MAGE-A4 Interacts with the Liver Oncoprotein Gankyrin and Suppresses Its Tumorigenic Activity
2003; Elsevier BV; Volume: 278; Issue: 12 Linguagem: Inglês
10.1074/jbc.m206104200
ISSN1083-351X
AutoresToshikazu Nagao, Hiroaki Higashitsuji, Kohsuke Nonoguchi, Toshiharu Sakurai, Simon Dawson, R. John Mayer, Katsuhiko Itoh, Jun Fujita,
Tópico(s)Cancer Research and Treatments
ResumoHepatocellular carcinoma ranks among the most common malignancies in Southeast Asia and South Africa. Although there are many modalities of treatment, the recurrence and metastasis rates are high, and the prognosis is unsatisfactory. Gankyrin, a recently found oncoprotein, is a promising target for drug therapy because it is overexpressed in all studied hepatocellular carcinomas. Gankyrin contains six ankyrin repeats and interacts with Rb, Cdk4, and the S6 ATPase of the 26 S proteasome. In this study, a yeast two-hybrid screen with gankyrin has identified MAGE-A4 as another interacting protein. The interaction, mediated by the C-terminal half of MAGE-A4, was reproduced in mammalian cells. The interaction was specific to MAGE-A4, because other MAGE family proteins structurally similar to MAGE-A4,i.e. MAGE-A1, MAGE-A2, and MAGE-A12, did not bind to gankyrin. MAGE-A4 partially suppressed both anchorage-independent growth in vitro and tumor formation in athymic mice of gankyrin-overexpressing cells. The ability of mutant MAGE-A4 to interact with gankyrin correlated with the ability to suppress the anchorage-independent growth. These results demonstrate that MAGE-A4 binds to gankyrin and suppresses its oncogenic activity. So far, the major focus of studies on the MAGE proteins has been on their potential for cancer immunotherapy. Our results may also shed light on novel functions for MAGE-A proteins. Hepatocellular carcinoma ranks among the most common malignancies in Southeast Asia and South Africa. Although there are many modalities of treatment, the recurrence and metastasis rates are high, and the prognosis is unsatisfactory. Gankyrin, a recently found oncoprotein, is a promising target for drug therapy because it is overexpressed in all studied hepatocellular carcinomas. Gankyrin contains six ankyrin repeats and interacts with Rb, Cdk4, and the S6 ATPase of the 26 S proteasome. In this study, a yeast two-hybrid screen with gankyrin has identified MAGE-A4 as another interacting protein. The interaction, mediated by the C-terminal half of MAGE-A4, was reproduced in mammalian cells. The interaction was specific to MAGE-A4, because other MAGE family proteins structurally similar to MAGE-A4,i.e. MAGE-A1, MAGE-A2, and MAGE-A12, did not bind to gankyrin. MAGE-A4 partially suppressed both anchorage-independent growth in vitro and tumor formation in athymic mice of gankyrin-overexpressing cells. The ability of mutant MAGE-A4 to interact with gankyrin correlated with the ability to suppress the anchorage-independent growth. These results demonstrate that MAGE-A4 binds to gankyrin and suppresses its oncogenic activity. So far, the major focus of studies on the MAGE proteins has been on their potential for cancer immunotherapy. Our results may also shed light on novel functions for MAGE-A proteins. hepatocellular carcinoma retinoblastoma tumor suppressor protein cyclin-dependent kinase melanoma antigen 3-aminotriazole hemagglutinin doxycycline enhanced green fluorescent protein fluorescein isothiocyanate tetramethylrhodamine isothiocyanate Gankyrin (gann ankyrin repeat protein, also known as PSMD10 and p28) is an oncoprotein, the expression of which is increased (1Higashitsuji H. Itoh K. Nagao T. Dawson S. Nonoguchi K. Kido T. Mayer R.J. Arii S. Fujita J. Nat. Med. 2000; 6: 96-99Crossref PubMed Scopus (281) Google Scholar, 2Fu X.Y. Wang H.Y. Tan L. Liu S.Q. Cao H.F. Wu M.C. World J. Gastroenterol. 