Loss of P21 Disrupts P14arf-Induced G1 Cell Cycle Arrest but Augments P14arf-Induced Apoptosis in Human Carcinoma Cells

Home , G1 phase, P14arf, P21, Restriction point

Loss of p21 disrupts p14ARF-induced G1 cell cycle arrest but augments p14ARF-induced apoptosis in human carcinoma cells

Philipp G Hemmati1,3, Guillaume Normand1,3, Berlinda Verdoodt1, Clarissa von Haefen1, Anne Hasenja¨ ger1, DilekGu¨ ner1, Jana Wendt1, Bernd Do¨ rken1,2 and Peter T Daniel*,1,2

1Department of Hematology, Oncology and Tumor Immunology, University Medical Center Charite´, Campus Berlin-Buch, Berlin-Buch, Germany; 2Max-Delbru¨ck-Center for Molecular Medicine, Berlin-Buch, Germany

The human INK4a locus encodes two structurally p16INK4a and p14ARF (termed p19ARF in the mouse), latter unrelated tumor suppressor proteins, p16INK4a and p14ARF of which is transcribed in an Alternative Reading Frame (p19ARF in the mouse), which are frequently inactivated in from a separate exon 1b (Duro et al., 1995; Mao et al., human cancer. Both the proapoptotic and cell cycle- 1995; Quelle et al., 1995; Stone et al., 1995). P14ARF is regulatory functions of p14ARF were initially proposed to usually expressed at low levels, but rapid upregulation be strictly dependent on a functional p53/mdm-2 tumor of p14ARF is triggered by various stimuli, that is, suppressor pathway. However, a number of recent reports the expression of cellular or viral oncogenes including have implicated p53-independent mechanisms in the E2F-1, E1A, c-myc, ras, and v-abl (de Stanchina et al., regulation of cell cycle arrest and apoptosis induction by 1998; Palmero et al., 1998; Radfar et al., 1998; Zindy p14ARF. Here, we show that the G1 cell cycle arrest et al., 1998). In turn, induction of p14ARF mediates the induced by p14ARF entirely depends on both p53 and p21 in accumulation of p53 via sequestration and subsequent human HCT116 and DU145 carcinoma cells. In con- degradation of its natural antagonist mdm-2 through trast, neither loss of p53 nor p21 impaired apoptosis the ubiquitin/proteasome pathway (Pomerantz et al., induction by p14ARF as evidenced by nuclear DNA frag- 1998; Stott et al., 1998). This prolongation of p53 half- mentation, phosphatidyl serine exposure, and caspase acti- life leads to the activation of its downstream target genes vation, which included caspase-3/7- and caspase-9-like such as p21 and Bax (Sherr, 2001). Thus, activation of activities. However, lack of functional p21 resulted in the the p14ARF/p53 signaling cascade is viewed as an accumulation of cells in G2/M phase of the cell cycle and important fail-safe mechanism that protects cells from markedly enhanced p14ARF-induced apoptosis that was, excessive and uncontrolled growth triggered through nevertheless, efficiently inhibited by the cell permeable hyperproliferative stimuli (Voorhoeve and Agami, 2003; broad-spectrum caspase inhibitor zVAD-fmk (valyl-alanyl- Zindy et al., 2003). Consequently, p14ARF was recog- aspartyl-(O)-methyl)-fluoromethylketone). Thus, loss of nized as potent tumor suppressor as functional inactiva- cell cycle restriction point control in the absence of p21 tion of p14ARF accelerates tumorigenesis and promotes may interfere with p14ARF-induced apoptosis. Finally, these chemoresistance by disabling p53 (Eischen et al., 1999; data indicate that the signaling events required for G1 cell Schmitt et al., 1999). cycle arrest and apoptosis induction by p14ARF dissociate In contrast to the initial notion that most if not all upstream of p53. biological activity of p14ARF depends on a functional Oncogene (2005) 24, 4114–4128. doi:10.1038/sj.onc.1208579 p53/mdm-2 signaling axis, a number of recent reports Published online 7 March 2005 indicate that p14ARF as well as its murine homologue p19ARF and p53 act in overlapping pathways rather than Keywords: p14ARF; p53; p21; cell cycle; apoptosis in a strictly sequential manner (Carnero et al., 2000). Consequently, murine p19ARF was shown to cause a G1 cell cycle arrest independently from p53, mdm-2, and Rb, (Weber et al., 2000) as well as p21 (Modestou et al., 2001) and p27 (Groth et al., 2000). In the same vein, we Introduction demonstrated that apoptosis induction by p14ARF is independent from p53 and Bax (Hemmati et al., 2002). The mammalian INK4a locus encodes two structurally Furthermore, a genome-wide screen identified a number unrelated proteins, the cyclin-dependent kinase inhibitor of genes putatively involved in the p53-independent regulation of cellular proliferation by p19ARF (Kuo et al., *Correspondence: P Daniel, Clinical and Molecular Oncology, 2003). Most intriguingly, it was reported recently that University Medical Center Charite´ , Campus Berlin-Buch, Linden- c-Myc, itself a potent trigger of p14ARF (p19ARF) berger Weg 80, Berlin-Buch 13125, Germany; expression, is part of a p53-independent inhibitory E-mail: [email protected] 3These authors contributed equally to this work feedbackloop (Qi et al., 2004). Received 14 June 2004; revised 26 January 2005; accepted 28 January Therefore, we aimed at further investigating and 2005; published online 7 March 2005 dissecting the signaling pathways involved in cell cycle p21 in p14ARF-induced cell cycle arrest and apoptosis PG Hemmati et al 4115 arrest and apoptosis induction by p14ARF. To this end, Whereas an arrest in G1 phase of the cell cycle upon we employed the colorectal cancer cell line HCT116 and expression of p14ARF entirely depends on the presence of its isogeneic sublines homozygously deleted for either p53 (and p21, see below), recent studies indicate that p53 p53 (HCT116-p53À/À) or p21 (HCT116-p21À/À), the is dispensable for the induction of p14ARF-triggered two key regulators of cell cycle arrest and cell death. apoptotic cell death (Hemmati et al., 2002; Eymin et al., We show here that the arrest in G1 phase of the cell cycle 2003). To demonstrate that apoptosis induction by upon expression of p14ARF strictly depends on the p14ARF is indeed independent from p53, apoptotic DNA presence of both p53 and p21. In contrast, p14ARF- fragmentation (Figure 2) and activation of caspases mediated apoptosis is not impaired upon deletion of (Figure 3) were studied in p53-proficient versus p53- either p53 or p21, involves the activation of caspases deficient HCT116 cells. To this end, HCT116-WT and including caspase-3/7- and caspase-9-like activities, and HCT116-p53À/À cells were transduced with Ad-p14ARF in is subject to inhibition by zVAD-fmk(valyl-alanyl- parallel with mock-treated (control) or control vector- aspartyl-(O)-methyl)-fluoromethylketone). Similarly, p53- infected (Ad-lacZ) cells and subjected to flow-cytometric mutated, p21-deficient DU145 prostate cancer cells were analysis of DNA fragmentation 72 h after infection. As not impaired in their ability to undergo apoptotic cell shown in Figure 2a, expression of p14ARF induced death upon p14ARF expression. Notably, loss of p21 apoptotic DNA fragmentation, that is, cells displaying resulted in the accumulation of cells in G2/M phase of the a sub-G1 DNA content, in both p53-proficient and cell cycle and markedly enhanced p14ARF-induced apop- p53-deficient HCT116-WT cells to an identical extent. tosis. This delineates a novel aspect of p14ARF-induced This effect increased in a dose-dependent manner cell death, which may be substantially augmented in the irrespective of the presence or absence of p53 absence of proper G1 restriction point control by loss (Figure 2b). Similarly, there was no difference in the of p21. induction of pancaspase activities by p14ARF in both HCT116 sublines (Figure 3a and b), indicating that p14ARF-induced apoptosis is independent from p53. To corroborate that apoptosis induction by p14ARF is p53 independent, DU145 prostate cancer cells, which Results are p53 mutated and p21 deficient, were infected with Ad-p14ARF and subjected to flow-cytometric detection of Apoptosis induction by p14ARF is not impaired by loss of p53 and/or p21 apoptotic DNA fragmentation (Figure 4a and b). In analogy to HCT116 cells, we observed a dose-dependent In contrast to the initial notion that most if not all increase in the number of apoptotic cells (Figure p14ARF activity strictly depends on a functional p53/ 4a), which was initiated at 48 h and reached substantial mdm-2 signaling axis, a number of recent reports levels at 72–96 h after transduction with Ad-p14ARF established that p14ARF is capable of mediating p53- (Figure 4b). In parallel, a dose-dependent increase of independent effects as well. To investigate this discre- pancaspase activities was detected in DU145 cells 72 h pancy and to further dissect the p14ARF signaling after infection (Figure 4c). As depicted in Figure 4d, cascade, we employed a previously constructed and caspase activation occurred between 48 and 72 h after functionally characterized adenoviral vector system transduction with Ad-p14ARF and further increased after (Ad-p14ARF) for the transient expression of p14ARF in a 96 h. This underscores the notion that p14ARF-induced number of genetically well-defined cell lines. These apoptosis is independent from p53. included HCT116 colorectal cancer cells lacking either p53 or p21 and DU145 prostate cancer cells carrying Loss of p21 sensitizes to apoptosis induction by p14ARF distinct defects in p53 and p21. To this end, HCT116 parental cells wild type for both Data obtained in p53-deficient DU145 and HCT116 p53 and p21 (HCT116-WT) and the isogeneic sublines cells indicated that p14ARF induces apoptosis irrespective homozygously deleted for either p53 (HCT116-p53À/À) of the presence or absence of functional p53. Further- or p21 (HCT116-p21À/À) were transduced with Ad- more, lackof p21 induction in either cell line suggests p14ARF and assayed for transgene expression and that p14ARF-triggered apoptosis is p21 independent. To activation of the p53 pathway (Figure 1). Expression further address this issue, we investigated p21-deficient of p14ARF was detectable by immunofluorescence in all HCT116 cell in parallel with p21-proficient HCT116- HCT116 sublines 24 h after infection with 25 MOI of WT cells. Interestingly, we observed that loss of p21 Ad-p14ARF (Figure 1a). While the localization within the strongly augmented the extent of apoptosis induction by cytoplasm was diffuse, a speckled nuclear expression p14ARF when nuclear DNA fragmentation was investi- pattern was detectable. Western blot analysis (Figure 1b) gated by flow cytometry (Figure 5a and b). At 100 MOI, showed an increase in the expression of p53 and p21 in p14ARF induced genomic DNA fragmentation in ap- HCT116-WT cells upon expression of p14ARF.In proximately 50% of the p21-proficient HCT116 wild- contrast, HCT116-p53À/À cells completely lacked the type cells, whereas in HCT116-p21À/À cells, a level of expression of p53 and induction of p21, respectively. 70% cells displaying fragmented nuclear DNA was Notably, HCT116-p21À/À cells showed a constitutively reached. When pancaspase activities were determined elevated p53 expression level but, as expected, were by the use of VAD-FITC (Figure 6a and b), approxi- completely void of p21. mately 60% of the HCT116 wild-type, but 90% of

