Chronic infection with hepatitis B virus (HBV) is a major risk factor for hepatocellular carcinoma. The HBV X protein (HBx) is thought to have oncogenic potential, although the molecular mechanism remains obscure. Pathological roles of HBx in the carcinogenic process have been examined using rodent systems and no report is available on the oncogenic roles of HBx in human cells in vitro. We therefore examined the effect of HBx on immortalization and transformation in human primary cells. We found that HBx could overcome active RAS-induced senescence in human immortalized cells and that these cells could form colonies in soft agar and tumors in nude mice. HBx alone, however, could contribute to neither immortalization nor transformation of these cells. In a population doubling analysis, an N-terminal truncated mutant of HBx, HBx-D1 (amino acids 51–154), which harbors the coactivation domain, could overcome active RAS-induced cellular senescence, but these cells failed to exhibit colonigenic and tumorigenic abilities, probably due to the low expression level of the protein. By scanning a HBx expression library of the clustered-alanine substitution mutants, the N-terminal domain was found to be critical for overcoming active RAS-induced senescence by stabilizing full-length HBx. These results strongly suggest that HBx can contribute to carcinogenesis by overcoming active oncogene-induced senescence. (Cancer Sci 2007; 98: 1540–1548)
sodium dodecylsulfate–polyacrylamide gel electrophoresis
Chronic infection with HBV is a major risk factor for HCC worldwide. HBV belongs to the Hepadnavirus family. Its genome is a 3.2-kb, circular, partially double-stranded DNA molecule with four overlapping open reading frames: PC-C, PS-S, P and X.(1) The HBV genome, which is converted to covalently closed circular DNA in the nucleus after infection, serves as the template for transcription, generating the four viral transcripts that encode the HBV core and polymerase polypeptides, the large surface antigen polypeptide, the middle and major surface antigen polypeptides, and the HBx polypeptide. HBV replicates by reverse transcription of viral pregenomic 3.5-kb RNA using the HBV polymerase that catalyzes RNA-dependent DNA synthesis and DNA-dependent DNA synthesis.(1,2) It is converted into the 3.2-kb partially double-stranded genomic DNA inside the viral capsid.
The critical role of HBV chronic infection in HCC has been well established etiologically, whereas the mechanism by which HBV causes transformation of hepatocytes remains unclear.(3–5) HBx has long been suspected of playing a positive role in hepatocarcinogenesis, as avian hepadnaviruses missing the X open reading frame seem not to be associated with HCC. HBx consists of 154 aa and is a multifunctional regulator that modulates many host cell functions through its interactions with a variety of host factors.(5) HBx consists of both a negative regulatory domain(6) and a coactivation domain that is required for the augmentation of virus and host genes.(7,8) HBx was reported to transform rodent immortal cells in vitro,(9,10) and a high incidence of HCC has been reported in transgenic mice overexpressing HBx.(11,12) However, the functional role of HBx in the transformation is still controversial. Some independent groups proposed collaborating roles of HBx in the hepatocarcinogenic process.(13–15) Although these reports are informative, all were experimentally assessed in rodent systems. Because mouse and human primary cells have different telomere biology,(16) DNA damage check point control mechanisms and cell cycle progression,(17,18) developing a human system to address the functional role of HBx is critically important. Here we report that we established human fibroblast cells stably expressing HBx protein and analyzed the effects of HBx expression on the ability to confer an immortal phenotype and tumorigenic potential.
