Review
INK4a/ARF: A multifunctional tumor suppressor locus

https://doi.org/10.1016/j.mrfmmm.2004.08.021Get rights and content

Abstract

The INK4a/ARF locus encodes two physically linked tumor suppressor proteins, p16INK4a and ARF, which regulate the RB and p53 pathways, respectively. The unusual genomic relationship of the open reading frames of these proteins initially fueled speculation that only one of the two was the true tumor suppressor, and loss of the other merely coincidental in cancer. Recent human and mouse genetic data, however, have firmly established that both proteins possess significant in vivo tumor suppressor activity, although there appear to be species- and cell-type specific differences between the two. For example, ARF plays a clear role in preventing Myc-induced lymphomagenesis in mice, whereas the role for p16INK4a is human carcinomas is more firmly established. In this review, I discuss the evolutionary history of the locus, the relative importance of these tumor suppressor genes in human cancer, and recent information suggesting novel biochemical and physiologic functions of these proteins in vivo.

Introduction

From the cancer geneticist's perspective, the INK4a/ARF locus would seem to be a design flaw. Cancer is a disease characterized by the genome-wide loss and gain of genetic material, and startling evidence from many systems has clearly demonstrated that somatic cells incur oncogenic mutations occur throughout the lifespan of mammals (reviewed in [1]). The persistent repression of these would-be tumors is required for long-term survival, yet the INK4a/ARF locus encodes two potent and distinct tumor suppressor genes juxtaposed within a single 3 × 104 bp segment of the 3 × 109 bp genome; an arrangement exceedingly vulnerable to a single, small genetic deletion. Because of this structure, many of us working in the field initially assumed that only one of the proteins represented the true ‘tumor suppressor’ while the other was merely co-deleted in tumorigenesis. Recent evidence, however, has lead to the unanticipated conclusion that both proteins are non-redundant tumor suppressors, and their dual inactivation in cancer is in fact a frequent path to malignancy. This observation suggests that either indeed the locus is a design flaw (that is, perhaps was not subject to selective pressures because cancer is of limited relevance in evolutionary terms), or that the structure of the locus encompasses functional benefits that are not readily obvious from the cancer geneticist's perspective. In this review, I will argue that the latter is the case, based on evolutionary, genetic and biochemical evidence.

The INK4a/ARF locus, situated on chromosome 9p21 in the human, is among the most frequent sites of genetic loss in human cancer. Deletion of this locus is seen with high frequency in a variety of malignancies including glioblastoma, melanoma, pancreatic adenocarcinoma, non-small cell lung cancer, bladder carcinoma and oropharyngeal cancer (Table 1, reviewed in [2], [3], [4]). Initially, two genes were identified as promising tumor suppressor candidates within the same 40 kbp stretch of genomic DNA at the locus, INK4b and INK4a [5]. These related genes encode proteins with a high degree of conservation at the amino acid level and appear to have resulted from a gene duplication event. At the time of their localization to the 9p21 region, the INK4-class of cell cycle inhibitors were already known to abrogate the kinase activity of cdk4 and 6 [6]; which in turn phosphorylate Rb to produce S-phase entry. Therefore, the finding of INK4a/b at a commonly deleted locus associated with cancer was promising, since Rb had been previously identified as a potent tumor suppressor gene. Additional refinement came from the finding that familial melanoma in several kindreds segregated with missense mutations in p16INK4a [7], [8], [9], [10]. Numerous subsequent analyses of sporadic primary tumors and derivative cell lines of many types have demonstrated p16INK4a inactivation by deletion, point mutation and promoter methylation (covered further below). These data initially pointed to p16INK4a as the principal tumor suppressor at 9p21.

This conclusion, however, was seriously questioned by the discovery of workers in the Sherr lab that an alternative first exon could be transcribed at the locus, producing alternate reading frame (ARF, also called p14ARF in the human and p19ARF in the mouse) [11]. This open reading frame has its own promoter and first exon (exon 1β, as opposed to exon 1α of p16INK4a), but splices into the common second and third exons shared with p16INK4a (Fig. 1). As the name implies, the proteins are encoded in alternative reading frames, and therefore are not isoforms and have no amino acid homology. While the encoding of open reading frames in overlapping stretches of the genome is common in viruses and bacteria, such genomic structure is practically unique in the mammalian genome. Evidence that ARF was also a tumor suppressor came from the observation that the protein was a potent cell cycle inhibitor, and that mice lacking exon 1β of ARF were highly prone to spontaneous and carcinogen-induced tumors, appearing to phenocopy the previously generated exons 2 and 3 deficient (Ink4a/Arf-null [12], that is lacking p16INK4a and p19ARF) mice [11], [13]. Therefore, the debate as to the relative importance of these proteins was set given the compelling in vitro and knock-out data from the mouse suggesting ARF was the tumor suppressor at 9p21.

