Potential conflict of interest: Nothing to report.
Ground glass hepatocytes (GGH) in chronic hepatitis B virus (HBV) infection harbor HBV pre-S deletion mutants in endoplasmic reticulum (ER) and exhibit complex biologic features such as ER stress, DNA damage, and growth advantage. The presence of pre-S mutants in serum has been shown to predict the development of hepatocellular carcinoma (HCC) in HBV carriers. GGHs hence represent a potentially preneoplastic lesion. Whether a specific growth factor is overexpressed and activated in GGHs remains to be clarified. In this study, growth factor(s) up-regulated by pre-S mutants was identified using a growth factor array in HuH-7 cells. Immunohistochemistry, reverse-transcriptase polymerase chain reaction, and Western blot analysis were performed to study the participation of these genes and their signal pathways in HuH-7 cells and liver tissues. We demonstrate that vascular endothelial growth factor-A (VEGF-A) was up-regulated by pre-S mutants in HuH-7 cells and further confirmed in GGHs by immunostaining. The VEGF-A up-regulation by pre-S mutants could be suppressed by vomitoxin, an ER stress inhibitor. Furthermore, pre-S mutants-expressed HuH-7 cells exhibited activation of Akt/mTOR (mammalian target of rapamycin) signaling and increased growth advantage, which could be inhibited by VEGF-A neutralization. Consistent with this notion, enhanced expression of VEGF-A and activation of Akt/mTOR signaling, comparable to the levels of paired HCC tissues, were also detected in HBV-related nontumorous livers. Conclusion: The enhanced expression of VEGF-A in GGHs provides potential mechanism to explain the progression from preneoplastic GGHs to HCC in chronic HBV infection. (HEPATOLOGY 2009;49:1962–1971.)
Hepatocellular carcinoma (HCC) is one of the most important causes of cancer death in the world, particularly in southeastern Asia and sub-Saharan Africa.1, 2 Epidemiological studies have established that chronic hepatitis B virus (HBV) infection is a major risk factor for HCC.3, 4 Several mechanisms have been proposed to explain the progression from chronic HBV infection to HBV-related HCC, including the insertional mutagenesis of HBV genomes, inflammation and regernerative hyperplasia initiated by an immune response to HBV infection, and the transactivator functions of HBV gene products such as HBx and truncated middle surface antigens, which may initiate the Ras/Raf-1/ERK signaling.5–8
HBV-related hepatocarcinogenesis involves multiple steps and the exact oncogenic mechanism for the progression from a benign precursor lesion to HCC remains to be established. Previously, we found that ground glass hepatocytes (GGHs) contain pre-S deletion mutants which are accumulated in the endoplasmic reticulum (ER) and induce ER stress.9 Two types of GGHs, designated Type I and Type II GGHs, have been identified and contain specific large HBV surface antigen (LHBs) mutants with deletions at either the pre-S1 (pre-S1 mutant, ΔS1) or pre-S2 (pre-S2 mutant, ΔS2) region, respectively.9–11 Type I GGHs usually distribute sporadically and express an inclusion-like pattern of surface antigens, whereas Type II GGHs consistently cluster in groups and occur at advanced stages of infection.10–12 Both types of pre-S mutants could initiate ER stress-dependent signals to induce oxidative DNA damage.13 The pre-S2 mutant could additionally induce an ER stress-independent signaling through JAB1/E2F/retinoblastoma/cyclin A signaling to promote hepatocyte proliferation.14 Transgenic mice harboring pre-S2 mutants could induce nodular dysplasia and HCC.15, 16 Furthermore, a prospective study revealed a predictive value of the pre-S deletion mutants in the development of HCC.17 Taken together, these findings suggest that GGHs, particularly the Type II GGHs, represent the preneoplastic lesions of HBV-associated HCC.18
HBV infection can induce chronic inflammation and contribute to HCC formation through the expression of cytokines, such as interleukin-6 and tumor necrosis factor-alpha.19, 20 HBx has been shown to alter cytokine expression to modulate the immune response and proliferation of hepatocytes.21–23 In transgenic mice of large surface antigens,6, 24 the overproduction of large envelope protein can cause inflammation and regenerative hyperplasia to induce HCC development. Therefore, inflammatory cytokines or growth factors may play a role in the disease progression from a precursor lesion to HCC through activation of growth factor/receptor signaling involving Akt/mTOR (mammalian target of rapamycin) or Ras/Raf-1/ERK.25 In this study we tested the hypothesis whether GGHs exhibit enhanced expression of specific cytokine or growth factor that may facilitate the progression from GGHs to HCC by using a cytokine/growth factor antibody array to screen for the candidate growth factors up-regulated by pre-S mutants in the HuH-7 cell line.