Materials and Methods
Retroviral vectors. All constructs for the expression of HBx (subtype adr) proteins, pNKF-HBx (aa 1–154), pNKF-HBx-D1 (aa 1–50) and pNKF-HBx-D5 (aa 51–154) have been described previously.(8) The retrovirus vectors pBabe-puro, hygro, puro-H-RASV12 hygro-hTERT and pWZL-blast were kindly provided by W. C. Hahn (Dana-Farber Cancer Institute, Harvard).(19,20) To construct pBabe-blast, the blasticidin S cDNA of pWZL-blast was used as a template to amplify the PCR products of blasticidin S with the primer set of AAGCTTACCATGGCCAAGCCTTTGT and ATCGATTTAGCCCTCCCACACATAA, generating an artificial HindIII site at the 5-end and a ClaI site at the 3′-end, respectively. The HBx cDNA of pNKF-HBx was used as a template to amplify the PCR products of HBx with a primer set of TGATCAATGGACTACAAAGACGAT and CTCGAGAGATCTTTAATTAATTAA, generating an artificial FbaI site at the 5-end and an XhoI site at the 3′-end, respectively. The PCR products were digested and inserted into the BamHI and SalI sites of the pBabe-blast vector. The EcoRI and BglII fragments of HBx-D1 and HBx-D5 from pNKF-HBx-D1 and pNKF-HBx-D5 were, respectively, inserted into the EcoRI and BglII sites of the pBabe-blast-HBx vectors. An alanine scanning method was applied to construct a series of HBx clustered alanine substitution mutants (designated ‘cm’) by site-directed mutagenesis. The mutagenesis was carried out using a splicing PCR method with all of the mutated oligonucleotide primer sets. The target sequence of seven aa residues was changed to AAASAAA, and all of the HBx-encoding DNA fragments bearing the clustered mutations were introduced into the EcoRI and BamHI sites of pNKFLAG, generating the pNKF-Xcm1 to pNKF-Xcm21 constructs. The EcoRI and BglII fragments of HBx-cm1 to HBx-cm21 from pNKF-Xcm1 to pNKF-Xcm21 were, respectively, inserted into the EcoRI and BglII sites of the pBabe-blast-HBx vectors. All of the constructs were sequenced by the dideoxy method using the Taq sequencing primer kit and a DNA sequencer (370A; Applied Biosystems).
Virus production and cell lines. Amphotropic retroviruses were produced by transfection of the 293T producer cell line with a retroviral vector and a vector encoding replication-defective helper viruses, pCL-Ampho (Imgenex), using FuGENE 6 transfection reagent (Roche Applied Science) according to the manufacturer's recommendations. Two days after the transfection, culture supernatants were collected, filtered, supplemented with 4 µg/mL polybrene, and used for infection. Two days after the infection, drug selection of infected cells was started, and the selected populations were used in all of the experiments. Infected cell populations were selected in puromycin (1.0 µg/mL), blasticidin S (4 µg/mL) and hygromycin (80 µg/mL) for up to 2 weeks.
Cell culture. Human lung fibroblasts (TIG3) from the Japanese Collection of Research Bioresources were maintained in DMEM with 10% heat-inactivated fetal bovine serum (JRH Biosciences). Human foreskin fibroblasts, BJ and BJ-hTERT-LT-ST-H-RASV12 cells were maintained as described previously.(19) These human fibroblasts were not clonal and were maintained as populations. BJ cells and TIG3 cells have a finite lifespan, and were used at PD between 25 and 35. PD were determined using the formula:
PD = Log(Nf/Ni)/Log2,
where Nf = the number of cells counted and Ni = the number of cells seeded. Comparisons of means and standard deviations were carried out using the unpaired t-test.
Western blot analysis. Cells were harvested, washed with phosphate-buffered saline (–), and sonicated in a lysis buffer (50 mM Tris-HCl [pH 7.4], 200 mM NaCl, 1 mM ethylenediaminetetra-acetic acid, 10% glycerol, 1 mM phenylmethylsulfonyl fluoride, 10 µg/mL leupeptin, 10 µg/mL aprotinin and 10 µg/mL dithiothreitol). Total lysates were fractionated by SDS-PAGE, transferred onto nitrocellulose membranes and subjected to western blot analysis with antibodies. Anti-FLAG M2 antibody and anti-β-actin antibody were from Sigma. Anti-RAS antibody F-235 (sc-29), anti-p53 antibody DO-1 (sc-126) and anti-p21 antibody F-5 (sc-6246) were from Santa Cruz. Anti-p16 antibody was from BD PharMingen. The proteins were visualized by enhanced chemiluminescence according to the manufacturer's instructions (Amersham).
Analysis of senescence. SA-β-Gal staining was carried out using the Senescence Detection Kit (Oncogene) as instructed by the manufacturer. For each sample, at least 200 cells were counted in randomly chosen fields.
Telomerase activity assays. Total lysates of cells were subjected to the telomerase repeat amplification protocol using a TRAPEZE kit (Intergen) according to the manufacturer's instructions.