The tumor suppressor proteins of the INK4a/ARF locus function in distinct anti-cancer pathways: p16INK4a regulates RB and ARF regulates p53 (Fig. 1). Inactivation of the p53 and RB pathways through a large variety of mechanisms occurs in the majority, if not all, of human cancers. Ample human and murine genetic data indicate that the pathways are non-overlapping, and that their concomitant inactivation is cooperative in terms of tumor predisposition. Although both the RB- and p53-pathways play important roles in differentiation, development and DNA repair; the members of the INK4a/ARF locus largely respond to aberrant growth or oncogenic stress. Therefore, the INK4a/ARF locus appears to function as a dual-pronged brake to malignant growth, which engages two potent anti-proliferative pathways represented by RB and p53. Both p16INK4a and ARF have been closely linked to the important cell biological phenotype of senescence, which is thought to serve as an important in vivo barrier to cancer, as well as a potential contributor to aging [14], [15].

In mammals, the INK4-class of cell cycle inhibitors consists of p15INK4b, p16INK4a, p18INK4c and p19INK4d (not to be confused with p19ARF). These proteins consist of four or more ankyrin repeats that are highly conserved. Evolutionarily, it appears that p15INK4b and p16INK4a share a common ancestor, as do p18INK4c and p19INK4d (Fig. 2). Fugu, the Japanese puffer fish, contains a single p15INK4b/p16INK4a-like member, and a single p18INK4c/p19INK4d-like member [16]. In mammals, p18INK4c and p19INK4d are highly expressed during development [17], while p15INK4b and p16INK4a are associated with tumor suppression. Mice lacking p18INK4c are clearly tumor prone and sensitive to carcinogens [18], [19], but a role for p18INK4c loss in human cancer has not been clearly established. All four INK4 members are well-conserved at the amino acid level, particularly in the second and third ankyrin binding domains, and little biochemical distinction has been made among members of this family. It appears that the functional differences ascribed to the various INK4 members relate to different patterns of expression in response to different genetic and environmental stimuli.

The INK4 proteins bind to cdk4 and cdk6 and inhibit their kinase activity for Rb, and perhaps other cdk targets. As opposed to the CIP/KIP inhibitors (e.g. p21 and p27) of cyclin dependent kinases, INK4 binding prevents the interaction of the cyclin dependent kinases with the D-type cyclins, which are required for catalytic activity [20]. The expression of p16INK4a or other INK4 members, therefore, produces decreased cdk4/6 kinase activity and Rb hypophosphorylation, which in turn leads to E2F repression and growth arrest. Although cdk4/6 may have targets other than Rb, genetic evidence suggests that RB is the major target for the tumor suppressor activity of p16INK4a. For example, in non-small cell lung cancer, both Rb loss and p16INK4a inactivation are frequently detected, but the lesions are generally mutually exclusive; suggesting that p16INK4a expression is unable to block tumorigenesis in Rb-deficient lung cancers [21], [22], [23]. These data have been supported by in vitro findings that p16INK4a cannot efficiently arrest Rb deficient cell lines [24]. Although Rb-independent functions for p16INK4a have been proposed – for example, as an inducer of apoptosis [25], [26], [27], [28], [29] – the in vivo significance of these effects are not presently known.

The initial clue to ARF's function came from the observation that its expression induced a cell-cycle arrest even in cells with enforced cyclin D expression [11]. Furthermore, it was noted that p19ARF loss obviated the requirement for p53 inactivation to immortalize murine embryo fibroblasts (MEFs) in vitro [13], and in tumors in vivo [30], suggesting ARF and p53 reside in a common genetic pathway. This view was refined by the observation that p19ARF inhibited the transformation of MEFs by Mdm2, but p19ARF had no effect on transformation in cells lacking p53 [31]; suggesting ARF might play a role in the regulation of the MDM2-mediated degradation of p53. These genetic results were strengthened biochemically by the demonstration that ARF binds to MDM2 and inhibits the ubiquitination of p53, thereby stabilizing this tumor suppressor protein [31], [32], [33], [34].

