These authors contributed equally to this work.
Association of human APOBEC3 cytidine deaminases with the generation of hepatitis virus B x antigen mutants and hepatocellular carcinoma†
Article first published online: 10 SEP 2007
Copyright © 2007 American Association for the Study of Liver Diseases
Volume 46, Issue 6, pages 1810–1820, December 2007
How to Cite
Xu, R., Zhang, X., Zhang, W., Fang, Y., Zheng, S. and Yu, X.-F. (2007), Association of human APOBEC3 cytidine deaminases with the generation of hepatitis virus B x antigen mutants and hepatocellular carcinoma. Hepatology, 46: 1810–1820. doi: 10.1002/hep.21893
Potential conflict of interest: Nothing to report.
- Issue published online: 28 NOV 2007
- Article first published online: 10 SEP 2007
- Manuscript Accepted: 25 JUN 2007
- Manuscript Received: 22 JAN 2007
- National Science Foundation of China. Grant Number: NSFC-30425012
- Cheung Kong Scholars Program Foundation of the Chinese Ministry of Education
- Research grants from the NIH. Grant Number: AI062644
Human APOBEC3 (apolipoprotein B mRNA editing enzyme, catalytic polypeptide 3) cytidine deaminases have been shown to be potent inhibitors of diverse retroviruses including Vif-deficient human immunodeficiency virus 1 (HIV-1), hepatitis virus B (HBV), adeno-associated virus, and endogenous retroelements. Despite the fact that these enzymes are known to be potential DNA mutators and to target retroviral DNA for cytidine deamination, the pathological effects of their deregulated expression in human diseases are not yet clear. Mutants of the viral HBx protein have been implicated in the carcinogenesis of hepatocellular carcinoma (HCC); however, little is known about how or why such mutants are generated in the human liver. Here, we report that a number of APOBEC3 deaminases preferentially edit the HBx region of HBV DNA and generate C-terminally truncated HBx mutants. Our functional studies indicated that APOBEC3-mediated HBx mutants, especially the C-terminally truncated mutants, cause a gain of function that enhances the colony-forming ability and proliferative capacity of neoplastic cells. Furthermore, we detected G-to-A hypermutation-mediated HBx mutants in preneoplastic liver tissues of selected patients with active chronic HBV infections. We also observed that the APOBEC3B (A3B) cytidine deaminase was widely up-regulated in HCC tumor tissues; it also promoted the growth of neoplastic human HepG2 liver cells and up-regulated heat shock transcription factor1 (HSF1) expression. Conclusion: These findings suggest that some of the APOBEC3 deaminases play a role in the carcinogenesis of HCC through the generation of HBx mutants, providing preneoplastic and neoplastic hepatocytes with a selective clonal growth advantage. Deregulated expression of A3B in liver tissues may also have the potential to promote genetic instability and tumorigenesis. (HEPATOLOGY 2007.)
Hepatocellular carcinoma (HCC) is one of the most prevalent of human malignancies, and individuals who are chronic hepatitis B virus (HBV) carriers have a greater than 100-fold increased relative risk of developing HCC.1–4 The HBV genome is composed of a partially double-stranded circular DNA that contains 4 overlapping genes: S/preS, C/preC, P, and X. The X gene, which encodes a 17-kDa protein termed HBx, is the most frequently integrated viral sequence found in HCC and has been strongly linked to the development of HCC.5–11 Studies have shown that wild-type HBx induces apoptosis of hepatocytes12, 13 and suppresses the focus formation induced by the cooperation of the ras and myc oncogenes,10 whereas C-terminally truncated HBx can enhance the transforming ability of ras and myc.10 Moreover, HBx mutants, and especially the C-terminally truncated mutants, have frequently been detected in the tumor tissues of HCC patients but only rarely in surrounding nontumor tissue.14–17 Taken together, these data suggest that HBx mutants, but not wild-type HBx, are associated with the carcinogenesis of HCC. However, the precise cause and underlying mechanisms responsible for the genesis of HBx mutants are still unclear.
