P21 Loss Cooperates with INK4 Inactivation Facilitating

Research Article p21 Loss Cooperates with INK4 Inactivation Facilitating Immortalization and Bcl-2–Mediated Anchorage- Independent Growth of Oncogene-Transduced Primary Mouse Fibroblasts Christopher J. Carbone,1 Xavier Gran˜a,1,2 E. Premkumar Reddy,1,2 and Dale S. Haines1,2

1Fels Institute for Cancer Research and Molecular Biology and 2Department of Biochemistry, Temple University School of Medicine, Philadelphia, Pennsylvania

Abstract These kinase complexes phosphorylate pocket proteins pRb, p107, and p130, leading to release of the E2F family of transcriptional The INK4 and CIP cyclin-dependent kinase (Cdk) inhibitors regulators and transcription of genes required for cell cycle (CKI) activate pocket protein function by suppressing Cdk4 progression and DNA synthesis. Cdk activity is negatively regulated and Cdk2,respectively. Although these inhibitors are lost in by the INK4 (p16INK4A, p15INK4B, p18INK4C, and p19INK4D) and CIP/ tumors,deletion of individual CKIs results in modest Kip (p21Waf1/Cip1, p27Kip1, and p57Kip2) family of inhibitors (2). proliferation defects in murine models. We have evaluated Induction of these proteins by stress, growth-inhibitory, or cooperativity between loss of all INK4 family members (using differentiation-inducing signals provides a means of halting cell cdk4r24c mutant alleles that confer resistant to INK4 cycle progression to allow for repair or promoting exit from the cell inhibitors) and p21Waf1/Cip1 in senescence and transformation cycle. of mouse embryo fibroblasts (MEF). We show that mutant It was initially thought that Cdk4/Cdk6 and Cdk2 work cdk4r24c and p21 loss cooperate in pRb inactivation and MEF sequentially on pRb and play essential and nonredundant roles immortalization. Our studies suggest that cdk4r24c mediates in promoting cell cycle progression (3). However, the generation of resistance to p15INK4B/p16INK4A that accumulates over pas- knockout mouse models has challenged this notion and recent sage,whereas loss of p21 suppresses hyperoxia-induced Cdk2 results indicate that, at least for Cdk4 and Cdk2, they play inhibition and pRb dephosphorylation on MEF explantation in V12 overlapping roles in pRb phosphorylation and modulation of E2F- culture. Although cdk4r24c and p21 loss cooperate in H-ras / dependent gene expression in mouse fibroblasts (4). Cdk inhibitors c-myc–induced foci formation,they are insufficient for (CKI) from the INK4 and CIP/Kip families also possess overlapping oncogene-induced anchorage-independent growth. Interest- À/À V12 activities in regulating pocket protein function, although in ingly, p21 ; cdk4r24c MEFs expressing H-ras and c-myc contrast to Cdks, they promote pocket protein dephosphorylation display detachment-induced apoptosis and are transformed V12 and positively regulate their activities (5). by c-myc, H-ras ,and Bcl-2. We conclude that the INK4 Considering the fundamental roles of Cdks and CKIs in family and p21 loss cooperate in promoting pRb inactivation, V12 proliferation control, it is not surprising to find that the expression cell immortalization,and H-ras /c-myc–induced loss of and activity of these proteins are deregulated in cancer cells. For contact inhibition. In addition,absence of pRb function H-rasV12 c-myc example, human tumors frequently overexpress cyclin D1, Cdk4, renders + –transduced fibroblasts prone to and/or cyclin E. apoptosis when deprived of the extracellular matrix,and Conversely, expression of CKIs p15INK4B, p16INK4A, p18INK4C, and oncogene-induced anchorage-independent growth of pocket p27Kip1 are down-regulated in tumors (6). Consistent with the protein–deficient cells requires apoptotic suppression. [Cancer expression data, transgenic mice harboring elevated levels of cyclin Res 2007;67(9):4130–7] D1 or cyclin E are tumor prone, whereas those lacking individual CKIs are predisposed to cancer (7–9). For CKIs, however, the Introduction cancer predisposition phenotypes of single gene knockout are Quiescent cells require exposure to mitogenic growth factors to generally very mild and more dramatic phenotypes are observed in enter the cell cycle and pass through the restriction point. The key mice lacking multiple CKIs (6, 10–12). These studies reinforce the effectors promoting cell cycle entry and progression through the idea of functional overlap between CKI proteins and suggest that G1-Stransition are the cyclin-dependent kinase (Cdk) 4/Cdk6- inactivation or down-regulation of multiple CKIs will be required cyclin D and Cdk2-cyclin E kinase complexes (1). Cdk4/Cdk6 for cellular transformation and tumor development. associates with mitogen-induced D-type cyclins in early G1, Because of the difficulties in studying the cellular transformation whereas Cdk2 pairs with cyclin Es that accumulate in late G1. process in vivo, cell-based models using primary mouse embryo fibroblasts (MEF) have been extensively used to define genes regulating ‘‘normal’’ cell proliferation and how gene disruption affects various growth characteristics in culture. Although there are Note: Supplementary data for this article are available at Cancer Research Online clear differences between the proliferative characteristics of human (http://cancerres.aacrjournals.org/). and murine fibroblasts, studies have defined genes that are Requests for reprints: Dale S. Haines, Fels Institute for Cancer Research and Molecular Biology and Department of Biochemistry, Temple University School of required for the escape of culture-induced growth arrest Medicine, Philadelphia, PA 19140. Phone: 215-707-5765; Fax: 215-707-2102; E-mail: (or senescence), oncogene-induced proliferation, and anchorage- [email protected]. À/À I2007 American Association for Cancer Research. independent growth (13). For example, p53 MEFs do not V12 doi:10.1158/0008-5472.CAN-07-0499 undergo culture or oncogenic H-ras –induced proliferation arrest

