Oncogene (2002) 21, 3207 ± 3212 ã 2002 Nature Publishing Group All rights reserved 0950 ± 9232/02 $25.00 www.nature.com/onc

Human p14ARF-mediated cell cycle arrest strictly depends on intact p53 signaling pathways

H Oliver Weber1,2, Temesgen Samuel1,3, Pia Rauch1 and Jens Oliver Funk*,1

1Laboratory of Molecular Tumor Biology, Department of Dermatology, University of Erlangen-Nuremberg, 91052 Erlangen, Germany; 2Regulation of Cell Growth Laboratory, National Cancer Institute, Frederick, Maryland, MD 21702-1201, USA

The tumor suppressor ARF is transcribed from the INK4a/ 4 and 6 activities, thus leading to decreased phosphor- ARF locus in partly overlapping reading frames with the ylation of RB and to G1 arrest. Cells that are de®cient CDK inhibitor p16Ink4a. ARF is able to antagonize the for RB are resistant to p16Ink4a-mediated cell cycle MDM2-mediated ubiquitination and degradation of p53, arrest (Sherr and Roberts, 1995; Weinberg, 1995). leading to either cell cycle arrest or apoptosis, depending on ARF is also a cell cyle inhibitor. It does not directly the cellular context. However, recent data point to inhibit CDKs but interferes with the function of additional p53-independent functions of mouse p19ARF. MDM2 to destabilize p53. ARF may be activated by Little is known about the dependency of human p14ARF aberrant activation of oncoproteins such as Ras function on p53 and its downstream genes. Therefore, we (Palmero et al., 1998), c-myc (Zindy et al., 1998), analysed the mechanism of p14ARF-induced cell cycle arrest E1A (de Stanchina et al., 1998), Abl (Radfar et al., in several human cell types. Wild-type HCT116 colon 1998) and E2F-1 (Bates et al., 1998). In these scenarios, carcinoma cells (p53+/+p21CIP1+/+ 14-3-3s+/+), but not ARF is believed to counteract the oncogenic hyper- 7/7 p53 counterparts, underwent G1 and G2 cell cycle arrest proliferative e€ect, predominantly by induction of both ARF CIP17/7 following infection with a p14 -adenovirus. In p21 G1 and G2 arrest (Quelle et al., 1995). ARF 7/ ARF cells, p14 did not induce G1 or G2 arrest, while 14-3-3s Human p14 is only 49% homologous to murine 7 ARF counterparts were mainly arrested in G1, pointing to p19 and 5 kDa smaller. Despite such high size and CIP1 ARF ARF essential roles of p21 in G1 and G2 arrest and cooperative sequence divergence, p19 and p14 share similar ARF ARF roles of p21 and 14-3-3s in ARF-mediated G2 arrest. Our properties. Both p19 and p14 can directly data demonstrate a strict p53 and p21CIP1 dependency of interact with MDM2 (Pomerantz et al., 1998; Zhang p14ARF-induced cell cycle arrest in human cells. et al., 1998). MDM2 binds to p53 leading to the Oncogene (2002) 21, 3207 ± 3212. DOI: 10.1038/sj/ ubiquitin-mediated degradation of p53. Since p53 onc/1205429 transactivates MDM2, this constitutes a negative feedback loop between MDM2 and p53 (Sherr, Keywords: INK4a/ARF; p53; p21; 14-3-3s; DNA 1998). The interaction of ARF with MDM2 prevents tumor virus the MDM2-mediated p53 degradation and subse- quently leads to the stabilization and activation of p53, which then induces either cell cycle arrest or Introduction apoptosis, depending on the cellular context. The INK4a/ARF gene locus is one of the most The INK4a/ARF gene locus encodes for two cell cycle frequently targeted genes in human carcinogenesis. regulatory proteins, p16Ink4a and ARF (alternative p16Ink4a is a bona ®de tumor suppressor, showing reading frame; human p14ARF, murine p19ARF), in frequent deletions, point mutations and hypermethyla- di€erent, overlapping reading frames. The two tran- tions exclusively targeting p16Ink4a in many human scripts have separate ®rst exons but share exons 2 and cancers (Chin et al., 1998a; Sharpless and DePinho, 3. Exon 2 is transcribed in distinct readings frames 1999). In contrast, the role of ARF in human such that there is no amino acid homology between the carcinogenesis is less clear. Many deletions target not two proteins (Chin et al., 1998b; Sharpless and only p16Ink4a, but also p14ARF. However, there are few DePinho, 1999). point mutations or deletions known so far to p16Ink4a belongs to the INK4 family of cyclin- exclusively target p14ARF (Chin et al., 1998b), and dependent kinase (CDK) inhibitors and inhibits CDK most point mutations in exon 2 render p16INK4A and not p14ARF unfunctional in in vitro assays (Quelle et al., 1997). On the other hand, the role of ARF as a tumor suppressor has been supported by a mouse *Correspondence; JO Funk; model in which exon 1b has exclusively been disrupted E-mail: [email protected] (Kamijo et al., 1997). The resulting p19ARF null mice 3Current address: The Burnham Institute, La Jolla, CA 92037, USA Received 11 December 2001; revised 23 January 2002; accepted 19 developed lymphomas and sarcomas at an early age, a February 2002 phenotype very similar to that of INK4a/ARF exon 2 ARF-mediated cell cycle arrest HO Weber et al 3208 knockout mice, which lack both p19ARF and p16Ink4a (Serrano et al., 1996). It is believed that the function of murine and human ARF to induce arrest mainly depends on the stabilization and activation of p53. However, the analysis of p19ARF in mouse embryo ®broblasts (MEFs) has pointed to additional p53-independent mechanisms of p19ARF function (Carnero et al., 2000; Weber et al., 2000). Furthermore, little is known about the downstream e€ectors once p53 is activated by ARF. In general, it is believed that the p53 function to arrest the cell cycle, i.e. following DNA damage, depends primarily on the transactivation of speci®c p53-responsive genes, such as p21CIP1, 14-3-3s,and others. (Levine, 1997; Vogelstein et al., 2000; Vousden, 2000). However, mouse p19ARF may inhibit the cell cycle through both p21CIP1-dependent and p21CIP1 - independent pathways (Modestou et al., 2001).