The exact mechanism whereby ARF stabilizes p53 is still a matter of controversy. By immunofluorescence, ARF localizes to the nucleolus, and appears to sequester MDM2 to that compartment [35], [36], [37]. The portions of the protein, however, that are required for nucleolar sequestration are poorly conserved across species, and differ between humans and mice (Fig. 3). Other groups, however, have suggested that ARF need not relocate MDM2 to the nucleolus for proper function; but rather only must inhibit the E3-ligase activity of MDM2 to stabilize p53 [38], [39]. There appears unanimity, however, that the conserved N-terminal 25 amino acid portion of human ARF is able to perform most, if not all, of the cell biological effects attributed to full-length ARF. This fact is further emphasized by the genetic finding that in the chicken, exon 1β splices to a different reading frame from that of mammals (and also different from p16INK4a), and there is no conservation outside of exon 1β between the chicken and mammals (Fig. 3) [40]. Therefore, the biochemical and genetic evidence agree that the in vivo function of ARF is entirely encoded by the exon 1β.

The regulation of ARF in vivo has been crucial to our understanding of the INK4a/ARF locus function. ARF was initially found to be a mediator of p53-dependent apoptosis in the lens in response to the cell cycle dysregulation associated with Rb deficiency [31], suggesting ARF linked oncogenic stress to p53. Several additional forms of oncogenic stress including expression of v-abl [41], c-myc [42], E1a [43], E2F [44], [45] or loss of RB [46] have also been demonstrated to induce ARF and thereby stabilize p53. Of interest, this response to oncogene expression depends on cellular context, as RAS potently induces ARF in murine, but not human, cells [47], [48], [49]. In murine embryo fibroblasts, the expression of p19ARF correlates with the onset of senescence, and cells lacking p19ARF do not senesce in culture [13], [17]. In contrast, p14ARF does not appear to play a major role in the replicative senescence of normal human cells [50], [51], [52], a phenotype principally mediated by p16INK4a and/or p53.

In addition to its well-defined role in stabilizing p53 by inhibiting Mdm2, several novel functions for ARF have been suggested. First, when overexpressed, ARF can induce a growth arrest in p53−/− cells, suggesting it has p53-independent functions [53], [54]. Additionally, ARF contributes to vascular regression in development of the eye, and this effect is p53-independent [55]. Interactions with E2F-1 [56], [57], [58], spinophilin [59], topoisomerase I [60], MdmX [61], HIF-1α [62], PXF/Pex19p [63] and nucleophosmin (NPM)/B23 [64], [65] have been reported. A caveat to these studies, however, is that the highly basic ARF may non-specifically interact with other proteins when overexpressed, and therefore in vitro interactions with ARF require in vivo confirmation under settings of physiologic expression. Of these putative interactions, the connection of ARF to ribosome biogenesis through an interaction with NPM is of particular interest. Two groups have demonstrated that ARF expression decreases the processing of ribosomal RNA, thereby limiting cell growth and inducing a cell cycle arrest [65], [66]. Although these biochemical observations are intriguing, however, it is noteworthy that null zygosity of p19ARF does not shorten the tumor latency of p53−/− animals [53], [67], whereas loss of p16INK4a does [68]. Therefore, the major tumor suppressor activity of ARF in these murine models depends on intact p53 function, but it is possible that subtle p53-independent effects of p19ARF on tumorigenesis are obscured by the overshadowing effect of germline p53 deficiency in these assays.

Section snippets

What represents the tumor suppressor activity at the INK4a/ARF locus?

As stated, the issue of what protein represents the 9p21 tumor suppressor has been the subject of some controversy, as initially the human (favoring p16INK4a) and murine (favoring ARF) genetic data appeared in conflict. Subsequent analyses have reconciled this seeming discrepancy as work by several groups has established that p14ARF can repress cancer in humans, while p16INK4a similarly does so in mice.

Other candidates?