Soft-agar colony formation assays. Soft-agar growth assays were carried out as described previously.(19) At the time of plating in soft agar, cultures were trypsinized and counted, and 5 × 103 or 5 × 104 total cells were mixed with 1.5 mL of 0.35% Noble agar-DMEM (top layer) and then poured on top of 5 mL of solidified 0.7% Noble agar-DMEM (bottom layer) in 6-cm-diameter dishes. After 3 weeks, colonies were counted, and pictures were taken.
Tumorigenicity assays. A total of 1 × 106 cells were resuspended in 50 µL Matrigel solution (BD Matrigel Basement Membrane Matrix HC; BD Biosciences) and immediately injected subcutaneously into 8-week-old female nude mice (BALB/cAnNCrl-nu BR). 2-D tumor sizes were measured once a week. The tumor volume (mm3) was calculated using the formula (length × width2)/2.(21)
Effect of HBx on cellular senescence of human primary cells. During immortalization, human cells differ from rodent cells in the regulation of telomere length(22,23) and cell cycle checkpoints.(24,25) Human cells must bypass two barriers to become immortalized: replicative senescence and crisis. Replicative senescence is characterized by an irreversible growth arrest but continued metabolic activity.(26) Crisis is characterized by widespread cell death.(26,27) By the introduction of hTERT, human primary cells avoid these two barriers and can become immortalized.(28–30)
It is possible that HBx contributes to the immortalization process of human primary cells, but not to the cellular transformation process. If so, it may facilitate cellular transformation indirectly by overcoming two crises, M1 and M2. To study whether this does facilitate cellular transformation, it is best to use human primary hepatocytes as HBV is a hepatotropic virus. However, human primary hepatocytes are almost impossible to obtain for such an experimental approach. HBx exhibits its transactivation function not only in hepatoma cell lines but also in various carcinoma and sarcoma cell lines. Under these situations, we addressed whether HBx contributes to the immortalization of human primary fibroblasts, BJ cells and TIG3 cells that have been well studied for cellular senescence and immortalization. We used hTERT-introduced BJ and TIG3 cells for positive controls of immortal cells.
The human primary fibroblasts, BJ cells and TIG3 cells were infected with the HBx-expression retroviruses and cultured in the presence of the selection drug, blastcidin S. The drug-resistant polyclonal cells were selected and characterized. Three different constructs of HBx were used to map the responsible domain: full-length HBx (HBx-wt), HBx-D1, which lacks the N-terminal negative regulatory domain, and HBx-D5, which lacks the coactivation domain (Fig. 1a). First we examined HBx expression in the primary human fibroblasts. We found that full-length HBx and HBx-D5 were highly but equally expressed, whereas expression of HBx-D1 was very weak in the blastcidin S-selected clones (Fig. 1b). We hypothesized that HBx expression may confer an immortal phenotype, which could contribute to cellular transformation and tumorigenesis, but we observed that the BJ cells expressing HBx proteins stopped dividing at PD 69.6 ± 0.9 (errors ± SD) (HBx-wt), PD 66.6 ± 1.6 (HBx-D5), PD 66.1 ± 1.4 (HBx-D1) and PD 60.5 ± 0.6 (control cells) (Fig. 1c). TIG3 cells, another human fibroblast, expressing HBx proteins stopped dividing at PD 77.2 ± 1.1 (HBx-wt), PD 75.1 ± 0.8 (HBx-D5), PD 75.1 ± 0.1 (HBx-D1) and PD 75.4 ± 0.2 (control cells) (Fig. 1d). Although a very minor extended lifespan (2–4 PD) was observed with HBx-wt-expressing primary human fibroblasts, the HBx protein could not elicit immortalization. We examined whether the effect of HBx on delay of cellular senescence was correlated with putative augmentation of telomerase activity in HBx-introduced BJ and TIG3 cells (Fig. 1e) as activation of the hTERT promoter was observed in hepatoma cell lines that were transiently cotransfected with the HBx expression vector and luciferase reporter vector of the hTERT promoter (S. Murakami et al. unpublished data, 2005). Telomerase activity in the extracts of cells expressing HBx-wt or HBx-D1 was slightly higher than that of cells expressing empty vector or HBx-D5 in both kinds of cells (Fig. 1e), but we failed to detect an increase in hTERT protein expression (data not shown). Therefore, the relevance of the weak augmentation of telomerase activity in the HBx-expressing primary cells remains unclear.