The plasmids of pIRES-WT, pIRES-ΔS1, and pIRES-ΔS2 expression of wild-type, pre-S1 mutant (nucleotide [nt] 3040–3111 deletion) and pre-S2 mutant (nt 4–57 deletion) large surface protein were established as described.15
The HuH-7 cell line was used in this study. All transfections were performed with Lipofectamine 2000 according to the manufacturer's instruction (Invitrogen Life Technologies, Carlsbad, CA).
Human Cytokine/Growth Factor Antibody Array.
The human cytokine/growth factor antibody array kit was purchased from RayBiotech (Norcross, GA) and used according to the manufacturer's instruction. Briefly, the membranes were incubated for 30 minutes in blocking buffer, and then 1 mL of culture medium from cells 48 hours after transfection was added and incubated at room temperature for 2 hours. After this, the membranes were washed and then incubated for 1 hour with biotin-conjugated antibody. After incubation the membranes were washed and then incubated for 2 hours with horseradish peroxidase (HRP)-conjugated streptavidin. The membranes were developed with the detection buffer, exposed to film, and processed by autoradiography. The intensities of signals were quantified with a densitometer.
Real-Time Polymerase Chain Reaction.
Total RNAs were extracted from cells 48 hours after transfection using REzol (PROtech Technologies, Taipei, Taiwan) reagent and used according to the manufacturer's instruction. Four μg RNA was converted to complementary DNA (cDNA) using oligo(dT) primers and SuperScript III reverse transcriptase (Invitrogen Life Technologies) and then polymerase chain reaction (PCR) was performed. The primers used were as follows: VEGF, 5′-AGGAGGAGGGCAGAATCATCA-3′ (sense), 5′-TCTCGATTGGATGGCAGTAGC-3′ (antisense); VEGFRII, 5′-TCAGCATAAGAAACTTGTAAACCGA-3′ (sense), 5′-ATGGACCCTGACAAATGTG-3′ (antisense); GAPDH, 5′-ACAGTCAGCCGCATCTTCTT-3′ (sense), 5′-GACAAGCTTCCCGTTCTCAG-3′ (antisense).
Western Blot Analysis.
Protein lysates were harvested from cells 48 hours after transfection, washed with phosphate-buffered saline (PBS), and lysed on ice in 1× RIPA buffer (Upstate Biotechnology, Lake Placid, NY) supplemented with complete protease inhibitor cocktail tables (Roche, Nutley, NJ). About 25–35 μg of protein were separated by SDS− polyacrylamide gel electrophoresis and then transferred to polyvinylidene difluoride membranes. The membranes were probed with primary antibody. After incubation with HRP-conjugated antibody, antibody detection was achieved by enhanced chemiluminescence (ECL; PerkinElmer Life Science, Boston, MA) according to the manufacturer's protocol. The primary antibodies used in this study were as follows: anti-VEGF, anti-VEGFRII, anti-PKB/Akt1, anti-PKB/Akt2, antiphospho-PKB/Akt (Ser-473), anti-mTOR, anti-ERK, and antiphospho-ERK antibodies, all from Santa Cruz Biotechnology (Santa Cruz, CA). Antiphospho-mTOR was purchased from Abcam (Cambridge, UK). Antiphospho-Raf was purchased from Cell Signaling Technology (Danvers, MA). Anti-Raf-1 was purchased from Epitomics (Burlingame, CA). Anti-HA-tag was from Zymed Laboratories (South San Francisco, CA). Anti-β-actin was from Chemicon (Temecula, CA).