HBV is an enveloped, partially double-stranded DNA virus that replicates by reverse transcription of a pregenomic RNA intermediate. Like retroviral DNA, the DNA of HBV exists in single-stranded form during reverse transcription. Several studies have shown that G-to-A hypermutations in HBV DNA are induced by APOBEC3 deaminases in human liver cell lines and serum samples from patients with HBV infection.18–24 These deaminases edit the retroviral genome, preferentially targeting the viral minus-strand DNA and producing G-to-A hypermutation in the plus-strand viral DNA during reverse transcription. These data led us to consider whether such G-to-A hypermutations are also present in HBx DNA in human preneoplastic liver tissues, and, if so, whether these hypermutations play a role in the generation of HBx mutants and the carcinogenesis of HCC.
Patients and Methods
Detection of HBx Mutants in Primary Liver Tissues.
Seven primary liver tissue specimens were obtained from patients with cirrhosis and active chronic HBV infections who had given informed consent and undergone liver biopsy. Total cellular DNA was extracted using the phenol/chloroform method. The full-length HBx gene was synthesized using a nested-primer polymerase chain reaction (PCR) procedure and specific primers for HBx. The first-round primers were 5′-cgcaaatatacatcgtatcca t-3′ (forward) and 5′-ttaggcagaggtgaaaaagttgcat-3′ (reverse), and the second-round primers were 5′-atggctgctargctgtgctgcca-3′ (forward) and 5′-ttaggcagaggtgaaaaagttgca T-3′ (reverse), where R is A/G. Hypermutated genomes were identified by using differential DNA denaturation PCR (3-dimensional PCR) in a 2-round procedure.25 The first-round reaction parameters were 94°C for 5 minutes; 35 cycles of 94°C for 30 seconds, 55°C for 30 seconds, and 72°C for 30 seconds; and finally 72°C for 7 minutes. Differential amplification was performed in the second round, using the equivalent of 0.5 μL of the first-round PCR reaction product as the template. Cycling conditions were 88°C for 5 minutes; 35 cycles of 88°C for 60 seconds; 45°C for 30 seconds, and 72°C for 30 seconds; and finally 72°C for 7 minutes. PCR products were purified from agarose gels (Qiaex II kit, Qiagen) and directly cloned into the pGEM-T Easy vector (Promega Corp., Madison, WI). DNA sequencing of individual clones was performed with a 377 DNA sequencer (Applied Biosystems Inc., Foster City, CA).
The pA3G-HA vector has been described.26 A3B-HA was amplified by reverse transcription PCR (RT-PCR) using messenger RNA (mRNA) from human primary hepatocellular carcinoma tissues, with the forward primer 5′-taagattatgaatccacagatcaga-3′ and reverse primer 5′-ttgcggccgctcaggcataatccggcacatcataagggtagttttcctgattctgga-3′ containing HindIII and NotI sites, respectively. The PCR product was cloned into pcDNA3.1(+) to generate pA3B-HA. The pCMVayw HBV vector, which expresses the pregenomic RNA under the control of a cytomegalovirus promoter, was provided by S. Wieland and F.V. Chisari (Scripps Research Institute, La Jolla, CA). The A3A-HA, A3C-HA, and A3F-HA expression vectors were provided by M. Malim.
Wild-type and mutant HBx genes were amplified by PCR from primary liver tissues obtained from patients with cirrhosis and active chronic HBV infection or from HBV-core DNA isolated from HepG2 cells cotransfected with pCMVayw HBV plus a pcDNA3.1 control vector or an expression vector encoding 1 of the APOBEC3 proteins. Primers specific for the HBx gene were used for PCR. The PCR products were cloned into 5′HA-pcDNA3.1.
Cells, Transfection, and Core-Associated HBV DNA Sequencing.