Oncogene p21 in p14ARF-induced cell cycle arrest and apoptosis PG Hemmati et al 4116

Figure 1 p14ARF expression in HCT116-WT, HCT116-p53À/À, and HCT116-p21À/À cells. (a) HCT116 cells were mocktreated or infected with the indicated adenoviral vectors at 5 MOI. Expression of p14ARF was visualized by immunofluorescence using a rabbit polyclonal anti-p14ARF antibody followed by Alexa fluor 488-conjugated goat-anti-rabbit IgG. Nuclei were counterstained with DAPI. Representative high-power fields from three independent experiments are shown. (b) HCT116 cells either wild type or homozygously deleted for p53 or p21 were mocktreated or transduced with the adenoviral constructs Ad-p14 ARF or Ad-lacZ as indicated at an MOI of 25. Cells were harvested 48 h after infection and 20 mg of total cellular proteins were subjected to SDS–PAGE and studied by Western blot analysis for the expression of p14ARF, p53, and p21. Equal loading was confirmed by reprobing with an antibody against b-actin

the HCT116-p21À/À cells displayed caspase activation Inhibition of p14ARF-mediated apoptosis by zVAD-fmk after infection with 100 MOI of Ad-p14ARF. These data suggest that p14ARF does not require p21 to induce To substantiate that the activation of caspases is apoptosis. In turn, loss of intact restriction point essential for p14ARF-induced cell death, apoptosis control, that is, in the absence of p21, appears to induction by p14ARF was studied in the presence or augment p14ARF-induced apoptosis. absence of the cell-permeable, broad-spectrum caspase

Oncogene p21 in p14ARF-induced cell cycle arrest and apoptosis PG Hemmati et al 4117 inhibitor zVAD-fmk(Figure 7). To corroborate that cell iodide (PI) double staining. As shown in Figure 7a and death is apoptotic, we employed measurement of phos- b, expression of p14ARF resulted in a substantial increase phatidyl serine exposure on the outer leaﬂet of the cell in the number of early apoptotic, that is, Annexin-V- membrane by the use of Annexin-V-FITC/propidium FITC-positive/PI-negative, and necrotic, PI-positive cells as compared to mock- (control) or control vector- infected (Ad-lacZ) cells 48 h after infection with Ad- p14ARF. As expected, cells lacking p21 were again more sensitive as compared to HCT116-WT cells. In the presence of zVAD-fmk, a marked decrease in phosphatidyl serine exposure (Figure 7b) was observed. Notably, zVAD-fmkalso reduced the occurrence of Annexin-V- FITC-positive/PI-positive cells, suggesting that these cells are late apoptotic rather than a priori necrotic. Finally, zVAD-fmk-mediated reduction in apoptotic DNA fragmentation (Figure 7c) to near background levels was observed in both HCT116-WT and HCT116- p21À/À. Taken together, these data indicate that p14ARF induces caspase-dependent apoptotic cell death, which is enhanced by loss of p21.