Enzyme-Linked Immunosorbent Assay.
The growth factors in culture supernatant were assayed by enzyme-linked immunosorbent assay (ELISA). Culture supernatants were collected 48 hours after transfection and then subjected to ELISA according to the manufacturer's instructions. The VEGF ELISA kit was purchased from Biosource International (Camarillo, CA).
Cell Proliferation Assay.
A total of 6 × 106 cells were transfected by microporator (Digital Bio Technology, Seoul, Korea) with 2.4 μg of DNA in 120 μL resuspension buffer. The electroporation settings were 1350 mV, 40 ms, and 1 pulse. Cells were then seeded into 96-well plates with 2000 cells per well 4 hours after transfection. After treatment with/without anti-VEGF antibody (R&D Systems, Minneapolis, MN) for 3 and 6 days, the culture mediums were removed. Then all cells were treated with MTT (Sigma, St. Louis, MO) for 3 hours. The medium was aspirated after centrifugation at 2500 rpm for 15 minutes and then the well was dried. After mixture with 100 μL of 0.1 N HCl/isopropanol into each well, the cell lysates were measured for absorbance at 570 nm.
Human Umbilical Vein Endothelial Cell Proliferation Assay.
To further test for the biologic activity of VEGF-A, the human umbilical vein endothelial cell (HUVEC) proliferation assay was used. The culture supernatants were collected at 48 hours after transfection of HuH-7 cells with a pre-S mutant and then were added to HUVECs seeded on gelatin-coated 96-well tissue culture plates containing M199 medium. Two days after treatment, HUVEC density was measured using cell titer 96 one aqueous cell proliferation kits (Promega, Madison, WI) according to the manufacturer's instruction.
Tissue samples were obtained from patients with HBV-positive HCC. The formalin-fixed, paraffin-embedded sections were used for human liver tissues. For immunohistochemical staining, the sections were incubated with rabbit anti-VEGF (Lab Vision, Fremont, CA) and mouse anti-HBsAg (hepatitis B surface antigen) (BioGenex, San Ramon, CA). The sections were subsequently assayed with the Super Sensitive Polymer-HRP IHC Detection System (BioGenex) according to the manufacturer's instructions. Nuclei were then stained with Mayer's hematoxylin in all sections.
Reverse Transcriptase PCR.
Total RNAs were extracted from cells 48 hours after transfection using REzol (PROtech Technologies) reagent and used according to the manufacturer's instruction. Four μg RNA was converted to cDNA using oligo (dT) primers and SuperScript III reverse transcriptase (Invitrogen Life Technologies) and then PCR was performed. The primers used were as follows: GRP78, 5′-GAAGACTTTGACCAGCGT-3′ (sense), 5′-CAATTTCAATTCTTGCTTGATGC-3′ (antisense); β-ACTIN, 5′-TGGAGAAAATCTGGCA-CCAC-3′ (sense), 5′-GAGGCGTACAGGGATAGCAC-3′ (antisense). The cycling conditions of PCR were 15 seconds at 95°C, 15 seconds at 58°C, and 15 seconds at 72°C.
Data were analyzed by analysis of variance (ANOVA) (m × n factorial design), or Student's t test (2-tailed) and shown with standard deviations (SDs). A P value < 0.05 was considered significant.
VEGF-A Was Identified as the Candidate Growth Factor Up-Regulated by Pre-S Mutants Using the Cytokine/Growth Factor Array Assay.
In this study the plasmid pIRES-hrGFP-2a was used as the expression vector to express HBV wild-type, ΔS1, and ΔS2 LHBs in HuH-7 cells, as shown previously (Fig. 1A).15 Cytokines and growth factor secretion were assessed using conditioned media after 48 hours posttransfection and arrayed by the RayBiotech system (Fig. 1B). The data showed that hepatocytes expressing pre-S1 and pre-S2 mutant proteins could enhance the secretion of several growth factors including VEGFs (indicated), fibroblast growth factor (FGFs), transforming growth factor (TGFs) (Fig. 1C), and hepatocyte growth factor (HGF) (Supporting Table 1). VEGF-A was selected for detailed study because of its angiogenesis and growth role in the early stage lesions of human carcinogenesis.26, 27 The production of VEGF-A in hepatocytes expressing pre-S2 mutant was slightly higher than that in the pre-S1 mutant.