Although data from the analyses of tumors from human and other species overwhelmingly demonstrate that both p16INK4a and ARF are tumor suppressor proteins, it is important to consider the possibility that there are still other tumor suppressor genes at 9p21. Certainly, the most obvious candidate is p15INK4b, located roughly 10 kbp centromeric to exon 1β of ARF. Co-deletion of p15INK4b with the INK4a/ARF locus is very common in human cancer, and therefore its inactivation is seen with high

Design flaw?

The aforementioned studies resolve the debate as to which protein encoded by the INK4a/ARF locus encodes tumor suppressor activity: both. At first glance, this design appears to make little sense in terms of cancer genetics, and suggests the question of whether evolution could really be so careless. Additionally, the locus structure cannot be attributed to functional redundancy, or that the possibility that p16INK4a and ARF function as tumor suppressors in different tissues, as there is ample

Conclusions

Although 9p21 may yet encode other tumor suppressor proteins, the genetic and biochemical data from humans, rodents and other species suggest that the INK4a/ARF locus encodes two potent tumor suppressor proteins, p16INK4a and ARF, which function in non-overlapping pathways to abrogate tumorigenesis. That this structure permits the attenuation of two tumor suppressor mechanisms from a common mutagenic event explains the high incidence of 9p21 deletion in human cancers. The evolutionary history

Acknowledgements

I would like to thank N. Bardeesy, M. Greenblatt and M. Ramsey for advice and ideas regarding the manuscript. This work was funded by the Sidney Kimmel Foundation for Cancer Research and through a Paul Beeson Scholar in Aging Research award, and by the NIH.

References (139)

  • Y. Zhang et al.

    Mutations in human ARF exon 2 disrupt its nucleolar localization and impair its ability to block nuclear export of MDM2 and p53

    Mol. Cell

    (1999)
  • M. Vivo et al.

    The human tumor suppressor arf interacts with spinophilin/neurabin II, a type 1 protein-phosphatase-binding protein

    J. Biol. Chem.

    (2001)
  • M.W. Jackson et al.

    MdmX binding to ARF affects Mdm2 protein stability and p53 transactivation

    J. Biol. Chem.

    (2001)
  • K. Fatyol et al.

    The p14ARF tumor suppressor protein facilitates nucleolar sequestration of hypoxia-inducible factor-1 alpha (HIF-1 alpha) and inhibits HIF-1-mediated transcription

    J. Biol. Chem.

    (2001)
  • T. Sugihara et al.

    Pex19p dampens the p19ARF-p53-p21WAF1 tumor suppressor pathway

    J. Biol. Chem.

    (2001)
  • K. Itahana et al.

    Tumor suppressor ARF degrades B23, a nucleolar protein involved in ribosome biogenesis and cell proliferation

    Mol. Cell

    (2003)
  • M. Sugimoto et al.

    Nucleolar Arf tumor suppressor inhibits ribosomal RNA processing

    Mol. Cell

    (2003)
  • R.M. Bachoo et al.

    Epidermal growth factor receptor and Ink4a/Arf. Convergent mechanisms governing terminal differentiation and transformation along the neural stem cell to astrocyte axis

    Cancer Cell

    (2002)
  • R. Meuwissen et al.

    Induction of small cell lung cancer by somatic inactivation of both Trp53 and Rb1 in a conditional mouse model

    Cancer Cell

    (2003)
  • N.E. Sharpless

    The persistence of senescence

    Sci. Aging Knowl. Environ.

    (2003)
  • A. Kamb et al.

    A cell cycle regulator potentially involved in genesis of many tumor types

    Science

    (1994)
  • M. Serrano et al.

    A new regulatory motif in cell-cycle control causing specific inhibition of cyclin D/CDK4

    Nature

    (1993)
  • L. Zuo et al.

    Germline mutations in the p16INK4a binding domain of CDK4 in familial melanoma

    Nat. Genet.

    (1996)
  • L. Liu et al.

    Germline p16INK4A mutation and protein dysfunction in a family with inherited melanoma

    Oncogene

    (1995)
  • E.A. Holland et al.

    Analysis of the p16 gene, CDKN2, in 17 Australian melanoma kindreds

    Oncogene

    (1995)
  • C.J. Hussussian et al.

    Germline p16 mutations in familial melanoma

    Nat. Genet.