Cancer Res 2007; 67: (9). May 1, 2007 4130 www.aacrjournals.org

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 2007 American Association for Cancer Research. cdk4r24c and p21 Loss Cooperate in MEF Transformation and are susceptible to H-rasV12–mediated anchorage-independent activity was determined from immunopurified complexes as described growth (14–16). Interestingly, deletion of pRb and p107, or all three previously (24). B h pocket proteins, allows MEFs to proliferate on introduction of Senescence-associated -galactosidase assay. -galactosidase activity mutant RASalone yet confers inefficient (at least when compared was detected as described previously (25) with slight modifications. MEFs À À from passage 10 of the 3T3 protocol were washed once with PBS, fixed with with p53 / MEFs) anchorage-independent proliferation after 0.5% glutaraldehyde (PBS), and washed in PBS supplemented with 1 mmol/L transduction with this and other oncogenes, including c-myc MgCl2. MEFs were stained with X-gal solution [1 mg/mL X-gal, 0.12 mmol/L (17). In terms of CKIs, MEFs lacking individual INKs or CIPs/Kips K3Fe(CN)6, 0.12 mmol/L Fe4(CN)6, and 1 mmol/L MgCl2 in PBS(pH 6.0)] at display only modest, if any, predisposition to immortalization in 37jC. V12 V12 culture, H-ras –induced proliferation, or H-ras + c-myc– Oncogene-induced foci and soft agar assays. Phoenix cells were mediated anchorage-independent growth (6, 10). Thus, a common transfected with 10 Ag of each retroviral plasmid using the CalPhos theme that emerges from these studies is that in murine Mammalian Transfection kit (BD Biosciences) and 25 Amol/L chloroquine fibroblasts, the CKIs are likely to play overlapping roles in (Sigma). The medium was changed 24 h after the transfection. Forty-eight suppressing culture and oncogene-induced proliferative arrest hours after transfection, the retrovirus-containing medium was collected, A and cells lacking pRb function will require inactivation or filtered, supplemented with 5 g/mL polybrene (Sigma), and used to infect early-passage MEFs growing in 3% oxygen/5% CO (done in duplicate for activation of other pathways to allow for oncogene-induced 2 each genotype). Infected cell populations were selected with 2 Ag/mL anchorage-independent proliferation. puromycin. For foci formation assays, 3 Â 105 puromycin-selected cells were Mutation of residue 24 from an arginine to cysteine in Cdk4 seeded into 10-cm plates. Cells were maintained for 10 days in a standard occurs in familial cases of melanomas and confers resistance of the tissue culture environment and the medium was changed every 2 days. For protein to inhibition by INK4 family members (18, 19). Mouse soft agar assays, MEFs were sequentially transduced with the indicated modeling experiments have shown that mice harboring a knock-in viruses, and puromycin-selected and 4 Â 104 cells were placed into 0.38% cdk4r24c mutation develop tumors at an accelerated rate (18–20). Nobel agar in DMEM containing 10% fetal bovine serum. This was overlaid Moreover, cdk4r24c MEFs are resistant to culture-induced senes- on solidified 0.7% Nobel agar containing culture medium in a 6-cm dish. cence and are prone to transformation by oncogenes. However, The cells were fed weekly and colonies were evaluated after 2 weeks. Viable these phenotypes are again mild, especially when compared with colonies were stained with 0.05% nitroblue tetrazolium. Detachment-induced apoptosis studies. The detachment-induced MEFs lacking all pocket proteins (21, 22), raising the possibility of apoptosis assays were done as described previously (26). Briefly, equal functional redundancy between INK4 proteins and the CIP family of numbers of cells were seeded onto 10-cm dishes containing solidified 0.7% inhibitors in these processes. The goal of studies described here was Nobel agar supplemented with culture medium. After 24 h, cells were either to investigate cooperativity between the cdk4r24c mutation and processed for Western blotting analysis using the anti-poly(ADP-ribose) loss of p21 in MEF senescence and oncogene-induced proliferation. polymerase (PARP) antibody or resuspended in 1Â binding buffer (PharMingen Annexin V-Phycoerythrin Apoptosis Detection kit). Cells were analyzed by flow cytometry. Materials and Methods Mice and MEF isolations. Mice harboring the cdk4r24c mutation (mixed 129SV, CD-1, and C57BL/6J) have been described previously (18). Results À/À À/+ p21 (FVB) and p53 mice (mixed 129SV and C57BL/6J) were obtained p21 deletion cooperates with cdk4r24c in suppression of from P. Leder (Harvard Medical School, Boston, MA) and S. Jones culture-induced senescence immortalization and pRb inacti- (University of Massachusetts, Worcester, MA), respectively. Homozygous À À À cdk4r24c were crossed with p21 / and p53 /+ animals. Pups were vation. The INK4 and CIP family of CKIs activate pocket protein genotyped by PCR and resulting heterozygous siblings were mated to function by suppressing the activities of Cdk4 and Cdk2, generate mice homozygous for one or two mutant alleles. MEFs were respectively. The goal of this study was to investigate cooperativity prepared from 13.5 days postcoitus embryos as described previously (23). between loss of INK4 family function (by using a knock-in mouse Cells were grown f1 to 2 days until the culture was 90% confluent, model where Cdk4 is replaced by a mutant cdk4r24c whose harvested, viably frozen, and labeled as passage 0. For experiments using encoded product is resistant to the action of all INK4 proteins) and À À cells cultured in 20% or 3% oxygen, torsos were dissected longitudinally and p21 (using p21 / mice) in MEF proliferation. It deserves to be cultured in either 3% oxygen/5% CO2 or 20% oxygen/5% CO2. noted that the other target of INK4 proteins, Cdk6, shows little, if Â 5 À À À À 3T3 assays. MEFs (3 10 ) of the indicated genotypes derived from any, expression in MEFs, and in contrast to Cdk2 / ;Cdk4 / passage 0 were plated in 10-cm dishes, grown for 3 days, trypsinized, and À/À À/À 5 MEFs, Cdk2 ;Cdk6 MEFs do not display cell cycle defects over counted, and 3 Â 10 cells were reseeded. This was repeated for each À À over Cdk2 / MEFs (1). Mice harboring mutant cdk4r24c alleles were passage. À/À À/À À/À Western blotting. Proteins were resolved by SDS-PAGE, transferred to bred with p21 mice to generate p21 ; cdk4r24c animals. p21 ; membranes, and probed with commercially available antibodies (see cdk4r24c MEFs were prepared from subsequent matings and Supplementary Materials and Methods). Proteins (except for p16INK4A) their growth characteristics in culture were compared with cdk4r24c, À À were detected using secondary antibodies [horseradish peroxidase (HRP)– p21 / , and WT fibroblasts (these experiments were done with MEFs linked sheep anti-mouse IgG (Amersham) for monoclonal antibodies and isolated from two independent embryos per genotype) using a 3T3 HRP-linked sheep anti-goat IgG (Chemicon) for the anti-actin polyclonal protocol. We used a 3T3 protocol (i.e., 3 Â 105 cells seeded every antibody] and reagents provided in a Western Lighting Chemiluminescence INK4A 3 days in 10-cm dishes) where cells are seeded at 30% to 40% Reagent kit from Amersham. p16 was detected using an antimouse confluency. This was done to accentuate any proliferation differences secondary antibody and solutions were provided in the Lumi-LightPLUS between the various genotypes that could be masked by MEFs Western blotting kit (Roche). Cell cycle analysis and Cdk2 kinase assays. Cells were harvested, fixed reaching contact inhibition. As shown in Fig. 1A, WT MEFs undergo with 70% ethanol, and incubated on ice for an hour. After centrifugation, senescence very rapidly on low-density plating and we terminated pellets were treated with 20 mg/mL RNase A and stained with 5 mg/mL culturing the cells when the number of cells harvested at day 3 was À/À propidium iodide. DNA content was determined using a flow cytometer and less than that initially plated. Interestingly, p21 MEFs possess an samples were analyzed with ModFit LT 3.1 SP2 software. Cdk2 kinase increased proliferative capacity over WT and even cdk4r24c cells at www.aacrjournals.org 4131 Cancer Res 2007; 67: (9). May 1, 2007

