Hepatitis B virus inhibits liver regeneration via epigenetic regulation of urokinase-type plasminogen activator

Authors

  • Eun-Sook Park,

    1. Department of Pharmacology and Center for Cancer Research and Diagnostic Medicine, IBST, Konkuk University School of Medicine, Seoul, Republic of Korea
    2. Research Institute of Medical Science, Konkuk University School of Medicine, Seoul, Republic of Korea
    Search for more papers by this author
  • Yong Kwang Park,

    1. Department of Pharmacology and Center for Cancer Research and Diagnostic Medicine, IBST, Konkuk University School of Medicine, Seoul, Republic of Korea
    Search for more papers by this author
  • Chan Young Shin,

    1. Department of Pharmacology and Center for Cancer Research and Diagnostic Medicine, IBST, Konkuk University School of Medicine, Seoul, Republic of Korea
    2. Research Institute of Medical Science, Konkuk University School of Medicine, Seoul, Republic of Korea
    Search for more papers by this author
  • Seung Hwa Park,

    1. Department of Anatomy, Konkuk University School of Medicine, Seoul, Republic of Korea
    Search for more papers by this author
  • Sung Hyun Ahn,

    1. Department of Pharmacology and Center for Cancer Research and Diagnostic Medicine, IBST, Konkuk University School of Medicine, Seoul, Republic of Korea
    Search for more papers by this author
  • Doo Hyun Kim,

    1. Department of Pharmacology and Center for Cancer Research and Diagnostic Medicine, IBST, Konkuk University School of Medicine, Seoul, Republic of Korea
    Search for more papers by this author
  • Keo-Heun Lim,

    1. Department of Pharmacology and Center for Cancer Research and Diagnostic Medicine, IBST, Konkuk University School of Medicine, Seoul, Republic of Korea
    Search for more papers by this author
  • So Young Kwon,

    1. Department of Internal Medicine, Konkuk University School of Medicine, Seoul, Republic of Korea
    Search for more papers by this author
  • Kwang Pyo Kim,

    1. Department of Molecular Biotechnology, WCU program, Konkuk University, Seoul, Republic of Korea
    Search for more papers by this author
  • Sung-Il Yang,

    1. Department of Pharmacology and Center for Cancer Research and Diagnostic Medicine, IBST, Konkuk University School of Medicine, Seoul, Republic of Korea
    Search for more papers by this author
  • Baik L. Seong,

    1. Department of Biotechnology, College of Life Science and Biotechnology, and Translational Research Center for Protein Function Control, Yonsei University, Seoul, Republic of Korea
    Search for more papers by this author
  • Kyun-Hwan Kim

    Corresponding author
    1. Department of Pharmacology and Center for Cancer Research and Diagnostic Medicine, IBST, Konkuk University School of Medicine, Seoul, Republic of Korea
    2. Research Institute of Medical Science, Konkuk University School of Medicine, Seoul, Republic of Korea
    • Address reprint requests to: Kyun-Hwan Kim, Department of Pharmacology, School of Medicine, Konkuk University, Seoul 143-701, Republic of Korea. E-mail: khkim10@kku.ac.kr; fax: +82-2-2049-6192.

    Search for more papers by this author

  • Potential conflict of interest: Nothing to report.

  • Supported in part by National Research Foundation of Korea Grant funded by the Korean Government (KRF-2008-314-C00269), a grant of the Korea Healthcare Technology R&D Project, Ministry of Health & Welfare, Republic of Korea (Grant No. A103001).

Abstract

Liver regeneration after liver damage caused by toxins and pathogens is critical for liver homeostasis. Retardation of liver proliferation was reported in hepatitis B virus (HBV) X protein (HBx)-transgenic mice. However, the underlying mechanism of the HBx-mediated disturbance of liver regeneration is unknown. We investigated the molecular mechanism of the inhibition of liver regeneration using liver cell lines and a mouse model. The mouse model of acute HBV infection was established by hydrodynamic injection of viral DNA. Liver regeneration after partial hepatectomy was significantly inhibited in the HBV DNA-treated mice. Mechanism studies have revealed that the expression of urokinase-type plasminogen activator (uPA), which regulates the activation of hepatocyte growth factor (HGF), was significantly decreased in the liver tissues of HBV or HBx-expressing mice. The down-regulation of uPA was further confirmed using liver cell lines transiently or stably transfected with HBx and the HBV genome. HBx suppressed uPA expression through the epigenetic regulation of the uPA promoter in mouse liver tissues and human liver cell lines. Expression of HBx strongly induced hypermethylation of the uPA promoter by recruiting DNA methyltransferase (DNMT) 3A2. Conclusion: Taken together, these results suggest that infection of HBV impairs liver regeneration through the epigenetic dysregulation of liver regeneration signals by HBx. (Hepatology 2013;58:762–776)

Abbreviations
5-AzaC

5-aza-2′-deoxycytidine

ChIP

chromatin immunoprecipitation

DNMT

DNA methyltransferase

HBV

hepatitis B virus

HCC

hepatocellular carcinoma

HGF

hepatocyte growth factor

rhHGF

recombinant human HGF

uPA

urokinase-type plasminogen activator.

