Hepatitis B virus pre-S2 mutant upregulates cyclin A expression and induces nodular proliferation of hepatocytes


  • Hui-Ching Wang,

    1. Division of Clinical Research, National Health Research Institutes, Tainan, Taiwan
    2. Graduate Institutes of Basic Medical Sciences, College of Medicine, National Cheng Kung University, Tainan, Taiwan
    Search for more papers by this author
  • Wen-Tsan Chang,

    1. Graduate Institutes of Biochemistry, College of Medicine, National Cheng Kung University, Tainan, Taiwan
    Search for more papers by this author
  • Wen-Wei Chang,

    1. Graduate Institutes of Basic Medical Sciences, College of Medicine, National Cheng Kung University, Tainan, Taiwan
    Search for more papers by this author
  • Han-Chieh Wu,

    1. Division of Clinical Research, National Health Research Institutes, Tainan, Taiwan
    Search for more papers by this author
  • Wenya Huang,

    1. Graduate Institutes of Medical Laboratory Science and Biotechnology, College of Medicine, National Cheng Kung University, Tainan, Taiwan
    Search for more papers by this author
  • Huan-Yao Lei,

    1. Graduate Institutes of Basic Medical Sciences, College of Medicine, National Cheng Kung University, Tainan, Taiwan
    2. Graduate Institutes of Microbiology and Immunology, College of Medicine, National Cheng Kung University, Tainan, Taiwan
    Search for more papers by this author
  • Ming-Derg Lai,

    1. Graduate Institutes of Biochemistry, College of Medicine, National Cheng Kung University, Tainan, Taiwan
    Search for more papers by this author
  • Nelson Fausto,

    1. Department of Pathology, University of Washington, Seattle, WA
    Search for more papers by this author
  • Ih-Jen Su

    Corresponding author
    1. Division of Clinical Research, National Health Research Institutes, Tainan, Taiwan
    2. Graduate Institutes of Basic Medical Sciences, College of Medicine, National Cheng Kung University, Tainan, Taiwan
    3. Graduate Institutes of Microbiology and Immunology, College of Medicine, National Cheng Kung University, Tainan, Taiwan
    • Division of Clinical Research, NHRI, 12C Ward, National Cheng Kung University Hospital, 138, Sheng-Li Rd, Tainan, Taiwan
    Search for more papers by this author
    • fax: (886) 6 3028046

  • Conflict of interest: Nothing to report.


Naturally occurring mutants with a deletion in the pre-S2 region of the large surface protein (ΔS2-LHBs) are prevalent in serum and livers of patients with chronic hepatitis B virus (HBV) infection associated with cirrhosis. The ΔS2-LHBs protein is retained in the endoplasmic reticulum (ER) and may induce ER stress. One interesting observation is the consistently clustered distribution of hepatocytes expressing ΔS2-LHBs. In this study, complementary DNA microarray analysis identified cyclin A and several groups of genes as being significantly upregulated by ΔS2-LHBs in the HuH-7 cell line. This observation was confirmed in liver tissues. The induction of cyclin A expression may occur via the specific transactivator function of ΔS2-LHBs independent of ER stress. In the presence of ΔS2-LHBs, hepatocytes sustained cyclin A expression and cell cycle progression under ER stress and displayed increased BrdU incorporation with multinuclear formation. Furthermore, ΔS2-LHBs could enhance anchorage-independent cell growth in a nontransformed human hepatocyte line and induced nodular proliferation of hepatocytes in transgenic mice. In conclusion, these in vitro and in vivo data support a role for ΔS2-LHBs in the hepatocyte hyperplasia and a likely role in the process of HBV-related tumorigenesis. (HEPATOLOGY 2005.)

Hepatitis B virus (HBV) is regarded as an etiological factor of hepatocellular carcinoma (HCC). Patients with chronic HBV infection carry a high risk of developing cirrhosis and HCC. The successful control of HCC in children after HBV vaccination further supports a role of chronic HBV infection in the development of HCC.1 Although the association of chronic HBV infection with HCC is well established, the underlying mechanism of HBV-related hepatocarcinogenesis still remains elusive.