Figure 1. p21 deletion extends MEF life span in culture and cooperates with mutant Cdk4 in suppression of culture-induced senescence. A, MEFs of the indicated genotypes (each derived from two independent embryos)were propagated in culture by seeding 3 Â 105 cells every 3rd day into 10-cm dishes. N3/NO, the number of cells present 3 d after seeding/3 Â 105. B, an in situ h-galactosidase assay was done with passage 10 MEFs of the indicated genotypes.

low passages when cells are seeded at low densities. However, their promoting pocket protein inactivation and escape from culture- proliferation rate declined over increasing passage. cdk4r24c MEFs induced senescence and immortalization. Moreover, unlike immor- À À À À also possess an extended life span when compared with WT MEFs, talized p21 / (10) and cdk4r24c MEFs, immortal p21 / ; cdk4r24c À À and unlike p21 / MEFs, their growth rate is consistent until passage MEFs retain WT p53 alleles and p19ARF (see Supplementary Fig. S1), 20, where immortalized variants begin to dominate the culture.3 indicating that p53 functions upstream of both Cdk2 and Cdk4. À À Figure 1A also shows that p21 / ; cdc4r24c MEFs do not undergo a Mutant Cdk4 renders cells resistant to p15INK4B and p16INK4A significant crises period, display a marked increased proliferative that accumulate on cell passaging,whereas loss of p21 Waf1/Cip1 capacity at all passages, and proliferate indefinitely. In addition, suppresses hyperoxia-induced Cdk2 down-regulation and h-galactosidase staining (a senescence marker) of cells at passage 10 pRb activation. We next investigated the mechanism of cooper- À À shows little, if any, h-galactosidase–positive cells in p21 / ; cdk4r24c ativity between the cdk4r24c mutation and p21 loss in MEF MEFs, whereas numerous h-galactosidase–positive cells were proliferation. It is well established that the levels of p15INK4B and À À detected in p21 / and cdk4r24c cultures (Fig. 1B). p16INK4A accumulate as WT MEF cells are propagated in culture À À À À To determine the affects of combined INK4 family and (27, 28). This is also the case for p21 / , cdk4r24c, and p21 / ; p21Waf1/Cip1 loss on pRb expression and activity over passage, we cdk4r24c MEFs (see Fig. 3A). Moreover, accumulation of p15INK4B assessed pRb phosphorylation status and levels of E2F target gene and p16INK4A is associated with a sharp decline (from passage 1 À À products cyclin A, p107, and E2F1 in WT, p21 / , cdk4r24c, and to 15) in pRb phosphorylation on a Cdk4-specific site (29, 30) in À À À À p21 / ; cdk4r24c MEFs. As shown in Fig. 2, the ratio of hyper- p21 / MEFs. Only a slight decrease was observed in cdk4r24c À À phosphorylated to hypophosphorylated pRb (as better evidenced by MEFs and no decline was evident in p21 / ; cdk4r24c MEFs (Fig. 3B). À À the short exposure) was dramatically elevated in p21 / ; cdk4r24c These, as well as the biological data presented in Fig. 1, suggest that À À MEFs when compared with p21 / and cdk4r24c fibroblasts, cdk4r24c renders MEFs resistant to the action of INK4 proteins that especially at later passages. We also noticed elevated levels of total accumulate on successive passaging of cells in culture. À À pRb in late passage p21 / ; cdk4r24c MEFs. The reason for the In contrast to p15INK4B and p16INK4A, p21Waf1/Cip1 is modestly increase is unclear and under investigation. Nevertheless, Fig. 2 induced, if at all, as MEFs are passaged in culture (10) and this is shows that hyperphosphorylated pRb correlates with higher levels also the case in cells harboring cdk4r24c (Fig. 3A). Thus, it is of E2F-regulated proteins cyclin A, p107, and E2F1 at passage 5 and possible that p21Waf1/Cip1 activity, but not expression, is differen- À À later in p21 / ; cdk4r24c/r24c MEFs. We conclude from these tially regulated. However, a previous study showed that p53- studies that loss of INK4 family and p21Waf1/Cip1 cooperates in responsive transcripts of the MDM2 promoter are robustly induced in WT MEFs immediately after placement in tissue culture, peak at or around passage one, and remain high over passaging (31). Considering this result and that p21 is also a p53 target gene (32), Waf1/Cip1 3 C.J. Carbone and D.S. Haines, unpublished data. we tested if p21 is induced on MEF explantation into tissue