We next confirmed the expression of VEGF-A using real-time PCR analysis. As compared to the control, the transcription level of VEGF-A was increased in hepatocytes expressing pre-S1 or pre-S2 mutant protein, being highest in pre-S2 mutant-expressing cells (Fig. 2A). Western blot assay confirmed that both pre-S mutants also increased the expression of VEGF-A up to 1.52-fold of the control in HuH-7 cells (Fig. 2B). In culture supernatants, the secretion of VEGF-A was also enhanced from hepatocytes expressing pre-S mutants by ELISA test, particularly for the pre-S2 mutant (Fig. 2C).
Expression of VEGF-A Was Up-Regulated by Pre-S Mutants Probably Through ER Stress.
Because the induction of ER stress could up-regulate the expression of VEGF-A,28, 29 we used ER stress inhibitor vomitoxin30 to clarify whether ER stress induced VEGF-A expression in this study. By reverse transcriptase PCR (RT-PCR), the expression of ER stress chaperone GRP78 was reduced after treatment with vomitoxin for 24 hours (Fig. 2D, lower panel). Western blot assay showed that the expression of VEGF-A was reduced in parallel by treatment with vomitoxin in hepatocytes expressing pre-S1 or pre-S2 mutant protein (Fig. 2D, upper panel).
Enhanced Expression of VEGF-A Promotes Cell Proliferation.
To elucidate whether up-regulated VEGF-A by pre-S mutants does have biologic activities to promote the growth of hepatocytes, we further performed cell proliferation assays on HuH-7 hepatocytes. The effect of VEGF-A on endothelial cells was run in parallel on HUVEC cells. The proliferation of hepatocytes was enhanced by pre-S mutants at day 3 in HuH-7 cells and even more profoundly at day 6 after change of the culture medium (Fig. 3A). In order to further clarify whether the enhanced proliferation of hepatocytes was related to an autocrine or paracrine effect of VEGF-A secreted in the culture supernatant, neutralizing VEGF antibody (1 μg/mL) was added to the culture supernatant. We observed that the enhanced growth of hepatocytes was significantly suppressed by VEGF neutralization, as compared with that of untreated cells (Fig. 3A). Subsequently, we detected whether the produced VEGF-A could also affect the growth of HUVECs. As shown in Fig. 3B, the culture supernatants of VEGF-A in hepatocytes expressing pre-S1 and pre-S2 mutant proteins significantly increased HUVEC proliferation by 20% as compared with the control M199 medium.
Activation of the VEGF-A Downstream Akt/mTOR Signaling Through VEGF Receptor-2 (VEGFR-2) by Pre-S Mutants in HuH-7 Cells.
To further study the mechanism of VEGF-A-induced cell growth in GGHs, we first assessed the expression of VEGF receptor-1 and -2 (VEGFR-1, -2), which are reported to function in the regulation of VEGF-A signaling. As compared to the control, VEGFR-2 mRNA was significantly up-regulated, predominantly by pre-S2 mutant, in the HuH-7 cell line (Fig. 4A), which was further confirmed by Western blot assay (Fig. 4B). The expression of VEGFR-1, however, was not increased.
Subsequently, we demonstrated the potentially enhanced expressions of VEGF-A downstream molecules Akt1, Akt2, and the phosphorylated Akt. The expression of mTOR and its phosphorylated form was elevated and activated in hepatocytes expressing pre-S mutants (Fig. 4C). Unexpectedly, another VEGF-A/VEGFR signaling pathway, Raf-1 and ERK, showed no enhanced expression or activation (Fig. 4D). The activation of Akt/mTOR signaling by VEGF-A induced by pre-S mutants was further supported by the attenuation of Akt/mTOR signaling by neutralization with anti-VEGF-A antiserum (1 μg/mL) (Fig. 4E), a finding consistent with the reduction of growth by VEGF-A neutralization in pre-S mutants-expressed HuH-7 cells shown in Fig. 3A.