Results

We set out to analyse the dependency of p14ARF- induced cell cycle arrest on p53 and the p53 target CIP1 ARF genes, p21 and 14-3-3s, in the human colorectal Figure 1 p14 -induced G1 and G2 arrest depends on p53. An +/+ adenovirus encoding for p14ARF (ad-ARF) was used to express cancer cell line, HCT116. This cell line (p53 ARF 7/7 CIP1+/+ +/+ p14 in wild-type and p53 HCT116 cells. As a negative p21 14-3-3s ) and various isogenic knockout control an adenovirus encoding for GFP (ad-GFP) was used. cell lines, derived from the parental HCT116 cells by Treatment with adriamycin (ADR) to induce DNA damage homologous recombination (Waldman et al., 1995), served as a control for the activation of p53. At 24 h post- have been successfully used in a number of cell cycle treatment either BrdU-labeled cells were harvested for ¯ow cytometry analysis and whole-cell extracts were made for analysis studies (Bunz et al., 1998; Chan et al., 1999; McShea et of the relevant cell cycle regulators by immunoblotting or al., 2000; Samuel et al., 2001; Waldman et al., 1996, determination of CDK activities 1997). Their phenotype, mostly in the context of a DNA damage response, has been extensively docu- mented. In HCT cells, the endogenous p14ARF gene is silenced by mutation of one and methylation of the of CDK2 activity and the subsequent hypophosphor- other allele, respectively (Burri et al., 2001). Therefore, ylation of RB, indicative of G1 arrest. The CDK2 the introduction of p14ARF into these cells substitutes inhibition was mediated by p21CIP1, since p21CIP1 was missing p14ARF expression instead of increasing already detected in anti-CDK2 immunoprecipitates both in the existing p14ARF expression. ad-ARF-infected and ADR-treated cells (data not The adenoviral pAdEasy system (He et al., 1998) shown). Consequently, the percentage of S phase cells, was used to express p14ARF in the cells. The newly as evaluated by BrdU incorporation and ¯ow generated adenovirus (ad-ARF) encoded for p14ARF cytometry analysis, was decreased (Figure 1). In and the GFP gene under the control of two separated comparison, ADR-treated HCT cells showed a pro- cytomegalovirus promotors. Therefore, the infected found p53 induction and subsequent induction of CIP1 cells could be traced throughout the experiment. As a p21 and 14-3-3s with concomitant G1 and G2 cell control, an adenovirus encoding for GFP only (ad- cycle arrest. GFP) was used. GFP ¯uorescence indicated that 100% CDK1/Cyclin B activity, the key regulatory CDK of the target cells were infected (data not shown). To complex governing G2/M, was also markedly reduced compare the activation of p53 in response to ARF with both in ad-ARF- and ADR-treated wild-type cells, its activation by DNA damage, all cells were treated indicative of G2 arrest. CDK1 activity is primarily with adriamycin (ADR). We analysed the onset of regulated by activating and inhibitory phosphoryla- ARF-mediated e€ects in a time frame of up to 48 h tions. Therefore, the phosphorylation status of CDK1 following infection. was analysed by immunoblotting. Interestingly, CDK1 Following infection with ad-ARF, the HCT cells showed a hyperphosphorylated pattern, indicative of showed a high level of ARF expression as early as 12 h inhibitory tyrosine 14/threonine 15 phosphorylation post-infection, which was stable for at least another (Herzinger et al., 1995), only in ADR-treated cells but 36 h (data not shown). In the wild-type HCT cells, ad- not in ad-ARF-infected cells (Figure 1). ARF infection led to the stabilization of p53 and The current model of ARF function involves the induction of the p53 targets MDM2, p21CIP1 and 14-3- direct association of ARF and MDM2 leading to the 3s (Figure 1). This coincided with the downregulation inhibition of MDM2-mediated ubiquitination of p53

