Resistance of primary cultured mouse hepatic tumor cells to cellular senescence despite expression of p16(Ink4a), p19(Arf), p53, and p21(Waf1/Cip1)

Primary cultured mouse hepatic cells become senescent within a short period, although rare cells form colonies from which continuously proliferating cell lines can be established. In contrast, hepatic tumor (HT) cells show little senescence and higher colony-forming capacity. To assess this difference, we investigated p16(Ink4a)/p19(Arf)/p53/p21(Waf1/Cip1) expression in primary normal and HT cells, together with cell lines established from both. In primary normal cells, p16(Ink4a)/p19(Arf) were expressed only in association with senescence and disappeared at later stages of colony formation. In contrast, primary HT cells showed sustained p16(Ink4a)/p19(Arf) expression from the beginning. No p16(Ink4a)/p19(Arf) alterations, such as deletion, mutations, or hypermethylation, were detected in the primary HT cells, although most cell lines derived from either normal or HT cell colonies lost p16(Ink4a) or p19(Arf) expression owing to hypermethylation or homozygous deletion of p16(Ink4a)/p19(Arf). On the other hand, primary normal and HT cells and most cell lines showed constitutively elevated expression of p53/p21(Waf1/Cip1), with a further increment after ultraviolet ir-radiation, indicating a functionally normal p53 pathway. These results indicate that primary HT cells are resistant to senescence despite retaining p16(Ink4a)/p19(Arf)/p53/p21(Waf1/Cip1) expression and that loss of p16(Ink4a)/p19(Arf) function is associated only with establishment of the cell lines.     

Both products of the mouse Ink4a/Arf locus suppress melanoma formation in vivo

Deletion of the INK4a/ARF locus at 9p21 is detected with high frequency in human melanoma. Within a short genomic distance, this locus encodes several proteins with established tumor-suppressor roles in a broad spectrum of cancer types. Several lines of evidence support the view that p16INK4a and p19ARF exert the tumor-suppressor activities of this locus, although their relative importance in specific cancer types such as melanoma has been less rigorously documented on the genetic level. Here, we exploit a well-defined mouse model of RAS-induced melanomas to examine the impact of germline p16INK4a or p19ARF nullizygosity on melanoma formation. We demonstrate that loss of either Ink4a/Arf product can cooperate with RAS activation to produce clinically indistinguishable melanomas. In line with the common phenotypic end point, we further show that RAS+ p16INK4a-/- melanomas sustain somatic inactivation of p19ARF-p53 and, correspondingly, that RAS+ p19ARF-/- melanomas experience high-frequency loss of p16INK4a. These genetic studies provide definitive proof that p16INK4a and p19ARF cooperate to suppress the development of melanoma in vivo.     

The role of Ink4a/Arf in ErbB2 mammary gland tumorigenesis

Most human tumors display inactivation of the p53 and the p16(INK4)/pRb pathway. The Ink4a/alternative reading frame (ARF) locus encodes the p16(INK4a) and p14(ARF) (murine p19(ARF)) proteins. p16(INK4a) is deleted in 40-60% of breast cancer cell lines, and p16(INK4a) inactivation by DNA methylation occurs in < or =30% of human breast cancers. In mice genetically heterozygous for p16(INK4a) or Ink4a/Arf, predisposition to specific tumor types is enhanced. Ink4a/Arf(+/-) mice have increased E micro -Myc-induced lymphomagenesis and epidermal growth factor receptor-induced gliomagenesis. ErbB2 (epidermal growth factor receptor-related protein B2) is frequently overexpressed in human breast cancer and is sufficient for mammary tumorigenesis in vivo. We determined the role of heterozygosity at the Ink4a/Arf locus in ErbB2-induced mammary tumorigenesis. Compared with mouse mammary tumor virus-ErbB2 Ink4a/Arf(+/-) mice, mouse mammary tumor virus-ErbB2 Ink4a/Arf(wt) mammary tumors showed increased p16(INK4a), reduced Ki-67 expression, and reduced cyclin D1 protein but increased mammary tumor apoptosis with no significant change in the risk of developing mammary tumors. These studies demonstrate the contribution of Ink4a/Arf heterozygosity to tumor progression is tissue specific in vivo. In view of the important role of Ink4a/Arf in response to chemotherapy, these transgenic mice may provide a useful model for testing breast tumor therapies.     