Infection by hepatitis B virus (HBV) causes a variety of liver diseases, including acute or chronic inflammation, cirrhosis, and hepatocellular carcinoma (HCC). Chronic infection of HBV, which is estimated to affect 400 million people worldwide, causes chronic inflammation in the liver resulting in liver damage.[1-3] Prolonged liver damage activates liver regeneration signals to compensate for the loss of hepatocytes. Immoderate liver regeneration provokes chronic and rapid turnover of hepatocytes, which may be an important factor for liver dysfunction and the eventual development of HCC.[1, 2] Since a well-balanced liver regeneration is essential for liver homeostasis, a disturbance in this process by virus infection could be the principal cause of various virus-mediated liver diseases.

HBV X protein (HBx), a pleiotropic protein encoded by HBV, is involved in gene activation,[4-6] apoptosis,[7, 8] and cell cycle control.[9, 10] The in vivo effect of HBx on hepatocyte proliferation and cell cycle progression has been examined in HBx-transgenic mice. HBx inhibits liver cell proliferation after partial hepatectomy and causes a delay in the G1/S transition.[11, 12] However, the mechanism of the HBx-mediated disturbance of liver regeneration is unclear.

Hepatic regeneration is a highly complex process involving a variety of growth factors that include hepatocyte growth factor (HGF), interleukin-6 (IL-6), tumor necrosis factor-alpha (TNF-α), and transforming growth factor-beta (TGF-β).[13-18] Among them, HGF is essential. HGF is synthesized and secreted as an inactive single chain molecule (pro-HGF). The proteolytic cleavage of pro-HGF results in the formation of the active two-chain form, which has high affinity for the c-met receptor.[19-22] Urokinase-type plasminogen activator (uPA) is a group of serine proteases that cleave and activate pro-HGF.[19, 20] uPA mediates a variety of biological activities at the cell surface, including plasminogen activation,[23] extracellular matrix remodeling,[24, 25] and growth-factor activation.[26] The plasminogen system plays an important role in the repair of a variety of tissues including the liver, kidney, and skin.[27-30] In addition, mice that are deficient in uPA or plasminogen are markedly impaired in liver regeneration.[28, 31]

The purpose of our study was to investigate the effect of HBV during liver regeneration in vivo. Host target molecules potentially involved in hepatocyte proliferation during liver regeneration after HBV infection have not been elucidated. The present studies demonstrate that the HBx protein directly regulates the expression of uPA, an essential enzyme for the activation of HGF, both in vitro and in vivo, through epigenetic regulation by way of hypermethylation of its promoter. We suggest that the impairment of liver regeneration by HBV infection can be attributed to the disturbance of liver regeneration signals by HBx.

Materials and Methods

Casein Zymography

The activity of uPA was analyzed by direct casein zymography as described.[47, 48] Briefly, samples were mixed with sample buffer without reducing reagent such as β-mercaptoethanol and run on a 10% polyacrylamide gel containing casein (1 mg/mL; Sigma, St. Louis, MO) and plasminogen (13 μg/mL; American Diagnostica, Stamford, CT). The caseinolysis band detected at 52 kDa was specific for uPA since it corresponded to the band of standard uPA (Chemicon, Temecula, CA).

Chromatin Immunoprecipitation (ChIP) Analysis

ChIP assay was performed using the ChIP assay kit (Millipore, Billerica, MA). Briefly, Huh7 cells were cross-linked with formaldehyde for 10 minutes at 37°C. After washing with phosphate-buffered saline (PBS), the collected cells were resuspended in sodium dodecyl sulfate (SDS) lysis buffer (Millipore). The cell lysates were sonicated seven times for 10 seconds each time at 30% power. The supernatants were precleared using protein G-agarose and incubated with anti-DNA methyltransferase (DNMT) 3A antibody (No. IMG-268; Imgenex) or normal mouse immunoglobulin G (IgG) as a negative control. The immunoprecipitated DNA fragments were detected by polymerase chain reaction (PCR) amplification using the specific primers shown in Supporting Table 1.

Hydrodynamic Injection of Plasmid DNA in Mice

Six to seven-week-old male BALB/c mice were hydrodynamically injected with 25 μg of plasmid DNA (pHBV1.2 [wt], pHBV1.2 [HBx-], pCMV-HBx-HA, and pCMV-HA) into the tail veins in a volume of PBS equivalent to 10% of mouse body weight.[32] The total volume of DNA was delivered into the vein with high pressure within 4-6 seconds (hydrodynamic in vivo transfection). For in vivo knockdown of uPA, mice were hydrodynamically injected with 25 μg of small interfering RNA (siRNA) targeting uPA (5′-CUCAUCUUGCACGAAUACU-3′). To restore uPA expression in the liver of HBV-expressing mice, 25 μg of mouse uPA vector was hydrodynamically injected with HBV1.2 (wt). After 2 days, the injection was repeated with the same amount of uPA vector. All procedures involving experimental animals were approved by the Animal Care Committee, Konkuk University.