The human HCC cell line HepG2 was maintained in Dulbecco's modified Eagle's medium supplemented with penicillin (100 units/mL), streptomycin (100 μg/mL), glutamine (2 mM), and 10% fetal bovine serum (FBS). HepG2 cells were cotransfected with pCMVayw HBV plus the pCDNA3.1 control vector or an expression vector encoding 1 of the APOBEC3 proteins. Cells were harvested 3 days after transfection. The input HBV plasmid DNA in transfected cells was removed by deoxyribonuclease I digestion, and the intracellular core-associated HBV DNA was extracted as described previously20: HepG2 cells were disrupted in lysis buffer (100 mM Tris-HCl, pH 8.0, with 0.2% Nonidet P-40). The cell lysate was clarified by centrifugation at 16,060 g for 1 minute to pellet nuclei and insoluble material. The supernatant was adjusted to 6 mM Mg(AC)2 and incubated for 2 hours at 37°C with 200 μg/mL of deoxyribonuclease I and 100 μg/mL RNase A. After digestion, the lysate was centrifuged for 1 minute at 16,060 g. The supernatant was incubated for 1 hour at 55°C after the addition of 10 mM ethylene diamine tetraacetic acid, 1% sodium dodecyl sulfate, 100 mM NaCl, and 200 μg/mL proteinase K. Finally, the DNA was extracted with a QIAamp DNA Blood Mini Kit (Qiagen Germany).
To determine the sequences of the HBx, S, and precore/core region DNA samples, we amplified the coding regions of the X, S (nucleotides 411–929), and precore/core (nucleotides 1562–2458) genes from cytoplasmic core-associated HBV DNA using PCR primers specific for the X gene (forward primer 5′-atggctgctaggctgtactgcca-3′, and reverse primer 5′-ttaggcagaggtgaaaaagttgcat-3′), the S region (forward primer 5′-tcctgctgctatgcctcatc-3′, and reverse primer 5′-tgtacaatatgatcctgtg-3′), and the precore/core (forward primer 5′-ctcatctgccggaccgtgtg-3′, and reverse primer 5′-ctaacattgagattcccgagattga-3′). The PCR products were directly cloned into the pGEM-T Easy vector (Promega). DNA sequencing of individual clones was performed with a 377 DNA sequencer (Applied Biosystems). The G-to-A hypermutants of the HBx gene in preneoplastic tissues were identifed using 3-dimensional PCR as previously described.25
Preparation of Liver Tumor Tissue Samples.
Liver tumor samples were obtained from 29 patients who had given informed consent and undergone curative hepatic resection for HCC at the Second Affiliated Hospital, School of Medicine, Zhejiang University, China. HCC tissue samples from all patients were selected immediately after surgical resection from the most viable areas of tumor to exclude areas of tissue necrosis and hemorrhage. Control nontumor specimens of surrounding liver tissue were obtained at a clear distance from the tumor edge (>3 cm) if there was no evidence of nearby tumor invasion. Tissues were snap-frozen immediately after resection and stored at −80°C until use. Microscopic examination was performed to define the histology of the HCC and normal tissue samples.
RNA Extraction and PCR Amplification of APOBEC Deaminases.
Total RNA was isolated from tumor or nontumor liver tissue using TRIzol reagent (Life Technologies) according to the manufacturer's instructions. Because there are multiple variants of APOBEC3-related transcripts, we used semiquantative RT-PCR to detect full-length mRNAs for APOBEC3 and further confirmed the identity of the PCR products by sequencing. Complementary DNA (cDNA) was synthesized from total RNA (3 μg) with oligo(dT)15 primers and Moloney murine leukemia virus reverse transcriptase (Promega). APOBEC1, A3B, A3F, A3G, and AID sequences were amplified using gene-specific primers (APOBEC1: forward, 5′-agagacagagcaccatgact-3′; reverse, 5′-gtacacacacggaatcatcc-3′; A3B: forward, 5′-gcgtatctaagaggctgaac-3′; reverse, 5′-ccttagagactgaggcccatc-3′; A3G: forward, 5′-atgaagcctcacttcagaaac-3′; reverse, 5′-tcagttttcctgattctggag-3′; A3F: forward, 5′-cctgtctttatcagaggtcc-3′; reverse, 5′-gaggcagaaagaggcctctgca-3′; AID: forward, 5′-gaagacactctggacaccac-3′; reverse, 5′-atccactgtcttcagcagag-3′). The thermal cycle conditions were: 95°C for 2 minutes; 35 cycles (25 cycles for β-actin) of 95°C for 20 seconds and 55°C for 20 seconds; 72°C for 90 seconds; and a final extension at 72°C for 5 minutes. The PCR product (10 μL) was electrophoresed on 1.5% agarose and visualized by ultraviolet absorption and ethidium bromide staining, with β-actin cDNA as the internal control. All the sequences amplified from cDNAs were analyzed by gel electrophoresis to confirm the predicted sizes, and selected sequences were confirmed by DNA sequencing. There was a good correlation between PCR product and cDNA template amounts of A3B cDNA under these conditions (Fig. 7B).