Mice with alterations in both p53 and Ink4a/Arf display a striking increase in lung tumor multiplicity and progression: differential chemopreventive effect of budesonide in wild-type and mutant A/J mice

p53 transgenic mice carrying a dominant negative mutation were crossed with Ink4A/Arf heterozygous-deficient mice to investigate whether there is a synergy between these two germ-line mutations in promoting carcinogen-induced lung tumor progression in mice. Mice with a p53 dominant negative mutation and Ink4A/Arf heterozygous deficiency exhibited >20-fold increase in tumor volume compared with approximately 4-fold increase in Ink4A/Arf heterozygous-deficient mice and a 9-fold increase in mice with only the p53 dominant negative mutation. The effect of Ink4A/Arf heterozygous deficiency on lung tumor progression occurred late in the carcinogenesis process (>30 weeks after carcinogen treatment). In addition, most of the lung tumors (approximately 80%) from mice with a p53 mutation and deletion of Ink4A/Arf were lung adenocarcinomas. In contrast, lung adenocarcinomas were seen in <10% of the lung tumors from the wild-type mice and approximately 50% of the lung tumors from Ink4a/Arf heterozygous-deficient or p53 mutant mice. These results indicate a significant synergistic interaction between the presence of a mutant p53 transgene and the Ink4A/Arf deletion during lung tumor progression (P < 0.01). The usefulness of this new mouse model in lung cancer chemoprevention was examined. The chemopreventive efficacy of budesonide was examined in wild-type mice, mice with Ink4A/Arf heterozygous deficiency, mice with a mutation in the p53 gene, or mice with both a mutation in the p53 gene and deletion in the Ink4A/Arf locus. Mice treated with budesonide displayed an average of 90% inhibition of lung tumor progression in a standard 18-week chemoprevention assay, regardless of p53 and/or Ink4A/Arf status. However, the efficacy of budesonide against lung tumor progression decreased from 94 to 77% (P = 0.07) in mice with alterations in both p53 and Ink4A/Arf in a 40-week chemoprevention assay. Similarly, when mice bearing established lung adenomas were treated with budesonide, genotype-dependent differential effects of budesonide in wild-type and mutant mice were clearly revealed with a 82, 64, 45, and 33% decrease in tumor volume in wild-type mice, p53(+/+)Ink4a/Arf(+/-) mice, p53(+/-)Ink4a/Arf(+/+) mice, and p53(+/-)Ink4a/Arf(+/-), respectively. Thus, mutant mice with alterations in p53 and/or Ink4A/Arf exhibited a significant resistance to chemoprevention by budesonide. Because p53 and Ink4a/Arf mutations are the most prevalent mutations in human lung cancers, the effectiveness of chemopreventive agents on the mutant A/J mice containing alterations with p53 and Ink4a/Arf is the best preclinical estimate of their efficacy in humans. Thus, the mutant A/J mouse model should prove useful for chemoprevention studies.     

Consistent inactivation of p19(Arf) but not p15(Ink4b) in murine myeloid cells transformed in vivo by deregulated c-Myc

Cyclin-dependent kinase inhibitors p16(INK4a) and p15(INK4b), encoded by the CDKN2A and B loci, play an important role in negative regulation of the cell cycle. Furthermore, p19(ARF) also encoded by the CDKN2A locus, has been shown to regulate positively the p53 pathway leading to growth arrest and apoptosis. All three genes have been inactivated in human tumors. In myeloid cells, p15(INK4b) mRNA is upregulated during cytokine-induced differentiation and/or growth arrest, and hypermethylation of the p15(INK4b) gene promoter region is a common event in acute myeloid leukemia. In the present study, we examined murine monocyte/macrophage tumors with deregulated c-myc for evidence of Ink4 gene inactivation. p15(Ink4b) mRNA and protein were detected in the majority of leukemias, and p16(Ink4a) mRNA and protein were highly expressed in two of them. pRb was in a hypophosphorylated state in most of the neoplasms indicating that the Cdk inhibitors that were expressed in the cells were functional. The observed expression of p15(Ink4b) is inconsistent with their proliferation state, although it might be expected to be expressed owing to the maturity of the cells. These data suggest, therefore, that deregulated c-Myc bypasses the pRb restriction point and cell cycle arrest in these tumors. An examination of p19(Arf) exons revealed deletions of the gene in up to 94% of the tumors. Since this gene shares exon 2 with p16(Ink4a), it is often difficult to determine which gene is the relevant tumor suppressor. However, the loss of only the p19(Arf)-specific exon 1 beta was observed in a tumor that had normal p16(Ink4a) protein expression. In addition, the p19(Arf)-specific exon was deleted in another tumor that expressed a functional chimeric protein, p15Ex1-p16Ex2-3; it was demonstrated here that this fusion protein is capable of inducing G1 arrest. These data overall supports the hypothesis that the critical inactivation event in these hematopoietic neoplasms is elimination of p19(Arf), and not Ink4 function.     