Expression of VEGF-A in Type I and Type II GGHs by Immunohistochemistry.
Whether VEGF-A was overexpressed in Type I and Type II GGHs was further evaluated and confirmed in liver tissues. Five surgically resected nontumorous livers obtained from HBV-related HCC patients were analyzed for the presence of GGHs and immunostained for the expression of VEGF-A and HBsAg in GGHs. Two normal liver tissues and three HCV-related nontumorous livers were included for comparison. As shown in Fig. 5 Aa,d, the hepatocytes in normal livers showed no detectable expression of HBsAg and VEGF-A. In the nontumorous livers of five HBV-related HCC cases, Type I GGHs displayed an inclusion-like HBsAg expression, scattered in distribution, and were observed in all five samples (Fig. 5 Ab). Type II GGHs, clustered in distribution (Fig. 5 Ac), were observed in only three of five samples. Both Type I and Type II GGHs had enhanced expression of VEGF-A (Fig. 5 Ae,f) as compared to the control livers. The three HCV-related nontumorous livers showed no detectable expression of VEGF except in the bile duct epithelial cells and a few stromal cells (data not shown). The expression of VEGF-A in HCC tissues was variable in expression intensity and pattern. In HBV-related cases, one case showing positive HBsAg immunostaining had enhanced intensities of VEGF-A expression (Fig. 5 Ba,c); three HBsAg-negative HBV-related HCC tissues also showed positive immunostaining for VEGF, one at the tumor margin (Fig. 5 Bd). The remaining tissue had no detectable expression of either HBsAg or VEGF in HCC tissues. The expression of VEGF-A in three HCV-related HCC tissues also showed variable intensity and patterns (data not shown). Collectively, GGHs, either Type I or Type II, had enhanced expression of VEGF-A, as compared to the control livers.
Expression of VEGF/Akt/mTOR Signaling in Paired Nontumorous and HCC Liver Tissues Obtained from HBV-Related HCC Patients by Western Blot Assay.
To ascertain whether the VEGF-A/Akt/mTOR signaling is associated with human HBV-related hepatocarcinogenesis, Western blot analyses were performed on 20 HBV-related paired nontumorous and HCC tissues for the expression of VEGF-A, p-Akt, and p-mTOR. As shown in Fig. 6, we demonstrated the consistently high level expression of VEGF-A and p-Akt/p-mTOR in the nontumorous livers, comparable to that in the HCC tissues. The VEGF-A expression was unexpectedly higher in 9 of 20 nontumorous tissues than those in HCC tissues, and its increase was positively correlated with the higher level expressions of p-Akt or p-mTOR in four of nine nontumorous liver tissues.
In this study we demonstrate for the first time that GGHs had enhanced expression of VEGF-A and activation of Akt/mTOR signaling, providing a potential mechanism for the stepwise progression from a benign lesion to HCC in HBV-related hepatocarcinogenesis. This finding further supports our previous observations that GGHs contain pre-S mutants in ER, induce ER stress and DNA damage, and represent potentially preneoplastic lesions in the liver of chronic HBV infection.18, 31
VEGF-A, one of the first angiogenesis factors identified, is the most important regulator of normal and tumor angiogenesis.27, 32, 33 Recent evidence indicates that VEGF-A may not only act as a paracrine growth factor on the vasculature but also as an autocrine growth factor in stimulating the proliferation of cancer cells expressing VEGF-A receptors.34, 35 Consistent with previous reports, our data in this study demonstrate that VEGF-A secreted from pre-S mutant-expressing HuH-7 hepatocytes could not only enhance the proliferation of endothelial cells, but also promote the proliferation of HuH-7 cells in an autocrine manner. VEGF-A has been reported to play an important role at the “early” stage of multistep tumor development through promoting angiogenesis and tumor growth in many tumor types.26, 27 In the preneoplastic or early stage of tumor development, oxidative stress, hypoxia, and deprivation of nutrients will collectively hamper the precursor lesion to form tumor.26–29, 36, 37 However, VEGF-A at this stage plays the crucial role of promoting angiogenesis and supporting sustained growth of these precursor lesions. Oxidative stress and hypoxia have been shown to induce VEGF-A expression at the early stage of HCC development.38–40 In this study, we demonstrate that the expression of VEGF-A could be inhibited by the ER stress inhibitor vomitoxin, and therefore ER stress response, induced by the accumulation of pre-S mutants in ER, which may play a role in the enhanced expression of VEGF-A in GGHs. In this sense, the enhanced expression of VEGF-A by pre-S mutants is nonspecific for pre-S mutants but related to ER stress. ER stress has also been shown to up-regulate cyclooxygensae-2 through NF-κB and p38 signaling and correlate with VEGF expression in HBV-associated HCC.41, 42 Therefore, ER stress signals, induced by the accumulation of pre-S mutants in ER, may potentially up-regulate VEGF-A in GGHs. Whether hypoxia plays a role in the induction of VEGF-A in this system remains to be clarified. The role of VEGF-A at the early stage of tumor development could be further supported by our data on the enhanced expression of VEGF-A in the majority of paired nontumorous livers. Similar findings were also true for the activation of Akt/mTOR signaling in the nontumorous livers. This observation again supports the potential role of VEGF-A and Akt/mTOR signaling at the “early” stage of HBV-related multistep tumorigenesis.