RT-PCR for Heat Shock Transcription Factor 1.
cDNA was synthesized from total RNA (3 μg) with oligo(dT)15 primers and Moloney murine leukemia virus reverse transcriptase (Promega). The heat shock transcription factor 1 (HSF1) sequence was amplified using gene-specific primers (forward: 5′-ccatcctgcgggagagtgaa-3′, and reverse: 5′-ggctccgagcctgtcagca-3′). The thermal cycle conditions were: 94°C for 5 minutes; 40 cycles of 94°C for 20 seconds, and 58°C for 30 seconds; 72°C for 40 seconds; and a final extension at 72°C for 7 minutes. The PCR product (10 μL) was electrophoresed on 1.5% agarose and visualized by ultraviolet absorption and ethidium bromide staining, with β-actin cDNA as the internal control.
Real-Time PCR for CD3 mRNA.
The following CD3 real-time PCR primers were used: forward 5′-ggcaagatggtaatgaagaaatgg-3′, and reverse 5′-agggcatgtcaatattactgtggtt-3′. The probe sequence was: FAM 5′-tggtattacacagacaccatataaagtctccatctctgg-3′. The PCR protocols were those of Pennington et al.27
Colony Formation Assay for HBx Mutants.
1 × 106 HepG2 cells were plated in a 24-well plate for 24 hours and then transfected with 4 μg HBx construct or empty pcDNA3.1 using Lipofectamine (Invitrogen Life Technologies) according to the manufacturer's instructions. Two days after transfection, the cells were subcultured at a ratio of 1:4 in G418 (Invitrogen Life Technologies) selective medium for 2 weeks (1,000 μg/mL). Three wells of drug-resistant colonies were fixed with methanol, stained with Giemsa staining solution (Sigma), and then scored (≥40 cells).
Colony Formation Assay for A3B.
5 × 106 HepG2 cells were plated in a 6-well plate for 24 hours and then transfected with 4 μg A3B expression plasmid or empty pcDNA3.1 using Lipofectamine according to the manufacturer's instructions. Five hours after transfection, the cells were subcultured at a ratio of 1:4 in G418 selective medium for 2 weeks (1,000 μg/mL). Three wells of drug-resistant colonies were fixed with methanol, stained with Giemsa staining solution, and then scored (≥40 cells).
Cell Cycle Analysis.
The drug-resistant colonies were used for cell cycle analysis with flow cytometry. Cells were trypsinized and fixed with ice-cold 70% ethanol. For DNA content analysis, propidium iodide was added at 40 μg/mL, and cells were incubated in the presence of RNase at 100 μg/mL for 30 minutes at 37°C. DNA content was determined in a Becton Dickinson FACAS. The percentage of cells in each of the phases of the cell cycle was calculated using the Cell Fit Program (Becton Dickinson).
Western Blot Analysis.
HepG2 cells were transfected with 4 μg HBx construct or empty pcDNA3.1 as described previously, and total cellular protein was extracted 2 days later using the M-PER Mammalian Protein Extraction Reagent (Pierce). Cellular protein (40 μg/well) was subjected to sodium dodecyl sulfate polyacrylamide gel electrophoresis (12% polyacrylamide gels) and then transferred to polyvinylidene difluoride (Bio-Rad) membranes. After blocking with Tris-buffered saline-Tween 20 containing 5% nonfat milk powder for 2 hours at room temperature, the polyvinylidene fluoride membranes were incubated with an anti-HA monoclonal antibody (Santa Cruz Biotechnology, Santa Cruz, CA; 1:500) overnight at 4°C. After 3 washes with Tris-buffered saline-Tween 20, the membranes were stained with a horseradish peroxidase–conjugated secondary antibody (1:5,000, Sigma) for 2 hours at room temperature and reacted with SuperSignal West Pico Chemiluminescent Substrate (Pierce Biotechnology, Rockford, IL).
