Although the p16INK4a and p21Waf1/Cip1 cyclin-dependent kinase (CDK) inhibitors are known to play key roles in cellular senescence in vitro, their roles in senescence remain rather poorly understood in vivo. This situation is partly due to the possibility of compensatory effect(s) between p16INK4a and p21Waf1/Cip1 or to the upregulation of functionally related CDK inhibitors. To directly address the cooperative roles of p16INK4a and p21Waf1/Cip1 in senescence in vivo, we generated a mouse line simply lacking both p16INK4a and p21Waf1/Cip1 genes [double-knockout (DKO)]. Mouse embryonic fibroblasts (MEF) derived from DKO mice displayed no evidence of cellular senescence when cultured serially in vitro. Moreover, DKO MEFs readily escaped Ras-induced senescence and overrode contact inhibition in culture. This was not the case in MEFs lacking either p16INK4a or p21Waf1/Cip1, indicating that p16INK4a and p21Waf1/Cip1 play cooperative roles in cellular senescence and contact inhibition in vitro. Notably, we found the DKO mice to be extremely susceptible to 7,12-dimethylbenz(a)anthracene/12-O-tetradecanoylphorbol-13-acetate–induced skin carcinogenesis that involves oncogenic mutation of the H-ras gene. Mechanistic investigations suggested that the high incidence of cancer in DKO mice likely reflected a cooperative effect of increased benign skin tumor formation caused by p21Waf1/Cip1 loss, with increased malignant conversion of benign skin tumors caused by p16INK4a loss. Our findings establish an intrinsic cooperation between p16INK4a and p21Waf1/Cip1 in the onset of cellular senescence and tumor suppression in vivo. Cancer Res; 70(22); 9381–90. ©2010 AACR.

The p16INK4a and p21Waf1/Cip1 cyclin-dependent kinase (CDK) inhibitors are known to play key roles in the onset of cellular senescence, a state of permanent cell cycle arrest in culture (16). The simultaneous induction of p21Waf1/Cip1and p16Ink4a expression cooperatively blocks the activation of both cyclin D kinase (CDK4/6) and cyclin E kinase (CDK2), allowing the accumulation of the dephosphorylated form of the retinoblastoma tumor suppressor protein (pRb) and thereby causing permanent cell cycle arrest (710). It has become apparent that cellular senescence can be induced by a variety of potentially oncogenic stimuli, such as telomere shortening, DNA damage, oxidative stress, or oncogene expression (1114), suggesting that cellular senescence is likely to act as a tumor suppression mechanism in vivo (15, 16). Although the roles of p16Ink4a and p21Waf1/Cip1 in cellular senescence are well documented in various cell culture studies, the in vivo roles of these CDK inhibitors are poorly understood (17). For example, mice lacking p16INK4a (18, 19) or p21Waf1/Cip1 (2022) exhibit only a little predisposition to spontaneous tumor formation. These observations raise the question of whether the results seen in cell culture studies truly reflect the physiologic roles of these CDK inhibitors in vivo. However, it is also possible that these weak phenotypes could be due to functional compensatory effect(s) between p16Ink4a and p21Waf1/Cip1.

To directly address the cooperative roles of p16INK4a and p21Waf1/Cip1in vivo, we generated a compound mouse line simply lacking both of the p16INK4a and p21Waf1/Cip1 genes [double-knockout (DKO) mouse] on a C57BL/6 background, which is known to be carcinogenesis resistant (2325), and evaluated the roles of these two critical senescence inducers in vitro and in vivo. Intriguingly, DKO mice are significantly more susceptible to chemical carcinogenesis compared with mice lacking either p16INK4a or p21Waf1/Cip1 alone. Moreover, mouse embryonic fibroblasts (MEF) derived from DKO mice are resistant to the onset and the maintenance of cellular senescence in culture. These results indicate that p16INK4a plays a critical role in cooperating with p21Waf1/Cip1 in tumor suppression in vivo.

Generation of p16−/−;p21−/− mice

p16−/− mice (C57BL/6) were kindly provided by N.E. Sharpless (University of North Carolina Lineberger Comprehensive Cancer Center, Chapel Hill, NC; ref. 19). p21−/− mice (C57BL/6) were kindly provided by P. Leder (Harvard Medical School, Boston, MA; ref. 20). p16−/−;p21−/− DKO mice were generated by crossing p16−/− mice and p21−/− mice. All animals were cared for by using protocols approved by the Committee for the Use and Care of Experimental Animals of the Japanese Foundation for Cancer Research.

Cell culture

MEFs were generated from E13.5 embryos of wild-type, p16−/−, p21−/−, or p16−/−;p21−/− DKO mice as previously described (26). MEFs were grown in DMEM supplemented with 10% fetal bovine serum in 3% O2/5% CO2 for 2 days, harvested, viably frozen, and labeled as passage 0. Serial 3T3 cultivation was done as described (27). Cells were counted in triplicate. The number of cells present on the third day (N3) was divided by the initial cell number (N0 = 3 × 105) and plotted as growth rate (N3/N0; Supplementary Fig. S1). The increase in the population doubling level (ΔPDL) was calculated according to the following formula: ΔPDL = log(Nf/N0)/log 2, where N0 is the initial number of cells (3 × 105) and Nf is the final number of cells.