Oncogene ARF-mediated cell cycle arrest HO Weber et al 3209 (Kamijo et al., 1998; Pomerantz et al., 1998; Sherr, associated kinase activity was only marginally altered, 1998; Stott et al., 1998; Zhang et al., 1998); however, suggesting that 14-3-3s is dispensable for p14ARF- the precise requirement of MDM2 relocation for p53 induced G1 but important for G2 arrest. Importantly, stabilization is under investigation (Korgaonkar et al., p21CIP17/7 14-3-3s7/7 cells showed identical resistance 2002; Llanos et al., 2001). To verify the functionality of to the ad-ARF infection as the p537/7 cells (Figure 2). ARF in HCT cells, coimmuniprecipitation experiments All cell types in this scenario did not undergo apoptosis were performed con®rming the binding of ARF to during the time frame of the analysis. MDM2 (data not shown). To con®rm these results using another expression Next, the p537/7 cells were analysed in an identical system, these experiments were also performed with a manner. Ad-ARF infection of these cells did not lead high-titer, pantropic retrovirus encoding for p14ARF. to induction of MDM2, p21CIP1, or 14-3-3s. Impor- Compared to the ad-ARF infection experiments, tantly, there was no indication of cell cycle arrest as identical results were achieved by infecting the cells judged by BrdU incorporation, RB phosphorylation, with the p14ARF retrovirus (data not shown). and CDK2 as well as cyclin B-associated kinase p14ARF e€ects were then analysed in the human activities (Figure 1). embryo kidney cell line 293. 293 cells have been Since p21CIP1 and 14-3-3s are key mediators of p53- transformed by adenovirus E1A and E1B oncogenes, induced G1 and G2 arrest, respectively, we then which has led to the inactivation of p53 and RB. To analysed single p21CIP17/7 and 14-3-3s7/7 cells as well analyse p14ARF-mediated e€ects in this cell line we as cells with combined de®ciency for p21CIP1 and 14-3- established an inducible system based on an ecdysone- 3s. Ad-ARF infection of p21CIP17/7 cells led to p53 inducible receptor (No et al., 1996) This receptor is induction with concomitant increase of MDM2 and 14- only activated in the presence of the hormone ecdysone 3-3s proteins (Figure 2), comparable to wild-type HCT or the ecdysone homologue ponasterone A leading to cells. Strinkingly, no signi®cant change in BrdU transactivation of the inserted gene. incorporation, RB phosphorylation, and CDK2 as Following induction by ecdysone, p14ARF protein well as cyclin B-associated kinase activities was level rised as early as 9 h and persisted for at least 48 h observed, suggesting a central role of p21CIP1 in (Figure 3a). Immuno¯uorescence analysis showed that ARF mediating G1 and G2 arrest following p14 -infection. 100% of the ecdysone-treated cells were induced to In 14-3-3s7/7 cells, p14ARF infection led to decreased express p14ARF, located to the nucleoli, whereas CDK2 activity and hypophosphorylation of RB untreated cells showed low p14ARF expression. No cell coinciding with p21 induction. However, cyclin B- cycle arrest in the ecdysone-treated 293 cells was observed as measured by BrdU incorporation (Figure 3a). No di€erence in the expression pattern of the relevant cell cycle regulators was seen. Importantly, CDK2 activity as well as cyclin B-associated kinase activity remained unchanged (Figure 3a).