Mounting evidence supports the involvement of HBV and its gene products in the multistep progression of liver tumorigenesis.2 A protein-designated HBx has been extensively studied; the data reveal a role of the protein as a transactivator involved in cell growth, apoptosis, DNA damage signals, mitogen-activated protein kinase, and JAK/STAT signaling pathways.3 Recently, the large surface protein (LHBs) and a C-terminally truncated middle surface protein (MHBst) have likewise been recognized as transactivators that share the same mechanism for transcriptional activation.4, 5

This group of activators may trigger a protein kinase C–dependent activation of the c-Raf-1/mitogen-activated protein kinase 2 signal transduction cascade, resulting in the activation of transcription factors such as activator protein 1 and nuclear factor κB. The functional activity of these activators is dependent on the cytoplasmic orientation of the pre-S2 region of MHBst and LHBs that is also related to their intracellular retention.6, 7 Besides the MHBst, we have previously identified a mutant large surface protein with an in-frame deletion over the pre-S2 region (ΔS2-LHBs) in the ground glass hepatocytes (GGHs) of livers with HBV-related cirrhosis. The deletion site of this protein defines a cytotoxic T-lymphocyte epitope, and ΔS2-LHBs may represent an immune escape mutant.8–10 These pre-S2 mutants are becoming increasingly prevalent in serum and liver tissues of patients with chronic HBV infection and HCC.8–10 ΔS2-LHBs is localized in the endoplasmic reticulum (ER) and has been implicated in the induction of ER stress responses.11

Interestingly, hepatocytes containing these mutants usually form clusters and undergo clonal proliferation,12 consistent with the suggestion that ΔS2-LHBs may confer a growth advantage on hepatocytes and play a role in HBV-related hepatocarcinogenesis. The present study was undertaken to elucidate the specific biological functions of this naturally occurring mutant protein and to explore its potential role in HBV-related hepatocarcinogenesis. A complementary DNA (cDNA) microarray analysis was first used to identify the candidate genes regulated by ΔS2-LHBs. Cell cycle regulation, transforming capability, and transgenic mice studies were then performed.


ΔS2-LHBs, mutant with a deletion in the pre-S2 region of the large surface protein; HBV, hepatitis B virus; ER, endoplasmic reticulum; HCC, hepatocellular carcinoma; GGH, ground glass hepatocyte; cDNA, complementary DNA; pon-A, ponasterone-A; HH4, nontransformed human hepatocyte cell line; BFA, brefeldin A; VT, vomitoxin; PCNA, proliferating cell nuclear antigen; CDK, cyclin-dependent kinase.

Materials and Methods

Immunohistochemical and Immunofluorescent Studies.

Histopathological studies were performed on resected liver tissue samples obtained from patients with HBV-related HCC. For immunohistochemical staining, 4-μm-thick paraffin sections were stained with mouse anti-preS1, MA18/7,13 or mouse anti-BrdU (DAKO, Carpinteria, CA). A biotinylated anti-mouse secondary antibody (DAKO) was then applied and incubated with peroxidase-conjugated streptoavidin, chromogenized by 3-amino-9-ethylcarbazo. For double immunofluorescence staining, sections were stained simultaneously with mouse anti-HBs (DAKO) and rabbit anti-cyclin A (Santa Cruz Biotechnology, Santa Cruz, CA). The slides were then incubated with fluorescein isothiocyanate and tetramethylrhodamine isothiocyanate–conjugated secondary antibodies. The slides were counterstained with Hoechst 33342 (Molecular Probes, Eugene, OR).

Establishment of Cell Lines Expressing ΔS2-LHBs.

A ponasterone-A (pon-A) inducible HuH-7 cell line containing the expression plasmids of wild-type HBV large surface gene and ΔS2-LHBs constructs was established.11 For the establishment of stable expression on the nontransformed human hepatocyte cell line (HH4), wild-type and ΔS2-LHBs genes were first subcloned into the retrovirus-based vector pLXSZ with Zeocine as the selectable antibiotic gene. Brefeldin A (BFA) (Sigma-Aldrich, St. Louis, MO) was used to treat cells as the positive control of ER stress induction. Vomitoxin (VT) (Sigma-Aldrich) was used as an ER stress inhibitor because it inhibits ER stress-induced activation of GRP78 and p58 IPK.14

cDNA Microarray Analysis and Northern Blot Hybridization.

