antigen staining is prevalent in β-catenin positive compared to negative HCCs (Fig. 2J), and that HBxAg stimulates the growth of Hep3B cells in serum free medium is associated with elevated levels of β-catenin (Fig. 6). This compliments earlier work that Wnt-1 over-expression cooperates with HBxAg in stimulating β-catenin signaling in hepatoma cells.39 Hence, β-catenin appears to be a target for HBxAg in HCC.

Previous work has shown that β-catenin is stabilized by mutation in approximately 20% of HCCs,23–25 while mutations in other β-catenin–binding proteins are relatively rare.37 The finding of β-catenin mutation within the GSK3β site in only one patient suggests that HBxAg upregulates and/or stabilizes wild-type β-catenin. The upregulated expression of β-catenin by HBxAg in Hep3B cells (Fig. 3), which only encodes the wild-type allele, strengthens this conclusion. The fact that HBxAg transcriptionally activates the β-catenin promoter (Fig. 5A) and stabilizes β-catenin by suppressing GSK3β (Fig. 5D), thereby blocking proteosome-mediated degradation of β-catenin, are two pathways whereby HBxAg may promote stabilization and/or accumulation of β-catenin. Inhibition of the proteasome with MG132 had little impact upon the levels of β-catenin (Fig. 5C), suggesting that most of the β-catenin stabilization was due to proteasome inhibition, which is a known property of HBxAg.40 Hence, HBxAg upregulates wild-type β-catenin transcriptionally and stabilizes it posttranslationally.

HBxAg also transcriptionally activates URG11 (Fig. 4), which acts like an oncoprotein by promoting hepatocellular growth and tumorigenesis.14 URG11 costains with HBxAg and β-catenin in tumorous and infected livers (Fig. 2), implying that URG11 may mediate the stabilization and/or upregulation of β-catenin by HBxAg. This was supported by findings that URG11 transcriptionally activates the β-catenin promoter (Fig. 5A), resulting in the accumulation of β-catenin (Fig. 3), and that URG11 stimulates growth of Hep3BX cells in serum-free medium through the upregulated expression of β-catenin (Fig. 6). The fact that URG11-specific siRNA inhibits the ability of HBxAg to stimulate the β-catenin promoter (Fig. 5A), coupled with the ability of HBxAg to stimulate Hep3B growth in serum free medium (Fig. 6), suggests that URG11 activation of β-catenin is part of the mechanism whereby HBxAg contributes to HCC. In such a model, HBxAg would transcriptionally activate URG11 and downregulate GSK3β, while URG11 would transcriptionally activate β-catenin and promote β-catenin–dependent growth in serum-free medium. Hence, HBxAg and URG11 both target β-catenin by different mechanisms, which may short-circuit normal β-catenin signaling in the chronically infected liver.

Alterations in β-catenin expression/signaling occur early in the pathogenesis of human36, 41 and murine22, 23 HCC, as well as during tumor progression and metastasis.42 The fact that HBxAg, URG11, and β-catenin are coexpressed in the liver (Fig. 2), and that HBxAg and URG11 target β-catenin activation (Figs. 3–5), suggests that these events occur before tumor development. The findings that c-src and/or related kinases are activated in almost half of early HCC,43 and that their activation by HBxAg44 is associated with the disruption of adherens junctions,45, 46 suggests that HBxAg may function through some of the same pathways as receptor tyrosine kinases. Since the latter participate in β-catenin activation through phosphorylation, and tyrosine-phosphorylated (activated) β-catenin has been reported in HBxAg-positive compared with HBxAg-negative cells,45, 46 it is concluded that β-catenin is likely to be a target for HBxAg. If so, it will be important to see whether HBxAg (or URG11) alters the expression/activity of E-cadherin, receptor tyrosine kinases, or other proteins (e.g., c-met and frizzled-7) that normally bind β-catenin, and whether any of these could be considered targets for the chemoprevention of HCC.

Ancillary