Portal de Periódicos da CAPES

Splicing into Senescence: The Curious Case of p16 and p19ARF

1997; Cell Press; Volume: 91; Issue: 5 Linguagem: Inglês

10.1016/s0092-8674(00)80441-9

1097-4172

Daniel A. Haber,

Microtubule and mitosis dynamics

The study of tumor suppressor genes generates much of its excitement from the convergence of experiments addressing the genetic basis of cancer, together with the cellular pathways that regulate cellular proliferation, immortality, and cell cycle progression. Among the most notable of these gene products is the cyclin-dependent kinase inhibitor p16 (also known as CDKN2, MTS-1, and INK4a), implicated in cell cycle regulation and cellular senescence, whose loss results in genetic predisposition to malignant melanoma and pancreatic cancer (for review, see3Hall M. Peters G. Adv. Cancer Res. 1996; 68: 67-108Crossref PubMed Google Scholar). In this issue of Cell, Kamijo and coworkers (1997) report that much of the phenotype ascribed to p16-null mice may in fact be attributed to disruption of p19ARF, an alternatively spliced transcript derived from genomic sequence shared with the p16 transcript but encoding a different open reading frame. These unexpected observations challenge previous assumptions about the unique role of p16 in cancer predisposition and cellular immortalization, suggesting distinct and potentially complementary functions for the two transcripts encoded by this complex locus. The discovery of p16 and its role in tumor predisposition resulted from advances in two disparate fields, cancer genetics and cell cycle regulation. The search for the familial melanoma gene led to the identification of genetic linkage to chromosome 9p, which was soon followed by the discovery of homozygous deletions at that locus in melanoma cell lines. Within the minimal deleted region lay p16 (6Kamb A. Gruis N.A. Weaver-Feldhaus J. Liu Q. Harshman K. Tavtigian S.V. Stockert E. Day III, R.S. Johnson B.E. Skolnick M.H. Science. 1994; 264: 436-440Crossref PubMed Scopus (2795) Google Scholar), previously identified by virtue of the interaction of its encoded protein with the cyclin-dependent kinase 4 (CDK4) (13Serrano M. Hannon G.J. Beach D. Nature. 1993; 366: 704-707Crossref PubMed Scopus (3298) Google Scholar). Like other members of the cyclin-dependent kinase inhibitor (CKI) family, p16 is a negative regulator of cyclin-CDK complexes (reviewed in17Sherr C.J. Roberts J.M. Genes Dev. 1995; 9: 1149-1163Crossref PubMed Scopus (3180) Google Scholar). p16 itself binds preferentially to CDK4 (hence the name inhibitor of kinase 4 or INK4) and CDK6, preventing their association with D-type cyclins and the ensuing phosphorylation of substrates such as the retinoblastoma protein pRB (see Figure 1). Expression of p16 thus maintains pRB in a hypophosphorylated and active form, leading to repression of E2F-DP complexes and G1 phase cell cycle arrest (reviewed in19Weinberg R.A. Cell. 1995; 81: 323-330Abstract Full Text PDF PubMed Scopus (4228) Google Scholar). The identification of p16 as a tumor suppressor gene was confirmed by the demonstration of intragenic mutations, both in the germline and in somatic tissues. The importance of this cell cycle regulatory pathway in the development of cancer has been underscored by alterations in other components of this pathway. For example, in addition to disruption of p16, human cancers may demonstrate mutational inactivation of RB, overexpression of D-type cyclins and CDKs, or a CDK4 mutation that abrogates inhibition of the gene product by p16. These genetic lesions often appear to be mutually exclusive within any given tumor, suggesting that they constitute equivalent steps in a single critical pathway (see16Sherr C.J. Science. 1996; 274: 1672-1677Crossref PubMed Scopus (4876) Google Scholar). While initially controversial, the role of p16 in familial melanoma is now firmly established. Analyses of melanoma pedigrees, defined by the presence of multiple affected individuals over multiple generations within a family, have demonstrated germline p16 mutations in approximately 50% of families (5Hussussian C.J. Struewing J.P. Goldstein A.M. Higgins P.A. Ally D.S. Sheahan M.D. Clark Jr., W.H. Tucker M.A. Dracopoli N.C. Nature Genet. 