Cancer Res 2007; 67: (9). May 1, 2007 4132 www.aacrjournals.org

Figure 2. p21 deletion cooperates with mutant Cdk4 in suppression of culture-induced pocket protein inactivation. Levels of pRb (both long and short exposures of the film are presented)and E2F target genes p107, E2F1, and cyclin A in the indicated MEFs and passages were determined by Western blotting. Hypophosphorylated and hyperphosphorylated of pRb (pRb-P). The actin blot was used as a loading control.

culture. As shown in Fig. 4A, p21Waf1/Cip1 levels markedly increase these two different oxygen concentrations. As shown in Fig. 4C, on introduction of cells into the standard tissue culture p21Waf1/Cip1 increases when MEFs are cultured at 20%, but not 3% environment. In contrast, the expression of other CKIs remains oxygen. In addition, we found that WT MEFs cultured in 3% oxygen constant. To verify that p21Waf1/Cip1 induction during this process displayed increased (a) proliferative capacity (Fig. 4D), (b) number Waf1/Cip1 is p53 dependent, we measured the amount of p21 in of cells in the S/G2-M phases of the cell cycle (Fig. 4E), (c) Cdk2 À À p53 / MEFs after explantation. As depicted in Fig. 4B, p21Waf1/Cip1 kinase activity (Fig. 4F), and (d) pRb phosphorylation and the À À is not induced in p53 / MEFs when placed into culture. We expression of E2F target gene products (Fig. 4G) when compared À À conclude from these experiments that p21Waf1/Cip1 is up-regulated with cells grown in 20% oxygen. In contrast to WT MEFs, p21 / via a p53-dependent manner on explantation of MEFs into cell MEFs displayed increases in these measured variables, regardless of culture. whether they were cultured in 3% or 20% oxygen. These studies Primary cells encounter oxidative stress when placed into the suggest that p21 loss confers resistance to hyperoxia-induced Cdk2 standard tissue culture environment of 20% oxygen. It is well down-regulation and pRb activation that occurs on explanation of known that p53 activity is induced by oxidative damage and recent MEFs in culture. work has shown that culturing of MEFs in atmospheric oxygen p21 loss cooperates with cdk4r24c in oncogene-induced foci promotes DNA damage and senescence (33). To determine if formation. Another characteristic of cells with altered prolifera- p21Waf1/Cip1 is differentially regulated when MEFs are cultured in tion control characteristics is that they undergo deregulated 20% atmospheric versus normoxic 3% oxygen, Western blotting was growth (as assessed by loss of contact inhibition or anchorage- done with extracts prepared from early-passage MEFs grown at independent growth) in response to oncogenes. For example,

Figure 3. Expression levels of p15INK4B, p16INK4A, p21Waf1/Cip1 and phosphorylated pRb at Thr826 in MEFs defective in INK4 and/or p21Waf1/Cip1 function. A, Western blotting for p21Waf1/Cip1, p16INK4A, and p15INK4B was done with extracts prepared from the indicated genotypes and passage. B, pRb phosphorylation at Thr826 was determined by Western blotting. The amount of extract analyzed was the same as in the previous figure where equivalent loading was verified.