Figure 3 Lack of cell cycle arrest in 293 cells upon induced expression of p14ARF.(a) An ecdysone-inducible system encoding for p14ARF was established in 293 cells (ARF293). Following ARF Figure 2 p14 -induced G1 and G2 arrest depends on the p53- treatment with 5 mM of the ecdysone homologue ponasterone A responsive genes p21 and 14-3-3s, respectively. An adenovirus the ARF293 cells were analysed at the respective time points as encoding for p14ARF (ad-ARF) was used to express p14ARF in the cells in Figure 1. (b) To verify the binding of p14ARF to p21CIP17/7, 14-3-3s7/7, and p21CIP17/7 14-3-3s7/7 HCT cells. MDM2 in the ARF293 cells, a coimmunoprecipitation analysis Identical analyses were performed as in Figure 1 was performed

Oncogene ARF-mediated cell cycle arrest HO Weber et al 3210 Furthermore, the association of p14ARF and MDM2 In light of these partially contradictory data, we was analysed under these conditions as an indication of aimed to investigate the dependency of human p14ARF ARF functionality. Immunoprecipitation experiments on p53. The p537/7 cells showed a complete defect in ARF with anti-ARF and anti-MDM2 antibodies con®rmed their G1 and G2 arrest response to p14 -infection. the binding of p14ARF to MDM2 and vice versa (Figure Moreover, p21CIP17/7 cells were resistant to p14ARF- 3b). Note that anti-ARF immunoprecipitates contained induced G1 and G2 arrest. Intriguingly, this further MDM2 protein already in the absence of ecdysone as a underscores that p21CIP1 functions as a central consequence of basal ARF expression. regulator both in G1 (Samuel et al., 2001; Waldman Genetic instability in cell lines might lead to et al., 1995) and G2 (Bunz et al., 1998; McShea et al., alterations of central signaling pathways. To exclude 2000; Waldman et al., 1996) by diverse mechanisms. such alterations as a cause of an impaired cell cycle On the contrary, 14-3-3s7/7 cells showed only ARF response, p14 -induced cell cycle arrest was also incomplete arrest in G2 following ARF infection, analysed in human primary foreskin keratinocytes pointing to at least a partial function of 14-3-3s in (HFKs). Following infection of the HFKs with ad- G2 (Chan et al., 1999; Hermeking et al., 1997; Samuel ARF, induction of p53, MDM2 and p21 as well as et al., 2001), also after ARF infection. Importantly, CIP17/7 7/7 concomitant G1 cell cycle arrest as judged by BrdU p21 14-3-3s cells behaved identical to the incorporation and the hypophosphorylation of RB p537/7 cells. This underscores the dependency of were detected (data not shown). In HFKs stably human p14ARF on the p53 pathway and suggests retrovirally transduced with the human papillomavirus cooperative functions of p21CIP1 and 14-3-3s in ARF type 16 E6 oncoprotein, which destabilizes p53, the ad- mediating p14 -induced G2 arrest. Furthermore, ARF infection did not lead to induction of p53 and these two p53-regulated genes appear to be predomi- p21CIP1, and there was no evidence of cell cycle arrest. nantly responsible for ARF-induced cell cycle arrest, which di€ers from the DNA damage-induced arrest. Few investigations have dealt with the pathway Discussion utilized by human p14ARF. The osteosarcoma cell line U2OS and the breast cancer cell line MCF7, both The majority of work dealing with ARF function was containing wild-type p53 and RB, were arrested 48 h done in rodent cells with di€erent genetic backgrounds. following exogenous p14ARF expression (Stott et al., Wild-type MEFs showed a decrease in the S phase 1998). In comparison, the osteosarcoma cell line population 48 h following p19ARF retroviral infection, SAOS2, which is defective for both p53 and RB in contrast to MEFs with sustained p53 mutations or function, and the p53-negative lung carcinoma cell line the p53-negative Balb 3T3 cells (Kamijo et al., 1997). H1299 were not arrested by p14ARF (Stott et al., 1998). Similarly, p537/7 MEFs were not arrested by a In line with these results, replicative and p14ARF-induced p19ARF-retrovirus (Modestou et al., 2001). In cotrans- senescence in human ®broblast strains was recently formation assays, foci formation of wild-type rat shown to depend on the presence of p21CIP1 (Wei et al., embryo ®broblasts (REFs) by Myc and Ras was 2001). In contrast, p14ARF transfection of SAOS2 cells inhibited by cotransfection of p19ARF (Pomerantz et and the p53-negative human bronchioalveolar carcino- al., 1998). In contrast, the foci formation of REFs ma cell line H358 followed by selection led to a decrease redendered p53-de®cient by either a dominant negative in the number of surviving cells, as compared to empty- p53-construct or the SV40 large T antigen was not vector transfection (Eymin et al., 2001). This approach inhibited by p19ARF (Pomerantz et al., 1998). was used as to determine ARF-dependent long-term Whereas these studies demonstrated a p53 depen- e€ects. In the latter study, p14ARF was found to interact dency of p19ARF, recent data provide evidence for p53- with E2F1 thereby inhibiting its transcriptional activity; independent functions in MEFs lacking not only p53 however, the functional role and importance of this but also MDM2 (Weber et al., 2000). p537/7 MDM27/ interaction remains to be elucidated. 7 MEFs showed a decrease of BrdU positive cells Taken together, the existing data point to di€erent following p19ARF retroviral infection, with a minimum mechanisms as to how ARF inhibits cellular prolifera- of labeled cells at 96 h post-infection, the mechanisms of tion. Mouse p19ARF appears to induce arrest in both a which are unknown. In another study, CRE recombi- p53-dependent and -independent manner, while human nase excisable p19ARF antisense constructs were used to p14ARF acts in a p53-dependent manner. Most decrease the endogenous p19ARF expression level in experiments showing a p53 dependency of p19ARF- MEFs. Immortalization both in a p53-dependent and induced arrest were done in a 24 ± 48 h-time frame p53-independent manner following excision of the (Kamijo et al., 1997; Modestou et al., 2001). However, p19ARF antisense constructs was shown (Carnero et al., once activated over a longer time period ARF might be 2000). In addition, the mechanism of p53-mediated cell able to induce arrest or apoptosis in a p53-negative cycle arrest once p53 is activated by p19ARF is unclear. background (Carnero et al., 2000; Eymin et al., 2001; p537/7 MEFs were not arrested by p19ARF. In contrast, Weber et al., 2000), the mechanisms of which are CIP17/7 ARF p21 MEFs underwent p19 -induced G1 and poorly understood. In conclusion, we demonstrate here ARF G2 arrest, suggesting p53-dependent, p21-independent that the onset of p14 -mediated G1 and G2 arrest in mechanisms (Modestou et al., 2001). To this end, the epithelial cells strictly depends on the p53 pathway and exact nature of these mechanisms has remained elusive. the downstream genes, p21CIP1 and 14-3-3s.