Western blot analysis

Proteins (40 μg) were analyzed by Western blotting, as previously described (6). The primary antibodies used were p16INK4a (11104 IBL), p21Waf1/Cip1 (sc-6246, Santa Cruz Biotechnology, Inc.), p15INK4b (sc-613, Santa Cruz Biotechnology), p53 (1C12, Cell Signaling), p19Arf [ab80 (Abcam) or sc-32748 (Santa Cruz Biotechnology)], H-Ras (sc-29, Santa Cruz Biotechnology), pRb (554136, BD Pharmingen), β-actin (AC-74, Sigma-Aldrich), and α-tublin (DM-1A, Sigma-Aldrich). Secondary antibodies were detected by enhanced chemiluminescence (Amersham).

Retrovirus infection and focus formation assay

Retroviral gene transfer was done by transient transfection of the LinxE ecotropic packaging cells with a pBabe-puro vector and the vector containing the human H-rasV12 oncogene, as previously described (6). Forty-eight hours after transfection, the retrovirus-containing medium was collected, filtered, supplemented with 8 μg/mL polybrene (Sigma), and used for multiple infection (twice a day for 2 days) into early-passage MEFs (P0–P1) growing in 3% O2/5% CO2. Infected cell populations were selected for 5 days in the presence of 1 μg/mL puromycin. For foci formation assays, 1.7 × 104 puromycin-selected immortalized cells were seeded into 6-cm-diameter dishes. Cells were maintained for 15 days in a 3% O2/5% CO2 culture environment, and the medium was changed twice a week until the cells were photographed and counted.

Senescence-associated β-galactosidase assay

Senescence-associated β-galactosidase assay was done as described previously (28).

Analysis of intracellular reactive oxygen species

To assess the generation of intracellular levels of reactive oxygen species (ROS), cells were incubated with 20 μmol/L 2′,7′-dichlorofluorescein diacetate (Calbiochem) for 20 minutes at 37°C. The peak excitation wavelength for oxidized 2′,7′-dichlorofluorescein was 488 nm and the emission wavelength was 525 nm as described previously (6).

Chemically induced skin tumor formation

The chemically induced skin tumor formation analysis was done as described previously (29). Briefly, mice of the wild-type, p16-KO, p21-KO, or DKO genotypes in the resting phase of the hair cycle (8 weeks old) were shaved and treated with 100 μg of 7,12-dimethylbenz(a)anthracene (DMBA) in 100 μL of acetone. One week after DMBA treatment, mice were subsequently treated twice a week with 12.5 μg of 12-O-tetradecanoylphorbol-13-acetate (TPA) in 100 μL of acetone for 20 weeks. Tumors developed in mice were analyzed for histology at 30 weeks after DMBA treatment.

Histologic analysis

Tissues were fixed in 10% buffered formalin, progressively dehydrated through gradients of alcohol, and embedded in paraffin. Samples were sectioned on a microtome at 4-μm thickness, deparaffinized in xylene, rehydrated, and then stained with H&E for histologic analysis.

p16INK4a;p21Waf1/Cip1 DKO MEFs proliferate without any induction of cellular senescence by serial passaging

Compound mice lacking both p16INK4a and p21Waf1/Cip1 on a C57BL/6 background (DKO mice) are fertile and born normally in the expected Mendelian ratio. MEFs derived from wild-type mice, p16INK4a knockout (p16-KO) mice, p21Waf1/Cip1 knockout (p21-KO) mice, and DKO mice were subjected to serial passage using the 3T3 protocol under standard culture conditions, which included atmospheric (20%) oxygen. Consistent with previous reports, p16-KO MEFs underwent cellular senescence in a kinetic pattern similar to that of wild-type MEFs (Fig. 1A; Supplementary Fig. S1A and C; refs. 18, 19). The level of p15INK4b expression was remarkably increased and the p21Waf1/Cip1 level remained high in senescent p16-KO MEFs, implying that a compensatory network among the CDK inhibitors indeed exists as previously suggested (refs. 3033; Fig. 1C, lanes 1–4 and 10–13). p21-KO MEFs proliferated without detectable senescence growth arrest, although they exhibited a slight reduction in proliferation rate at passage 4, as previously reported (ref. 26; Fig. 1A; Supplementary Fig. S1A and C). Interestingly, MEFs lacking both p16INK4a and p21Waf1/Cip1(DKO MEFs) grew slightly but consistently better than p21-KO MEFs, suggesting that p16INK4a also plays a role, at least to some extent, in the cellular senescence provoked by the serial passage of MEFs (Fig. 1A; Supplementary Fig. S1A).

Figure 1.