Induction of caspase activation by p14ARF Apoptosis induction by p14ARF involves the activation of caspases as indicated by a dose-dependent increase in pancaspase activities and inhibition of apoptosis induction by the broad-spectrum caspase inhibitor zVAD-fmk. To differentiate pancaspase activities, HCT116-WT, HCT116-p53À/À, and HCT116-p21À/À cells were infected with 100 MOI Ad-p14ARF in parallel with mock-treated (control) or control vector-infected (Ad-lacZ; 100 MOI) cells and assayed for the activation of caspase-3/7-like caspases by the use of a fluorescent derivative of the caspase-3/7-like caspase inhibitor DEVD- fmk(aspartyl-glutamyl-valyl-aspartyl-( O)-methyl)-fluoromethyl-ketone) (Figure 8). The topoisomerase IIa inhibitor epirubicin served as a positive control and was used at a final concentration of 2 mg/ml.AsdepictedinFigure8a and b, p14ARF induced caspase-3/7-like activities in all HCT116 sublines 72 h after infection. In analogy to the cell death data shown in Figures 2 and 3, this effect occurred in a p53-independent manner but was enhanced in the absence of p21. In contrast, epirubicin-induced activation of caspase-3/7-like activity was reduced in p53-deficient, but not altered in p21-deleted HCT116 cells. Similarly,

Figure 2 Induction of apoptosis by p14ARF is independent from p53. (a) DNA fragmentation. HCT116 colorectal cancer cells either wild type (HCT116-WT) or bi allelically deleted for p53 (HCT116- p53À/À) were mocktreated (control) or infected with the indicated adenoviral vectors at an MOI of 100. Apoptosis was analysed on the single-cell level by ﬂow-cytometric measurement of nuclear DNA contents 72 h after infection. The relative number of cells displaying an apoptotic, sub-G1 DNA content is given between the marker bars. DNA histograms are representative of three independent experiments. (b) Dose response. HCT116-WT (p53 WT) or HCT116-p53À/À (p53 KO) cells were mocktreated (control) or infected with either Ad-lacZ (lacZ) at an MOI of 100 or increasing MOIs of Ad-p14ARF for 72 h. Apoptosis was determined by ﬂow-cytometric detection of nuclear DNA fragmentation. Bars represent the mean of cells displaying a sub-G1 DNA content7s.d. from three independent experiments (gray bars: p53 WT; black bars: p53 KO)

Oncogene p21 in p14ARF-induced cell cycle arrest and apoptosis PG Hemmati et al 4118 caspase-9-like LEHDase activity was detected in all caspase-9-like LEHDase activity was markedly reduced in HCT116 sublines upon expression of p14ARF (Figure 8c p53-deficient versus p53-proficient cells. These data suggest and d). Whereas the p14ARF-induced activation of caspase- that activation of caspase-3/7 (and other DEVDases) as 9-like activity did not differ between p53-proficient and well as caspase-9 (and other LEHDases) plays an p53-deficient cells, biallelic deletion of p21 strongly important role in mediating p14ARF-induced apoptosis. enhanced LEHDase activity. Again, epirubicin-induced p14ARF-induced G1 arrest depends on p53 and p21 To determine the impact of p14ARF on cell cycle arrest, HCT116-WT, HCT116-p53À/À, and HCT116-p21À/À cells were transduced with 25 MOI Ad-p14ARF and assayed for cell cycle distribution after 48 h. As shown in Figure 9a, expression of p14ARF induced an arrest in G1 phase of the cell cycle in cells proficient for both p53 and p21, that is, HCT116-WT. In contrast, HCT116- p53À/À and HCT116-p21À/À cells did not arrest in G1 phase of the cell cycle. Whereas HCT116-WT cells showed a corresponding decrease in the number of cells in S phase, the proportion of S phase cells remained mostly unchanged in cells deficient for either p53 or p21. Consequently, the relative number of cells in G2/M phase of the cell cycle increased in the absence of either p53 or p21. The accumulation of cells in G1 phase of the cell cycle upon expression of p14ARF occurred in a dose-dependent manner only in p53/p21- proficient HCT116-WT cells (Figure 9b). In contrast, a dose-dependent decrease of cells in G1 phase of the cell cycle upon expression of p14ARF was observed in cells homozygously deleted for either p53 or p21. These results clearly indicate that the induction of G1 cell cycle arrest by p14ARF depends on a functional p53/p21 signaling axis.

Inhibition of DNA synthesis by p14ARF is independent from p53 and p21 The results described above strongly suggest that p14ARF-induced G1 arrest strictly depends on the presence of both p53 and p21. Yet, several lines of evidence indicate that p14ARF is capable of inhibiting cell cycle progression as well via p53-independent mechanisms (Eymin et al., 2001; Martelli et al., 2001). Therefore, we tested the impact of p14ARF expression on DNA synthesis in cells either proﬁcient (HCT116-WT) or deﬁcient for p53 (HCT116-p53À/À) or p21 (HCT116- p21À/À). To this end, HCT116-WT, HCT116-p53À/À, and

Figure 3 Induction of caspase activities by p14ARF is p53 independent. (a) Caspase activation. HCT116-WT and HCT116- p53À/À cells were infected with Ad-p14ARF (100 MOI) or control virus (Ad-lacZ; 100 MOI) in parallel with mock-treated cells (control). Prior to harvest, cells were incubated with an FITC- labeled conjugate of the cell-permeable pancaspase inhibitor VAD- fmkfor the detection of activated caspases in situ. Percentages of cells displaying pancaspase activities after 72 h are given between marker bars. Representative histograms from three independent experiments are shown. (b) Dose response. HCT116-WT (p53 WT) or HCT116-p53À/À (p53 KO) cells were mocktreated (control) or transduced with either Ad-lacZ (lacZ) at an MOI of 100 or increasing MOIs of Ad-p14ARF for 72 h. Activation of caspases was determined by ﬂow cytometry as described in (a). Bars represent the mean of cells displaying pancaspase activities7s.d. from three independent experiments (gray bars: p53 WT; blackbars: p53 KO)

Oncogene p21 in p14ARF-induced cell cycle arrest and apoptosis PG Hemmati et al 4119

Figure 4 Apoptosis induction and activation of caspases by p14ARF in p53-mutated, p21-deficient DU145 cells. (a) Dose response. DU145 prostate cancer cells mutated for p53 and deficient for p21 expression were mocktreated (control) or infected with either Ad-lacZ (lacZ) at an MOI of 100 or transduced with increasing MOIs of Ad-p14ARF for 72 h. Cells were analysed for apoptotic DNA fragmentation by flow cytometry. Bars represent the mean7s.d. of three independent experiments. (b) Time course. DU145 cells were mocktreated (control) or infected with Ad-lacZ or Ad-p14ARF at an MOI of 50. Cells were harvested at the indicated time points after infection and apoptotic DNA fragmentation was determined by flow cytometry. Means of apoptotic cells7s.d. from three independent experiments are given (control: white squares; Ad-lacZ: white circles; Ad-p14ARF: blackcircles). ( c) Caspase activation. DU145 cells were treated as described in (a) and incubated with an FITC-labeled conjugate of the cell-permeable pancaspase inhibitor VAD-fmkfor in situ detection of activated caspases at the end of the cultured period. Caspase activation was determined by flow cytometry and bars represent the mean7s.d. of three independent experiments. (d) Time course. DU145 cells were either mocktreated (control) or infected with the indicated adenoviral vectors at an MOI of 100. In situ activation of caspases was determined at the indicated time points. Bars shown represent the mean7s.d. of three independent experiments (control: white squares; Ad-lacZ: white circles; Ad-p14ARF: blackcircles)