1994; 8: 15-21Crossref PubMed Scopus (1108) Google Scholar). Such pedigrees are rare, but 5%–10% of all melanoma patients have at least one other affected family member; ∼20% of individuals with this more common and moderate family history have a germline mutation in p16 (2FitzGerald M.G. Harkin D.P. Silva-Arrieta S. MacDonald D.J. Lucchina L.C. Unsal H. O'Neill E. Koh J. Finkelstein D.M. Isselbacher K.J. et al.Proc. Natl. Acad. Sci. USA. 1996; 93: 8541-8545Crossref PubMed Scopus (187) Google Scholar). Germline mutations in p16 are typically point mutations or small intragenic deletions within the first exon (1α) or second exon. Rare cases of familial melanoma have also been linked to a specific point mutation in CDK4 (called R24C) (20Zuo L. Weger J. Yang B. Goldstein A.M. Tucker M.A. Walker G.J. Hayward N. Dracopoli N.C. Nature Genet. 1996; 12: 97-99Crossref PubMed Scopus (656) Google Scholar), which renders the encoded protein insensitive to inhibition by p16, supporting the relevance of this functional pathway in conferring genetic predisposition to melanoma. The identification of p16 as a tumor suppressor gene implicated in familial melanoma was associated with the discovery of genomic deletions at that locus in a large number of cell lines derived from many tumor types (leading to the name multiple tumor suppressor 1 or MTS-1). In primary tumors, somatic inactivation of p16 is noted primarily in pancreatic cancer, esophageal cancer, glioblastoma, and acute T-cell lymphoblastic leukemia, and may occur by a number of different mechanisms. In pancreatic cancer, p16 is disrupted in the majority of cases, either by small intragenic mutations or by large genomic deletions that span the p16 locus. Tumors without p16 mutations frequently show hypermethylation of a CpG island that extends from the promoter into exon 1α, associated with silencing of transcription (9Merlo A. Hermann J.G. Mao L. Lee D.J. Gabrielson E. Burger P.C. Baylin S.B. Sidransky D. Nature Med. 1995; 1: 686-692Crossref PubMed Scopus (1852) Google Scholar) (see Figure 2). The frequent occurrence of gross genomic deletions spanning the p16 locus in primary tumors and in tumor-derived cell lines raised the possibility that additional genes might be disrupted. It is possible that these deletions simply reflect the presence of sequence-specific elements that favor such genomic rearrangements, but they may also be selected as an effective mechanism for inactivating two contiguous genes in a single genetic event. Adjacent to p16 lies the gene encoding one of its homologs, p15 (also known as INK4b), which demonstrates comparable binding and inhibition of CDK4 and CDK6 activity. p15 is frequently deleted along with p16 in tumors, but it has not been found to be a specific target for deletions or definitive point mutations. Nonetheless, the p15 promoter has been found to be silenced by hypermethylation in cases of leukemia and gliomas that do not also demonstrate hypermethylation of the p16 promoter, raising the possibility of an independent contribution in some human cancers. The striking frequency of p16 inactivation in tumor-derived cell lines suggests that inactivation of this protein plays an important role in the cellular immortalization events that accompany the establishment of growth in culture. Consistent with this model are the observations that cultured primary cells express increasing amounts of p16 as they approach the limit of their lifespan in vitro (4Hara E. Smith R. Parry D. Tahara H. Stone S. Peters G. Mol. Cell. Biol. 1996; 16: 859-867Crossref PubMed Scopus (634) Google Scholar) and that forced expression of activated Ha-ras in primary cells induces senescence that is relieved following inactivation of either p16 or p53 (15Serrano M. Lin A.W. McCurrach M.E. Beach D. Lowe S.W. Cell. 1997; 88: 593-602Abstract Full Text Full Text PDF PubMed Scopus (3776) Google Scholar). The apparent link between loss of p16 and cellular immortality and tumorigenesis was recently addressed by the targeted disruption of exon 2 of p16 in mice (14Serrano M. Lee H.-W. Chin L. Cordon-Cardo C. Beach D. DePinho R.A. Cell. 1996; 85: 27-37Abstract Full Text Full Text PDF PubMed Scopus (1378) Google Scholar). p16-null mice developed fibrosarcomas and lymphomas (70% developed tumors at a mean age of 29 weeks), which were enhanced by topical application of the carcinogen DMBA combined with ultraviolet irradiation (60% of homozygous null mice developed tumors with a comparable histology by 9 weeks and 17% of heterozygous mice developed less aggressive tumors at 15 weeks). Mouse embryo fibroblasts (MEFs) grew for multiple passages in culture without passing through a characteristic crisis and were readily transformed by Ha-ras alone, a property of cells that have already sustained an immortalizing event (14Serrano M. Lee H.