2002; 8: 638-643Crossref PubMed Scopus (90) Google Scholar) in hepatocellular carcinomas (HCCs).1 Gankyrin consists of six ankyrin repeats and a 38-amino acid N-terminal extension and binds to the retinoblastoma tumor suppressor protein (Rb), the S6 ATPase subunit of the 26 S proteasome (PSMC4, RPT3, TBP7), and cyclin-dependent kinase 4 (Cdk4) (1Higashitsuji H. Itoh K. Nagao T. Dawson S. Nonoguchi K. Kido T. Mayer R.J. Arii S. Fujita J. Nat. Med. 2000; 6: 96-99Crossref PubMed Scopus (281) Google Scholar, 3Dawson D. Apcher S. Mee M. Higashitsuji H. Baker R. Uhle S. Dubiel W. Fujita J. Mayer R.J. J. Biol. Chem. 2002; 277: 10893-10902Abstract Full Text Full Text PDF PubMed Scopus (113) Google Scholar, 4Li J. Tsai M.-D. Biochemistry. 2002; 41: 3977-3983Crossref PubMed Scopus (83) Google Scholar). Overexpression of gankyrin increases both the phosphorylation and degradation of Rbin vivo and oncogenically transforms NIH/3T3 cells. Gankyrin binds to Cdk4 and counteracts the inhibitory function of the tumor suppressors p16INK4A and p18INK4C (4Li J. Tsai M.-D. Biochemistry. 2002; 41: 3977-3983Crossref PubMed Scopus (83) Google Scholar). In a rodent model of hepatocarcinogenesis, gankyrin is overexpressed from the earliest stage of tumor development (5Park T.J. Kim H.S. Byun K.H. Jang J.J. Lee Y.S. Lim I.K. Mol. Carcinog. 2001; 30: 138-150Crossref PubMed Scopus (55) Google Scholar). These findings suggest that gankyrin is a major player in cell cycle control and tumorigenesis in HCCs. The MAGE (melanoma antigen) genes were initially identified because they encode tumor antigens that can be recognized by cytolytic T lymphocytes derived from the blood lymphocytes of cancer patients (6van der Bruggen P. Traversari C. Chomoz P. Lurquin C. De Plaen E. van den Eynde B. Knuth A. Boon T. Science. 1991; 254: 1643-1647Crossref PubMed Scopus (3126) Google Scholar). The MAGE gene family is composed of more than 25 genes in humans and are classified as type I MAGE genes (including MAGE-A, MAGE-B, and MAGE-Cgenes) and type II MAGE genes, which include those that reside outside of the MAGE-A, MAGE-B, andMAGE-C genomic clusters (7Chomez P. De Backer O. Bertrand M. De Plaen E. Boon T. Lucas S. Cancer Res. 2001; 61: 5544-5551PubMed Google Scholar, 8Baker P.A. Salehi A. J. Neurosci. Res. 2002; 67: 705-712Crossref PubMed Scopus (263) Google Scholar). The MAGE-Asubfamily comprises 12 genes (MAGE-A1 toMAGE-A12), and is expressed in various types of tumors but not in normal adult tissues, except for testis and placenta. The MAGE-A antigens are of particular interest for antitumor immunotherapy because they are strictly tumor specific and are shared by many tumors. Despite the isolation of growing numbers ofMAGE genes, their function in normal tissues remains mostly unknown. To further characterize the molecular mechanism underlying the oncogenic activity of gankyrin and facilitate development of a therapeutic agent against HCCs, we have used a yeast two-hybrid screen to identify further gankyrin interactions. We report here that MAGE-A4 binds to human gankyrin and suppresses its oncogenicity. A full-length human gankyrin cDNA was cloned into pAS2–1 vector (BD Biosciences) and co-transformed into Y190 yeast cells with a placenta or a U-2 OS cDNA library in pACT2 vector (BD Biosciences). Yeast clones containing interacting proteins were identified by growth on media lacking tryptophan, leucine, and histidine, followed by assaying for β-galactosidase activity. The interaction was confirmed by co-transformation into Y190 cells with pAS2–1-gankyrin and their growth on the selection medium with 25 mm 3-aminotriazole (3AT). For analysis of interactions, full-length and various mutant cDNAs of gankyrin (GenBankTM accession number D83197) and MAGE-A4 (GenBankTM accession number U10687) were generated by the polymerase chain reaction and cloned into a pAS2–1 or pACT2 vector, respectively. cDNAs corresponding to the C-terminal regions of MAGE-A1 (amino acids 203–309, GenBankTMaccession number NM_004988), MAGE-A2 (amino acids 210–314, GenBankTM accession number NM_005361), and MAGE-A12 (amino acids 210–314, GenBankTM accession number XM_010079) were