www.aacrjournals.org 4133 Cancer Res 2007; 67: (9). May 1, 2007

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 2007 American Association for Cancer Research. Cancer Research introduction of activated Ras (i.e., H-rasV12) into primary WT Bcl-2 confers anchorage-independent growth of oncogene- À/À fibroblasts induces senescence, although it promotes loss of transduced p21 ; cdk4r24c MEFs. Besides doing foci formation contact inhibition in cells lacking pocket protein function and assays, we also tested if stably transduced H-rasV12 and c-myc À À both loss of contact inhibition and anchorage-independent growth p21 / ; cdk4r24c MEFs are capable of anchorage-independent À À À À in p53 / MEFs (17, 34). In addition, p21 / MEFs lose contact growth. Growth in soft agar was not apparent in H-rasV12 + c-myc– À À À À inhibition when transduced with H-rasV12 and c-myc oncogenes transduced p21 / ; cdk4r24c, p21 / ,orcdk4r24c MEFs À À À À (35). Thus, we next determined how p21 / ; cdk4r24c MEFs (see Supplementary Fig. S2). In contrast, control p53 / MEFs respond upon H-rasV12 and c-myc transduction. As shown in formed colonies in soft agar on introduction of H-rasV12 and À À Fig. 5, the proliferation of p21 / ; cdk4r24c MEFs transduced with H-rasV12 + c-myc (Supplementary Fig. S2). Interestingly, when we H-rasV12 was not restrained when cells reached confluency and looked at these plates, 1 week after seeding, we noticed colonies of cells continued to proliferate to form multiple layers or foci. very small size (8–10 cells) and the comprised cells displayed visible À À Similarly, p21 / ; cdk4r24c MEFs transduced with c-myc alone apoptotic characteristics.3 Based on this observation, we wanted to À À resulted in foci composed of rounded, refractory cells (Fig. 5). test if oncogene-transduced p21 / ; cdk4r24c cells display À À p21 / ; cdk4r24c cells containing c-myc and H-rasV12 exhibited a increased apoptosis when deprived of attachment and if suppres- profound increase in proliferation, as evidenced by the presence of sion of an apoptotic response via the introduction of the multiple layers of cells in over 50% of the tissue culture dish (Fig. 5). antiapoptotic oncogene Bcl-2 enhances transformation by À À These phenotypes were not observed in any of the WT, p21 / , and H-rasV12 + c-myc. Figure 6A shows that when compared with À À cdk4r24c transductions with individual or both oncogenes. These WT MEFs, p21 / ; cdk4r24c MEFs transduced with H-rasV12 + results show that loss of INK4 and p21Waf1/Cip1 cooperate in c-myc undergo increase apoptosis as assessed by PARP cleavage oncogene-induced loss of contact inhibition. (Fig. 6A, left) and Annexin staining (Fig. 6A, right) when cells are

Figure 4. p21Waf1/Cip1 is induced via a p53-dependent mechanism and p21 loss confers resistance to hyperoxia-induced Cdk2 down-regulation and pRb activation on explantation of MEFs into culture. A, CKI levels were assessed by Western blotting before placing cells in culture (0) and 24 and 48 h later. B, p21Waf1/Cip1 and actin were measured by Western blotting before (0)and 4 and 8 h after being placed into culture. C, p21Waf1/Cip1 and actin were assessed by Western blotting in MEFs before placing cells in culture (0) and at the indicated time points after exposure to 3% or 20% oxygen. D, passage 1 MEFs (derived from two independent embryos)were plated at 2 Â 105 cells per 60-mm diameter dish (done in triplicate)and grown in 20% or 3% oxygen. Cells were counted 1, 2, and 3 d after plating. E, fluorescence-activated cell sorting analysis of DNA content was done with the cells harvested 3 d after plating. F, Cdk2 kinase activity was measured in WT and p21À/À MEFs 3 d after platting. Bottom, input of histone H1 and equivalent amounts of IgG in each immunoprecipitation. G, pRb phosphorylation status and levels of E2F-responsive proteins cyclin A, p107, and E2F1 were assessed by Western blotting using the same extracts used in the kinase assays. All of the above described experiments were done with MEFs from multiple embryos (at least three from each genotype)and show reproducible results.

Cancer Res 2007; 67: (9). May 1, 2007 4134 www.aacrjournals.org

Figure 5. Loss of p21Waf1/Cip1 and INK4 function allows for unrestrained proliferation in MEFs on oncogene transduction. MEFs of the indicated genotypes were transduced with a recombinant H-rasV12 and/or c-myc retroviruses. Cells were seeded and grown in standard tissue culture conditions for 10 d. Black outline, cells that are present in layers.

À À deprived of anchorage. Moreover, p21 / ; cdk4r24c fibroblasts are cells resistant to p15INK4B and p16INK4A that are up-regulated as efficiently transformed on sequential transduction with c-myc, cells are propagated in culture. However, these cells are sensitive to HrasV12, and Bcl-2 (Fig. 6B). These results suggest that suppression the growth-inhibitory activity of p21Waf1/Cip1 that is induced by À À of apoptotic pathways is required for oncogene-induced anchor- culturing of cells in atmospheric oxygen. In contrast, p21 / MEFs age-independent proliferation of cells lacking INK4 and p21Waf1/Cip1 are resistant to hyperoxia-induced inhibition and proliferate well at functions. early passages. Yet, they likely become sensitive to the action of high levels of p15INK4B and p16INK4A at later passages and lose their proliferative capacity. Cells lacking both INK4 and p21Waf1/Cip1 Discussion functions are resistant to growth-inhibitory signals that arise early In this study, we show for the first time the cooperativity between and late during the culturing process; thus, they display an overall loss of INK4 family function (via the use of cells that harbor the increased proliferative and immortalization rates over cells that cdk424c mutation) and p21 in culture-induced senescence, pRb harbor single alterations. This model does not rule out the possi- inactivation, and immortalization of MEFs. In addition, we show bility that p18INK4C, p19INK4D, and even p27Kip1 also cooperate in that p21Waf1/Cip1 is robustly induced via a p53-dependent manner the culture-induced senescence process because these proteins are on explantation of MEFs into the standard tissue culture expressed in MEFs. Notwithstanding, it will be of interest to define environment, but not when MEFs are grown in physiologic oxygen how deletion of individual INK4s and CIPs in varying combinations concentration. The atmospheric hyperoxia-induced p21Waf1/Cip1 affects MEF proliferation. The large number of INK4s and CIPs and expression inhibits Cdk2 activity, pRb phosphorylation, and their disparity in regulation provide a means for precisely E2F-dependent gene expression. In addition, although INK4 and regulating proliferation rates in context of numerous signals. p21 loss also cooperate in H-rasV12/c-myc–induced foci formation, The ‘‘cooperativity-based on expression model’’ may explain why these alterations are not sufficient to confer H-rasV12 + c-myc– synergy of loss between various INK4s and CIPs is tissue specific induced anchorage-independent growth, as lack of attachment (6). It could also reconcile differences in the relative contribution of triggers apoptosis. In keeping with these observations, INK4 and individual proteins to the senescence process in human versus À À p21 / MEFs are transformed by expression of H-rasV12, c-myc, murine cells and even the disparate results obtained when the cells and the antiapoptotic oncogene Bcl-2. of the same type and species are cultured differently. For example, These studies provide a working model for mechanisms of although the cdk4r24c mutation promotes escape from senescence cooperativity between INK4 and CIP family members in regulation in murine cells, it does not have a major affect in human cells (36). of pocket protein function and perhaps more generalized In contrast, p21 loss has a dramatic effect on senescence in human proliferation control. Because INK4s (via down-regulation of fibroblasts (37). Considering that the stress signals that drive Cdk4/Cdk6) and CIPs (via suppression of Cdk2) work upstream senescence in human versus mouse cells are quite different, it is of pocket proteins, cooperativity between individual CKIs on likely that the relative contribution of CKIs are dictated by the type regulating their function will be defined by their relative and strength of stress signals that induce their expression in expression/activity within individual cell types or during specific human and mouse cells. Similarly, whereas a previous study has À À biological processes. For example, we propose based on our shown that p21 / MEFs proliferate similarly to WT MEFs using a expression studies that absence of INK4 activity renders cdk4r24c 3T3 protocol where cells are seeded at high densities, we show here www.aacrjournals.org 4135 Cancer Res 2007; 67: (9). May 1, 2007