-W. Chin L. Cordon-Cardo C. Beach D. DePinho R.A. Cell. 1996; 85: 27-37Abstract Full Text Full Text PDF PubMed Scopus (1378) Google Scholar). This view of p16 and its direct role in cellular immortalization and tumorigenesis is now challenged by the experiments of Kamijo and coworkers, who suggest that the knockout phenotype may be explained in large part by disruption not of p16, but of p19ARF, an alternatively spliced transcript whose product is completely distinct from p16 in its sequence and functional properties. Studies of the INK4a locus in both human and mouse cells revealed a novel transcript with 3′ sequence identical to p16 but with a unique 5′ end (1Duro D. Bernard O. Della Valle V. Berger R. Larsen C.-J. Oncogene. 1995; 11: 21-29PubMed Google Scholar, 8Mao L. Merlo A. Bedi G. Shapiro G.I. Edwards C.D. Rollins B.J. Sidransky D. Cancer Res. 1995; 55: 2995-2997PubMed Google Scholar, 11Quelle D.E. Zindy F. Ashmun R.A. Sherr C.J. Cell. 1995; 83: 993-1000Abstract Full Text PDF PubMed Scopus (1287) Google Scholar, 18Stone S. Jiang P. Dayananth P. Tavtigian S.V. Katcher H. Parry D. Peters G. Kamb A. Cancer Res. 1995; 55: 2988-2994PubMed Google Scholar). This transcript, called p19ARF (for alternative reading frame) is derived from a distinct first exon (exon 1β) that is 13–20 kb centromeric to the first exon of p16 (exon 1α). Exon 1β is spliced to exon 2, which is shared with p16, but this occurs in a different reading frame and hence p16 and p19ARF have no amino acid sequence similarity. The open reading frame of the p19ARF transcript is terminated within exon 2, with exon 3 comprising an untranslated 3′ exon (see Figure 2). p19ARF is predicted to encode a basic polypeptide that has no sequence homology to other known proteins, and shows ∼50% identity overall between human and mouse (compared with 65% amino acid identity for p16). The protein is nuclear, with localization to subnuclear dots or speckles. Expression of the p19ARF transcript is ubiquitous in postnatal tissues, in contrast to the more restricted expression pattern of p16, and it appears to be elevated in cells lacking p53. While p19ARF does not bind to CDKs or inhibit the activity of cyclin-CDK complexes, overexpression results in a cell cycle arrest, both in G1 and G2 (11Quelle D.E. Zindy F. Ashmun R.A. Sherr C.J. Cell. 1995; 83: 993-1000Abstract Full Text PDF PubMed Scopus (1287) Google Scholar). However, p19ARF does not appear to be specifically targeted by mutations in human cancer. Mutations have not been reported in the unique exon 1β, either in the germline of melanoma patients or in sporadic tumors. While mutations in the shared exon 2 may disrupt both p16 and p19ARF, no definitive mutations have been found in p19ARF that do not also disrupt p16. Chain-terminating mutations in p16 that encode a missense mutation in the p19ARF reading frame do not appear to alter the function of p19ARF, and in fact, exon 1β alone appears sufficient to mediate cell cycle arrest by p19ARF (12Quelle D.E. Cheng M. Ashmun R.A. Sherr C.J. Proc. Natl. Acad. Sci. USA. 1997; 94: 3436-3440Crossref PubMed Scopus (169) Google Scholar). Promoter hypermethylation has not been reported for p19ARF, which appears to be regulated by a promoter distinct from that directing expression of p16. It is with this background that Kamijo and coworkers specifically targeted exon 1β by homologous recombination and observed a p19ARF-null phenotype that appears to recapitulate much of that observed following disruption of exon 2, which is common to both p16 and p19ARF (14Serrano M. Lee H.-W. Chin L. Cordon-Cardo C. Beach D. DePinho R.A. Cell. 1996; 85: 27-37Abstract Full Text Full Text PDF PubMed Scopus (1378) Google Scholar). Thirty percent of mice developed spontaneous lymphomas and sarcomas by 24 weeks, and 80% of DMBA-treated mice developed tumors between 9 and 20 weeks. Embryonic fibroblasts derived from p19ARF-null mice failed to undergo crisis after multiple passages in vitro and were efficiently transformed by Ha-ras. Furthermore, MEFs derived from p19ARF heterozygotes lost the remaining wild-type allele as they passed through crisis, supporting the importance of homozygous loss of p19ARF in cellular immortalization. Taken together, these observations indicate that the phenotype resulting from disruption of both p16 and p19ARF is comparable to that produced by disruption of p19ARF alone. The targeting of adjacent genes, let alone different transcripts from the same gene, can be a tricky business. This is well illustrated by the variable phenotypes of mice with different deletions of the myogenic factor MRF4 gene, which are attributed to altered expression of the neighboring gene Myf5 (see10Olson E.N. Arnold H.