also cloned into pACT2. FLAG-tagged MAGE-A4, HA-tagged MAGE-A4, and HA-tagged gankyrin cDNAs were cloned into the eukaryotic expression vectors pMKit-neo, pMKit-hygro, and pCMV4. pEGFP-C1 vector (BD Biosciences) was used to express EGFP-S6 and EGFP-MAGE-A4 fusion proteins. For conditional expression of FLAG-MAGE-A4, we used the tetracycline-regulated system as described (9Higashitsuji H. Higashitsuji H. Nagao T. Nonoguchi K. Fujii S. Itoh K. Fujita J. Cancer Cell. 2002; 2: 335-346Abstract Full Text Full Text PDF PubMed Scopus (61) Google Scholar). The pMACS4-IRES vector (Miltenyi Biotec GmbH, Bergisch Gladbach, Germany) was used to express MAGE-A4 or C-terminal 107 or 55 amino acids of MAGE-A4 with truncated human CD4. The CD4-expressing transfectants were enriched by using the MACSelect 4 system (Miltenyi Biotec GmbH). Mouse NIH/3T3 embryonal fibroblasts and its derivatives, monkey COS-7 kidney cells, human U-2 OS osteosarcoma cells, Huh-7 hepatoma cells, PLC/PRF/5 hepatoma cells, and HEK293 human embryonal kidney cells and its derivatives were cultured in Dulbecco's modified Eagle medium supplemented with 10% calf or fetal calf serum at 37 °C in a humidified atmosphere of 5% CO2 in air. Transfection was performed by the calcium phosphate method. pl16-GK-6 cells were generated from NIH/3T3 cells so that they constitutively expressed HA-tagged human gankyrin and the tetracycline activator, rtTA (9Higashitsuji H. Higashitsuji H. Nagao T. Nonoguchi K. Fujii S. Itoh K. Fujita J. Cancer Cell. 2002; 2: 335-346Abstract Full Text Full Text PDF PubMed Scopus (61) Google Scholar). In the presence of 2 μg/ml doxycycline (Dox) in culture medium, they expressed FLAG-tagged MAGE-A4. GK-S25 cells were derived from NIH/3T3 cells and constitutively expressed gankyrin. To assess the anchorage dependence of growth in vitro, 5 × 102 cells were plated in 0.3% agarose on top of 0.6% agarose in the presence or absence of Dox in a 35-mm dish. Four weeks later, the number of colonies (>20 cells) was counted microscopically. Statistical differences between sample means were calculated by analysis of variance, followed by an unpaired Student's ttest. Western blot analysis and immunoprecipitation were performed as described previously (1Higashitsuji H. Itoh K. Nagao T. Dawson S. Nonoguchi K. Kido T. Mayer R.J. Arii S. Fujita J. Nat. Med. 2000; 6: 96-99Crossref PubMed Scopus (281) Google Scholar, 9Higashitsuji H. Higashitsuji H. Nagao T. Nonoguchi K. Fujii S. Itoh K. Fujita J. Cancer Cell. 2002; 2: 335-346Abstract Full Text Full Text PDF PubMed Scopus (61) Google Scholar). Antibodies used were mouse monoclonal anti-HA antibody (Roche Diagnostics), anti-FLAG antibody (Sigma), anti-GFP antibody (BD Biosciences), anti-actin antibody (Chemicon International, Inc., Temecula, CA), rabbit polyclonal anti-gankyrin antibody (Santa Cruz Biotechnology), anti-MAGE antibody (FL-309; Santa Cruz Biotechnology), horseradish peroxidase-conjugated goat anti-mouse antibody (DAKO, Kyoto, Japan), and anti-rabbit antibody (DAKO). To analyze interaction between endogenous MAGE-A4 and endogenous gankyrin, U-2 OS cell lysates were immunoprecipitated using rabbit anti-MAGE antibody (FL-309) immobilized to the protein G support (Seize X mammalian immunoprecipitation kit; Pierce). The antibody is broadly reactive with all MAGE family members according to the manufacturer. The immobilized antibody was incubated consecutively with four aliquots of cell lysates (total of 1 ml), the first aliquot at 4 °C overnight and the others at 20 °C sequentially for 1 h. The bound proteins were eluted with the elution buffer supplied in the kit. The elution was repeated three times, and each fraction was analyzed by Western blotting using rabbit polyclonal anti-gankyrin antibody (Santa Cruz Biotechnology.). Interaction of transiently expressed exogenous MAGE-A4 or its mutant with exogenous gankyrin was examined by co-transfecting COS-7 or U-2 OS cells with plasmids expressing FLAG-tagged