Figure 5 Induction of apoptosis by p14ARF is enhanced in p21- HCT116-p21À/À cells were infected with 50 MOI of Ad- deficient HCT116 cells. HCT116-WT (p21 WT) and HCT116-p21À/À p14ARF for 48 h, pulse labeled with 5-bromodeoxyuridine (p21 KO) were infected with recombinant adenoviruses as (BrdU) for 30 min, and subsequently analysed by flow indicated at an MOI of 100. Apoptosis was determined by flow- cytometric analysis of genomic DNA fragmentation 72 h after cytometry in parallel with mock-treated (control) and infection. Cells displaying a sub-G1 DNA content were identified control vector-infected (Ad-lacZ; 50 MOI) cells. as apoptotic. (a) One representative of three independent experi- (Figure 10). The quantification of BrdU-positive cells ments is shown. (b) Dose response. Cells were cultured and treated representing cells in S phase confirmed that inhibition of as in (a). Mean values7s.d. of three independent experiments are DNA synthesis upon expression of p14ARF did not differ shown (gray bars: p21 WT; blackbars: p21 KO) between wild-type cells or cells homozygously deleted

Oncogene p21 in p14ARF-induced cell cycle arrest and apoptosis PG Hemmati et al 4120 for either p53 or p21 as indicated by the similar decrease in percentages of BrdU positivity. Nevertheless, loss of either p53 or p21 impairs S-phase control per se as evidenced by the overall higher percentages of BrdU positivity in cells void of either p53 or p21. Thus, p14ARF-mediated G1 arrest is paralleled by inhibition of DNA synthesis, which is independent from the presence or absence of functional p53 or p21.

Discussion

It is still a matter of debate whether the induction of cell cycle arrest and apoptosis by p14ARF equally depend on an intact p53/mdm-2 signaling axis. Data from initial reports indicated a functional role of both p53 and p21 as a strict prerequisite for 14ARF-mediated cell cycle arrest. Nevertheless, a number of follow-up studies clearly demonstrated that this is not entirely the case. In particular, distinct members of E2F family of transcription factors were suggested to be critical downstream targets of p53-independent growth arrest programmes engaged by p14ARF (Eymin et al., 2001; Martelli et al., 2001; Mason et al., 2002; Rowland et al., 2002). Furthermore, a delay in S-phase progression with consecutive accumulation of cells in G1 phase of the cell cycle upon expression of p14ARF was shown to occur independently of p53 (Yarbrough et al., 2002). Similarly, others have shown that a replicative senescence-like type of G1 arrest may be induced by the prolonged expression of p19ARF in mouse embryonic fibroblasts regardless of p53 and p21 function (Carnero et al., 2000; Weber et al., 2000), and in the absence of mdm-2. In the same vein, we recently showed that forced expression of p14ARF induces apoptosis in p53/Bax-deficient cells (Hemmati et al., 2002). In contrast, genetic evidence from murine knockout models strongly suggested that p19ARF is capable of regulating cellular proliferation in a p53-independent manner (Eischen et al., 1999; Weber et al., 2000). Most intriguingly, the c-Myc oncoprotein was recognized recently as a key factor in regulating p19ARF-mediated proliferative effects in a p53-independent manner (Qi et al., 2004). By serving as a direct transcriptional target, p19ARF differentially regulates the c-Myc’s transactivating and transrepressing function. Although not fully under- stood yet, it is now widely recognized that both the induction of cell cycle arrest and apoptosis by p14ARF involve a much more complex networkof intracellular signaling events than initially thought (Cleveland and Sherr, 2004). In particular, the fact that an arrest in G1 phase of the cell cycle is entirely dependent on Figure 6 Induction of pancaspase activities by p14ARF is functional p53 (Weber et al., 2002), whereas the induction augmented in p21-deficient HCT116 cells. HCT116-WT (p21 WT) and HCT116-p21À/À (p21 KO) were mocktreated or infected of apoptotic cell death is not suggests that the two with recombinant adenoviral vectors as indicated at an MOI of signaling cascades are, at least in part, independent. 100. Activation of caspases was determined by flow cytometry 72 h In the present work, we show that p14ARF triggered an after infection. (a) Representative histograms of three independent arrest in the G1 phase of the cell cycle that strictly experiments are shown. The relative number of cells displaying depends on both functional p53 and p21. In contrast, pancaspase activity is given between the marker bars. (b) Dose ARF response. Cells were cultured and treated as described in (a). Mean p14 -induced apoptosis was not impaired in tumor values7s.d. of three independent experiments are shown (gray cells with disrupted p53/p21 but was even enhanced bars: p21 WT; blackbars: p21 KO) upon targeted knockout of the p21 gene, that is, in

Oncogene p21 in p14ARF-induced cell cycle arrest and apoptosis PG Hemmati et al 4121

Figure 7 Inhibition of p14ARF-mediated apoptosis by caspase inhibitors. HCT116-WT and HCT116-p21À/À cells were either mock treated (control) or infected with recombinant adenoviruses at a MOI of 100. (a) Annexin-V-FITC. Apoptosis was determined by flow- cytometric detection of Annexin-V-FITC-positive/PI-negative cells 72 h after infection. For inhibition of apoptosis, the broad- spectrum caspase inhibitor zVAD-fmkwas added at a final concentration of 40 mM. Representative histograms from three independent experiments are shown. The relative number of cells in each quadrant is given in percent. (b) Cells were treated as described in (a). Bars represent the mean7s.d. from three independent experiments. (c) DNA fragmentation. Cells were treated as in (a) and assayed for apoptotic DNA fragmentation by flow cytometry 72 h after infection. Bars represent the mean7s.d. from three independent experiments (white bars: control; hatched bars: Ad-lacZ; gray bars: Ad-p14ARF; blackbars: Ad-p14 ARF þ zVAD-fmk(40 mM))

HCT116 -p21À/À cells. This indicates that cell cycle arrest cell cycle within 24–48 h upon expression of p14ARF.In and apoptosis induction by p14ARF proceed through contrast, isogeneic HCT116 cells homozygously deleted distinct signaling events that dissociate upstream of p53. for either p53 or p21 did so to a far lesser extent but, Regarding cell cycle regulation, HCT116 wild-type instead, accumulated in G2/M. To conﬁrm these results, cells rapidly arrested preferentially in G1 phase of the we employed the prostate cancer cell line DU145, which