CBX7 controls the growth of normal and tumor-derived prostate cells by repressing the Ink4a/Arf locus

Control of cell proliferation by Polycomb group proteins (PcG) is an important facet of cellular homeostasis and its disruption can promote tumorigenesis. We recently described CBX7 as a novel PcG protein controlling the growth of normal cells. In an attempt to identify a putative role of CBX7 in tumorigenesis, we analysed CBX7 expression in a panel of cancer cell lines and primary tissues. CBX7 was highly expressed in three different prostate cancer cell lines and present at elevated levels in normal prostate. Ablation of CBX7 expression using short hairpin RNAs (shRNA) resulted in upregulation of p16Ink4a and p14Arf in both LNCaP and PC-3 prostate cell lines. CBX7 knockdown caused an impairment of cell growth that was dependent on the status of the p14Arf/p53 and p16Ink4a/Rb pathways in both normal and cancer prostate cells. CBX7 overexpression in LNCaP cells resulted in a slight growth advantage in both androgen-dependent and -independent conditions. Moreover, CBX7 expression cooperated with c-Myc in rendering LNCaP cells insensitive to growth arrest by androgen receptor inhibition. Together, these data suggest that CBX7 represses p16Ink4a and p14Arf expression in normal and tumor-derived prostate cells, affecting their growth depending on the status of the p16Ink4a/Rb and the p14Arf/p53 pathways.     

Spectral karyotyping of sarcomas and fibroblasts derived from Ink4a/Arf-deficient mice reveals chromosomal instability in vitro

The Ink4a/Arf locus is functionally linked to the Rb and p53 pathways through the action of its two gene products. Mouse models null for this locus show rapid onset of cancer with a preponderance of lymphomas and sarcomas. We report on a study of cell lines derived from sarcomas arising in Ink4a/Arf null mice. The cytogenetics of these lines was monitored over the course of serial passage. Results indicate that early passage cells are relatively normal. However, after multiple passages chromosomal instability becomes apparent as evidenced by increasing tetraploidy and aneuploidy, and the concomitant loss of clonality. To further evaluate the effect of Ink4a/Arf-deficiency on chromosomal stability in vitro, we isolated Ink4a/Arf deficient primary murine embryonic fibroblasts (MEFs), serially passaged them, and analyzed their chromosomal stability by spectral karyotyping (a 24-color chromosome paint-FISH technique). We found that chromosomal instability in Ink4a/Arf deficient MEFs developed with the same timing as seen in cell lines derived from Ink4a/Arf deficient sarcomas. Thus, chromosomal instability seen in Ink4a/Arf deficient tumors in vitro may be unrelated to the original phenotype of the tumor in vivo. Therefore, interpretation of cytogenetic data from cell lines derived from Ink4a/Arf deficient tumors should be done on early passage cells.     