Partial Hepatectomy

A 70% partial hepatectomy model was used to induce physiological liver regeneration in mice. Four days after hydrodynamic injection, about 70% of the total liver mass was resected after the intraperitoneal administration of xylazine (10 μg/kg body weight) and Zoletil 50 (10 mg/kg body weight). In brief, the left lobe and a part of the median lobe of each mouse liver were removed through a midabdominal incision.[33] After suturing the abdominal wall, the animals were returned to their cages and permitted to feed ad libitum. Liver regeneration was evaluated by the ratio of body weight and the liver weight 8 days after the partial hepatectomy. The animals were sacrificed at the indicated times and the liver tissues and blood samples were collected and frozen at −70°C until assayed.

Results

HBV Down-Regulates uPA Expression and Its Activity Through Epigenetic Regulation of uPA Promoter by HBx

The plasminogen system is involved in the repair of liver tissues.[28] uPA is a serine protease that regulates the activation of HGF.[13, 16] Therefore, to investigate whether HBV infection alters the uPA expression level we examined the messenger RNA (mRNA) level of uPA in Huh7 cells transfected with HBx-expressing plasmid or the replication-competent HBV genome. As shown in Fig. 1A, the wild-type (wt) 1.2mer HBV (HBV1.2(wt)) strongly inhibited the transcription of endogenous uPA, whereas HBx-null HBV (HBV1.2(HBx-)) did not. These findings suggest that the HBx protein, among various HBV-encoded proteins, is responsible for uPA down-regulation. This was confirmed by transfection with HBx-GFP plasmid (Fig. 1A).

Figure 1.

HBV down-regulates uPA expression through epigenetic regulation on the uPA promoter by HBx. (A) The mRNA levels of uPA as determined by real-time PCR (left) and semiquantitative RT-PCR (right). (B) Diagrams of uPA promoter-luciferase reporter gene constructs (left panel). Determination of uPA promoter activity by luciferase reporter assay (right). (C) Relative luciferase activity of the uPA promoter (uPA-luc [−2064 to +24]) in Huh7 cells transiently (left and middle) or stably (right) transfected with indicated plasmids. Data shown are mean ± SD of three independent experiments. (D) RT-PCR and western blot analysis of uPA expression in Huh7 cells stably expressing the HBx protein. (E) Analysis of uPA activity in cell lysates and culture medium using casein zymography. (F) Bisulfite sequencing analysis of genomic uPA promoter in HBx transient and stable Huh7 cells. Percentages were calculated by methylated CpG per total CpG of the uPA promoter region (−230 to +24) from at least three independent experiments. TSS, transcription start site; open circle, unmethylated CpG; black circle, methylated CpG.

To investigate the mechanism involved in the reduction of endogenous uPA transcription, we performed a reporter assay using the promoter region of uPA. Since the region about 2 kb upstream of the start codon of the uPA gene is reported to have promoter activity,[34] the upstream region of the uPA gene (up to −2064 bp) was subcloned into the luciferase reporter vector, pGL3basic (Fig. 1B). For a fine mapping of the critical region of uPA gene transcription, we constructed serial deletion mutants of the uPA promoter region and each promoter activity was measured using the luciferase assay system (Fig. 1B). The regions of the uPA promoter containing −2064 to −1689 and −542 to +24 showed strong promoter activity (Fig. 1B), suggesting that the two regions are independently involved in uPA transcription. Percentage of the two regions showed additive effect on the transcription.

Using the promoter clone that contained up to −2064 bp upstream sequence, we tested whether HBV suppresses the uPA promoter. The uPA promoter activity was significantly reduced after transfection with HBx-GFP or HBV1.2(wt), but not with HBx-deficient HBV1.2(HBx-) (Fig. 1C). Similarly, the uPA promoter activity was almost completely blocked in the HBx-stable cell line (Fig. 1C). These data indicate that the down-regulation of uPA is controlled at the transcriptional level by HBx.

To examine whether the production of uPA protein is down-regulated in HBx-expressing cells, we performed both reverse-transcription PCR (RT-PCR) and western blot analysis in parallel. Both uPA mRNA and protein levels were significantly decreased by HBx (Fig. 1D). To determine whether the decreased protein level of uPA by HBx is linked to the decrease of uPA activity, we measured the uPA activity by way of the casein zymography assay from the cell media and the lysates of transiently transfected HBx cells. A significant decrease in uPA activity was observed in both samples as compared to the control cells, suggesting that both the level of secreted and cytoplasmic uPA was significantly reduced in HBx-expressing cells (Fig. 1E).