Oncogene p21 in p14ARF-induced cell cycle arrest and apoptosis PG Hemmati et al 4122

Figure 8 Induction of caspase-3-like and caspase-9-like activities by p14ARF. HCT116-WT, HCT116-p53À/À (p53 KO), and HCT116- p21À/À (p21 KO) were infected with 100 MOI of Ad-p14ARF in parallel with mock-treated (control) or control-vector (Ad-lacZ; 100 MOI) infected cells. Treatment with the topoisomerase IIa inhibitor epirubicin at a final concentration of 2 mg/ml served as a positive control. Cells were incubated with FAM-DEVD-fmk, a fluorescein-conjugated derivative of the cell-permeable caspase-3/7- like inhibitor DEVD-fmk 72 h after the onset of treatment. Activation of caspases in situ was determined by flow cytometry on the single-cell level. (a) One representative histogram of three independent experiments is shown. The relative number of cells displaying caspase-3/7-like (DEVDase) activities is given between the marker bars. (b) Cells were treated as indicated and analysed for the activation of caspase-3/7-like activities as described above. Bars represent the mean7s.d. of three independent experiments. (c) Cells were treated as described in (a) and incubated with FAM-LEHD-fmk, a fluorescein-coupled derivative of the caspase-9-like substrate LEHD-fmk. Caspase activation was determined by flow cytometry 72 h after treatment. Representative histograms are shown. (d) Cells were treated as described in (c) and bars represent the mean7s.d. of cells displaying caspase-9-like (LEHDase) activities from three independent experiments (white bars: HCT116-WT; gray bars: p53 KO; blackbars: p21 KO)

is characterized by mutated, functionally inactive p53, conﬁrm the notion that the p14ARF-induced G1 arrest is resulting in a failure to induce p21 expression (Hemmati entirely dependent on the presence of functional p53 and et al., 2002). Similar to HCT116 cells that are homo- its downstream effector p21. Furthermore, our data zygously deleted for either p53 or p21, DU145 cells indirectly suggest that the absence of p53 cannot be failed to arrest in G1 phase of the cell cycle after readily compensated by other p53 homologues such as expression of p14ARF. Instead, like HCT116-p53À/À and p63 and p73. Moreover, this indicates that the HCT116-p21À/À cells, the accumulation of cells in propensity to undergo G2/M arrest neither relies on G2/M phase of the cell cycle was observed. These data p53 nor p21. In contrast, it was recently suggested that

Oncogene p21 in p14ARF-induced cell cycle arrest and apoptosis PG Hemmati et al 4123

Figure 9 p14ARF induces G1 arrest in p53/p21-proficient HCT116 cells. (a) HCT116 cells wild type (HCT116-WT) or homozygously deleted for either p53 (HCT116-p53À/À) or p21 (HCT116-p21À/À) were infected with 25 MOI of Ad-p14ARF. In parallel, cells were either mocktreated (control) or infected with an adenoviral control vector (Ad-lacZ) at an MOI of 25. Cell cycle distribution was assayed 48 h after infection by flow cytometry. Histograms shown are representative of three independent experiments. The relative number of cells in each phase of the cell cycle (G1, S, G2/M) is given in percent. (b) HCT116-WT, HCT116-p53À/À, and HCT116-p21À/À were infected with an increasing MOI of Ad-p14ARF in parallel with mock- (Co) or control vector-infected cells (lac; 100 MOI) and assayed for cell cycle distribution 48 h after infection. Bars represent the mean7s.d. of cells in G1 phase of the cell cycle as determined in three independent experiments the p14ARF-induced G2 arrest occurs through a p21- the p53-mediated induction of p21 represents the dependent pathway in p53-deficient cells (Eymin et al., dominant arrest program triggered by p14ARF, whereas 2003). Our data indicate, however, that a G1 arrest via G2/M arrest and the previously described S-phase delay

Oncogene p21 in p14ARF-induced cell cycle arrest and apoptosis PG Hemmati et al 4124

Figure 10 Inhibition of DNA synthesis by p14ARF in HCT116 cells. (a) HCT116-WT, HCT116-p53À/À, and HCT116-p21À/À cells were either mocktreated (control) or infected with adenoviral vectors as indicated at an MOI of 25. At the end of the culture period, cells were pulse labeled with BrdU for 30 min before harvesting. DNA content and BrdU incorporation were analysed by ﬂow cytometry. The number of BrdU-positive cells representing cells in S phase of the cell cycle is given in percent. Representative histograms are shown. (b) Quantiﬁcation of BrdU-positive cells. Means of three independent experiments7s.d. are shown. White bars: control (Co); gray bars: Ad-lacZ (lacZ); blackbars: Ad-p14 ARF (ARF)

may serve as fail-safe mechanisms once p53 and/or p21 et al., 2004). Whereas these data indicate that p14ARF become functionally inactivated. Furthermore, the data regulates cdc2-kinase activity at the post-translational obtained from HCT116-p53À/À, HCT116-p21À/À,and level, that is, through interference with protein turnover DU145 indicate that loss of p53 and p21 does not affect or subcellular localization, further studies are necessary the propensity of p14ARF to arrest cells in G2/M phase of to clarify the mechanism of cdc2 targeting by p14ARF the cell cycle. Consequently, we recently showed that that, however, does not appear to involve direct physical this G2 arrest induced by p14ARF is mediated by interaction with and sequestration via p14ARF. inhibition of the the p34cdc2 (cdk-1) kinase through an In contrast, recent reports have suggested other entirely p53/p21-independent mechanism (Normand mechanisms by which p14ARF may bypass p53-induced