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 2007 American Association for Cancer Research. Cancer Research using a different 3T3 protocol where cells are seeded at lower models, it is tempting to speculate that pocket protein loss renders densities that p21 loss promotes proliferation. Thus, it is likely that H-rasV12 + c-myc–transduced fibroblasts prone to apoptosis when p21Waf1/Cip1 plays a more dominant role (either alone or in deprived of signals provided by the extracellular matrix. If this cooperation with other factors that are differentially present) when speculation is correct, it would explain why pocket protein– À À cells are seeded at lower confluency. deficient fibroblasts are not as sensitive as p53 / MEFs (defective Besides showing cooperativity in suppression of MEFs senes- on both G1-Scheckpoint and apoptotic pathways) to oncogene- cence, we also show that cdk4r24c and p21 loss cooperate in induced anchorage-independent proliferation. oncogene-induced proliferation and allow cells to overcome Interestingly, p21Waf1/Cip1 has been implicated as both a positive contact inhibition. Previous studies have shown that H-rasV12 up- and a negative regulator of cell proliferation. In addition, mouse regulates INK4s and p21Waf1/Cip1, raising the likelihood that loss of modeling experiments have convincingly showed that p21 loss can À À the function of the proteins in p21 / ; cdk4r24c MEFs promotes accelerate or delay tumorigenesis (38). Results presented here H-rasV12–induced proliferation (34). In addition, H-rasV12 has been suggest that under conditions where the actions of the INK4 family À À shown to override c-myc–induced apoptosis in p21 / MEFs, of inhibitors are perturbed, p21Waf1/Cip1 functions as a potent À À pointing to synergy between H-rasV12 + c-myc in proliferation of growth inhibitor. Interestingly, p21 / ; cdk4r24c MEFs do display À À À À p21 / ; cdk4r24c MEFs (35). Interestingly and in contrast to p53 / an increased apoptotic index compared with WT MEFs when À À MEFs, H-rasV12 + c-myc–transduced p21 / ; cdk4r24c MEFs do not deprived of attachment in the presence of oncogenes and studies of À À À À proliferate in soft agar. These results are similar to what has been MEFs deficient of pRb / , p107 / , p130 have shown that they reported with MEFs lacking multiple pocket proteins (17). Thus, undergo increased apoptosis on serum withdrawal (39). Thus, although deregulation of pRb or pocket protein function (via loss of although it is conceivable in certain situations, p21Waf1/Cip1 does the proteins themselves or the function of upstream regulators, not directly promote proliferation, but absence of its function leads such as those reported here) confers loss of contact inhibition by to a defective G1-Scheckpoint and enhances cell death to stimuli, these oncogenes, our results are consistent with the idea that resulting in a lower tumor incidence. An obvious question that will pocket protein loss is not sufficient for oncogene-induced arise from these studies is will the major conclusions put forth be anchorage-independent proliferation. Data presented here indicate restricted to cell culture models. It is well appreciated that genes À À that p21 / ; cdk4r24c cells harboring H-rasV12 + c-myc are prone to (e.g., p53 and pRb) involved in mediating culture-induced detachment-induced apoptosis and are transformed on cotrans- proliferative arrest and suppression of oncogene-mediated transduction with Bcl-2. Although the mechanisms of cooperativity formation also possess critical tumor suppressor properties in vivo. between cdk4r24c, p21 loss, Bcl-2, and H-rasV12 + c-myc obviously As mentioned above, p21 disruption does predispose mice to need to be better defined and extended to other pRb-deficient tumors late in life and facilitates tumorigenesis provoked by

Figure 6. p21À/À; cdk4r24c MEFs undergo enhanced detachment-induced apoptosis and proliferate in soft agar on transduction with c-myc, H-rasV12, and Bcl-2. A, MEFs stably transduced with H-rasV12 and c-myc oncogenes were harvested and either pelleted (0)or placed in dishes containing solidified Nobel agar for 24 h. MEFs were collected and apoptosis was assessed by PARP cleavage or Annexin V positivity. B, p21À/À; cdk4r24c and WT MEFs were sequentially transduced with empty vector (P)or the indicated oncogenes. After selection, cells were suspended in Nobel agar and the colonies were visualized by staining 2 wks later.