p16(Ink4a) in melanocyte senescence and differentiation

BACKGROUND: The Ink4a-Arf tumor suppressor locus encodes two growth inhibitors, p16 and Arf, both of which are also implicated as effectors in cellular senescence. Because human germline defects in the INK4A-ARF locus are associated with familial melanoma, melanocytes may have unusual INK4A-ARF functions or controls of cell senescence. Because senescence is believed to be an anticancer mechanism, we investigated the role of Ink4a-Arf and its individual components in melanocyte senescence. METHODS: Melanocytes were cultured from littermate mice with zero, one, or two functional copies of the Ink4a-Arf locus. Senescence was evaluated by cumulative population doubling curves and by the assessment of acidic beta-galactosidase (an indicator of senescence) expression. Pigmentation and cell size were evaluated by spectrophotometry and microscopy. p16 and Arf expression in primary and spontaneously immortalized melanocyte or melanocyte precursor cell lines were evaluated by immunoblotting. Retroviral vectors containing normal p16 and Arf complementary DNAs were used to restore expression of these genes in Ink4a-Arf(-/-) melanocytes. RESULTS: Wild-type melanocytes (i.e., Ink4a-Arf(+/+)) senesced within 4-5 weeks of culture. Ink4a-Arf(-/-) melanocytes did not senesce and readily became immortal. Ink4a-Arf(+/-) melanocytes showed defective senescence. Senescent Ink4a-Arf(+/+) melanocytes were heavily pigmented, but Ink4a-Arf(+/-) and Ink4a-Arf(-/-) melanocytes were less pigmented. All of six spontaneously immortalized melanocyte or melanocyte precursor lines from Ink4a-Arf(+/+) mice lacked p16 protein expression, although most retained Arf protein expression. After restoration of p16 but not Arf expression, Ink4a-Arf(-/-) melanocytes stopped growing, became highly melanized, and expressed acidic beta-galactosidase. By contrast, restoration of Arf but not p16 expression led to cell death without evidence of senescence. CONCLUSION: Normal mouse melanocyte senescence and associated pigmentation require both copies of Ink4a-Arf and appear to depend more on p16 than on Arf function. Mutations of the INK4A-ARF locus may favor tumorigenesis from melanocytes by impairing senescence, cell differentiation, and (where ARF is disrupted) cell death.     

Malignant melanoma: cross-reacting (common) tumor rejection antigens

The expression of tumor-associated transplantation antigens (TATA) by 3 different murine melanomas was examined. A comparison was made between different modes of inducing tumor-rejection activity, including immunization with irradiated cells from tissue culture lines, with irradiated cells from solid tumor lines, and with viable cells growing in footpads (followed by amputation). Melanoma cell lines examined included the spontaneous B16 melanoma, the ultraviolet-light-induced K1735 melanoma, and the dimethylbenzanthracene-induced JB/RH melanoma. The data presented demonstrate that not only do all 3 melanoma lines studied express cell surface antigens sufficient to elicit immune response which result in tumor-rejection activity, but that these antigens show crossreactivity among the 3 melanoma lines studied. The specificity of the TATA appear to be restricted to the melanomas, since crossreactivity was not observed with 2 different fibrosarcoma cell lines, or with 2 sarcoma cell lines. In addition, it was found that both the JB/RH and K1735 melanoma cells release (or shed) cell surface antigens which can elicit tumor rejection activity, and that these antigens can be extracted with aqueous butanol, as has been demonstrated with B16 melanoma.     

Serological characterization of a shared melanoma-associated antigen of mouse melanomas: relationship to the B700 glycoprotein.

A group of murine melanomas, consisting of the C57BL/6 melanomas JB/RH and B16 F10, and the C3H/He melanoma K1735, have been shown to be cross-immunogenic in tumour rejection assays, and to be antigenically distinct from the DBA/2 melanoma S91. In addition, the M(r) 65,000 melanoma-associated glycoprotein, B700, isolated from the B16F10 melanoma, was shown to induce a pattern of cross-immunity in semi-syngeneic mice, which was identical to that obtained with the melanomas. An antigen expressed by the JB/RH melanoma has been serologically defined by complement-dependent cytotoxic antibodies present in the sera of semi-syngeneic mice hyperimmunized against this melanoma. This antigen, designate JB/RH antigen, also was detected on JB/MS, B16 F10 and K1735 melanomas, but not on S91 melanoma. The cytotoxic antibodies defining the JB/RH antigen could be absorbed by the B700 glycoprotein isolated from B16 F10 melanoma, but not from S91 melanoma. Monoclonal antibodies were generated and shown to recognize a M(r) 65,000 antigen expressed by B16 F10 melanoma, but not S91 melanoma, suggesting that they have a specificity similar to that of the anti-JB/RH serum.     

Ink4a and Arf are crucial factors in the determination of the cell of origin and the therapeutic sensitivity of Myc-induced mouse lymphoid tumor