Oncogene p21 in p14ARF-induced cell cycle arrest and apoptosis PG Hemmati et al 4125 G1 cell cycle arrest. These include the binding, sequestration, and functional inactivation of E2F transcription factors necessary for cells to enter and progress through S phase (Eymin et al., 2001; Martelli et al., 2001; Rowland et al., 2002) among other factors involved in DNA replication (Yarbrough et al., 2002). Indeed, by inhibition of S-phase entry and progression, cells may accumulate in G1 phase of the cell cycle. This may serve well to explain the fact that in some studies a G1 arrest was observed in cells lacking functional p53 or carrying a disruption of the Rb pathway. However, the relevance of such p53-independent modes of p14ARF-induced G1 arrest remains to be established. In contrast to p14ARF-induced cell cycle arrest in G1, apoptosis is triggered through a distinct, p53- and p21-independent mechanism. The mitochondrial pathway of apoptosis is initiated by a number of different stimuli including DNA damage or the activation of Figure 11 Dissociation of cell cycle arrest and apoptosis induction cellular and viral oncogenes (Daniel, 2000; Daniel et al., by p14ARF occurs upstream of p53. The presence of both p53 and p21 is strictly required to arrest cells in G1 phase of the cell cycle 2003; Gu¨ ner et al., 2003). In contradiction to the initial ARF ARF upon expression of p14 . In contrast, induction of apoptosis by notion that apoptosis induction by p14 relies on a p14ARF is not impaired upon loss of p53 or p21. Previous data functional p53/mdm-2 rheostat, we recently showed established that this type of cell death is independent from Bax. that p14ARF is capable of triggering the mitochondrial However, loss of p21 enhances apoptosis sensitivity upon expres- pathway of apoptosis followed by the activation of sion of p14ARF. This delineates that the activation of p21-mediated checkpoint control by p14ARF interferes with p14ARF-induced caspases, independently from the presence or absence apoptosis signaling of functional p53 and Bax (Hemmati et al., 2002). In continuation of this work, we show here that induction of apoptosis by p14ARF is not hampered by disruption of p53 or p21. This was evidenced by nuclear DNA in higher levels of p53 protein (Figure 1). However, both fragmentation, which was preceded by caspase acti- p53-dependent and p53-independent apoptosis were vation and phosphatidyl serine exposure on the cell shown to be sensitized for by loss of p21. Indeed, surface, that is, Annexin-V positivity. Moreover, we defects in p53-dependent induction of p21, repression establish that loss of p21 substantially augments p14ARF- of p53-dependent induction of p21 transcription, or induced apoptosis, suggesting that p21 is not required caspase-mediated cleavage of p21 may result in in- for apoptosis induction as has been reported in other creased p53-dependent apoptosis (Bissonnette and experimental systems before (Fotedar et al., 1999). In Hunting, 1998; Zhang et al., 1999; Kokontis et al., view of the strict dependence of p14ARF on p53 and p21 2001). Therefore, increased expression of p53 may to mediate G1 arrest, our data delineate that the sensitize for apoptosis induction by p14ARF, that is, via induction of apoptosis and cell cycle arrest by p14ARF activation of downstream target genes. In turn, a occurs through separate signaling pathways that dis- number of reports delineate that p21 is a negative sociate upstream of p53 (Figure 11). Thus, loss of p53 regulator of p53-independent apoptosis. Here, enhance- and its transcriptional target p21 does not impair the ment of apoptosis, that is, as induced by TGF-b, TNF- induction of apoptosis but abrogates p14ARF-induced G1 a, histone deacetylase inhibitors, and IFN-g was arrest. Nevertheless, p21 appears to play a regulatory observed by loss of p53-independent transctivation of role in p14ARF-induced cell death. Hence, the accelera- p21 (Donato and Perez, 1998; Zhu et al., 1998; Mahyar- tion of p14ARF-induced apoptosis in the absence of p21 Roemer and Roemer, 2001; Pennington et al., 2001). may indicate a critical role for cell cycle arrest programs Experiments are therefore underway to further dissect in p14ARF-triggered cell death. Such an inhibitory the mechanisms of p21-enhanced apoptosis induction by function of p21 was described earlier for DNA damage p14ARF in relation to p53 and cell cycle checkpoint responses, where loss of p21 sensitized carcinoma cells disruption. Nevertheless, p53-independent induction of for apoptosis both in vitro (Waldman et al., 1995) and in apoptosis by p14ARF has now been shown not only by vivo in human disease (Rau et al., 2003). This, in fact, our group (Hemmati et al., 2002) but also by others makes sense when considering the induction of cell (Eymin et al., 2003). death by p14ARF during oncogenic transformation. Apoptosis induction by p14ARF involves a breakdown There, disruption of p53 and consecutive impairment of mitochondrial membrane potential followed by the of p21 signaling may have deleterious consequences activation of caspases, which is independent from the and sensitization for p14ARF-induced apoptosis upon presence or absence of functional p53 and Bax loss of p21 might counterbalance such events. (Hemmati et al., 2002). Here, we provide additional It is worth noting that expression data in the p21À/À evidence that p14ARF-induced apoptosis activates the and p53À/À cells seem to indicate that p21 and p53 may mitochondrial apoptosis pathway as indicated by the also negatively regulate each other as loss of p21 results activation of caspase-9-like LEDHase activities. In