Finally, to examine whether the HBx-mediated down-regulation of uPA is controlled by epigenetic regulation, we investigated the methylation status of the uPA promoter region in Huh7 cells. Bisulfite sequencing revealed that the CpG region of the uPA promoter (−230 nt to +24 nt) was significantly hypermethylated in HBx-expressing cells (51.2% and 99.7% for HBV1.2(wt) and HBx-stable cells, respectively) compared to the control cells (Fig. 1F). The CpG island on the uPA promoter in HBx stable cells was almost completely methylated. These results suggest the possibility that HBx down-regulates uPA expression through hypermethylation of the uPA promoter regions. In contrast to the methylation status, the levels of mRNA and protein expression were not completely down-regulated, which is probably due to the presence of an upstream promoter activity (Fig. 1B).

Taken together, we identified the two crucial promoter regions involved in uPA transcription, and the proximal promoter region was subjected to epigenetic dysregulation by HBx.

DNMT3A2 Is Responsible for the HBx-Induced uPA Down-Regulation

To investigate effect of different types of DNMT on the HBx-induced down-regulation of uPA expression, we cotransfected puPA-Luc (−2064 to +24) with expression vector containing different types of DNMT. The activity of the uPA promoter was repressed only by DNMT3A2 (40.41% ± 0.17%), while DNMT1 (75.75% ± 0.31%) had very little effect (Fig. 2A). The results were further confirmed by the mRNA and protein levels of uPA upon transfection with different types of DNMTs. Transfection of DNMT3A2 significantly down-regulated both mRNA and protein levels of uPA commensurate to HBV1.2(wt) transfection (Fig. 2B). We then examined whether the knockdown of DNMT3A2 restored the HBx-induced uPA suppression. The mRNA and protein levels of DNMT3A2 were specifically reduced by siRNA for DNMT3A2 (Fig. 2C). Under this condition, the HBx-induced down-regulation of uPA was significantly restored to the control level (Fig. 2C), confirming that DNMT3A2 is directly involved in the HBx-induced repression of uPA expression. These results demonstrated that DNMT3A2, among other DNMTs, is responsible for the HBx-induced uPA down-regulation.

Figure 2.

DNMT3A2 is responsible for the HBx-induced uPA down-regulation and recruitment onto the uPA promoter by binding with HBx. (A) Relative luciferase activity assay using uPA-luc (−2064 to +24) in Huh7 cells cotransfected with each pDNMT expression vector. (B) Real-time quantification and western blot analysis of uPA after transfection with the indicated plasmids in Huh7 cells. (C) Quantitative RT-PCR and western blot analysis of DNMT3A in Huh7 cells transfected with the indicated siRNAs. (D) Relative luciferase activity assay of uPA-luc in Huh7 cells transfected with pHBV1.2(wt) or pHBV1.2(HBx-) after treatment with 5′-AzaC, a DNA methyltransferase inhibitor. Expression of uPA mRNA after 5′-AzaC treatment in Huh7-HBx stable cells. (E) HBx coimmunoprecipitated with endogenous DNMT3A in Huh7 cells stably expressing HBx-HA. (F) Schematic representation of the uPA promoter region for ChIP assay. The transcription initiation site is denoted as +1. (F) ChIP assay of the genomic uPA promoter regions (R1 and R2) using anti-DNMT3A antibody in HBx-stable Huh7 cells.

Finally, to validate the involvement of DNMT activity in the HBx-induced uPA suppression, we treated cells with 5-aza-2′-deoxycytidine (5-AzaC), a DNMT inhibitor. Luciferase assays revealed that the uPA promoter activity was restored by 5-AzaC treatment (Fig. 2D). Semiquantitative RT-PCR analysis also showed that 5-AzaC treatment significantly restored uPA RNA transcription (Fig. 2D). These results demonstrate that the suppression of uPA by HBx is controlled by the epigenetic regulation of the uPA promoter.

HBx Recruits DNMT3A2 to the uPA Promoter

Knock-down of DNMT3A2 did not affect endogenous uPA expression when HBx was not expressed (Fig. 2C), suggesting that HBx is probably able to recruit DNMT3A2 onto the promoter region of uPA. To address this possibility, the interaction between DNMT3A2 and HBx was investigated by a coimmunoprecipitation assay. As evident in Fig. 2E, HBx coimmunoprecipitated with endogenous DNMT3A2, supporting the idea that HBx interacts with DNMT3A2.

We then further examined if the HBx-DNMT3A2 complex is capable of binding to the uPA promoter by the ChIP assay. As shown in Fig. 2F, the hypermethylated promoter region (−230 to +52, R1) was analyzed in the ChIP assay and a part of exon 1 (+367 to +656, R2) was used as a negative control. DNMT3A2 bound only to the de novo methylated R1 in HBx-expressing Huh7 cells (Fig. 2F). These data, combined with the result of Fig. 2C, demonstrated that HBx binds with DNMT3A2 and recruits this enzyme complex to the CpG island of the uPA promoter, resulting in the hypermethylation-mediated suppression of uPA.