Cancer Res 2007; 67: (9). May 1, 2007 4136 www.aacrjournals.org

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 2007 American Association for Cancer Research. cdk4r24c and p21 Loss Cooperate in MEF Transformation oncogene activation and loss of other tumor suppressors, including Acknowledgments pRb (38, 40, 41). In addition, mice harboring the cdk4r24c mutation are prone to a wide range of tumors and display cooperativity with Received 2/5/2007; accepted 2/14/2007. Grant support: AG22022 (E.P. Reddy) and CA095569 (E.P. Reddy, X. Gran˜a, and D.S. p27 deletion in formation of pituitary tumors (18, 19). In addition, Haines). C.J. Carbone is a recipient of Department of Defense Breast Cancer Research loss of both INK4 (via mutation or promoter methylation) and Program predoctoral fellowship grant DAMD17-03-1-0412. Tobacco Settlement Funds, p21Waf1/Cip1 (via p53-dependent and p53-independent mechanisms) Fels Foundation, and the W.W. Smith Charitable Trust (D.S. Haines). The costs of publication of this article were defrayed in part by the payment of page function is a relatively common event in human malignancies. charges. This article must therefore be hereby marked advertisement in accordance Thus, we predict based on these data and the ex vivo studies with 18 U.S.C. Section 1734 solely to indicate this fact. À À À presented here that cdk4r24c and p21 loss will cooperate in We thank P. Leder for p21 / mice, Steve Jones for p53 /+ mice, Scott Lowe (Cold Spring Harbor Laboratories, Cold Spring Harbor, NY) for pBabe-puro-H-rasV12, promoting tumorigenesis in vivo, and this will be accelerated in D. Liebermann (Temple University, Philadelphia, PA) for MSCVpac, and G. Nolan models that are deficient in apoptotic pathways. (Stanford University, Stanford, CA) for Phoenix cells.