Total RNA was extracted using the Oligotex messenger RNA kit (QIAGEN, GmbH, Hilden, Germany) from cells 20 hours after 8-μmol/L pon-A induction. The conversion of 1 μg messenger RNA to biotin-labeled cDNA and the cDNA hybridization were performed according to previously published protocols.15 The labeled probes were hybridized to the MillenniaChip version 1 (obtained from Dr. Ching Li, Department of Molecular Medicine, National Cheng Kung University, Taiwan), which contains 884 selected genes with 13 controls. The hybridization signals were recorded with a high-resolution scanner, and the digitalization of the signals for the subsequent expression analysis was performed.16 For Northern blot hybridization, total RNA (2 μg) was resolved using a 1% MOPS/formaldehyde gel, followed by transfer to a nylon membrane. Hybridization was performed with biotin-labeled DNA probes. DNA templates of proliferating cell nuclear antigen (PCNA), cyclin A2, E2F transcription factor, and cyclin G1 were further labeled using a random primer DNA biotinylation kit (Kirkeggard & Perry Laboratories, Gaithersburg, MD). Gene transcripts were detected by using an RNADetector kit (Kirkeggard & Perry Laboratories). The probes for ER stress signals, GRP78 and GRP94, were reverse-transcriptase polymerase chain reaction products.11 The signal intensity of the messenger RNA was analyzed via densitometer scanning of the autoradiographs. The levels of GAPDH messenger RNA were used as an internal control.


Plasmids p(3A)SAg-WT and p(3A)SAg-ΔS2 expressing wild-type and pre-S2 mutant (nt 4-57 deletion) large surface protein have been previously described.11 The pre-S regions (pre-S1/S2) of wild-type and pre-S2 mutant were generated via polymerase chain reaction with p(3A)SAg and p(3A)SAg-ΔS2 as templates using primers 5′-TAGAGCTCGCCGTGGGAGGTT GGTCATCAAA-3′ and 5′-CTCTCGAGGTTCGT CACAGGGTCCCCAG-3′. The plasmid pIRES-preS-WT and pIRES-preS-ΔS2 were constructed by subcloning the previous polymerase chain reaction products into SacI and XhoI restriction sites of the pIRES-hrGFP2a vector (Stratagene, La Jolla, CA). These two plasmids thus contained only the pre-S region of the large surface gene lacking the small S region. For reporter assay of cyclin A, the cyclin A wild-type reporter plasmid was used (kindly provided by Dr. Manfioletti Guidalberto,17 University of Trieste, Trieste, Italy).

Cyclin A Reporter Assay.

For luciferase assay, HuH-7 cells were transfected with 1 μg of the cyclin A wild-type reporter plasmid, 1 μg of expression vectors, and 0.1 μg of pSG5-LacZ expression vector to normalize transfection efficiencies. HuH-7 cells treated with 20 ng/mL epidermal growth factor were used as the positive control of cyclin A expression. For comparison, two ER regulators, BFA and VT, were included. BFA was treated at 2 μg/mL; VT was treated at 125 ng/mL. All assays were performed three times using the luciferase reporter assay system (Promega, Madison, WI) according to the manufacturer's instructions.

Western Blot Hybridization.

Protein lysates were harvested in 1× RIPA buffer via alternate freezing and thawing. Twenty micrograms of total protein were resolved on SDS-polyacrylamide gels and transferred to polyvinylidene fluoride membranes. The membrane was incubated with the primary antibody followed by a second, horseradish peroxidase–conjugate antibody. Proteins were visualized using the ECL chemiluminescence kit (Perkin Elmer Life Science, Boston, MA). The primary antibodies used in this study were anti-preS1 and MA18/713; anti-GRP78 (BD Biosciences, Palo Alto, CA); and anti-GRP94, anti-cyclin A, anti-cyclin D1, and anti-CDK4 (Santa Cruz Biotechnology).

Flow Cytometry Analysis of Cell Cycle Regulation.

DNA content was assessed via propidium iodide staining of ethanol-fixed cells, followed by flow cytometry analysis (FACS Calibur; Becton Dickinson Immunocytometry Systems, San Jose, CA). The percentage of cells in the G0/G1, S, and G2/M phases of the cell cycle was assessed using ModFIT software (Verity Software House, Topsham, ME).

Colony Formation Assay.

A base agar layer in complete culture medium was prepared at a final concentration of 0.7%; 5 mL were placed in a 10-cm2 plate. A layer of agar (0.35%) mixed with 1 × 105 HH4 cells in 3 mL was subsequently positioned on the surface and was topped with 2 mL of 0.7% agar; 3 mL of cell culture medium was added to the surface of the agar every 3 days. Three replicates were made, and the plates were incubated at 37°C in 5% CO2 for 3 weeks. The number and size of cell colonies were determined with a dissecting microscope.

BrdU Incorporation Assay.