Cancer Signal Transduction Pathway cDNA Microarray.
For these experiments, we used the Human Cancer Pathway Finder II Gene Array (SuperArray Bioscience Corp.) to screen for genes that were differentially expressed between A3B plasmid- and pcDNA3.1-transfected 293T cells. The 293T cells were transfected with the A3B expression plasmid or pcDNA3.1 for 48 hours, then collected for RNA extraction and cDNA microarray assays as previously described.28
We analyzed the results by t-tests using SPSS 11.5 software. P-values less than 0.05 were considered to be statistically significant.
APOBEC3 Cytidine Deaminases Preferentially Edit the X Region of HBV DNA.
APOBEC3 proteins prefer single-stranded DNAs as their targets. HBx is therefore a likely target for APOBEC3 deaminases because the X gene that encodes HBx is the last region to be completed during the synthesis of HBV minus-strand DNA and therefore exists for an extended period in single-stranded form. Thus, to determine whether APOBEC3 cytidine deaminases edit the HBx region of HBV, we co-transfected HepG2 cells with an APOBEC3 expression vector or control vector and an HBV-producing plasmid. Cytoplasmic core-associated HBV DNAs were amplified by PCR and sequenced. A low level of G-to-A mutations in the HBx region was detected in samples collected from HepG2 cells transfected with control vector (Fig. 1). However, APOBEC3 proteins A3B, A3C, A3F, and A3G all produced an increase in G-to-A mutations in the HBx region of HBV DNA (Fig. 1). A3G induced the highest number of G-to-A mutations, followed by A3B and A3C (Fig. 1). The frequency of C-to-T mutations was not altered in the presence of the APOBEC3 proteins (Fig. 1); also, A3A showed no increase in mutation rate when compared with the control (data not shown).
APOBEC3 members have been found to exhibit a sequence preference with regard to human T-lymphotropic virus genomes.29 To determine whether various regions of HBV DNA show different frequencies of APOBEC3-induced G-to-A mutations, we analyzed the HBs region of the HBV DNA in the presence of A3G or A3B, using the same core-associated HBV DNA samples employed for the HBx gene analysis. Unlike the HBx region (Fig. 2A), few G-to-A hypermutations were detected in the HBs region (positions 411–929) of HBV DNA in the presence of A3G (Fig. 2B, 22 clones). Similarly, A3B induced more G-to-A mutations in the HBx region (Fig. 2C) than in the HBs region (Fig. 2D). We have also sequenced the precore/core region (positions 1562–2458) in the presence of A3G and detected only one G-to-A mutation in 1 of the 24 clones sequenced. Therefore, APOBEC3-induced G-to-A mutations in HBV DNA seem to be region-dependent.
APOBEC3 Deaminases Generate C-Terminally Truncated HBx Mutants.
To determine whether APOBEC3-induced G-to-A mutations produce mutated forms of HBx, we next analyzed the HBx products expressed in HepG2 cells after co-transfection with HBV expression plasmids and APOBEC3-expresing plasmids. We identified a number of HBx mutants (Figs. 3 and 4), most of which harbored multiple point mutations or a combination of C-terminal truncation mutations. G-to-A mutations at positions 359 and 360 (TGG to TAA, TAG, or TGA) generated a premature stop codon at position 120aa in the HBx gene, resulting in the synthesis of a truncated HBx protein missing the last 35 amino acids (Fig. 3). The C-terminally truncated HBx mutations could be generated by 4 of the APOBEC3 (A3B, A3C, A3F, and A3G) enzymes tested (Fig. 3). Of the 36 HBx clones edited by A3G, 10 contained premature stop codon mutations. The C-terminally truncated HBx mutants lost the growth suppressive effect domain (Fig. 4) of HBx.10, 11 A3B and A3G had a potent effect on the HBx gene and generated 10 and 13 hot-spot mutations, respectively, in the HBx gene (Fig. 4). Of the HBx mutants produced by A3B, 2 involved mutational hot-spots (A23T, D; S41P, Y) located in the transforming domain of HBx, and 2 (I88M, K95R) lay within the p53-binding domain. A3G produced 6 mutants within the transforming domain and 1 within the p53-binding domain.