Effect of HBx on immortalized BJ-hTERT cells. Next, we addressed whether HBx facilitates the cellular transformation process using human immortal cells. For this purpose, we used BJ-hTERT cells – these were BJ-derived cells immortalized by the introduction of hTERT, as characterized previously.(19) HBx-wt as well as its truncated mutants had no effect on cell proliferation, telomerase activity or cell transformation. Using the newly established TIG3-hTERT cells, we confirmed that the stable expression of HBx, XD1 or XD5 did not affect cell proliferation or cell transformation (data not shown). These results indicate the inability of HBx alone to transform these human immortalized cells.
Ability of HBx to overcome H-RASV12-induced senescence in BJ cells immortalized by hTERT Seeing as HBx did not exhibit the ability to immortalize primary human fibroblasts or to elicit transformation into hTERT-induced immortal primary human fibroblasts, we considered whether HBx functioned together with an oncogene and induced cell transformation. Senescence induced by active oncogene expression (OIS), such as oncogenic RAS, is one of the anticancer processes in which tumor suppressors and their related networks are involved, as demonstrated in vitro and recently also in vivo.(31,32) Overcoming OIS is critical for cellular transformation in vitro and cancerous cell proliferation in vivo.(31) Therefore, we addressed whether HBx has a collaborating role in transforming cells in the presence of oncogenic RAS or in overcoming RAS-induced senescence.
To examine the effect of HBx on RAS-induced senescence-like growth arrest, we introduced H-RASV12 into BJ-hTERT, BJ-hTERT-HBx-wt, BJ-hTERT-HBx-D1 and BJ-hTERT-HBx-D5 cells using a retrovirus (Fig. 2d). BJ-hTERT cells expressing H-RASV12 stopped proliferating within several days of RAS introduction. In contrast, BJ-hTERT cells expressing both H-RASV12 and HBx-wt (BJ-hTERT + H-RASV12+ HBx-wt) continued to proliferate to more than 80 PD (Fig. 2a). Although HBx-D1 also demonstrated the ability to overcome active RAS-induced senescence, HBx-D5 failed to overcome OIS (Fig. 2a). We also found that the growth rate of BJ-hTERT + H-RASV12 + HBx-wt cells was much higher than that of BJ-hTERT + H-RASV12 + HBx-D1 cells, probably reflecting the fact that some portion of the latter cells were positive for SA-β-gal (Fig. 2b,c). Consistent with this result, cells staining positive for SA-β-gal were significantly fewer in BJ-hTERT + H-RASV12 + HBx-wt than in BJ-hTERT + H-RASV12 + HBx-D1 (Fig. 2c). These results indicate that HBx-wt has the ability to overcome RAS-induced senescence. HBx-D1, the coactivator domain of HBx, seems to be indispensable and sufficient for overcoming RAS-induced senescence analyzed by the PD analysis, although HBx-D1 did not show the same ability as HBx-wt. The incomplete ability of HBx-D1 may be due to the low expression of HBx-D1 in the blastcidin S-selected clones in BJ-hTERT cells, as observed with the BJ cells (see Discussion).
HBx protein is required for anchorage-independent growth and tumor formation in nude mouse in response to H-RASV12. HBx can overcome RAS-induced senescence (examined by the PD analysis) and can indicate that HBx and RAS can induce cell transformation. Therefore, we examined whether BJ-hTERT + H-RASV12 + HBx-wt and BJ-hTERT + H-RASV12 + HBx-D1 cells can form colonies in soft agar. We found that BJ-hTERT + H-RASV12 + HBx-wt cells showed cell number-dependent formation of colonies, which were much smaller size than those of control cells, BJ-hTERT + H-RASV12 + SV40 LT + ST(20,33) (Fig. 3a,b). In contrast, BJ-hTERT + H-RASV12 + HBx-D1 cells could not form colonies in soft agar (Fig. 3a), although these cells overcame RAS-induced senescence. This result strongly suggests that HBx-D1 is not equivalent to HBx-wt in its ability to make colonies in soft agar.
Next we tested the tumor-forming ability of BJ-hTERT + H-RASV12 + HBx-wt or HBx-D1 cells in nude mice. BJ-hTERT + H-RASV12 + HBx-wt cells were found to form tumors in four of eight mice, although these tumors grew much more slowly and were much smaller than those formed by BJ-hTERT + H-RASV12 + SV40 LT + ST cells (eight of eight animals) (Fig. 3c). In contrast, BJ-hTERT + H-RASV12 + HBx-D1 cells did not generate tumors in nude mice (Fig. 3c), consistent with the results of the soft-agar assay. These results indicate that HBx contributes to cellular transformation by collaborating with active RAS in human immortalized cells. To our knowledge, this is the first report showing that HBx plays a critical role in cellular transformation, collaborating with active RAS in human immortalized cells.