For detecting cell proliferation on HH4 cells, cells were cultured in a six-well culture dish for at least 24 hours before staining. Cells were labeled by the addition of BrdU to a final concentration of 100 μmol/L followed by incubation at 37°C for 4 hours. BrdU-incorporated cells were detected via immunohistochemical staining with mouse anti-BrdU antibody. For the detection of BrdU incorporation on transgenic mice, 7-day-old mice received an intraperitoneal injection of BrdU (50 mg/kg body weight). Liver samples were fixed and embedded in paraffin.

Transgenic Mice Study.

The plasmids p(3A)SAg-WT and p(3A)SAg-ΔS2 were used to make transgenic mice. The plasmids were microinjected into embryos of FVB/NJ mice. All microinjected embryos were implanted into pseudopregnant female mice. Transgenic mice carrying the transgene were selected via polymerase chain reaction confirmation of the pre-S gene. Mice with the HBV transgene were next tested for surface protein expression in blood and liver tissue via enzyme-linked immunosorbent assay.9


As previously reported,11 hepatocytes containing ΔS2-LHBs consistently clustered in the livers of patients with chronic HBV infection associated with cirrhosis (Fig. 1A). An enhanced expression of HBV surface antigen was evident at the cell periphery. Figure 1B shows the identified pre-S2 mutant gene with a deletion over the pre-S2 region. Western blot hybridization analysis of pon-A inducible HuH-7 cell lines (Fig. 1C) revealed two positive protein bands of LHBs: one of these (gp42) was glycosylated; the other (p39) was nonglycosylated.

Figure 1.

Tissue and cell line expression of ΔS2-LHBs. (A) Liver section obtained from a 58-year-old man with cirrhosis positive for hepatitis B surface antigen and negative for hepatitis C virus. ΔS2-LHBs accumulated in the marginal region of GGHs, which clustered in groups. (B) The pre-S2 mutant identified from GGHs of this patient contained an in-frame deletion at nucleotides 4-57 and a point mutation at the start codon of pre-S2, which results in a loss of amino acid 124-142 and a defect of middle surface protein expression. The hatched boxes represent CD8-specific T-cell epitopes. (C) Western blot analysis of pon-A–induced LHBs expression on HuH-7 cells. The identified pre-S2 mutant gene was cloned into a pon-A–inducible expression vector. A stable inducible cell line expressing the wild-type and pre-S2 mutant protein was established on HuH-7 cells as previously described.11 After induction with 8 μmol/L pon-A, wild-type LHBs and ΔS2-LHBs were detected via Western blot hybridization. The bands labeled with gp42 and p39 represent glycosylated and unglycosylated wild-type LHBs, respectively. ΔS2-LHBs showed two bands with smaller size compared with wild-type LHBs. ΔS2-LHBs, mutant with a deletion in the pre-S2 region of the large surface protein; Ctrl, control; WT, wild-type.

Gene Expression Profile Induced By ΔS2-LHBs.

To identify genes differentially expressed and regulated by ΔS2-LHBs, we performed cDNA microarray studies. Given that the hepatocytes harboring ΔS2-LHBs were usually distributed in clusters, we paid particular attention to genes related to cell cycle regulation and proliferation. The data indicated that cyclin A2, PCNA, and cyclin-dependent kinase (CDK) 5 were upregulated by ΔS2-LHBs by as much as fivefold. Other upregulated genes included transcriptional factor E2F and genes involving DNA damage (rec A and human 8-hydroxyguanine glycosylate/lyase). The expression of the genes encoding cyclin A2, PCNA, E2F, and cyclin G1 were confirmed via Northern blot hybridization (Fig. 2). As previously reported,11 ΔS2-LHBs accumulated in the ER, and the ER stress signals GRP78 and GRP94 were upregulated by ΔS2-LHBs (see Fig. 2).

Figure 2.

Northern blot analysis of cellular genes in pon-A inducible HuH-7 cell lines. Two micrograms of total RNA were extracted from control, wild-type, and ΔS2-LHBs cells 20 hours after induction with 8 μmol/L pon-A. The expressions of surface gene, cyclin A2, PCNA, E2F, cyclin G1, GRP78, and GRP94 are shown. Cyclin A2, PCNA, and E2F genes were induced by ΔS2-LHBs, while cyclin G1 was downregulated by wild-type and ΔS2-LHBs. The transcript of GRP78 was profoundly induced by ΔS2-LHBs, while the induction of GRP94 was mild compared with vector control and wild-type. Quantitative results of densitometry intensity are shown in the right column, with the actin and GAPDH genes as internal controls. Standard deviations were indicated for experiments repeated three times. Ctrl, control; WT, wild-type; ΔS2, mutant with a deletion in the pre-S2 region of the large surface protein; PCNA, proliferating cell nuclear antigen.