Oncogene p21 in p14ARF-induced cell cycle arrest and apoptosis PG Hemmati et al 4126 analogy to DNA fragmentation and pancaspase activ- Immunofluorescence ity, these caspase-9-like activities were greatly enhanced Immunofluorescence was performed as described previously in cells lacking functional p21. In contrast, activation of (Weber et al., 2002). Briefly, 1.5 Â 105 cells were seeded on caspase-3/7 and caspase-9-like activities by the topo- sterile coverslips, grown for 24 h, and infected with adenoviral isomerase IIa poison epirubicin was substantially vectors at the indicated MOI for 24 h. After three washes with reduced in p53-deficient cells, underscoring the notion PBS, cells were fixed with 50% acetone/methanol (v/v) for that drug-induced apoptosis, at least in part, depends on 15 min and air-dried. After permeabilization with 0.1% NP-40 functional p53, whereas p14ARF-induced apoptosis is in PBS for 10 min, slides were incubated with blocking solution entirely p53 independent. (PBS supplemented with 2% BSA) for 40 min prior to incubation with an anti-p14ARF antibody in blocking solution at room temperature on a shaker for 1 h. After three washes in PBS, slides were incubated with an Alexa fluor 488-labeled Materials and methods anti-rabbit antibody (Molecular Probes, Leiden, The Nether- lands) for 2 h. Finally, slides were washed and mounted with Cell culture Mowiol solution (Merck, Bad Soden, Germany) containing DAPI (Sigma-Aldrich) at a final concentration of 0.5 mg/ml. HEK293 and DU145 were obtained from the ATCC (Manassas, VA, USA) or the DSMZ (Braunschweig, Ger- many). HCT116 wild-type cells and their isogeneic knockout Immunoblotting sublines HCT116-p21À/À (Waldman et al., 1995) and HCT116- Cells were harvested by trypsinization, washed twice with ice- p53À/À (Bunz et al., 1998) were kindly provided by Dr Bert cold PBS, and lysed in appropriate amounts of lysis buffer Vogelstein, Johns Hopkins Cancer Center, Baltimore, MD, (10 mM Tris-HCl, pH 7.5, 300 mM NaCl, 1% Triton X-100, À/À USA. For the HCT116-p53 , we employed clone 379.2 that, 2mM MgCl2,5mM EDTA) supplemented with protease À/À in contrast to another HCT116-p53 clone employed in a inhibitors (1 mM pepstatin, 1 mM leupeptin, and 0.1 mM previous report of our group, has the propensity to undergo phenylmethylsulfonyl fluoride) for 30 min on ice. Samples apoptotic DNA fragmentation. HEK293 and DU145 cells were then centrifuged at 15 000 g for 15 min at 41C and the were grown in DMEM/high glucose (4.5 g/l) medium supple- concentration of total cellular proteins from the supernatants mented with 10% fetal calf serum (FCS), 100 U/ml penicillin, was determined using the bicinchoninic acid assay (Pierce, and 0.1 mg/ml streptomycin (all from Invitrogen, Karlsruhe, Rockford, IL, USA). Thereafter, samples were mixed with Germany). HCT116 cells were cultured in McCoy’s 5A sample buffer (125 mM Tris-HCl, 288 mM b-mercaptoethanol, medium (Invitrogen) supplemented with 10% FCS, 100 U/ml 20% glycerol, 2% SDS, 10 mg/ml bromophenol blue), boiled penicillin, and 0.1 mg/ml streptomycin. for 5 min, and equal amounts of protein (20 mg) were separated by SDS–PAGE using 16 or 10% gels, respectively. Immuno- Construction of recombinant adenoviral vectors blotting and visualization of the proteins using enhanced chemiluminescence were performed as described (Hemmati Ad5-CMVp14ARF (Ad-p14ARF) was constructed as described et al., 2002). For control of equal protein loading, membranes previously (Hemmati et al., 2002) and high-titer stocks of were reprobed with an antibody against b-actin. recombinant adenovirus were generated according to standard procedures as described (Gillissen et al., 2003). An adenovirus Cell cycle analysis and measurement of apoptotic cell death for the expression of b-galactosidase (Ad-lacZ) was used as a control. Transduction efficiency was tested in a number of In all, 5 Â 105 cells were seeded per 75 cm2 flask(Costar, cell lines (including DU145 and HCT116) by infection with Cambridge, MA, USA), cultured overnight, and infected with the Ad-lacZ control virus. Cells were maintained at 371C adenoviral vectors as indicated. At the indicated time points, with 5% CO2 in a fully humidified atmosphere and infected cells were harvested and cell cycle analysis was performed on with adenoviral vectors diluted in DMEM/high glucose in the the single-cell level by measuring the DNA content of absence of FCS or antibiotics at the indicated MOI for 2 h at individual cells by flow cytometry as described (Daniel et al., 371C or were mocktreated with serum-free medium only. 1999). Analysis was carried out with a linear amplification in the FL-2 channel of a FACScan flow cytometer (BD- Pharmingen) equipped with the CellQuestPro software and Antibodies and reagents quantified by the use of ModFit. Apoptosis was determined on Mouse monoclonal anti-p14ARF antibody (clone 14P02, raised the single-cell level by measuring the DNA content of against full-length recombinant human p14ARF protein) was individual cells with a logarithmic amplification in the FL-3 purchased from NeoMarkers (Freemont, CA, USA). Anti- channel of a FACScan flow cytometer (BD-Pharmingen) bodies against p21 (clone 6B6, raised against a fusion protein equipped with the CellQuestPro software as described between GST and full-length human p21) and p53 (clone DO- (Hemmati et al., 2002). Data are given in percent hypoploidy 1, reactive against amino acids (aa) 1–45 of human p53) (i.e. the percentage of cells with a sub-G1 DNA content), were both murine monoclonal antibodies purchased from which reflects the percentage of apoptotic cells with fragmen- BD-Pharmingen (Heidelberg, Germany) used at dilution of ted genomic DNA. In addition, apoptosis was determined by 1 : 1000. An antibody against b-actin (Sigma-Aldrich, St Louis, flow-cytometric detection of Annexin-V-FITC (BD-Pharmin- MI, USA, reactive against an 11 aa C-terminal actin fragment) gen, San Diego, CA, USA) binding to phosphatidyl serine was a rabbit polyclonal antibody used at a dilution of 1 : 500. groups exposed at the outer leaflet of the cell membrane as All antibodies were diluted in phosphate-buffered saline (PBS) described (Scholz et al., 2002). Treatment with the topoisome- supplemented with 0.05% Tween-20 (PBS-T), 3% nonfat dry rase IIa inhibitor epirubicin (Pharmacia, Erlangen, Germany) milk, and 0.1% NaN3. Secondary goat-anti-mouse IgG and at a final concentration of 2 mg/ml was used as a positive goat-anti-rabbit IgG antisera coupled to horseradish perox- control for apoptosis induction. For inhibition of apoptosis, idase from Promega (Madison, WI, USA) were used at a the cell permeable broad-spectrum caspase inhibitor zVAD- dilution of 1 : 5000 in PBS-T. fmk(Bachem, Bubendorf, Switzerland) was added at a final

Oncogene p21 in p14ARF-induced cell cycle arrest and apoptosis PG Hemmati et al 4127 concentration of 40 mM to the individual samples 2 h after logies, MN, USA. Caspase-9-like caspase activities (LEH- adenoviral infection. Dases) were detected by using a cell-permeable, FAM-labeled derivative of LEHD-fmk(leucyl-glutamyl-histidyl-aspartyl- BrdU incorporation (O-methyl)-fluoromethyl-ketone) (Immunochemistry Techno- logies). After infection with recombinant adenovirus, cells were 5 Cells were seeded into six-well plates at 1 Â 10 cells/well, harvested and washed twice in PBS. Equal numbers of cells cultured overnight, and infected with adenoviral vectors. Prior (5 Â 105) were resuspended in 100 ml PBS supplemented with to harvest, cells were pulse labeled with BrdU (Sigma-Aldrich) FITC-VAD-fmk, FAM-DEVD-fmk, or FAM-LEHD-fmk at at a final concentration of 10 mM in the culture medium for a final concentration of 10 mM for 30 min at 371C with 30 min. Analysis of DNA synthesis, that is, incorporation of moderate shaking. Thereafter, cells were collected by centri- BrdU across the cell cycle, was performed as described (Craig fugation, washed twice with PBS, and finally resuspended in et al., 1998) using a fluorescein-isothiocyanate (FITC)-labeled 200 ml PBS. Caspase activation was quantified by flow- monoclonal antibody against BrdU (BD-Pharmingen). The cytometric detection of cells with increased fluorescence, that percentage of BrdU-positive cells is given and represents cells is, through binding of FITC-VAD-fmk, FAM-DEVD-fmk, or in S phase. FAM-LEHD-fmkto activated caspases in situ.

Detection of caspase activation on the single-cell level Measurement of caspase activation by the use of an FITC- Acknowledgements labeled conjugate of the cell-permeable pancaspase inhibitor This workwas supported by the Deutsche Krebshilfe Grant VAD-fmkfrom Promega was performed as described (Hem- 10-2088-Da3 to PTD and PGH. We would like to thank Antje mati et al., 2002). Activation of caspase-3/7-like caspases and Anja Richter and Jana Rossius for expert technical (DEVDases) was detected by the use of a carboxyﬂuorescein assistance. HCT116 cells were generously provided by Dr Bert (FAM)-labeled derivative of the cell permeable caspase-3/7- Vogelstein, Johns Hopkins Cancer Center, Baltimore, MD, like inhibitor DEVD-fmk from Immunochemistry Techno- USA.