To verify that the predicted C-terminally truncated HBx mutants actually produce C-terminally truncated HBx proteins, we constructed pcDNA3.1-HA-HBx mutant and wild-type HBx expression vectors and used them to transfect 293T cells. HA-HBx proteins in 293T cells were detected by western blotting with anti-HA antibody. The results confirmed that both wild-type HBx and C-terminally truncated HBx proteins were expressed in 293T cells (Fig. 5A). Wild-type HBx protein was expressed at higher levels than the mutant HBx protein in the transfected 293T cells (Fig. 5A). However, this expression difference disappeared in the presence of the proteasome inhibitor MG132 (Fig. 5A), suggesting that the mutant HBx proteins were more rapidly degraded by the proteasome.
APOBEC3-Induced HBx Mutants Enhance the Colony-Forming Ability and Proliferative Capacity of Neoplastic Cells.
Previous studies have shown that naturally occurring C-terminally truncated mutants of HBx derived from hepatocellular carcinomas confer a selective clonal advantage on preneoplastic or neoplastic hepatocytes by abrogating p53-mediated apoptosis, and thereby contribute to hepatocellular carcinogenesis.10, 11 We therefore examined the effects of the C-terminally truncated HBx mutants induced by APOBEC3 enzymes on the colony-forming ability of neoplastic hepatocytes using a colony formation assay. HepG2 cells were transfected with the various HBx expression plasmids, and 2 weeks after G418 selection, the number of colonies was scored. As previously reported,10, 11 the expression of wild-type HBx caused a substantial reduction in the number of colonies when compared with the control cells transfected with the empty vector pcDNA3.1 (Figs. 5B, C). Full-length HBx has been reported to induce apoptosis and inhibit hepatocyte growth.12, 13 Wild-type HBx could have inhibited cell growth by nonspecific overexpression in HepG2 cells. In contrast to wild-type HBx, the APOBEC3-induced C-terminally truncated HBx mutants induced an increase in the number of colonies (Figs. 5B, C). More importantly, the size of the colonies induced by HBx mutants was larger than that of cells transfected with pcDNA3.1 (Fig. 5B). This finding is consistent with the observation that HBx mutants induced a transition into S phase (Fig. 5D).
To further substantiate our observation that C-terminally truncated mutants of HBx can enhance colony formation by neoplastic hepatocytes, we assessed the cell cycle status of HepG2 cells transfected with the various C-terminally truncated HBx mutants or control pcDNA 3.1. Flow cytometry analysis confirmed that a higher percentage of the C-terminally truncated HBx mutant-expressing cells were in S-phase: S-phase cells constituted 33.0% and 31.12% of the HBx-mutant 1 and HBx-mutant 2 cells, respectively, as compared with 22.29% of the pcDNA3.1-transfected control cells (Fig. 5D). These data indicate that the APOBEC3-induced C-terminally truncated HBx mutants could promote the proliferation of neoplastic cells.
Naturally Occurring C-Terminally Truncated HBx Mutants Are Present in Preneoplastic Liver TissueIn Vivo.
Although APOBEC3-induced G-to-A mutations in HBV DNA have been detected in cell culture and in HBV particles in the serum of HBV-infected individuals, detection of G-to-A-mutated HBV DNA from liver tissues had not been reported previously. We therefore investigated the prevalence of HBx mutants in primary liver tissue from patients with cirrhosis and active chronic HBV infection. Using a 3-dimensional PCR method,25 we identified G-to-A hypermutations in HBx DNA from 3 of 7 patients with chronic active HBV infection (Fig. 6A). G-to-A mutations at positions 359 and 360 (TGG to TAA), which resulted in a premature stop codon and generated C-terminally truncated HBx mutants, were observed in most HBx genes from the 3 patients displaying G-to-A hypermutation (Fig. 6B). In addition, HBx point mutations were also frequently observed. Because hepatic cirrhosis has been considered to be a potential preneoplastic factor, these results suggest that naturally occurring C-terminally truncated HBx mutants are present in preneoplastic liver tissues of at least some patients with active chronic HBV infection.
A3B Deaminase Is Widely Up-Regulated in HCC Tumor Tissues.