Oncogene ARF-mediated cell cycle arrest HO Weber et al 3211 bu€er containing 50 mM HEPES pH 7.5, 450 mM NaCl, Materials and methods 1mM EDTA, 2.5 mM EGTA, 10 % glycerol, 1 mM NaF, 0.1 m sodium orthovanadate, 10 mM beta-glycerophosphate Cell lines and transfections M and 0.1% Tween 20. For Immunoblotting, 20 mg of whole- HCT116 and various knockout clones of HCT116 derived by cell extract were loaded. Immunoprecipitations were carried homologous recombination were a generous gift from B out using 200 or 1500 mg whole cell extract and band Vogelstein and H Hermeking, and cultured as described intensity compared to the signal obtained from 20 mgof (Bunz et al., 1998; Chan et al., 1999; Waldman et al., 1995). crude extract. In each case two independently derived knockout clones were used. Transient transfections were made using lipofectamine Antibodies (Gibco) according to standard protocols. Antibodies were sourced as follows: anti-cdc2, anti-p21, anti- p53 (Oncogene Science), anti-cyclin B1 (Pharmingen, GNS- Generation of adenoviruses and retroviruses 1), anti-p21 (C-19), anti-CDK2, anti-MDM2 (Santa Cruz), Full-length p14ARF generated by PCR was subcloned into anti-RB (Becton-Dickinson). Anti-14-3-3s was a gift from H pAdTrack-CMV (Qbiogene). Recombination with the pA- Hermeking (Chan et al., 1999). A polyclonal antibody against dEasy vector (provided by B Vogelstein; He et al., 1998) a C-terminal p14ARF peptide was generated in rabbits. To containing the adenoviral genes was performed in bacteria as test this antibody, a pCMV vector encoding for the human described. The resulting plasmid was then transfected into p14ARF protein was transfected into HCT cells. p14ARF was 293 cells to produce the complete Ad-ARF adenovirus. The detected in immunoblots of transfected cell extracts, and the MOI to reach 100% of the cells was optimized by counting C-terminal p14ARF peptide blocked this reactivity (data not GFP-positive cells. LXSN-ARF was a gift from Jennifer shown). Benanti and Denise A Galloway. LXSN retroviruses and generation of packaging cells have been described (Funk et FACS analysis al., 1997). The env protein of this retrovirus has been changed to the VSVg protein which mediates viral entry through lipid Samples were ®xed in 70% ethanol in PBS after BrdU binding and plasma membrane fusion. High titer virus was labeling according to Funk et al. (1997), stained with 50 mg/ produced by ultracentrifugation. ml Pl and analysed on a Becton Dickinson FacScan II.