MAGE-A4, FLAG-tagged C-terminal 107 amino acids of MAGE-A4, HA-tagged full-length gankyrin, FLAG alone, or HA alone in various combinations. Interaction of stably transfected gankyrin with MAGE-A4 was examined by using pl16-GK-6 cells, which were derived from mouse NIH/3T3 cells, constitutively expressed HA-tagged human gankyrin, and inducibly expressed FLAG-tagged MAGE-A4 in the presence of Dox. To analyze the effects of MAGE-A4 on binding of gankyrin to Rb, S6, and Cdk4, U-2 OS cells and 293T cells were co-transfected with plasmids expressing FLAG-gankyrin, HA-tagged Rb, EGFP-S6 fusion protein, HA-tagged Cdk4, HA-tagged MAGE-A4, EGFP-MAGE-A4 fusion protein, EGFP alone, or FLAG alone in various combinations. Immunofluorescence staining was performed essentially as described (9Higashitsuji H. Higashitsuji H. Nagao T. Nonoguchi K. Fujii S. Itoh K. Fujita J. Cancer Cell. 2002; 2: 335-346Abstract Full Text Full Text PDF PubMed Scopus (61) Google Scholar, 10Hopwood D. Celis J.E. Cell Biology, A Laboratory Handbook. 2nd Ed. Academic Press, San Diego, CA1998: 221-231Google Scholar). COS-7 cells were replated on chamber slides after transfection, fixed with phosphate-buffered saline containing 4% paraformaldehyde for 30 min, and then rendered permeable with phosphate-buffered saline containing 0.2% Triton X-100 for 30 min at room temperature. After blocking nonspecific antibody-binding sites with bovine serum albumin, the cells were incubated with mouse monoclonal anti-FLAG antibody (Sigma) and rabbit polyclonal anti-HA antibody (Berkeley Antibody, Richmond, CA). Then the bound antibodies were reacted with FITC-linked anti-mouse and TRITC-linked anti-rabbit IgGs (Amersham Biosciences) and observed under a confocal laser microscope (Olympus, Tokyo, Japan). Twenty-four female BALB/c Slc-nu/nu athymic mice (4 weeks old) were injected subcutaneously with pl16-GK-6 cells or GK-S25 cells (8 × 106 cells each) and divided into two groups. Twelve were given Dox (2 mg/ml) in the drinking water ad libitum. Tumor size was calculated by measuring the length, width, and thickness with calipers. We performed yeast two-hybrid assays to search for proteins capable of physically interacting with human gankyrin. Using full-length gankyrin as a bait, we identified 18 clones of 3.5 × 106 yeast clones transformed with human placenta or U-2 OS cell cDNA libraries. Each clone proliferated on media containing the histidine inhibitor 3AT and was positive for β-galactosidase staining (Fig. 1 A, and data not shown). DNA sequencing analysis of the rescued plasmids revealed that two of them encoded the C-terminal 107 amino acids of MAGE-A4. Consistent with our previous findings (3Dawson D. Apcher S. Mee M. Higashitsuji H. Baker R. Uhle S. Dubiel W. Fujita J. Mayer R.J. J. Biol. Chem. 2002; 277: 10893-10902Abstract Full Text Full Text PDF PubMed Scopus (113) Google Scholar), the remaining 16 clones encoded different C-terminal sequences of the S6 ATPase subunit of the 26 S proteasome. MAGE-A4 did not interact with the GAL4 DNA binding domain alone (Fig. 1 A), indicating that it interacted with gankyrin. To confirm that gankyrin interacts with the isolated MAGE-A4 fragment in mammalian cells, COS-7 cells were co-transfected with plasmid DNAs expressing HA-tagged human gankyrin and FLAG-tagged C-terminal 107 amino acids of MAGE-A4. When cell lysates were immunoprecipitated with an anti-FLAG antibody, HA-gankyrin was detected in them but not in precipitates from cells co-transfected with parental FLAG vector and HA-gankyrin (Fig. 1 B, left panels). Reciprocally, truncated MAGE-A4 was detected in the anti-HA immunoprecipitates from cells co-transfected with plasmids expressing HA-gankyrin and FLAG-tagged truncated MAGE-A4 (Fig. 1 B,right panels). We next examined whether full-length MAGE-A4 interacts with gankyrin in COS-7 cells. As shown in Fig. 1 C, HA-tagged gankyrin was immunoprecipitated with anti-FLAG antibody in lysates from cells co-transfected and expressing FLAG-tagged