Cyclin A Induction By ΔS2-LHBs is ER Stress–Independent.

Because ΔS2-LHBs accumulated in the ER and induced ER stress, we wanted to clarify whether the upregulation of cyclin A by ΔS2-LHBs was mediated through either ER stress signals or an ER stress–independent pathway. Western blot hybridization analysis revealed that the expression of cyclin A was not induced by the ER stress inducer BFA, suggesting that ER stress may not be involved in the induction of cyclin A expression. However, the ER chaperone GRP78 appeared to play a role in the regulation of ΔS2-LHBs–induced cyclin A expression, because the induction of cyclin A by ΔS2-LHBs could be inhibited by VT (Fig. 3A). To solve this dilemma, we performed the cyclin A reporter assay to determine if ΔS2-LHBs could transcriptionally activate cyclin A. Using HuH-7 cells treated with epidermal growth factor as a positive control for cyclin A induction, neither the ER stress inducer (BFA) nor the ER stress inhibitor (VT) had any effect on cyclin A expression (Fig. 3B). Furthermore, VT showed no inhibitory effect on cyclin A induction by ΔS2-LHBs (see Fig. 3B). The cytoplasmic pre-S region of the LHBs protein has been reported to serve as a gene transactivator in enhancing cellular gene expression.18 To further clarify the transactivator function of the pre-S protein, we established new constructs expressing only the pre-S region (pre-S1/S2) of wild-type and ΔS2-LHBs proteins that lack the small S region. The pre-S proteins encoded by these new constructs were localized in the cytoplasm because of a lack of the ER transmembrane domain on the S region (data not shown).19 The results again showed that only the pre-S region of ΔS2-LHBs, not the wild-type, could enhance cyclin A expression (see Fig. 3B). We concluded that ΔS2-LHBs may activate the gene expression of cyclin A via an ER stress–independent pathway, and ΔS2-LHBs may play a specific transactivating role. This induction of cyclin A by ΔS2-LHBs could be further confirmed by the double immunofluorescence staining on liver tissues. In the livers of patients with chronic HBV infection associated with cirrhosis, the ΔS2-LHBs–expressing hepatocytes showed a much higher staining intensity of cyclin A in their cytoplasm than their adjacent, normal hepatocytes (Fig. 3C).

Figure 3.

Induction of cyclin A in marginal type GGHs and ΔS2-LHBs–expressing cells. (A) Cyclin A expression in pon-A–inducible ΔS2-LHBs–expressing HuH-7 cells. Protein lysates were collected after treatment of 2 μg/mL BFA, 8 μmol/L pon-A, and 125 ng/mL VT. BFA and VT alone could not increase the protein expression of cyclin A. Cyclin A, however, was increased after induction with 8 μmol/L pon-A. The pon-A–incucible expression of cyclin A could be inhibited by VT. (B) Cyclin A reporter assay. HuH-7 cells were transiently transfected with 1 μg of cyclin A wild-type reporter construct, 1 μg of vector control or the indicated expression plasmids, and 0.1 μg pSG5-LacZ for internal control of transfection efficiency. (C) Immunofluorescence staining of cyclin A and HBV surface proteins on GGHs in liver of a 58-year-old man with cirrhosis associated with chronic HBV infection. Cell nuclei were counterstained with Hoechst 33342. A marked increase of cyclin A expression in HBs-expressing hepatocytes was observed. (Original magnification ×100). ΔS2, mutant with a deletion in the pre-S2 region of the large surface protein; VT, vomitoxin; BFA, brefeldin A; pon-A, ponasterone-A; CycAwt, cyclin A wild-type; Ctrl, control; EGF, epidermal growth factor; Hoechst, Hoechst 33342; HBs, hepatitis B surface antigen.

ΔS2-LHBs Influence Cell Cycle Progression Under ER Stress.

ER stress responses include the unfolded protein and ER overload responses.20 The activation of unfolded protein response attenuates cyclin D1 and causes G1 phase arrest.21 To determine if ΔS2-LHBs sustained cell cycle progression under ER stress, we treated cells with BFA and assayed cell cycle progressions. Treatment with pon-A did not result in the G1 arrest of cells, whereas BFA treatment did (Fig. 4A). Over 80% of cells in each group were arrested in G1 phase after BFA treatment. In contrast to the control and wild-type groups, cells expressing ΔS2-LHBs had a delayed response to BFA treatment. The percentage of ΔS2-LHBs–expressing cells arrested in the G1 phase only slightly increased to 56% after 24 hours of BFA treatment, suggesting that ΔS2-LHBs may sustain cell cycle progression under ER stress.