HBx Blocks uPA-Mediated HGF Activation, Inhibiting HGF-Induced Liver Cell Proliferation

To investigate whether uPA down-regulation by HBx can directly regulate the extracellular modulation of HGF activation, we first analyzed the uPA activity using a zymography assay. The activity of uPA was significantly decreased in the media of HBx-expressing Huh7 cells (Fig. 3A). To examine the effect of HBx-induced suppression of uPA activity upon HGF activation, we incubated the pro-form of recombinant human HGF (rhHGF) with the cell supernatants and analyzed the cleavage product, the active HGF (a-HGF), by western blot. A pronounced decrease in the level of pro-form was accompanied with a simultaneous increase of active form of HGF in the medium of control Huh7-pcDNA cells, whereas such a change was not observed in the medium of Huh7-HBx stable cells (Fig. 3B). These data suggest that the down-regulation of uPA by HBx is functionally related to the inhibition of HGF activation.

Figure 3.

HBx inhibits the uPA-mediated HGF activation and the HGF-induced liver cell proliferation. (A) uPA activity analysis by casein zymography using the culture medium of HBx stable Huh7 cells. (B) Inhibition of pro-HGF processing in HBx cells. Cell culture media was changed with serum-free media for 6 hours and treated with recombinant pro-HGF for 1 hour. The media were harvested and concentrated under vacuum and resolved by 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) to detect the cleaved (active) HGF protein. (C) Cell proliferation assay with/without pro-HGF treatment. The medium was replaced with fresh medium and cells were cultured in the presence or absence of pro-HGF (50 ng/mL). Viable cells were counted using Trypan blue. Each plot represents the mean of three independent experiments.

Next, to determine whether the inhibition of HGF activation has an influence on hepatocyte proliferation, control and HBx-expressing Huh7 cells were cultured with or without the pro-form of rhHGF (pro-HGF), and the subsequent cell proliferations were examined. As shown in Fig. 3C, significant retardation of cell proliferation was observed only in HBx-expressing liver cells when stimulated by pro-HGF, whereas both cell lines showed similar proliferation rates when cells were not stimulated by pro-HGF. These results indicated that the HBx-induced repression of uPA is functionally linked to the retardation of hepatocyte proliferation under HGF stimulation.

HBV Inhibits Liver Regeneration by the Epigenetic Regulation of uPA In Vivo

To investigate whether the inhibition of liver cell proliferation by HBV infection or HBx can also be observed in vivo, we established a mouse model of acute HBV infection through hydrodynamic injection of viral DNA.[35] Plasmids of pHBV1.2(wt) or pHBV1.2(HBx-) were hydrodynamically injected into the mice tail vein. The levels of HBV replication in liver tissues were analyzed by Southern blot analysis and the expression levels of viral proteins by western blot, enzyme-linked immunosorbent assay (ELISA), and immunohistochemistry using liver tissues and sera.

As shown in Fig. 4A, HBV DNA replication was dominantly observed only in wt HBV mice, whereas only a marginal replication level was detected in HBx-null mice. The expression levels of viral proteins (hepatitis B surface antigen [HBsAg] and hepatitis B core antigen [HBcAg]) were not significantly changed in either wt and HBx-null mice. Moreover, partial hepatectomy did not affect the expression of viral protein (Fig. 4C). Given that the function of HBx is critical for HBV replication,[35] the data demonstrate the suitability of the mouse model of acute HBV infection for the functional study of HBV in vivo.

Figure 4.

HBx blocks liver regeneration in a mouse model of HBV infection. (A-C) Validation of the mouse model of HBV infection. HBV replication by Southern blot analysis (A) and immunohistochemical staining of HBsAg and HBcAg (B) in mouse liver tissues. HBsAg expression by western blot were analyzed using the liver tissues obtained at 4 days after hydrodynamic injection (C, top) and measurement of HBsAg levels in mouse sera by ELISA after hydrodynamic injection of DNA as well as partial hepatectomy (PH) (C, bottom). Time schedule for partial hepatectomy is indicated. (D,E) Liver regeneration is inhibited by HBx in the HBV mouse model. Macroscopic evaluation of the regenerated livers of HBV or HBx mouse model 8 days after partial hepatectomy. Boxplots of liver weight/body weight 8 days after partial hepatectomy. (F) Liver function test after partial hepatectomy were performed as described in the Supporting Materials and Methods.

To explore the possibility that the liver expression of HBx affects liver regeneration, we examined the status of a regenerated liver after partial hepatectomy of HBV model mice according to the time schedule as shown in Fig. 4C. Importantly, a marked reduction of the regenerated liver at 8 days after partial hepatectomy was observed upon wt HBV infection, in contrast to the mice of HBx-null HBV infection where normal liver regeneration was apparent (Fig. 4D). Similar results were also observed from the mouse model of CMV-driven HBx expression (Fig. 4E). The results demonstrated that the expression of HBx inhibits liver regeneration in partial hepatectomy-triggered liver damage. There were no significant differences in the kinetics of serum aspartate transaminase, alanine transaminase, total protein, and albumin after partial hepatectomy among the tested groups at all timepoints examined, suggesting that HBV infection or HBx expression do not significantly affect basic biochemical function of the liver (Fig. 4F).