Ecdysone-inducible system In 293 cells, an ecdysone-inducible system was established according to the manufacturer's instructions (Invitrogen). Acknowledgments Following cloning of the p14ARF gene into the pIND We thank Bert Vogelstein and Heiko Hermeking for cell inducible expression vector the resulting pIND-ARF was lines and reagents; Denise Galloway and Jennifer Benanti stably transfected into EcR-293 cells that already expressed for LXSN vectors; Karen Vousden for suggestions and the functional ecdysone receptor from pVgRXR. discussion; Frank McCormick, Heiko Hermeking, Susanna Trapp, Andreas Baur, Berlinda Verdoodt, Tino Blazek, and Alexander Steinkasserer for comments; and Gerold Immunoblotting, immunoprecipitations, and kinase assays Schuler for continuing support and encouragement. This These analyses were performed as described previously (Funk work was supported by grants from the ELAN program of et al., 1997; Herzinger et al., 1995; Samuel et al., 2001). the University of Erlangen-Nuremberg and the Deutsche Whole-cell lysates were prepared by brief sonication in lysis Forschungsgemeinschaft to JO Funk.

References

Bates S, Phillips AC, Clark PA, Stott F, Peters G, Ludwig Eymin B, Karayan L, Seite P, Brambilla C, Brambilla E, RL and Vousden KH. (1998). Nature, 395, 124 ± 125. Larsen CJ and Gazzeri S. (2001). Oncogene, 20, 1033 ± Bunz F, Dutriaux A, Lengauer C, Waldman T, Zhou S, 1041. Brown JP, Sedivy JM, Kinzler KW and Vogelstein B. Funk JO, Waga S, Harry JB, Espling E, Stillman B and (1998). Science, 282, 1497 ± 1501. Galloway DA. (1997). Genes Dev., 11, 2090 ± 2100. Burri N, Shaw P, Bouzourene H, Sordat I, Sordat B, Gillet He TC, Zhou S, da Costa LT, Yu J, Kinzler KW and M, Schorderet D, Bosman FT and Chaubert P. (2001). Vogelstein B. (1998). Proc. Natl. Acad. Sci. USA, 95, Lab. Invest., 81, 217 ± 229. 2509 ± 2514. Carnero A, Hudson JD, Price CM and Beach DH. (2000). Hermeking H, Lengauer C, Polyak K, He TC, Zhang L, Nat. Cell Biol., 2, 148 ± 155. Thiagalingam S, Kinzler KW and Vogelstein B. (1997). Chan TA, Hermeking H, Lengauer C, Kinzler KW and Mol. Cell, 1, 3 ± 11. Vogelstein B. (1999). Nature, 401, 616 ± 620. Herzinger T, Funk JO, Hillmer K, Eick D, Wolf DA and Chin L, Merlino G and DePinho RA. (1998a). Genes Dev., Kind P. (1995). Oncogene, 11, 2051 ± 2056. 12, 3467 ± 3481. Kamijo T, Zindy F, Roussel MF, Quelle DE, Downing JR, Chin L, Pomerantz J and DePinho RA. (1998b). Trends Ashmun RA, Grosveld G and Sherr CJ. (1997). Cell, 91, Biochem. Sci., 23, 291 ± 296. 649 ± 659. de Stanchina E, McCurrach ME, Zindy F, Shieh SY, Kamijo T, Weber JD, Zambetti G, Zindy F, Roussel MF and Ferbeyre G, Samuelson AV, Prives C, Roussel MF, Sherr Sherr CJ. (1998). Proc. Natl. Acad. Sci. USA, 95, 8292 ± CJ and Lowe SW. (1998). Genes Dev., 12, 2434 ± 2442. 8297.