Figure 4.

ΔS2-LHBs influences cell cycle regulation on BFA-treated cells. (A) HuH-7 cell lines were untreated or treated with pon-A for 24 hours before treatment with 2 μg/mL BFA. The percentages of cells in the G1 phase were analyzed with ModFIT software. Cells in the G1 phase increased from 44% to 47% to over 75% to 80% after BFA treatment on the vector control and the wild-type groups, while ΔS2-LHBs–expressing cells increased only slightly to 56%. (B) Cell cycle–associated genes regulated by ER stress were analyzed via Western blot analysis under 2 μg/mL BFA treatment on pon-A–inducible HuH-7 cell lines. GRP78 and GRP94 were upregulated by BFA, while cyclin A, cyclin D1, and CDK4 were downregulated. CDK2 and eukaryotic translation initiation factor 2α showed no change. In cells expressing ΔS2-LHBs, cyclin A and cyclin D1 were not affected by BFA, while CDK4 showed a delayed response compared with the control and wild-type groups. pon-A, ponasterone-A; BFA, brefeldin A; Ctrl, control; WT, wild-type; ΔS2, mutant with a deletion in the pre-S2 region of the large surface protein; PI, propidium iodide; CDK, cyclin-dependent kinase.

ΔS2-LHBs Protect ER Stress–Induced Degradation of Cyclin D1 and CDK4.

ER stress may lead to cell cycle arrest through the inhibition of protein synthesis of cyclin D1, which involves the activation of ER resident kinase, eukaryotic translation initiation factor 2α phosphorylation, and CDK2 degradation.21 We therefore examined cyclin D1 expression in ΔS2-LHBs–expressing HuH-7 cells under ER stress induction. Downregulation of cyclin D1 and cyclin A could be observed in the control cells and cells expressing wild-type LHBs, but not in cells expressing the ΔS2-LHBs protein (Fig. 4B). The ΔS2-LHBs–expressing cells showed a delayed response of CDK4 degradation, while the expression of CDK2 and eukaryotic translation initiation factor 2α were not affected in HuH-7 cells under ER stress induction.

Growth Stimulation Effects of Pre-S2 Mutant Protein.

To understand the mechanism of ΔS2-LHBs stimulation of cell growth, we performed BrdU incorporation and soft agar assays on HH4. HH4 cells expressing ΔS2-LHBs showed a higher BrdU incorporation rate (36%) compared with the control and wild-type groups (28% and 30%, respectively) (Fig. 5A). Interestingly, multinucleated BrdU-positive nuclei were observed in as many as 35% of the ΔS2 cells, while fewer than 5% and 10% were observed in the control and wild-type, respectively. To test the cell-transforming ability of ΔS2-LHBs, we performed a soft agar assay of HH4 cells after 24 passages. Our results showed that the colony number was elevated by twofold in the wild-type LHBs and sixfold in ΔS2-LHBs (Fig. 5B). The colony size of ΔS2-LHBs–expressing hepatocytes also increased fivefold compared with the controls. Therefore, ΔS2-LHBs showed a stronger transforming ability on HH4.

Figure 5.

Influence of ΔS2-LHBs on HH4 cell growth. (A) BrdU incorporation assay on HH4 cells. BrdU-positive cells were visualized and quantitated via microscopy. (B) Soft agar assay on HH4 stable clones. The colony numbers increased twofold for wild-type–expressing cells and sixfold for ΔS2-LHBs–expressing cells. The HH4 colonies formed on the wild-type and ΔS2-LHBs groups were markedly increased in size. (Original magnification ×200.) Ctrl, control; WT, wild-type; ΔS2, mutant with a deletion in the pre-S2 region of the large surface protein; HH4, nontransformed human hepatocyte cell line.

Pre-S2 Mutant Protein Studies in Transgenic Mice.