We then examined the expression of mouse uPA (muPA) in liver tissues by immunohistochemistry and western blot analysis. The expression level of muPA was significantly decreased in liver tissues of wtHBV-expressing mice as compared to that of HBx-null HBV mice (Fig. 5A). Western blot analysis with different antibodies (capable of detecting reduced and nonreduced muPA) also revealed that the muPA protein in liver tissues was reduced in wt HBV mice. The level of muPA reduction was more prominent in liver tissues of CMV-driven HBx expression (Fig. 5B). The level of the muPA mRNA was also decreased in HBx liver tissues, whereas the expression of active HGF receptor, c-met, remained constant (Fig. 5C). The results confirmed that the muPA was down-regulated in the liver by HBx, further extending similar results in vitro (Fig. 1D,E).

Figure 5.

uPA is down-regulated in the liver of HBV or HBx model mice through epigenetic control on the uPA promoter. (A) Mouse uPA (muPA) expression by immunohistochemical staining in mouse liver tissues. (B) muPA protein expression by western blot analysis in the mouse liver under nonreducing and reducing conditions. (C) Semiquantitative RT-PCR analysis of muPA, c-met, and viral RNA in the mouse liver. (D) Methylation analysis of muPA promoter by bisulfide sequencing using the genomic DNA of HBV model mice. Percentages were calculated by methylated CpG/total CpG of the muPA promoter region with at least three independent experiments. Open circle, unmethylated CpG; black circle, methylated CpG. (E) Ki 67 (marker of cell proliferation) expression in the formalin-fixed liver tissues obtained at 48 hours after partial hepatectomy. (F) Percentage of Ki67-positive hepatocytes was calculated using at least three independent experimental sets.

To investigate whether the expression of muPA is regulated by epigenetic control, we analyzed the methylation status of the muPA promoter. Bisulfite sequencing revealed that the CpG island of the muPA promoter (−385 nt to −220 nt) is highly methylated in the livers of wt HBV-expressing mice (42.6%) compared to those of HBx-null mice (4.8%) (Fig. 5D). Confirming and further extending similar results in human liver cells (Fig. 1), these results suggest that HBx represses the uPA expression in the mouse liver through epigenetic control of the uPA promoter.

Next, we examined whether the rate of hepatocyte proliferation is affected by HBx by counting the Ki67-positive cells (marker of cell proliferation) after partial hepatectomy (Fig. 5E). While there was no sign of cell death or apoptosis, the Ki67-positive cells by immunohistochemical staining were markedly reduced in wt HBV-expressing mouse liver tissues as compared to the control or HBx-null liver tissues (Fig. 5F). These results collectively suggest that HBx inhibits liver regeneration by way of epigenetic control of uPA in vivo.

To investigate the kinetic change of uPA by HBx in mouse liver, we determined the expression and activity of uPA at several timepoints after partial hepatectomy. As shown in Fig. 6A,B, wt HBV significantly suppressed and delayed the expression and the activity of uPA.

Figure 6.

Kinetic analysis of uPA, IL-6, and TNF-α expression after partial hepatectomy. (A) Kinetic analysis of uPA expression in liver tissues after partial hepatectomy by western blot. The relative intensity of each band was quantitated using Bio-Imaging analyzer (LAS-3000). *P < 0.05, **P < 0.005, ***P < 0.001. (B) Kinetic analysis of uPA activity in liver tissues after partial hepatectomy by Casein-plasminogen zymography at the indicated timepoints. Kinetic analysis of serum IL-6 (C) and TNF-α (D) after partial hepatectomy at the indicated timepoints. (E) Knockdown of uPA expression in mouse liver by hydrodynamic injection of siRNA for uPA. (F) HBx effect on liver regeneration in uPA knockdown mice. Partial hepatectomy were performed 4 days after co-hydrodynamic injection of HBV plasmid and siuPA. At least three mice were used for the analysis of each group.

We also examined the potential role of other proregenerative factors such as IL-6 and TNF-α in this model system. The results showed that, except for only a partial elevation of TNF-α at early timepoints, there were no significant change in the cytokine profile among the experimental groups (Fig. 6C,D), suggesting that the uPA pathway is dominantly responsible for this system. This conclusion was further strengthened in the uPA knockdown mice, where HBx exerted no appreciable effects on liver in uPA knockdown background (Fig. 6E,F). Collectively, our results demonstrated that the uPA pathway is mainly involved in HBV-mediated inhibition of liver regeneration.