The cell of origin of tumors and the factors determining the cell of origin remain unclear. In this study, a mouse model of precursor B acute lymphoblastic leukemia/lymphoma (pre-B ALL/LBL) was established by retroviral transduction of Myc genes (N-Myc or c-Myc) into mouse bone marrow cells. Hematopoietic stem cells (HSCs) exhibited the highest susceptibility to N-Myc-induced pre-B ALL/LBL versus lymphoid progenitors, myeloid progenitors and committed progenitor B cells. N-Myc was able to induce pre-B ALL/LBL directly from progenitor B cells in the absence of Ink4a and Arf. Arf was expressed higher in progenitor B cells than Ink4a. In addition, N-Myc induced pre-B ALL/LBL from Arf(-/-) progenitor B cells suggesting that Arf has a predominant role in determining the cell of origin of pre-B ALL/LBL. Tumor cells derived from Ink4a/Arf(-/-) progenitor B cells exhibited a higher rate of proliferation and were more chemoresistant than those derived from wild-type HSCs. Furthermore, the Mdm2 inhibitor Nutlin-3 restored p53 and induced massive apoptosis in mouse pre-B ALL/LBL cells derived from Ink4a/Arf(-/-) cells and human B-ALL cell lines lacking Ink4a and Arf expression, suggesting that Mdm2 inhibition may be a novel therapeutic approach to the treatment of Ink4a/Arf(-/-) B-ALL/LBL, such as is frequently found in Ph(+) ALL and relapsed ALL. Collectively, these findings indicate that Ink4a and Arf are critical determining factors of the cell of origin and the therapeutic sensitivity of Myc-induced lymphoid tumors.     

Genetic and epigenetic alterations in lung tumors from bitransgenic Ki-rasG12C expressing mice

Mutations in Ki-ras occur in approximately 30-50% of patients with adenocarcinoma (AC) of the lung. We previously reported the development of a bitransgenic mouse model that expressed the human Ki-ras(G12C) allele in a lung-specific, tetracycline-inducible manner and gave rise to benign lung tumors. In the current study, these benign tumors, which represent relatively early lesions in neoplastic progression, were analyzed for molecular alterations secondary to mutant Ki-ras expression to determine the gene(s) that contribute to adenoma (AD) development. Tumors were removed following doxycycline (DOX) treatment for 9 and 12 mo and examined for alterations in cell-cycle regulatory genes. Quantification of mRNA expression for cyclin D1, retinoblastoma, p16(Ink4a), p19(Arf), and survivin was carried out by real-time PCR. All of the tumors examined exhibited a mean reduction of approximately fivefold for the retinoblastoma gene (P < 0.02). Increased expression of both p19(Arf) and survivin were detected in a majority of the tumors examined (P < 0.01 and 0.001, respectively), but no change in cyclin D1 RNA expression was observed. A subset of the lung tumors (8/28) displayed reduced levels of p16(Ink4a) expression (P = 0.02). Immunohistochemical analysis confirmed the upregulation of p19(Arf) and survivin in all 10 of the lung tumors examined. However, increased staining for cyclin D1 was observed in the tumor tissue. In addition, increased levels of activated p53 were found in lung tumor tissues stained with an anti-phospho-p53 antibody, while an absence of staining was observed with an anti-phospho-pRb antibody in both normal control and tumor tissue. Analysis of the methylation status of p16(Ink4a) by methylation-specific PCR (MSP) demonstrated that seven of eight tumors exhibiting decreased expression of p16(Ink4a) had at least partial methylation of the promoter region. Single stranded conformational polymorphism (SSCP) analysis demonstrated that neither exons 1 or 2 of p16(Ink4a) nor exons 5-8 of p53 exhibited mutations. These data thus identify alterations in specific genes and pathways that combine with the mutation in Ki-ras to promote the formation of benign lung tumors and suggest potential targets for the development of novel chemotherapeutic and chemopreventive agents during the early stages of lung tumor progression.     

Loss of p53 and Ink4a/Arf cooperate in a cell autonomous fashion to induce metastasis of hepatocellular carcinoma cells

Hepatocellular carcinoma (HCC) is a leading cause of cancer-related death worldwide. HCC patients frequently present with disease that has metastasized to other regions of the liver, the portal vein, lymph nodes, or lungs, leading to poor prognoses. Therefore, model systems that allow exploration of the molecular mechanisms underlying metastasis in this disease are greatly needed. We describe here a metastatic HCC model generated after the somatic introduction of the mouse polyoma virus middle T antigen to mice with liver-specific deletion of the Trp53 tumor suppressor locus and show the cell autonomous effect of p53 loss of function on HCC metastasis. We additionally find that cholangiocarcinoma also develops in these mice, and some tumors display features of both HCC and cholangiocarcinoma, suggestive of origin from liver progenitor cells. Concomitant loss of the Ink4a/Arf tumor suppressor locus accelerates tumor formation and metastasis, suggesting potential roles for the p16 and p19 tumor suppressors in this process. Significantly, tumor cell lines isolated from tumors lacking both Trp53 and Ink4a/Arf display enhanced invasion activity in vitro relative to those lacking Trp53 alone. Thus, our data illustrate a new model system amenable for the analysis of HCC metastasis.     