In this study the expression of VEGFR-2, but not VEGFR-1, was enhanced in hepatocytes expressing pre-S mutants. This observation is unexpected and interesting. The expression level of VEGFR-2 was not affected by VEGF-A neutralization and may relate to a transcriptional mechanism through NF-κB signaling in stress conditions or activated by viral proteins.43, 44 The enhanced expression of VEGFR-2 may synergize with the enhanced VEGF-A production to facilitate the growth of hepatocytes and angiogenesis in an autocrine or paracrine manner, as shown in Fig. 3. Because neutralization of VEGF-A could inhibit the activation of Akt/mTOR signaling and abolish the growth of HuH-7 cells induced by pre-S mutants, VEGF-A likely engages VEGFR-2 to initiate the Akt/mTOR signaling in this system. Consistent with the in vitro data, we demonstrated the enhanced expression of VEGF-A and activation of Akt/mTOR signaling in the majority of HCC tissues, as previously reported in HCC tissues.45–48 The activation of this signal pathway has been reported in many types of human cancers, especially at the early stage of tumorigenesis.49, 50 Cytokines such as tumor necrosis factor-alpha or growth factors secreted from the inflammatory tumor microenvironment may play a promoting role for tumor development through the mTOR pathway.51 Unexpectedly, however, we did not observe the activation of Raf-1/ERK signaling induced by pre-S mutants in this study, although activation of Raf-1/ERK signaling has been reported in HBx and truncated middle surface protein-initiated signaling.5, 6 The Raf-1/ERK signaling is actually activated in the majority (80%) of HCC tissues but only activated in 30% of the paired nontumorous livers in our series (Teng and Su, unpubl. data), suggesting that the activation of Raf-1/ERK signaling may represent a relatively late event in HBV hepatocarcinogenesis.
As shown in Fig. 1, VEGF-A is not the only growth factor activated in GGHs or induced by pre-S mutants because other growth factors such as HGF, FGF, and TGF-β were also identified to be up-regulated in our cytokine/growth factor array (Supporting Table 1). The other growth factors we identified in this study may represent good candidates to investigate in the future. Finally, the mechanisms of HBV-related hepatocarcinogenesis may be complex and GGHs may only represent one potential role, because pre-S mutants could only be identified in around 60% of HBV-related HCC tissues (Wu and Su, unpublished data, 2008). Other factors such as the transactivators of HBx or truncated middle S may interplay and this remains to be clarified.5–8
In conclusion, this study demonstrates for the first time that GGHs exhibit enhanced VEGF-A expression, probably through ER stress induced by the accumulation of pre-S mutants in ER. The enhanced expression and secretion of VEGF-A could activate Akt/mTOR signaling to promote HBV-related hepatocarcinogenesis through VEGFR-2, thereby providing a potential mechanism for the progression from a benign precursor lesion of GGHs to HCC, providing a potential target for chemoprevention in the high-risk group of patients with chronic HBV infection.