The possibility of a link between APOBEC3 deaminases and human HCC was further analyzed by examining HCC samples (T) and their matched surrounding nontumor (NT) samples (29 pairs). No expression of APOBEC1 (A1) or AID mRNA was detected in any of the samples tested. Low levels of A3G or A3F were sporadically detected in some HCC or NT samples; an increased expression of A3G or A3F in HCC samples (over that in NT samples) was observed in only a small fraction of the samples (Fig. 7A). In sharp contrast, A3B transcripts were significantly elevated in 24 of 29 HCC samples (Fig. 7A). The proportion of A3B transcripts that were elevated in HCC samples was 0.83 [95% confidence interval (0.64, 0.94), maximum likelihood estimate, P = 0.0005]. Selected sample pairs were also analyzed for APOBEC2, A3A, and A3C mRNA expression: Unlike the result obtained for A3B, fewer than half of the sample pairs showed an elevated expression of these cytidine deaminases in HCC samples (data not shown). Therefore, only A3B was selectively increased in HCC.
The elevated expression of A3B in HCC samples is unlikely to reflect lymphoid cell contamination because A3B proteins are usually not detectable in lymphoid cells.29, 30 Conversely, A3G and A3F are highly expressed in lymphoid cells,29, 30 but they were not detected in many HCC samples (Fig. 7A). Furthermore, the detection of CD3 mRNA, a marker for T lymphoid cells, was generally lower in tumor tissues than in NT tissues (Fig. 7C), despite the fact that A3B levels were higher in tumor tissues than in NT tissues (Fig. 7A).
A3B Deaminase Enhances the Colony-Forming Ability of Neoplastic Cells.
To determine whether A3B directly influences the colony-forming ability of neoplastic cells, human HepG2 cells were transfected with A3B expression plasmids or pcDNA3.1 vector, and 2 weeks after G418 selection, the number of colonies was scored. The enforced expression of A3B in human HepG2 cells caused an increase in the number of colonies when compared with control cells transfected with the empty vector pcDNA3.1 (Fig. 8).
A3B Up-Regulates Heat Shock Transcription Factor 1 (HSF1).
To understand the possible mechanism by which A3B enhances cell colony–forming ability, we used a cDNA microarray monitoring the activation of 15 signal transduction pathways to identify the genes that are differentially expressed between A3B plasmid-transfected and pcDNA3.1-transfected 293T cells. These data indicated that A3B up-regulated HSF1 (Fig. 9A), and this A3B-mediated increase in HSF1 expression was confirmed by RT-PCR analysis (Fig. 9B). The RT-PCR results showed that the expression of HSF1 was also increased in liver tumor tissues when compared with NT tissues (Fig. 9C).
In this study we have investigated the molecular determinants and mechanisms that are responsible for the generation of HBx mutants associated with HCC. We hypothesized that if HBx mutants play a role in the carcinogenesis of HCC, they should be present in HBV infected-preneoplastic liver tissues. To validate this hypothesis, we analyzed HBx mutants, and specifically C-terminally truncated HBx mutants, in preneoplastic liver tissue from patients with cirrhosis and active chronic HBV infection. As expected, such C-terminally truncated HBx mutants, as well as other HBx mutants, were found in primary liver tissues from these patients. Because hepatic cirrhosis is thought to represent a preneoplastic stage, these results suggest that naturally occurring C-terminally truncated HBx mutants are present in preneoplastic liver tissues of some patients with active chronic HBV infection. DNA sequence analysis showed that such HBx mutants are generated by G-to-A hypermutations similar to those produced by APOBEC3 cytidine deamineases. G-to-A mutations at positions 359 and 360 (TGG to TAA) generated a premature stop codon in most HBx genes from the 3 patients found to exhibit G-to-A hypermutations. These findings suggest that human APOBEC3 members may be involved in the generation of HBx mutants in hepatocytes.