    (1994)
  • D.E. Quelle et al.

    Alternative reading frames of the INK4a tumor suppressor gene encode two unrelated proteins capable of inducing cell cycle arrest

    Cell

    (1995)
  • N.E. Sharpless et al.

    Telomeres, stem cells, senescence, and cancer

    J. Clin. Invest.

    (2004)
  • J. Gilley et al.

    One INK4 gene and no ARF at the Fugu equivalent of the human INK4A/ARF/INK4B tumour suppressor locus

    Oncogene

    (2001)
  • F. Zindy et al.

    Expression of the p16INK4a tumor suppressor versus other INK4 family members during mouse development and aging

    Oncogene

    (1997)
  • D.S. Franklin et al.

    Functional collaboration between different cyclin-dependent kinase inhibitors suppresses tumor growth with distinct tissue specificity

    Mol. Cell Biol.

    (2000)
  • A.A. Russo et al.

    Structural basis for inhibition of the cyclin-dependent kinase Cdk6 by the tumour suppressor p16INK4a

    Nature

    (1998)
  • H. Tanaka et al.

    Disruption of the RB pathway and cell-proliferative activity in non-small-cell lung cancers

    Int. J. Cancer

    (1998)
  • G.I. Shapiro et al.

    Reciprocal Rb inactivation and p16INK4 expression in primary lung cancers and cell lines

    Cancer Res.

    (1995)
  • G.A. Otterson et al.

    Absence of p16INK4 protein is restricted to the subset of lung cancer lines that retains wildtype RB

    Oncogene

    (1994)
  • J. Lukas et al.

    Retinoblastoma-protein-dependent cell-cycle inhibition by the tumour suppressor p16

    Nature

    (1995)
  • C. Lagresle et al.

    Transgenic expression of the p16(INK4a) cyclin-dependent kinase inhibitor leads to enhanced apoptosis and differentiation arrest of CD4-CD8-immature thymocytes

    J. Immunol.

    (2002)
  • Z. Sachs et al.

    p16Ink4a interferes with abelson virus transformation by enhancing apoptosis

    J. Virol.

    (2004)
  • T. Plath et al.

    A novel function for the tumor suppressor p16(INK4a): induction of anoikis via upregulation of the alpha(5)beta(1) fibronectin receptor

    J. Cell Biol.

    (2000)
  • I. Tamm et al.

    Adenovirus-mediated gene transfer of P16INK4/CDKN2 into bax-negative colon cancer cells induces apoptosis and tumor regression in vivo

    Cancer Gene Ther.

    (2002)
  • L. Chin et al.

    Cooperative effects of INK4a and ras in melanoma susceptibility in vivo

    Genes Dev.

    (1997)
  • T. Kamijo et al.

    Functional and physical interactions of the ARF tumor suppressor with p53 and Mdm2

    Proc. Natl. Acad. Sci. U.S.A.

    (1998)
  • F.J. Stott et al.

    The alternative product from the human CDKN2A locus, p14(ARF), participates in a regulatory feedback loop with p53 and MDM2

    EMBO J.

    (1998)
  • J.D. Weber et al.

    Nucleolar Arf sequesters Mdm2 and activates p53

    Nat. Cell Biol.

    (1999)
  • J.D. Weber et al.

    Cooperative signals governing ARF-mdm2 interaction and nucleolar localization of the complex

    Mol. Cell Biol.

    (2000)
  • C. Korgaonkar et al.

    ARF function does not require p53 stabilization or Mdm2 relocalization

    Mol. Cell Biol.

    (2002)
  • S. Llanos et al.

    Stabilization of p53 by p14ARF without relocation of Mdm2 to the nucleolus

    Nat. Cell Biol.

    (2001)
  • S.H. Kim et al.

    Absence of p16INK4a and truncation of ARF tumor suppressors in chickens

    Proc. Natl. Acad. Sci. U.S.A.

    (2003)
  • A. Radfar et al.

    p19(Arf) induces p53-dependent apoptosis during abelson virus-mediated pre-B cell transformation

    Proc. Natl. Acad. Sci. U.S.A.

    (1998)
  • F. Zindy et al.

    Myc signaling via the ARF tumor suppressor regulates p53-dependent apoptosis and immortalization

    Genes Dev.

    (1998)
  • Cited by (0)

    View full text