Oncogene ARF-mediated cell cycle arrest HO Weber et al 3212 Korgaonkar C, Zhao L, Modestou M and Quelle DE. (2002). Sharpless NE and DePinho RA. (1999). Curr. Opin. Genet. Mol. Cell Biol., 22, 196 ± 206. Dev., 9, 22 ± 30. Levine AJ. (1997). Cell, 88, 323 ± 331. Sherr CJ and Roberts JM. (1995). Genes Dev., 9, 1149 ± 1163. Llanos S, Clark PA, Rowe J and Peters G. (2001). Nat. Cell Sherr CJ. (1998). Genes Dev., 12, 2984 ± 2991. Biol., 3, 445 ± 452. Stott FJ, Bates S, James MC, McConnell BB, Starborg M, McShea A, Samuel T, Eppel JT, Galloway DA and Funk JO. Brookes S, Palmero I, Ryan K, Hara E, Vousden KH and (2000). J. Biol. Chem., 275, 23181 ± 23186. Peters G. (1998). EMBO J., 17, 5001 ± 5014. Modestou M, Puig-Antich V, Korgaonkar C, Eapen A and Vogelstein B, Lane D and Levine AJ. (2000). Nature, 408, Quelle DE. (2001). Cancer Res., 61, 3145 ± 3150. 307 ± 310. No D, Yao TP and Evans RM. (1996). Proc. Natl. Acad. Sci. Vousden KH. (2000). Cell, 103, 691 ± 694. USA, 16, 3346 ± 3351. Waldman T, Kinzler KW and Vogelstein B. (1995). Cancer Palmero I, Pantoja C and Serrano M. (1998). Nature, 395, Res., 55, 5187 ± 5190. 125 ± 126. Waldman T, Lengauer C, Kinzler KW and Vogelstein B. Pomerantz J, Schreiber-Agus N, Liegeois NJ, Silverman A, (1996). Nature, 381, 713 ± 716. Alland L, Chin L, Potes J, Chen K, Orlow I, Lee HW, Waldman T, Zhang Y, Dillehay L, Yu J, Kinzler K, Cordon-Cardo C and DePinho RA. (1998). Cell, 92, 713 ± Vogelstein B and Williams J. (1997). Nat. Med., 3, 723. 1034 ± 1036. Quelle DE, Zindy F, Ashmun RA and Sherr CJ. (1995). Cell, Weber JD, Je€ers JR, Rehg JE, Randle DH, Lozano G, 83, 993 ± 1000. Roussel MF, Sherr CJ and Zambetti GP. (2000). Genes Quelle DE, Cheng M, Ashmun RA and Sherr CJ. (1997). Dev., 14, 2358 ± 2365. Proc. Natl. Acad. Sci. USA, 94, 669 ± 673. Wei W, Hemmer RM and Sedivy JM. (2001). Mol. Biol. Cell, Radfar A, Unnikrishnan I, Lee HW, DePinho RA and 21, 6748 ± 6757. Rosenberg N. (1998). Proc. Natl. Acad. Sci. USA, 95, Weinberg RA. (1995). Cell, 81, 323 ± 330. 13194 ± 13199. Zhang Y, Xiong Y and Yarbrough WG. (1998). Cell, 92, Samuel T, Weber HO, Rauch P, Verdoodt B, Eppel JT, 725 ± 734. McShea A, Hermeking H and Funk JO. (2001). J. Biol. Zindy F, Eischen CM, Randle DH, Kamijo T, Cleveland JL, Chem., 276, 45201 ± 45206. Sherr CJ and Roussel MF. (1998). Genes Dev., 12, 2424 ± Serrano M, Lee H, Chin L, Cordon-Cardo C, Beach D and 2433. DePinho RA. (1996). Cell, 85, 27 ± 37.

Oncogene