full-length MAGE-A4 and HA-tagged gankyrin. Reciprocally, FLAG-tagged MAGE-A4 was immunoprecipitated with an anti-HA antibody. Similar interactions were observed in U-2 OS cells as well (data not shown). Furthermore, in the COS-7 cells expressing FLAG-MAGE-A4 and HA-gankyrin, double immunofluorescence staining showed that both proteins were co-localized in the cytoplasm (Fig. 1 D). COS-7 cells are well known to vastly over-express transfected proteins, which makes it a problem to conclude that physiologically relevant associations occur between such expressed proteins. Therefore, we generated an NIH/3T3-derived clone (pl16-GK-6) in which HA-gankyrin was stably overexpressed and FLAG-MAGE-A4 was inducibly expressed in the presence of Dox. As shown in Fig.2 A, Dox increased the level of FLAG-MAGE-A4 but not that of HA-gankyrin in pl16-GK-6 cells. Only in the presence of Dox was FLAG-MAGE-A4 co-immunoprecipitated with HA-gankyrin by the anti-HA antibody and HA-gankyrin co-immunoprecipitated with FLAG-MAGE-A4 by anti-FLAG antibody (Fig.2 A, bottom two panels). About 40% of total MAGE-A4 and 5% of total gankyrin present in pl16-GK-6 cells after transfection were estimated to be in a complex (data not shown). Western blot analysis demonstrated that the level of stably overexpressed gankyrin in pl16-GK-6 cells was less than those observed in human U-2 OS osteosarcoma and Huh-7 hepatoma cells (Fig.2 B, top panels). The level of MAGE-A4 in the presence of Dox was probably less than that in U-2 OS osteosarcoma and PLC/PRF/5 hepatoma cells (Fig. 2 B,bottom panels). Further confirmation is necessary for this observation, because the anti-MAGE antibody used was reactive with MAGE proteins other than MAGE-A4. We next examined the interaction between endogenous gankyrin and MAGE-A4 in U-2 OS cells from which we isolated the original truncated MAGE-A4 cDNA clones. As shown in Fig. 2 C, the endogenous gankyrin was co-immunoprecipitated with MAGE proteins by anti-MAGE antibody. Taken together, these results strongly suggest that gankyrin and MAGE-A4 interacts in human cancer cells. Because gankyrin binds to Rb, the S6 subunit of the 26 S proteasome, and Cdk4 (1Higashitsuji H. Itoh K. Nagao T. Dawson S. Nonoguchi K. Kido T. Mayer R.J. Arii S. Fujita J. Nat. Med. 2000; 6: 96-99Crossref PubMed Scopus (281) Google Scholar, 3Dawson D. Apcher S. Mee M. Higashitsuji H. Baker R. Uhle S. Dubiel W. Fujita J. Mayer R.J. J. Biol. Chem. 2002; 277: 10893-10902Abstract Full Text Full Text PDF PubMed Scopus (113) Google Scholar, 4Li J. Tsai M.-D. Biochemistry. 2002; 41: 3977-3983Crossref PubMed Scopus (83) Google Scholar), the effects of overexpression of MAGE-A4 on the binding of gankyrin to these proteins were examined. When U-2 OS cells were co-transfected with plasmids expressing FLAG-tagged gankyrin and HA-tagged Rb, HA-Rb was immunoprecipitated with anti-FLAG antibody as expected (Fig. 3 A). Overexpression of EGFP-MAGE-A4 fusion protein by co-transfection did not affect the amount of immunoprecipitated HA-Rb. When U-2 OS cells were co-transfected with plasmids expressing FLAG-tagged gankyrin and EGFP-S6 fusion protein, EGFP-S6 was immunoprecipitated with anti-FLAG antibody as expected (Fig. 3 B). Overexpression of HA-tagged MAGE-A4 by co-transfection did not affect the amount of immunoprecipitated EGFP-S6. Similarly, no effect of EGFP-MAGE-A4 was observed on the binding of gankyrin to Cdk4 (data not shown). Similar results were obtained with 293T cells (data not shown). To characterize the interacting domains of gankyrin and MAGE-A4, we made various deletion mutants and analyzed their interactions in the yeast two-hybrid system. We demonstrated previously that full-length is necessary for the interaction with Rb (1Higashitsuji H. Itoh K. Nagao T. Dawson S. Nonoguchi K. Kido T. Mayer R.J. Arii S. Fujita J. Nat. Med. 2000; 6: 96-99Crossref PubMed Scopus (281) Google Scholar). Similarly, no deletion mutant of gankyrin interacted with MAGE-A4, indicating that