To assess the biologic effects of ΔS2-LHBs in vivo, we established a transgenic mouse model of wild-type LHBs and ΔS2-LHBs. In contrast to the smooth surface of livers in wild-type and control nontransgenic mice, the external surface of livers from ΔS2-LHBs mice aged 0.75 to 18 months had a nodular appearance (Fig. 6A). Both wild-type LHBs and ΔS2-LHBs were expressed mainly around the central vein. Histologically, liver cell dysplasia could be observed in the majority of ΔS2-LHBs mice beyond 3 months of age but only infrequently in the wild-type mice beyond 6 months of age (Table 1). Liver cell dysplasia was diagnosed according to the morphological criteria,22 including cellular enlargement, nuclear pleomorphism, hyperchromatism, multinucleation, normal or slightly increased nucleus/cytoplasmic ratio, and nuclear maximum diameter exceeding 11 μm. ΔS2-LHBs expression on transgenic livers typically coincided with the dysplastic changes of hepatocytes. Hepatocyte multinucleation usually appeared later in ΔS2-LHBs mice, at 12 months of age. No tumor formation was observed in either group for up to 18 months. Compared with male mice, females showed a relatively mild change of nuclear morphology, but with increased numbers of lipid-containing vehicles (data not shown). Livers from nontransgenic and wild-type mice showed a histologically normal architecture, and no nuclear dysplastic change was observed (see Fig. 6A).

Figure 6.

ΔS2-LHBs–induced liver changes in transgenic mice. (A) The external surface of a ΔS2-LHBs transgenic liver showing nodular appearance, which was also highlighted in liver histology (HE-1, original magnification ×40) and immunohistochemical studies (HBs-IHC) in the ΔS2-LHBs group. Both the wild-type and ΔS2-LHBs proteins expressed mainly around the central vein. There was a sharp transition between ΔS2-LHBs–expressing and nonexpressing hepatocytes. ΔS2-LHBs–expressing cells showed larger size and nuclear polymorphism compared with wild-type and control nontransgenic mouse liver (HE-2, original magnification ×400). Immunohistochemical staining of cyclin A showed predominantly cytoplasmic expression on the ΔS2-LHBs transgenic hepatocytes, whereas nuclear staining of cyclin A was primarily observed in wild-type and control mouse liver. (B) BrdU labeling index of male transgenic and control nontransgenic mice. (C) ER stress induction on ΔS2-LHBs transgenic mice. The protein expression level of GRP78 was significantly higher in ΔS2-LHBs transgenic mice compared with wild-type and control mice. Ctrl, control; WT, wild-type; ΔS2, mutant with a deletion in the pre-S2 region of the large surface protein; IHC, immunohistochemical; HE, hematoxylin-eosin; HBs, hepatitis B surface antigen.

Table 1. Histological Observations of Male Transgenic and Nontransgenic (Control) Mice
  1. Abbreviations: NI, necrotic inflammation; LCD, liver cell dysplasia.

 94001 (25%)0
Wild-type≤192 (22.2%)2 (22.2%)00
 3201 (50%)00
 6404 (100%)1 (25%)0
 9402 (50%)00
 >12301 (33%)1 (33%)0
S2≤175 (71.4%)5 (71.4%)1 (14.3%)0
 353 (60%)3 (60%)3 (60%)0
 632 (66.7%)2 (66.7%)3 (100%)0
 983 (37.5%)2 (25%)6 (75%)0
 >1264 (66.7%)3 (50%)5 (83.3%)0

The immunohistochemical expression of cyclin A was distinctly different, with nuclear expression on the wild-type and control mice liver and almost exclusively cytoplasmic on the ΔS2-LHBs transgenic livers (see Fig. 6A). This cytoplasmic expression of cyclin A in ΔS2-LHBs transgenic livers correlated well with the expression pattern in human liver tissues harboring ΔS2-LHBs.

To test hepatocyte proliferation on transgenic mice, we next detected BrdU-positive nuclei to quantitate the percentages of hepatocytes in the S phase (Fig. 6B). ΔS2-LHBs mice showed a higher cell proliferation index. The expression of ER stress chaperone GRP78 in ΔS2-LHBs mice increased by 1.4-fold compared with wild-type and control mice (Fig. 6C).


This study was undertaken to better understand why hepatocytes expressing ΔS2-LHBs form clusters. We demonstrated the upregulation of cyclin A by ΔS2-LHBs, which was confirmed in liver tissues with chronic HBV infection. This finding is interesting and important, given that cyclin A is overexpressed in HCC tissues.23–26