References

Bissonnette N and Hunting DJ. (1998). Oncogene, 16, Hemmati PG, Gillissen B, von Haefen C, Wendt J, StarckL, 3461–3469. Gu¨ ner D, Do¨ rken B and Daniel PT. (2002). Oncogene, 21, Bunz F, Dutriaux A, Lengauer C, Waldman T, Zhou S, Brown 3149–3161. JP, Sedivy JM, Kinzler KW and Vogelstein B. (1998). Kokontis JM, Wagner AJ, O’Leary M, Liao S and Hay N. Science, 282, 1497–1501. (2001). Oncogene, 20, 659–668. Carnero A, Hudson JD, Price CM and Beach DH. (2000). Nat. Kuo ML, Duncavage EJ, Mathew R, den Besten W, Pei D, Cell Biol., 2, 148–155. Naeve D, Yamamoto T, Cheng C, Sherr CJ and Roussel Cleveland JL and Sherr CJ. (2004). Cancer Cell, 6, 309–311. MF. (2003). Cancer Res., 63, 1046–1053. Craig C, Kim M, Ohri E, Wersto R, Katayose D, Li Z, Choi Mahyar-Roemer M and Roemer K. (2001). Oncogene, 20, YH, Mudahar B, Srivastava S, Seth P and Cowan K. (1998). 3387–3398. Oncogene, 16, 265–272. Mao L, Merlo A, Bedi G, Shapiro GI, Edwards CD, Rollins Daniel PT. (2000). Leukemia, 14, 2035–2044. BJ and Sidransky D. (1995). Cancer Res., 55, 2995–2997. Daniel PT, Schulze-Osthoff K, Belka C and Gu¨ ner D. (2003). Martelli F, Hamilton T, Silver DP, Sharpless NE, Bardeesy N, Essays Biochem., 39, 73–88. Rokas M, DePinho RA, Livingston DM and Grossman SR. Daniel PT, Sturm I, Ritschel S, Friedrich K, Do¨ rken B, Bendzko (2001). Proc. Natl. Acad. Sci. USA, 98, 4455–4460. P and Hillebrand T. (1999). Anal. Biochem., 266, 110–115. Mason SL, Loughran O and La Thangue NB. (2002). de Stanchina E, McCurrach ME, Zindy F, Shieh SY, Ferbeyre Oncogene, 21, 4220–4230. G, Samuelson AV, Prives C, Roussel MF, Sherr CJ and Modestou M, Puig-Antich V, Korgaonkar C, Eapen A and Lowe SW. (1998). Genes Dev., 12, 2434–2442. Quelle DE. (2001). Cancer Res. 61, 3145–3150. Donato NJ and Perez M. (1998). J. Biol. Chem., 273, Normand G, Hemmati PG, Verdoodt B, von Haefen C, Wendt 5067–5072. J, Gu¨ ner D, May E, Do¨ rken B and Daniel PT. (2004). J. Duro D, Bernard O, Della Valle V, Berger R and Larsen CJ. Biol. Chem., 6 Dec 2004; [Epub ahead of print]. (1995). Oncogene, 11, 21–29. Palmero I, Pantoja C and Serrano M. (1998). Nature, 395, Eischen CM, Weber JD, Roussel MF, Sherr CJ and Cleveland 125–126. JL. (1999). Genes Dev., 13, 2658–2669. Pennington KN, Taylor JA, Bren GD and Paya CV. (2001). Eymin B, Karayan L, Seite P, Brambilla C, Brambilla E, Mol. Cell. Biol., 21, 1930–1941. Larsen CJ and Gazzeri S. (2001). Oncogene, 20, 1033–1041. Pomerantz J, Schreiber-Agus N, Liegeois NJ, Silverman A, Eymin B, Leduc C, Coll JL, Brambilla E and Gazzeri S. Alland L, Chin L, Potes J, Chen K, Orlow I, Lee HW, (2003). Oncogene, 22, 1822–1835. Cordon-Cardo C and DePinho RA. (1998). Cell, 92, 713–723. Fotedar R, Brickner H, Saadatmandi N, Rousselle T, Qi Y, Gregory MA, Li Z, Brousal JP, West K and Hann SR. Diederich L, Munshi A, Jung B, Reed JC and Fotedar A. (2004). Nature, 431, 712–717. (1999). Oncogene, 18, 3652–3658. Quelle DE, Zindy F, Ashmun RA and Sherr CJ. (1995). Cell, Gillissen B, Essmann F, Graupner V, StarckL, RadetzkiS, 83, 993–1000. Do¨ rken B, Schulze-Osthoff K and Daniel PT. (2003). Radfar A, Unnikrishnan I, Lee HW, DePinho RA and EMBO J., 22, 3580–3590. Rosenberg N. (1998). Proc. Natl. Acad. Sci. USA, 95, Groth A, Weber JD, Willumsen BM, Sherr CJ and Roussel 13194–13199. MF. (2000). J. Biol. Chem., 275, 27473–27480. Rau B, Sturm I, Lage H, Berger S, Schneider U, Hauptmann Gu¨ ner D, Belka C and Daniel PT. (2003). Curr. Med. Chem. S, Wust P, Riess H, Schlag PM, Do¨ rken B and Daniel PT. Anti-Cancer Agents, 3, 319–326. (2003). J. Clin. Oncol., 21, 3391–3401.

Oncogene p21 in p14ARF-induced cell cycle arrest and apoptosis PG Hemmati et al 4128 Rowland BD, Denissov SG, Douma S, Stunnenberg HG, Weber HO, Samuel T, Rauch P and FunkJO. (2002). Bernards R and Peeper DS. (2002). Cancer Cell., 2, 55–65. Oncogene, 21, 3207–3212. Schmitt CA, McCurrach ME, de Stanchina E, Wallace- Weber JD, Jeffers JR, Rehg JE, Randle DH, Lozano G, Brodeur RR and Lowe SW. (1999). Genes Dev., 13, Roussel MF, Sherr CJ and Zambetti GP. (2000). Genes Dev., 2670–2677. 14, 2358–2365. Scholz C, Starck L, Willimsky G, Blankenstein T, Do¨ rken B Yarbrough WG, Bessho M, Zanation A, Bisi JE and Xiong Y. and Daniel PT. (2002). Gene. Ther., 9, 1438–1446. (2002). Cancer Res., 62, 1171–1177. Sherr CJ. (2001). Nat. Rev. Mol. Cell. Biol., 2, 731–737. Zhang Y, Fujita N and Tsuruo T. (1999). Oncogene, 18, Stone S, Jiang P, Dayananth P, Tavtigian SV, Katcher H, 1131–1138. Parry D, Peters G and Kamb A. (1995). Cancer Res., 55, Zhu J, Jiang J, Zhou W and Chen X. (1998). Cancer Res., 58, 2988–2994. 5061–5065. Stott FJ, Bates S, James MC, McConnell BB, Starborg M, Zindy F, Eischen CM, Randle DH, Kamijo T, Cleveland Brookes S, Palmero I, Ryan K, Hara E, Vousden KH and JL, Sherr CJ and Roussel MF. (1998). Genes Dev., 12, Peters G. (1998). EMBO J., 17, 5001–5014. 2424–2433. Voorhoeve PM and Agami R. (2003). Cancer Cell, 4, 311–319. Zindy F, Williams RT, Baudino TA, Rehg JE, Skapek SX, Waldman T, Kinzler KW and Vogelstein B. (1995). Cancer Cleveland JL, Roussel MF and Sherr CJ. (2003). Proc. Natl. Res., 55, 5187–5190. Acad. Sci. USA, 100, 15930–15935.

Oncogene