Effects of HBx on p16 and p21 expression and the ability of HBx to overcome RAS-induced senescence. Overexpression of RAS causes oncogene-induced premature senescence in normal human fibroblasts (Fig. 4c) and hTERT-immortalized human fibroblasts (Fig. 2a), but RAS failed to induce premature senescence in HBx-wt- or HBx-D1-introduced BJ-hTERT cells (Fig. 2a). We next examined the effect of stable expression of HBx in BJ cells with or without expression of hTERT, as interference with both the p53 and pRb pathways is necessary to avoid RAS-induced cellular senescence, in which p16 and p21 are the critical downstream effectors of pRb and p53, respectively. Expression of p16 and p21 was upregulated in HBx-wt- or HBx-D1-introduced BJ-hTERT cells; however, HBx-D5 has no ability to induce the expression of these genes. The presence of H-RASV12 resulted in downregulation of the augmented expression of p16 and p21 in HBx-wt- or HBx-D1-introduced BJ cells and BJ-hTERT cells (Fig. 4a,b). These results suggest that HBx can suppress expression of p53, p16 and p21 in H-RASV12-introduced cells, contributing to overcoming RAS-induced senescence. Next we examined whether HBx-wt and H-RASV12 not immortalized by hTERT were sufficient for cellular transformation. We introduced H-RASV12 into BJ-HBx-wt, BJ-HBx-D1 and BJ-HBx-D5 cells and analyzed them by PD analysis and soft-agar colony assay. In the PD analysis, H-RASV12-introduced BJ-HBx-wt and BJ-HBx-D1 cells did overcome RAS-induced cellular senescence but stopped cell division at PD 62, which is approximately the cellular senescence of BJ cells (Figs 1c,4c), whereas H-RASV12-introduced BJ-HBx-D5 did not overcome senescence and stopped cell division. These results suggest that HBx can overcome RAS-induced senescence but can not immortalize the cells (Fig. 4c). In the soft-agar colony formation assay, BJ-HBx-wt-H-RASV12 and BJ-HBx-D1-H-RASV12 could but BJ-HBx-D5-H-RASV12 could not form very tiny colonies, suggesting that HBx-wt and H-RASV12 in the absence of hTERT may enable the cells to proliferate in an anchorage-independent manner (data not shown).
As HBx-D1, which was very weakly expressed, exhibited almost the same ability as HBx-wt to upregulate the tumor suppressor genes and to overcome RAS-induced senescence in these cells, we wondered whether HBx-D1 missing the N-terminal domain may have some negative effect on cell proliferation. Because the transient expression level of HBx-D1 in BJ cells was similar to those in HepG2 cells, as reported previously (Fig. 4d),(8) it was not due to the construct design of the vector. The expression of HBx-D1 was slightly lower than those of HBx-wt and HBx-D5 on day 5 after selection, much lower on day 10 after selection (data not shown). On day 14 after selection, the expression of HBx-D1 reached the lowest level, and after day 14 that expression level was kept (Figs 1b,4d). HBx-D1-introduced BJ cells grew slower than HBx-wt- or HBx-D5-introduced BJ cells (data not shown) and contained more SA-β-Gal-positive cells during proliferation (Fig. 4e). These results suggest that cells expressing lower levels of HBx-D1 proliferated more than cells expressing higher levels of HBx-D1, due to some toxic or antiproliferative effect of the coactivation domain of HBx in the human primary cells (see Discussion).