Cyclin A plays an important role in the S and G2/M phases of the cell cycle. Modification of cyclin A expression by HBV DNA integration in HCC may result in a strong expression of HBV–cyclin A hybrid transcripts, which may affect cell cycle regulation.23, 27, 28 In the present cell cycle analysis, hepatocytes expressing ΔS2-LHBs protein were resistant to cell cycle arrest under BFA-induced ER stress conditions. Because the expression levels of cyclin A remained sustained throughout the cell cycle, distinct from other cell cycle regulators that declined under ER stress conditions, the sustained expression of cyclin A by ΔS2-LHBs may explain the presently observed cell cycle progression under ER stress conditions. Similar to another gene transactivator, MHBst, a tumor promoter retained in ER,6 the induction of cyclin A by ΔS2-LHBs is likely to be mediated via gene transactivation, which is independent of ER stress signals. This induction of cyclin A may be cell type–specific, because Northern blot analysis and cyclin A reporter assay on lung cancer cell line A549 showed no such effect (data not shown). Although BFA could not enhance the expression of cyclin A, treatment of the GRP78 inhibitor can lead to cyclin A downregulation (see Fig. 3A). Therefore, the ER chaperone GRP78 may play a role in maintaining cyclin A turnover at the protein level.

One interesting finding in the cDNA microarray study is the concurrent upregulation of E2F by ΔS2-LHBs. E2F is a transcription factor that can activate cyclin A. Whether the upregulation of cyclin A is mediated via the Rb/E2F signaling has yet to be investigated. Besides cyclin A and E2F, the proliferation-related gene PCNA was also upregulated by ΔS2-LHBs. The expression of PCNA was found to be higher in ΔS2-LHBs expressing Huh7 cells, as well as in marginal types of GGHs (data not shown). Cyclin G1 is involved in G2/M arrest in response to DNA damage, and has growth inhibitory activity that is mechanistically linked to the ARF-p53 and pRb tumor suppressor pathways.29 The inhibition of cyclin G1 by ΔS2-LHBs may contribute to cell survival in response to DNA damage. In concurrence with one previous study,30 several molecules involved in DNA damage and repair (e.g., 8-hydroxyguanine glycosylase/lyase, DEK, and rec A) were also upregulated by ΔS2-LHBs in this study (data not shown), probably through the induction of reactive oxidative intermediates by ER stress. The potential DNA damages by the ER accumulation of mutant surface proteins in GGHs is interesting, because the hepatocytes in nodules associated with cirrhosis have been shown to have chromosomal instability and probably represent a preneoplastic lesion.

The potential role of pre-S mutants in HBV-related hepatocarcinogenesis obtains further support from the present findings of the enhancing transforming capability of ΔS2-LHBs and the nuclear dysplastic changes in transgenic mice. ΔS2-LHBs enhanced the colony-forming ability and colony size of the nontransformed HH4 cell lines in soft agar. Furthermore, data from transgenic mice indicated a role for ΔS2-LHBs in the induction of nodular proliferation of hepatocytes with dysplastic changes. Although we did not detect tumor formation at up to 18 months, the observed presence of nuclear hyperchromasia, irregular nuclear contour, cytoplasmic clear cell change, and increase of mitotic figures on transgenic mice is consistent with the emergence of a premalignant condition induced by ΔS2-LHBs.31 A second event may be required for full development of HCC.

Another noteworthy observation from the present study concerns the cytoplasmic localization of cyclin A in type II GGHs and ΔS2-LHBs transgenic livers. Under physiological conditions, cyclin A is predominantly localized in the nucleus from the G1/S transition onward. It is possible that there exists another isoform of cyclin A, which may have distinct biological function and subcellular localization. The appearance of truncated cyclin A in mouse embryonic stem cells correlates with cell differentiation.32 In addition, high cell density induces cyclin A cleavage in cultured mammalian cells.33 The nondegradable fragment of cyclin A residing in the ER membrane increases centrosome overduplication, nuclear polyplolidy, and cell transformation.34 Furthermore, overexpression of cyclin A in human HeLa cells induces detachment of kinetochores and spindle pole/centrosome overreplication.35 In this study, the multinucleation of HH4 cells expressing ΔS2-LHBs protein might be related to the aberrant expression of cyclin A. The exact nature of cytoplasmic localization of cyclin A and its regulation by ΔS2-LHBs deserves further investigation.

In conclusion, the presently documented aberrant expression of cyclin A and nodular proliferation of hepatocytes in livers with chronic HBV infection and in ΔS2-LHBs transgenic mice is consistent with a potential role of pre-S mutants in HBV-related hepatocarcinogenesis. Recently, the prevalence of pre-S mutants has been correlated to the risk of hepatocellular carcinoma in patients with chronic HBV infection.36–38 The data we presented in this study will provide valuable information on the potential function of pre-S mutants in HBV-related hepatocarcinogenesis.