all ankyrin repeats and the N-terminal extension are necessary for the binding (Fig.4 A, and data not shown). The MAGE-A4 clone originally isolated by the two-hybrid screen contained only the C-terminal 107 amino acids (position 211 to 317, Fig. 1, A and B, and Fig. 4 B). Further N-terminal truncation up to residue 226 did not prevent the interaction with gankyrin. By contrast, a truncation of 10 amino acids from the C terminus of MAGE-A4 abolished its binding to gankyrin (Fig.4 B). The interaction was specific to MAGE-A4, because the corresponding regions of other MAGE proteins, although structurally quite similar to MAGE-A4, did not interact with gankyrin (Fig. 4,C and D). These results indicate that the C-terminal region of MAGE-A4 containing the HLA-A2-presented peptides (11Duffour M.-T. Chaux P. Lurquin C. Cornelis G. Boon T. van der Bruggen P. Eur. J. Immunol. 1999; 29: 3329-3337Crossref PubMed Scopus (67) Google Scholar, 12Graff-Dubois S. Faure O. Gross D.-A. Alves P. Scardino A. Chouaib S. Lemonnier F.A. Kosmatopoulos K. J. Immunol. 2002; 169: 575-580Crossref PubMed Scopus (66) Google Scholar) (GVYDGREHTV and YLEYRQVPV) specifically interacts with gankyrin. To investigate the biological effects of the interaction of gankyrin and MAGE-A4, we used the NIH/3T3-derived pl16-GK-6 cells in which gankyrin was stably overexpressed and MAGE-A4 expression could be induced with Dox (Fig. 2 A). These gankyrin-transformed cells formed colonies in soft agar (Fig.5 A). When the cells were cultured in the presence of Dox, the numbers of colonies was decreased to 65% of those in the absence of Dox (Fig. 5 B). When the mouse NIH/3T3-derived GK-S25 cells stably overexpressing gankyrin were transfected with plasmids expressing the C-terminal 55 amino acids of MAGE-A4 that did not bind to gankyrin (Fig. 4 B), no decrease in the number of colonies were observed (Fig. 5 C). By contrast, the C-terminal 107 amino acids of the MAGE-A4 that binds to gankyrin inhibited the colony formation, suggesting that the ability to bind to gankyrin correlates with the ability to suppress anchorage-independent growth in vitro. The anchorage-independent phenotype in cell culture has been closely correlated with the ability of cells to form tumors in animals (13Shin S.-I. Freedman V.H. Risser R. Pollack R. Proc. Natl. Acad. Sci. U. S. A. 1975; 72: 4435-4439Crossref PubMed Scopus (656) Google Scholar). We therefore evaluated the effect of MAGE-A4 on tumor formation of gankyrin-transformed cells in athymic nude mice. After being subcutaneously inoculated with pl16-GK-6 cells, the animals were divided into two groups, one of which was administered Dox in drinking water. In the absence of Dox, all mice developed tumors 5 weeks after the inoculation of pl16-GK-6 cells (Fig.6). By contrast, the tumors appeared later and grew slower in mice given Dox. Dox by itself showed no suppressive effects on tumor formation by GK-S25 cells (data not shown). Taken together, these results demonstrate that MAGE-A4 directly binds to gankyrin and suppresses its tumorigenic activity. Rb and p53 play critical roles in transducing a variety of growth inhibitory signals to the cell cycle control machinery via distinct mechanisms (14Vogelstein B. Lane D. Levine A.J. Nature. 2000; 408: 307-310Crossref PubMed Scopus (5822) Google Scholar, 15Evan G.I. Vousden K.H. Nature. 2001; 411: 342-348Crossref PubMed Scopus (2711) Google Scholar, 16Hahn W.C. Weinberg R.A. Nat. Rev. Cancer. 2002; 2: 331-341Crossref PubMed Scopus (831) Google Scholar). The concurrent inactivation of these two pathways occurs frequently in human cancers and argues that unscheduled entry into the cell cycle and escape from cell cycle arrest/apoptosis are two critical events that a cell cycle requires to become cancerous. Gankyrin destabilizes Rb and is commonly overexpressed in HCCs (1Higashitsuji H. Itoh K. Nagao T. Dawson S. Nonoguchi K. Kido T. Mayer R.J. Arii S. Fujita J. Nat. Med. 2000; 6: 96-99Crossref PubMed Scopus (281) Google Scholar, 2Fu X.Y. Wang H.Y. Tan L. Liu S.Q. Cao H.F. Wu M.C. World J. Gastroenterol. 