References 15. Harvey M, Sands AT, Weiss RS, et al. In vitro growth 29. Paternot S, Arsenijevic T, Coulonval K, Bockstaele L, characteristics of embryo fibroblasts isolated from p53- Dumont JE, Roger PP. Distinct specificities of pRb 1. Malumbres M, Barbacid M. Mammalian cyclin-depen- deficient mice. Oncogene 1993;8:2457–67. phosphorylation by CDK4 activated by cyclin D1 or dent kinases. Trends Biochem Sci 2005;30:630–41. 16. Sherr CJ. Tumor surveillance via the ARF-p53 cyclin D3: differential involvement in the distinct 2. Sherr CJ, Roberts JM. CDK inhibitors: positive and pathway. Genes Dev 1998;12:2984–91. mitogenic modes of thyroid epithelial cells. Cell Cycle negative regulators of G1 phase progression. Genes Dev 17. Peeper DS, Dannenberg JH, Douma S, te Riele H, 2006;5:61–70. 1999;13:1501–12. Bernards R. Escape from premature senescence is not 30. Takaki T, Fukasawa K, Suzuki-Takahashi I, et al. 3. Sherr CJ, Roberts JM. Living with or without cyclins sufficient for oncogene transformation by Ras. Nat Cell Preferences for phosphorylation sites in the retinoblas- and cyclin-dependent kinases. Genes Dev 2004;18: Biol 2001;3:198–203. toma protein of D-type cyclin-dependent kinases, Cdk4 2699–711. 18. Rane SG, Cosenza SC, Mettus RV, Reddy EP. Germ and Cdk6, in vitro. J Biochem 2005;137:381–6. 4. Berthet C, Klarmann KD, Hilton MB, et al. Combined line transmission of the Cdk4 (R24C) mutation facili- 31. Mendrysa SM, Perry ME. The p53 tumor suppressor loss of Cdk2 and Cdk4 results in embryonic lethality and tates tumorigenesis and escape from cellular senes- protein does not regulate expression of its own Rb hypophosphorylation. Dev Cell 2006;10:563–73. cence. Mol Cell Biol 2002;22:644–56. inhibitor, MDM2, except under conditions of stress. 5. Grana X, Garriga J, Mayol X. Role of the retinoblas- 19. Sotillo R, Dubus P, Martin J, et al. Wide spectrum of Mol Cell Biol 2000;20:2023–30. toma protein family, pRB, p107, and p130 in the negative tumors in knock-in mice carrying a Cdk4 protein 32. el-Deiry WS, Tokino T, Velculescu VE, et al. WAF1, a control of cell growth. Oncogene 1998;17:3365–83. insensitive to INK4 inhibitors. EMBO J 2001;20:6637–47. potential mediator of p53 tumor suppression. Cell 1993; 6. Ortega S, Malumbres M, Barbacid M. Cyclin D- 20. Sotillo R, Garcia JF, Ortega S, et al. Invasive 75:817–25. dependent kinases, INK4 inhibitors, and cancer. Bio- melanoma in Cdk4-targeted mice. Proc Natl Acad Sci 33. Parrinello S, Samper E, Krtolica A, Goldstein J, Melov chim Biophys Acta 2002;1602:73–87. U SA 2001;98:13312–7. S, Campisi J. Oxygen sensitivity severely limits the 7. Bortner DM, Rosenberg MP. Induction of mammary 21. Dannenberg JH, van Rossum A, Schuijff L, te Riele H. replicative lifespan of murine fibroblasts. Nat Cell Biol gland hyperplasia and carcinomas in transgenic mice Ablation of the retinoblastoma gene family deregulates 2003;5:741–7. expressing human cyclin E. Mol Cell Biol 1997;17: G(1) control causing immortalization and increased cell 34. Serrano M, Lin AW, McCurrach ME, Beach D, Lowe 453–9. turnover under growth-restricting conditions. Genes SW. Oncogenic ras provokes premature cell senescence 8. Robles AI, Larcher F, Whalin RB, et al. Expression of Dev 2000;14:364. associated with accumulation of p53 and p16INK4a. Cell cyclin D1 in epithelial tissues of transgenic mice 22. Sage J, Mulligan GJ, Attardi LD, et al. Targeted 1997;88:593–602. results in epidermal hyperproliferation and severe disruption of the three Rb-related genes leads to loss of 35. Carnero A, Beach DH. Absence of p21WAF1 cooper- thymic hyperplasia. Proc Natl Acad Sci U S A 1996;93: G(1) control and immortalization. Genes Dev 2000;14: ates with c-myc in bypassing Ras-induced senescence 7634–8. 3037–50. and enhances oncogenic cooperation. Oncogene 2004; 9. Wang TC, Cardiff RD, Zukerberg L, Lees E, Arnold 23. KamijoT,ZindyF,RousselMF,etal.Tumor 23:6006–11. A, Schmidt EV. Mammary hyperplasia and carcinoma suppression at the mouse INK4a locus mediated by 36. Wei W, Herbig U, Wei S, Dutriaux A, Sedivy JM. Loss in MMTV-cyclin D1 transgenic mice. Nature 1994;369: the alternative reading frame product p19ARF. Cell of retinoblastoma but not p16 function allows bypass of 669–71. 1997;91:649–59. replicative senescence in human fibroblasts. EMBO Rep 10. Pantoja C, Serrano M. Murine fibroblasts lacking p21 24. Calbo J, Parreno M, Sotillo E, et al. G1 cyclin/cyclin- 2003;4:1061–6. undergo senescence and are resistant to transformation dependent kinase-coordinated phosphorylation of en- 37. Brown JP, Wei W, Sedivy JM. Bypass of senescence by oncogenic Ras. Oncogene 1999;18:4974–82. dogenous pocket proteins differentially regulates their after disruption of p21CIP1/WAF1 gene in normal 11. Nakayama K, Ishida N, Shirane M, et al. Mice lacking interactions with E2F4 and E2F1 and gene expression. diploid human fibroblasts. Science 1997;277:831–4. p27(Kip1) display increased body size, multiple organ J Biol Chem 2002;277:50263–74. 38. Rowland BD, Peeper DS. KLF4, p21 and context- hyperplasia, retinal dysplasia, and pituitary tumors. Cell 25. Dimri GP, Lee X, Basile G, et al. A biomarker that dependent opposing forces in cancer. Nat Rev Cancer 1996;85:707–20. identifies senescent human cells in culture and in aging 2006;6:11–23. 12. Kiyokawa H, Kineman RD, Manova-Todorova KO, skin in vivo. Proc Natl Acad Sci U S A 1995;92:9363–7. 39. Foijer F, Wolthuis RM, Doodeman V, Medema RH, te et al. Enhanced growth of mice lacking the cyclin- 26. Zugasti O, Rul W, Roux P, et al. MEK-Erk cascade in Riele H. Mitogen requirement for cell cycle progression dependent kinase inhibitor function of p27(Kip1). Cell anoikis is controlled by Rac1 and Cdc42 via Akt. Mol in the absence of pocket protein activity. Cancer Cell 1996;85:721–32. Cell Biol 2001;21:6706–17. 2005;8:455–66. 13. Itahana K, Campisi J, Dimri GP. Mechanisms of 27. Palmero I, McConnell B, Parry D, et al. Accumulation 40. Brugarolas J, Bronson RT, Jacks T. p21 is a critical cellular senescence in human and mouse cells. Bioger- of p16INK4a in mouse fibroblasts as a function of CDK2 regulator essential for proliferation control in Rb- ontology 2004;5:1–10. replicative senescence and not of retinoblastoma gene deficient cells. J Cell Biol 1998;141:503–14. 14. Harvey DM, Levine AJ. p53 alteration is a common status. Oncogene 1997;31:495–503. 41. Martin-Caballero J, Flores JM, Garcia-Palencia P, event in the spontaneous immortalization of primary 28. Gil J, Peters G. Regulation of the INK4b-ARF-INK4a Collado M, Serrano M. Different cooperating effect of BALB/c murine embryo fibroblasts. Genes Dev 1991;5: tumour suppressor locus: all for one or one for all. Nat p21 or p27 deficiency in combination with INK4a/ARF 2375–85. Rev Mol Cell Biol 2006;7:667–77. deletion in mice. Oncogene 2004;23:8231–7.

www.aacrjournals.org 4137 Cancer Res 2007; 67: (9). May 1, 2007

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 2007 American Association for Cancer Research. p21 Loss Cooperates with INK4 Inactivation Facilitating Immortalization and Bcl-2−Mediated Anchorage-Independent Growth of Oncogene-Transduced Primary Mouse Fibroblasts

Christopher J. Carbone, Xavier Graña, E. Premkumar Reddy, et al.

Cancer Res 2007;67:4130-4137.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/67/9/4130

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2007/05/01/67.9.4130.DC1

Cited articles This article cites 41 articles, 17 of which you can access for free at: http://cancerres.aacrjournals.org/content/67/9/4130.full#ref-list-1

Citing articles This article has been cited by 1 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/67/9/4130.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/67/9/4130. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.