-H. Rigby P.W.J. Wold B.J. Cell. 1996; 85: 1-4Abstract Full Text Full Text PDF PubMed Scopus (428) Google Scholar). However, Kamijo and coworkers demonstrate that expression of the p16 transcript is not lost following disruption of exon 1β (if anything, the normal induction of p16 expression following multiple passages in culture is enhanced, an observation attributed to the loss of competition for processing with the p19ARF transcript). The p16 transcript in p19ARF-null mice is wild-type, and the encoded protein demonstrates the expected binding to CDK4. Thus, barring any subtle deregulation of p16, the phenotype of mice lacking exon 1β appears to be attributable directly to loss of the p19ARF transcript. Conversely, mice lacking exon 2 would be predicted to have lost both p16 and p19ARF, although in vitro studies have suggested that, unlike exon 1β, exon 2 may not be critical to the function of p19ARF (12Quelle D.E. Cheng M. Ashmun R.A. Sherr C.J. Proc. Natl. Acad. Sci. USA. 1997; 94: 3436-3440Crossref PubMed Scopus (169) Google Scholar). Clearly, specific targeting of the unique p16 exon 1α will be required to determine definitively the relative contributions of these two transcripts to the knockout phenotype. How can the consequences of inactivating p19ARF in the mouse, namely tumorigenesis and cellular immortalization, be reconciled with the compelling evidence derived from human cancers which points to p16, rather than p19ARF, as the critical target? There is certainly ample precedent for striking differences between mouse and human phenotypes, particularly with respect to cancer predisposition. However, given the functional consequences of disrupting p19ARF in the mouse, it would not be unreasonable to expect that its loss may contribute to the transformed phenotype in the many tumors with genomic deletions that span the p16-p19ARF locus (an argument that might be extended to p15 as well). While p16 may be the essential target for malignant transformation in a subset of human cancers, the loss of these adjacent genes may lead to additional phenotypes that may be better understood once their specific functional pathways have been defined. The pendulum now swings back to defining the specific role of p16 in the mouse. It is possible that disruption of exon 1α will have no phenotype, which would attribute cellular immortalization and tumor development uniquely to the loss of p19ARF, and leave us scrambling to explain yet another difference between mouse and man. Alternatively, specific disruption of p16 may lead to a phenotype similar to that reported following targeting of exon 2, since that exon is critical for p16 function and potentially dispensable for the mediation of cell cycle arrest by p19ARF (12Quelle D.E. Cheng M. Ashmun R.A. Sherr C.J. Proc. Natl. Acad. Sci. USA. 1997; 94: 3436-3440Crossref PubMed Scopus (169) Google Scholar). The possibility that both p16 and p19ARF might play independent roles in the regulation of cell cycle progression and cellular senescence is particularly intriguing. Cell cycle arrest by p19ARF appears to be abolished in cells lacking p53, suggesting that it somehow acts upstream of p53 (7Kamijo T. Zindy F. Roussel M.F. Quelle D.E. Downing J.R. Ashmun R.A. Grosveld G. Sherr C.J. Cell, this issue. 1997; 91: 649-659Scopus (1359) Google Scholar). The exact mechanism by which p19ARF interacts with p53 is uncertain. Unlike the exclusive inactivation of p16 and RB, the observation that some tumors arising in p19ARF-null mice have mutations in p53 indicates that inactivation of these two genes is not functionally equivalent. Nonetheless, like RB the p53 pathway has been implicated in cell cycle progression and senescence. And here lies the most startling implication to emerge from the work of Kamijo and colleagues, the possibility that a single gene “p16-p19ARF” may reside at a crossroad of regulation for both RB and p53, through the use, unprecedented in mammalian cells, of overlapping distinct reading frames.