2002; 8: 638-643Crossref PubMed Scopus (90) Google Scholar). Gankyrin therefore plays an important role in hepatocarcinogenesis and is a promising target for therapeutic drug development. Here, we demonstrated that MAGE-A4 binds to gankyrin and suppresses its oncogenic activity. The interaction was identified by the yeast two-hybrid assay and reproduced in mammalian cells. Because endogenous gankyrin is associated with endogenous MAGE protein(s) in human cancer cells in which MAGE-A4 transcripts were detected, it is highly likely that the interaction is physiologically relevant. MAGE-A4 is a member of the MAGE family, a large group of proteins that contain a well conserved ∼200 amino acid region known as the MAGE homology domain (7Chomez P. De Backer O. Bertrand M. De Plaen E. Boon T. Lucas S. Cancer Res. 2001; 61: 5544-5551PubMed Google Scholar, 8Baker P.A. Salehi A. J. Neurosci. Res. 2002; 67: 705-712Crossref PubMed Scopus (263) Google Scholar). Although its general function is unknown, the MAGE homology domain of necdin binds to E2F-1 (17Taniura H. Matsumoto K. Yoshikawa K. J. Biol. Chem. 1999; 274: 16242-16248Abstract Full Text Full Text PDF PubMed Scopus (122) Google Scholar). In the present study, the C terminus of MAGE-A4, in addition to the C-terminal half of MAGE homology domain, was found to be necessary for MAGE-A4 to bind to gankyrin. This binding was specific for MAGE-A4, because similar amino acid sequences from other MAGE-A family members did not bind to gankyrin. Despite their normally restricted physiological expression, the MAGE genes are expressed in a wide variety of cancer cells. In the case of HCC, more than 60% of the carcinomas expressMAGE-A1 and/or MAGE-A3 transcripts, which have been regarded as tumor-specific markers (18Tahara K. Mori M. Sadanaga N. Sakamoto Y. Kitano S. Makuuchi M. Cancer. 1999; 85: 1234-1240Crossref PubMed Scopus (75) Google Scholar, 19Mou D.C. Cai S.L. Peng J.R. Wang Y. Chen H.S. Pang X.W. Leng X.S. Chen W.F. Br. J. Cancer. 2002; 86: 110-116Crossref PubMed Scopus (68) Google Scholar). Expression of otherMAGE family members, including MAGE-A4, has also been detected in HCCs (20Chen C.H. Huang G.T. Lee H.S. Yang P.M. Yan M.D. Chen D.S. Sheu J.C. Liver. 1999; 19: 110-114Crossref PubMed Scopus (33) Google Scholar). When overexpressed in the cytosol of cancer cells, MAGE proteins are proteolytically processed, transported to the endoplasmic reticulum, and then presented on the cell surface as antigenic major histocompatibility complex-associated peptides (8Baker P.A. Salehi A. J. Neurosci. Res. 2002; 67: 705-712Crossref PubMed Scopus (263) Google Scholar). Thus, attention has been focused on the potential of MAGE as a target for cancer immunotherapy. The C-terminal region of MAGE-A4 that binds to gankyrin contains decapeptides presented by HLA-A2 (11Duffour M.-T. Chaux P. Lurquin C. Cornelis G. Boon T. van der Bruggen P. Eur. J. Immunol. 1999; 29: 3329-3337Crossref PubMed Scopus (67) Google Scholar, 12Graff-Dubois S. Faure O. Gross D.-A. Alves P. Scardino A. Chouaib S. Lemonnier F.A. Kosmatopoulos K. J. Immunol. 2002; 169: 575-580Crossref PubMed Scopus (66) Google Scholar). Whether binding of gankyrin affects degradation of MAGE-A4 and/or the presentation of antigenic peptide remain to be investigated. The physiological roles played by the MAGE gene family are unknown with a few exceptions (8Baker P.A. Salehi A. J. Neurosci. Res. 2002; 67: 705-712Crossref PubMed Scopus (263) Google Scholar). Necdin has an important role in development and/or maintenance of discrete cells in the nervous system. Overexpression of necdin causes cell cycle arrest through mechanisms that may involve physical interactions with E2F-1 or p53 (17Taniura H. Matsumoto K. Yoshikawa K. J. Biol. Chem. 1999; 274: 16242-16248Abstract Full Text Full Text PDF PubMed Scopus (122) Google Scholar). NRAGE (MAGE-D1), which binds to several proteins including the p75 neurotrophin receptor and inhibitor of apoptosis proteins (IAP), blocks cell cycle progression and enhances apoptosis (21Salehi A.H. Roux P.P. Kubu C.J. Zeindler C. Bhakar A. Tannis L.L. Verdi J.M. Barker P.A. Neuron. 