Important region of HBx for overcoming cellular senescence and anchorage-independent growth. As HBx exhibited the ability to overcome active RAS-induced senescence, we next tried to identify the critical regions of HBx for overcoming cellular senescence. BJ-hTERT cells were infected with retroviruses expressing one of the clustered alanine-substituted mutants covering all parts of HBx,(34) and a series of cell clones stably expressing these HBx-cm mutants, BJ-hTERT-HBx-cm, was established (Fig. 5). H-RASV12 was then introduced into BJ-hTERT-HBx-cm1 to BJ-hTERT-HBx-cm21 cells and cell proliferation was examined. The regions covering HBx-cm8 to HBx-cm10, and those covering HBx-cm19 to HBx-cm21 were found to be not critical for overcoming active RAS-induced senescence and anchorage-independent growth as the BJ-hTERT-RAS clones expressing these HBx-cm mutants proliferated and formed colonies in soft agar, similar to BJ-hTERT-HBx-wt-H-RASV12 cells. The BJ-hTERT-RAS clones expressing HBx-cm1 to HBx-cm7, and those expressing HBx-cm14 to HBx-cm16, were like BJ-hTERT-HBx-D1-RAS, which can grow but at a much reduced rate compared with BJ-hTERT-HBx-RAS cells. The HBx regions covering HBx-cm11 to HBx-cm13, HBx-cm17 and HBx-cm18 were found to be critical for overcoming active RAS-induced senescence as the BJ-hTERT-RAS clones expressing these HBx-cm mutants failed to proliferate, meaning that these had no ability to overcome active RAS-induced cellular senescence at all (Fig. 6) (Table 1). Among the BJ-hTERT-HBx-cm cells, expression levels of HBx-cm1 to HBx-cm6 were very weak, like that of HBx-D1. Furthermore, the protein bands of HBx-cm1 to HBx-cm5 migrated slightly slower than those of HBx-cm6 and the other HBx-cm mutants in the coactivation domain in SDS-PAGE analysis (see Discussion).
Table 1. Degree of proliferation of H-RASV12-introduced BJ-hTERT-HBx-cm cells
Degree of proliferation
HBx-cm1–21 in this table represent HBx-cm1–21-introduced BJ-hTERT-H-RASV12 cells.
Hepatitis B virus X protein has long been suspected to be positively involved in HBV-associated HCC, but its molecular role in hepatocarcinogenesis remains unclear. Although HBx is involved directly in the transformation of immortal rodent cells in vitro and in tumor formation in the livers of nude mice, the oncogenic activity of HBx itself remains to be elicited as the reproducibility of these experiments has been seriously controversial.(5) Furthermore, the positive role of HBx has not been addressed with human primary cells or human immortal cells. To our knowledge, our report is the first to show that HBx retains the ability to overcome RAS-induced senescence of immortalized human cells, although it is not sufficient for immortalizing human primary cells or transforming human immortal cells. hTERT-immortalized human cells stably expressing HBx-wt and RAS can form colonies in soft agar and tumors in nude mice in a cell-number-dependent manner. HBx can overcome RAS-induced senescence of BJ cells, but HBx-wt and active RAS could not immortalize the human fibroblasts. Although our findings are different to a report showing that HBx itself retains the transforming ability in NIH3T3 cells,(9) they are similar to results in rodent immortal embryonic fibroblast cells.(10)
To determine the region of HBx responsible for the ability to overcome RAS-induced senescence, we used two truncation mutants: HBx-D1 (aa 51–154), which exhibits transcriptional coactivation function and augments HBV transcription and replication,(8) and HBx-D5 (aa 1–50), which harbors the negative regulatory domain of transcriptional modulation.(6) When HBx-D1 and H-RASV12 were introduced into BJ-hTERT cells, HBx-D1 was similar to wild-type HBx in overcoming RAS-induced senescence in the PD analysis and in SA-β-gal staining. Therefore, HBx-D1 alone seems to be sufficient for overcoming active RAS-induced senescence and for anchorage-independent growth, but it is not sufficient for BJ-hTERT + H-RASV12 + HBx-D1 cells to form visible colonies in soft agar and tumors in nude mice. HBx alone may be sufficient for overcoming RAS-induced senescence, but hTERT is required for immortal proliferation of the transformed cells with H-RASV12 and HBx. As HBx-D1 exhibits a similar ability to HBx-wt in overcoming RAS-induced senescence and anchorage-independent growth, but not in immortalizing human fibroblasts, HBx-D1 may harbor all of the critical abilities of HBx. However, HBx-D1 is different from HBx-wt in the ability to form visible colonies in soft agar and to form tumors in nude mice.