Using co-transfections and DNA sequence analyses, we have demonstrated that although A3A does not show an increased level of mutation when compared with control samples, A3B, A3C, A3F, and A3G, all cause an increase in the number of G-to-A mutations in the HBx region of HBV DNA. However, we also observed that these 4 deaminases differ in the level of hypermutation that they induce: A3G induced the largest number of G-to-A mutations, followed by A3B and A3C. Thus, 4 known APOBEC3 members are capable of inducing G-to-A hypermutations in HBV DNA, and, according to our sequence preference analysis, they preferentially edit the HBx region.
We also addressed the question of whether these G-to-A hypermutations induced by APOBEC3 deaminases are responsible for the generation of C-terminally truncated HBx mutants in human hepatocytes. By analyzing the amino acid sequences of HBx genes with G-to-A hypermutations, we found that the APOBEC3-induced G-to-A hypermutations of HBx DNA did indeed generate a number of HBx mutants in human hepatocytes, most of which harbored multiple point mutations or a combination of C-terminal truncation mutations. Like the HBx mutants detected in preneoplastic liver tissues, the G-to-A mutations at positions 359 and 360 generated a premature stop codon at position 120aa in the HBx gene, leading to the synthesis of a truncated HBx protein from which the last 35 amino acids are missing. All 4 APOBEC3 enzymes, A3B, A3C, A3F, and A3G, produced a higher frequency of C-terminally truncated HBx mutations that resulted in a loss of the growth suppressive effect domain of HBx. The A3B and A3G deaminases exerted a particularly potent mutating effect on the HBx gene. In HBx mutants produced by A3B, 2 mutational hot-spots (A23T, D; S41P, Y) were located in the transforming domain of HBx, and 2 (I88M, K95R) within the p53-binding domain. Among the A3G-induced HBx mutants, 6 were within the transforming domain and 1 within the p53-binding domain. Western blot analyses confirmed that the C-terminally truncated HBx mutants indeed produced truncated HBx proteins.
C-terminally truncated naturally occurring HBx mutants found in HCCs have been shown to confer a selective clonal advantage on preneoplastic or neoplastic hepatocytes by abrogating p53-mediated apoptosis, thereby contributing to hepatocellular carcinogenesis.10, 11 In agreement with these observations, we have demonstrated that APOBEC3-generated C-terminal truncated HBx mutants enhanced the colony-forming ability and proliferative capacity of neoplastic HepG2 cells. In fact, several tumor-associated domains have been mapped to HBx: the transforming domain (aa 1–50), the p53-mediated repression domain (aa 88–97), and the growth-suppressive effect domain (aa 142–154)(10). A deletion at the C-terminal end may alter the balance of HBx functional domains in regulating cell proliferation, viability, and transformation, and the abrogation of p53-mediated apoptosis by HBx mutants may provide a selective clonal advantage for preneoplastic or neoplastic hepatocytes and contribute to hepatocellular carcinogenesis.
A3B was widely up-regulated in HCC tumor tissues when compared with matched NT tissues. This increased expression of A3B could contribute to the generation of Hbx mutants, including C-terminally truncated HBx mutants that promote hepatocellular carcinogenesis. A3B also has the potential to directly enhance colony-forming ability and promote the growth of human hepatocytes by mechanisms that are as yet unknown. Our preliminary results showed that A3B could up-regulate HSF1, and the expression of HSF1 was increased in liver tumor tissues when compared with nontumor tissues. HSF1 is a major transactivator of heat shock protein induction and is thought to affect tumor initiation and progression.28, 31, 32
Because of their potent DNA-mutating activity,33, 34 the abnormal expression of APOBEC3 deaminases in the nucleus also may contribute to aberrant editing of cellular genomic DNA, leading to genomic instability. Further study is needed to determine whether APOBEC3 enzymes, and especially A3B, which has been shown to be predominantly localized to the nucleus,22, 35, 36 can cause mutations in mammalian cellular DNA or edit RNA, and, if so, what their targets might be in hepatocytes.
In conclusion, this study provides the first evidence that a number of APOBEC3 deaminases can generate C-terminally truncated mutants of HBx, causing a gain of function that enhances the colony-forming ability and proliferative capacity of neoplastic cells. These events are likely to contribute to the initiation and promotion of HCC oncogenesis. Further studies are required to delineate the precise role of these APOBEC3 deaminases in the development of HCC.
The authors thank Z. He for technical assistance and D. McClellan for editorial assistance.