Treatment of DNMT Inhibitor Restores Liver Regeneration In Vivo

We examined whether the inhibition of DNMT can compensate HBx and restore the liver regeneration. The DNMT activity was chemically inhibited by injection of 5′AzaC in mouse tail vein following the time schedule shown in Fig. 7A. The expression level of uPA, which was down-regulated in the HBx containing mouse liver, was restored by the treatment of 5′AzaC (Fig. 7B). Importantly, the liver regeneration was almost completely restored by the treatment of DNMT inhibitor (Fig. 7C). The results are consistent with and further extend the observation made with the tissue culture system (Fig. 2) that the activity of DNMT is essential for HBV-mediated inhibition of liver regeneration.

Figure 7.

Inhibition of DNMT activity restores liver regeneration in vivo. (A) Experimental scheme for 5′AzaC treatment (5 mg/kg injection before partial hepatectomy, and 1 mg/kg after partial hepatectomy). (B) Analysis of uPA expression in liver tissues after 5′AzaC injection as determined by western blot. (C) Boxplots of relative liver regeneration at 8 days after partial hepatectomy before and after 5′AzaC injection. At least three mice were used for the analysis of each group. HI, hydrodynamic injection; PH, partial hepatectomy.

Reconstitution of Aberrant uPA Expression Restores Liver Regeneration In Vivo

As a direct proof of down-regulation of uPA as a major molecular event for HBV-mediated inhibition of liver regeneration, we examined whether the supplementation of exogenous uPA could rescue the HBx-mediated inhibition of liver regeneration. uPA was supplemented with HBV by hydrodynamic injection of the muPA gene according to the experiment schedule shown in Fig. 8A.

Figure 8.

Supplementation of exogenous uPA restores liver regeneration in vivo. (A) Experimental scheme for supplement of mouse uPA and partial hepatectomy. (B) Macroscopic evaluation of the regenerated livers 8 days after partial hepatectomy. Values represent mean ± SD of liver weight. RML, right median lobe; RLL, right liver lobe; PC, posterior caudate lobe; AC, anterior caudate lobe. (C) Boxplots of liver weight / body weight 8 days after partial hepatectomy. (D) Ki67 staining in liver tissues at 48 hours after partial hepatectomy. mRNA (E) and protein (F) expression of muPA was determined by RT-PCR and western blot analysis using mouse liver tissues.

In contrast to a marked inhibition of liver regeneration by HBV, a robust and marked liver regeneration was observed 8 days after partial hepatectomy when uPA was supplemented (Fig. 8B,C). The expression level of Ki67 as a marker of cell proliferation (Fig. 5E) was also restored in the mouse livers upon uPA supplement (Fig. 8D).

The expression of uPA was confirmed by semiquantitative RT-PCR and western blot using the partial hepatectomized liver tissues. As shown in Fig. 8E,F, the expression levels of uPA mRNA and protein were down-regulated by wt HBV. Yet, supplementation with the uPA gene restored the expression of uPA mRNA and protein. These results suggest that the down-regulated uPA is the main molecular cause of the HBV-mediated inhibition of liver regeneration.

Discussion

 Infection by HBV causes acute or chronic liver damage by dysregulation of hepatocyte proliferation and liver regeneration. Among various HBV proteins, the HBx protein is suggested to be associated with the inhibition of liver cell proliferation triggered by partial hepatectomy.[11, 12] However, until now the mechanism of the HBx-mediated disturbance of liver regeneration has not been elucidated. In this study we investigated the molecular mechanism underlying the interference of liver regeneration pathways due to HBV infection using a mouse HBV infection model. Our study revealed that HBx inhibits liver regeneration by way of suppression of uPA expression through epigenetic control of the uPA promoter by recruiting DNMT3A2. The involvement of uPA in the HBV infection model was further confirmed by the functional analysis of HGF activation (Fig. 3) and in vivo reconstitution experiment of uPA (Fig. 8).

The liver is a unique organ, with an ability to regenerate itself against the loss of hepatocytes by environmental factors including drugs, ingested toxins, or pathogen infections.[13, 16, 46] After losing a considerable mass of the liver, mature hepatocytes begin to proliferate by the stimulation of regenerative factors released from parenchymal (hepatocytes) and nonparenchymal liver cells (including Kupffer cells and stellate cells). Several growth factors and cytokines have been shown to regulate hepatocyte proliferation. Among them, HGF and cytokines, including IL-6 and TNF-α, are major factors activated during liver regeneration triggered by partial hepatectomy or liver damage.[14, 16-18] IL-6 and TNF-α are secreted from Kupffer cells by the activation of gut-derived factors such as lipopolysaccharides. HGF is a main initiator of liver regeneration, which is released as a precursor, pro-HGF, from the stellate cells and is cleaved by uPA, which is secreted by the hepatocytes at the site of HBV infection.[19-21] Our data showed that HBx blocks the synthesis and secretion of uPA, thereby leading to the inhibition of uPA-mediated HGF activation and liver cell proliferation. Our results showed that the level of methylation on the uPA promoter is significantly higher in human liver cells compared to that of mouse liver cells (Figs. 1H, 5D). While our study does not formally exclude the possibility of potential involvement of cytokines such as IL-6 and TNF-α, our data convincingly indicates that uPA is a major factor, as shown by a direct rescue of HBx-mediated inhibition of liver regeneration by uPA (Fig. 8).