The differential impact of p16(INK4a) or p19(ARF) deficiency on cell growth and tumorigenesis

Mounting genetic evidence suggests that each product of the Ink4a/Arf locus, p16(INK4a) and p19(ARF), possesses tumor-suppressor activity (Kamijo et al., 1997; Krimpenfort et al., 2001; Sharpless et al., 2001a). We report the generation and characterization of a p19(ARF)-specific knockout allele (p19(ARF)-/-) and direct comparison with mice and derivative cells deficient for p16(INK4a), both p16(INK4a) and p19(ARF), and p53. Like Ink4a/Arf-/- murine embryo fibroblasts (MEFs), p19(ARF)-/- MEFs were highly susceptible to oncogenic transformation, exhibited enhanced subcloning efficiency at low density, and resisted both RAS- and culture-induced growth arrest. In contrast, the biological profile of p16(INK4a)-/- MEFs in these assays more closely resembled that of wild-type cells. In vivo, however, both p19(ARF)-/- and p16(INK4a)-/- animals were significantly more tumor prone than wild-type animals, but each less so than p53-/- or Ink4a/Arf-/- animals, and with differing tumor spectra. These data confirm the predominant role of p19(ARF) over p16(INK4a) in cell culture-based assays of MEFs, yet also underscore the importance of the analysis of tumor suppressors across many cell types within the organism. The cancer-prone conditions of mice singly deficient for either p16(INK4a) or p19(ARF) agree with data derived from human cancer genetics, and reinforce the view that both gene products play significant and nonredundant roles in suppressing malignant transformation in vivo.     

Molecular analysis of the Ink4a/Rb1-Arf/Tp53 pathways in radon-induced rat lung tumors

Inhalation of radon is closely associated with an increased risk of lung cancers. While the involvement of Ink4a in lung tumor development has been widely described, the tumor suppressor gene has not been studied in radon-induced lung tumors. In this study, loss of heterozygosity (LOH) analysis of the Cdkn2a locus, common to the Ink4a and Arf genes, was performed on 33 radon-induced rat lung tumors and showed a DNA loss in 50% of cases. The analysis of p16(Ink4a) protein expression by immunohistochemistry revealed that 50% of the tumors were negative for this protein. Looking for the origin of this lack of expression, we observed a low frequency of homozygous deletion (6%), a lack of mutation, an absence of correlation between promoter methylation and Ink4a mRNA expression and no correlation between LOH and protein expression. However, a tendency for an inverse correlation between p16(Ink4a) and pRb protein expression was observed. The expressions of p19Arf, Mmd2 and Mdm4 were not deregulated and only 14% of the tumors were mutated for Tp53. These results indicated that Ink4a/Cdk4/Rb1 pathway deregulation, more than Arf/Mdm2/Tp53 pathway, has a major role in the development of these tumors through p16(Ink4a) deregulation. However, all known mechanisms of inactivation of the pathway do not play a recurrent role in these tumors and the actual origin of the lack of p16(Ink4a) protein expression remains to be established.     

p19Arf inhibits the invasion of hepatocellular carcinoma cells by binding to C-terminal binding protein

The INK4A/ARF tumor suppressor locus is frequently inactivated in hepatocellular carcinoma (HCC), yet the consequences of this remain unknown. We recently described a HCC mouse model in which loss of the Ink4a/Arf locus accelerates the development of metastasis and enhances tumor cell migration and invasion in cell culture assays. We show here that knockdown of p19Arf in an HCC cell line increases invasion in cell culture assays. Furthermore, reintroduction of p19(Arf) into HCC cell lines lacking Ink4a/Arf inhibits tumor cell invasion, without affecting cell proliferation, or cell transformation as measured by soft agar colony formation. Inhibition of cell invasion by p19(Arf) was dependent on its C-terminal binding protein (CtBP) interaction domain but independent of Mdm2 binding and nucleolar localization. Indeed, RNA interference-mediated knockdown of CtBP1 or CtBP2 decreased cell invasion, and ectopic expression of CtBP2 enhanced tumor cell migration and invasion. Thus, our data indicate a novel role for the Arf tumor suppressor protein in regulating phenotypes associated with tumor progression and metastasis in HCC cells.