The coactivation function was recently mapped by scanning a HBx library of clustered alanine substitution mutants (HBx-cm library), and two separate sequences in HBx-D1 were found to be critical.(8) Using the same HBx-cm library, we attempted to map the sequences critical for overcoming RAS-induced senescence. We have identified three different phenotypes among the HBx-cm mutants: those phenotypes are like HBx-wt, HBx-D1 and HBx-D5 (Fig. 6). HBx-cm mutations within the D5 region, cm1 to cm7, have the ability to partially overcome OIS, whereas those within the D1 region (cm8–10, cm14–16 and cm 19–21) fail to exhibit the overcoming ability. The HBx-D5 phenotype is even found among the HBx-cm mutants (cm13, cm17 and cm18) that are defective in the coactivation function.(8) These results indicate that the ability to fully overcome OIS requires two putative functions carried by the D1 and D5 regions of the HBx protein. Because HBx-D5 does not have a positive or negative effect on RAS-induced senescence (Figs 2,3,4c), the negative regulatory domain may be active only in full-length HBx. The very low expression of HBx-D1 in human primary cells and hTERT-immortalized cells may be due to the selection result of clones, reflecting that a high level of HBx-D1 protein was eliminated due to a toxic effect of the coactivation domain,(5) or due to deletion of the N-terminal domain that has some critical role in stabilizing HBx in the expression system. Both of these may actually occur. The former is supported by the enrichment of cells expressing HBx-D1 during the early stages of drug selection. The latter is highly possible as expression levels of HBx-cm1 to HBx-cm6 covering most of the N-terminal domain were very low, as for HBx-D1. Pang et al. recently reported a stabilization mechanism of HBx through direct interaction with Pin1,(35) which binds phosphorylated serine and the next proline. The target serine residue is within the N-terminal domain or within the region covered by HBx-cm6. Interestingly, the HBx-cm1 to HBx-cm5 bands migrated more slowly than the HBx-cm6 band (Fig. 5b), supporting the possibility that the N-terminal domain may be critical for Pin1 binding to stabilize HBx. One interesting possibility that remains to be tested is that activation of the degradation pathway of HBx causes the toxic effect on cell proliferation. This possibility may explain the low expression of HBx-D1 and the cm mutants in the N-terminal domain. In this context, it remains unclear at present the reason for the rather stable expression of two bands of HBx-cm7 that seem to confer the same phenotype as HBx-D1 in the characterization of the cells.
The region of D1 that is responsible for overcoming RAS-induced senescence should be defined. Because some HBx-cm mutants defective in coactivation function still exhibit the ability to overcome OIS, it seems that the coactivation function is dispensable for the role. More than a dozen host factors have been reported to interact directly with the HBx-D1 region, including p53,(36,37) Smad4,(38) DDB1,(39,40) and two core subunits of the proteasome.(5) It is especially important to determine whether the binding of HBx to p53 is responsible for the ability to overcome RAS-induced senescence, as the direct binding of p53 to HBx was found to suppress p53-dependent gene activation.(5,37)
Although we have shown here that the D5 region of HBx has an indispensable biological role in anchorage-independent cell growth, the critical role of the D5 region in overcoming OIS remains obscure. The ability of the D5 region in full-length HBx to support anchorage-independent growth will provide a good experimental system for revealing the function of the negative regulatory domain of HBx, as no host factor has been reported to interact specifically with the D5 region.
Our results clearly indicate that HBx retains the ability to overcome RAS-induced senescence in human cells immortalized by hTERT, although HBx alone could neither immortalize nor transform human cells. The ability of HBx to collaborate with active RAS in cell transformation may explain its role in hepatocellular carcinogenesis. Our findings, however, were obtained using an experimental model with immortalized cells derived from human fibroblasts. Our results may not reflect the role of HBx in HBV-infected liver, as overcoming the processes of OIS seems to vary with tissue and tumor type.(41) The role of HBx should therefore be addressed using human hepatocytes and immortalized human hepatocytes. The former, however, are quite difficult to obtain whereas the latter are available at present. It had been immortalized by introducing the other viral oncogene SV LT.(42,43)
We thank K. Masutomi and W. C. Hahn for kindly providing the retroviral vectors and human immortal cell lines, T. B. S. Yen (UCSF, USA) and H. Tang (Sichuan University, China) and the members of the Division for critical discussions, and Ms Yasukawa and Ms Kuwabara for excellent technical assistance. This work was supported in part by a Grant-in-aid for Scientific Research and Development (B) (1790031) and a Grant-in-aid for Scientific Research on Priority Areas (C) (12213050 and 17013035) from the Ministry of Education, Culture, Sports, and Technology.