2000; 27: 279-288Abstract Full Text Full Text PDF PubMed Scopus (256) Google Scholar, 22Jordan B.W. Dinev V. LeMellay V. Troppmair J. Gotz R. Wixler L. Sendtner M. Ludwig S. Rapp U.R. J. Biol. Chem. 2001; 276: 39985-39989Abstract Full Text Full Text PDF PubMed Scopus (79) Google Scholar). Magphinin (MAGE-D4) is suggested as regulating cell proliferation during gametogenesis, and its ectopic expression suppresses cell proliferation (23Saburi S. Nadano D. Akama T.O. Hirama K. Yamanouchi K. Naito K. Tojo H. Tachi C. Fukuda M.N. J. Biol. Chem. 2001; 276: 49378-49389Abstract Full Text Full Text PDF PubMed Scopus (39) Google Scholar). In contrast to NRAGE, necdin, and magphinin, which belong to the type II MAGE genes, MAGE-A4 did not affect the cell cycle progression, proliferation rate, nor apoptosis by itself. 2T. Nagao and J. Fujita, unpublished observation. However, MAGE-A4showed anti-tumorigenic effects on gankyrin-overexpressing cells. Because this effect was observed in vitro as well asin vivo in athymic nude mice, it is cell autonomous and not mediated by T cells. In mammals, the execution of G1 Cdks, namely Cdk4, Cdk6, and Cdk2, is essential for onset of the S phase (24Weinberg R.A. Cell. 1995; 81: 323-330Abstract Full Text PDF PubMed Scopus (4313) Google Scholar). A major target for the G1 Cdks is Rb. Phosphorylation of Rb leads to activation of the E2F-DP1 transcriptional factor complex that controls the expression of genes essential for onset of the S phase (14Vogelstein B. Lane D. Levine A.J. Nature. 2000; 408: 307-310Crossref PubMed Scopus (5822) Google Scholar). Although there appears to be cell type- and genotype-specific differences in the control of anchorage-dependent growth, a recent study has demonstrated that the G1 Cdks and Cdc6 constitute major cell cycle targets for the regulation of the G1-S transition by anchorage and oncogenic stimulation (25Jinno S. Yageta M. Nagata A. Okayama H. Oncogene. 2002; 21: 1777-1784Crossref PubMed Scopus (18) Google Scholar). Gankyrin binds to Cdk4, evades inhibition by p16INK4A(3Dawson D. Apcher S. Mee M. Higashitsuji H. Baker R. Uhle S. Dubiel W. Fujita J. Mayer R.J. J. Biol. Chem. 2002; 277: 10893-10902Abstract Full Text Full Text PDF PubMed Scopus (113) Google Scholar, 4Li J. Tsai M.-D. Biochemistry. 2002; 41: 3977-3983Crossref PubMed Scopus (83) Google Scholar), increases phosphorylation and degradation of Rb, and transforms cells to grow in an anchorage-independent manner (1Higashitsuji H. Itoh K. Nagao T. Dawson S. Nonoguchi K. Kido T. Mayer R.J. Arii S. Fujita J. Nat. Med. 2000; 6: 96-99Crossref PubMed Scopus (281) Google Scholar). Because the suppressive effect of MAGE-A4 on anchorage-independent growth correlated with its binding to gankyrin, it is possibly mediated by gankyrin and/or dependent on association with gankyrin. However, we were unable to detect effects of MAGE-A4 on the degradation of Rb. 3T. Nagao, unpublished observation. MAGE-A4 did not reduce the binding of gankyrin to Rb, S6, or Cdk4. Molecules involved in the anti-tumorigenic activity of MAGE-A4 are yet to be elucidated. Recently, continued activity of a specific oncogene has been found to be necessary to maintain the cancer phenotype in some cancer cells, which suggests a possibility that suppression of gankyrin alone has a therapeutic effect in HCCs (26Weinstein I.B. Science. 2002; 297: 63-64Crossref PubMed Scopus (1484) Google Scholar). Further clarification of the mechanisms underlying the effects of MAGE-A4 on gankyrin and anchorage-independent growth will facilitate development of novel therapeutics against HCCs and also shed light on normal physiological functions of MAGE-A proteins. We thank Dr. Manabu Sugai for helpful suggestions.
Referência(s)