High levels of pro-HGF have been reported in the plasma of patients with liver diseases such as acute/chronic hepatitis and cirrhosis, but not in the plasma of normal subjects.[36] Considering that the majority of patients with acute/chronic hepatitis and cirrhosis are infected with HBV worldwide, the high level of pro-HGF in those patients is probably due to the down-regulation of serum uPA by HBV infection, as shown in Fig. 3. In addition, active HGF was shown as a negative growth regulator for HCC cells and proposed as a tumor suppressor during the early stages of liver carcinogenesis, although it was also shown to stimulate normal hepatocytes.[37-39] Very recently, it was demonstrated that Gab1 plays a central role in the regulation of HGF-mediated hepatoma growth inhibition.[40] In our study, pro-HGF was accumulated, and the active form of HGF was barely detectable in HBx-expressing hepatocytes (Fig. 3). Therefore, the accumulation of unprocessed pro-HGF resulting in a low level of active HGF may be associated with HBV-mediated liver diseases such as acute/chronic hepatitis, cirrhosis, and HCC development.

Liver damage by HBV infection and subsequent activation of liver regeneration signals to compensate for the loss of hepatocytes may enforce a rapid turnover of hepatocytes, which has been suggested as a risk factor for HCC development.[1, 2] A study reported a premature cell cycle entry by the HBx protein during compensatory liver regeneration.[41] As such, the retardation of liver regeneration by HBV infection may hyperactivate alternative (nongrowth factor) regeneration signal. It is worth reminding that one of the liver regeneration signals, TNF-α, can also sensitize the apoptosis of HBx-expressing liver cells by interacting with antiapoptotic c-FLIP.[8] This may in turn trigger a compensatory, or possibly an excessive, liver regeneration. This malignant loop may accelerate the dysregulation of hepatic signals and hepatocyte turnover, eventually leading to the development of HBV-mediated liver diseases. Further studies are needed to address the complex role of HBV between liver regeneration and disease development.

There is abundant evidence that the plasminogen activator system plays an important role in tissue repair,[27-30] including the liver, kidney, and skin. Previous studies have reported a retardation of liver regeneration in the uPA-deficient (uPA−/−) mice after liver injury.[28, 31] Shimizu et al.[28] reported that an impaired activation of HGF after Fas-mediated massive hepatic apoptosis was associated with retarded liver regeneration in uPA−/− mice. Our results bridges and reconciles seemingly disparate observations independently made from HBV-transgenic[11, 12] and uPA−/− mice,[28] and provides a rationale at the molecular level for the inhibition of liver regeneration.

Recently, HBx has been shown to induce epigenetic regulation of both the host genes and its own viral cccDNA (covalently closed circular HBV DNA), which serves as a template for the transcription of all viral RNAs.[42-45] For instance, HBx expression increased the total DNMT activity by up-regulating DNMT1, DNMT3A1, and DNMT3A2 and selectively promoting regional hypermethylation of a specific tumor suppressor gene, insulin-like growth factor-3.43 Moreover, HBx activated the expression of DNMT1, which leads to inhibition of expression of tumor suppressor genes, such as p16INK4a and E-cadherin, by promoter hypermethylation.[42] Interestingly, Zheng et al.[44] have shown through ChIP experiments that HBx can interact directly with DNMT3A and histone deacetylase 1 (HDAC1) and recruit DNMT3A to the regulatory promoters of IL-4 receptor and metallothionein-1F and subsequently silences their transcription by de novo DNA methylation.

More recently, it was shown that HBx is recruited to the HBV minichromosome (a complex of cccDNA with histones and nonhistone viral and cellular proteins) and regulates the epigenetic control of cccDNA function and the eventual viral replication in HBV replicating cells.[45] Here we further extend the HBx-mediated epigenetic target, the uPA promoter, where HBx recruits the DNA modifying enzyme facilitating the promoter hypermethylation and repression of transcription.

In conclusion, the HBx protein of HBV directly regulates the expression of the uPA gene in vitro and in vivo through epigenetic regulation by hypermethylation of the uPA promoter. As it is an essential enzyme for HGF activation, the down-regulation of uPA subsequently results in hypoactivation of pro-HGF, eventually hampering liver regeneration. The present report provides a molecular basis for the impairment of liver regeneration by HBV infection. Unraveling the molecular mechanisms underlying the dysregulation of liver regeneration by HBV infection will not only extend the understanding of pathophysiology of virus-mediated liver diseases, but ultimately provide new treatment options for patients with liver diseases.

Ancillary