Serial 3T3 cultivation of primary MEFs of different genotypes. A and B, MEFs derived from wild-type (WT), p16-KO, p21-KO, and DKO mice were cultivated according to the 3T3 protocol. The graphs represent the accumulated population doublings at every passage (A, in 20% oxygen; B, in 3% oxygen). The means ± SD of triplicate experiments are shown. C, protein levels of the cell cycle inhibitors p16INK4a, p21Waf1/Cip1, p15INK4b, and p19Arf. Note that passage 20 of the wild-type and p16-KO MEFs in 20% oxygen culture could not be analyzed because a very long period of time was required to generate an immortalized culture. To compare the levels of proteins across the gels, the same sample (immortalized MEFs) was loaded on every gel (lanes 1, 10, 19, and 30). β-Actin was used as a loading control. Representative data of two independent experiments are shown. Uncropped images are shown in Supplementary Data.

Figure 1.

Serial 3T3 cultivation of primary MEFs of different genotypes. A and B, MEFs derived from wild-type (WT), p16-KO, p21-KO, and DKO mice were cultivated according to the 3T3 protocol. The graphs represent the accumulated population doublings at every passage (A, in 20% oxygen; B, in 3% oxygen). The means ± SD of triplicate experiments are shown. C, protein levels of the cell cycle inhibitors p16INK4a, p21Waf1/Cip1, p15INK4b, and p19Arf. Note that passage 20 of the wild-type and p16-KO MEFs in 20% oxygen culture could not be analyzed because a very long period of time was required to generate an immortalized culture. To compare the levels of proteins across the gels, the same sample (immortalized MEFs) was loaded on every gel (lanes 1, 10, 19, and 30). β-Actin was used as a loading control. Representative data of two independent experiments are shown. Uncropped images are shown in Supplementary Data.

Close modal

Because primary MEFs have been shown to be exquisitely sensitive to oxidative stress in culture (34), we next cultured these MEFs with 3% oxygen, which is known to be similar to the physiologic oxygen condition in vivo (34). In contrast to the result with the 20% oxygen condition, we were unable to see any significant difference in the cell proliferation rate between wild-type MEFs and MEFs lacking the p16INK4a or p21Waf1/Cip1 gene (Fig. 1B; Supplementary Fig. S1B and D), although the DKO MEFs proliferated slightly better than the MEFs of other types (Fig. 1B; Supplementary Fig. S1B). As seen in the 20% oxygen condition, the level of p15INK4b expression, but not the expression of the other p16INK4a family members (data not shown), was substantially increased in both p16-KO MEFs and DKO MEFs under the 3% oxygen condition (Fig. 1C, lanes 1–18 and 30–40), suggesting that p15INK4b may play a compensatory role in the absence of p16INK4a regardless of the oxygen conditions in vitro, although this level of p15INK4b expression does not seem to be high enough to block cell proliferation (Fig. 1B; Supplementary Fig. S1B). Furthermore, because the p19Arf gene shares a locus with p16INK4a and is also implicated in tumor suppression (35, 36), we examined the level of p19Arf in these MEFs. However, we did not see any substantial difference in p19Arf expression in MEFs lacking the p16INK4a and/or p21Waf1/Cip1 gene compared with wild-type MEFs (Fig. 1C).

DKO MEFs expressing oncogenic Ras escape contact inhibition and form foci

Because cellular senescence can also be induced by activated ras oncogene expression, we next asked if the status of the p16INK4a and/or p21Waf1/Cip1 gene affects Ras-induced senescence under physiologic (3%) oxygen conditions. In contrast to passage-induced senescence, MEFs of all the genotypes exhibited features of cellular senescence, including cell cycle arrest, a flat and enlarged morphology, and the accumulation of the dephosphorylated form of pRb within 10 days of infection with retrovirus encoding oncogenic Ras (Fig. 2). Notably, increased expression of p15INK4b as well as p53 and p19Arf upregulation was observed in Ras-induced senescence (Fig. 2D). The depletion of p15INK4b by small interference RNA abrogated the appearance of senescence markers in Ras-infected DKO MEFs (Supplementary Fig. S2). It is therefore likely that the upregulation of p15INK4b expression plays an important role, at least partly, in oncogenic Ras–induced senescence in DKO MEFs. However, although the levels of senescence-associated β-galactosidase activity (Fig. 2B) and the intracellular levels of ROS (Fig. 2C), which are markers of cellular senescence, were increased in DKO MEFs, the level of induction was less evident compared with the MEFs of other genotypes. These results raise the possibility that p16INK4a and p21Waf1/Cip1 are both required for the full induction of oncogenic Ras–induced senescence in culture. Indeed, DKO MEFs expressing oncogenic Ras readily escaped the senescent state and reinitiated proliferation more rapidly than the MEFs of other genotypes under the 3% oxygen condition (Fig. 3A). Note that the saturation density of the DKO MEFs expressing oncogenic Ras was remarkably higher than that of the other cell types (Fig. 3B). Interestingly, significant numbers of foci were also observed in the confluent culture of the DKO MEFs expressing oncogenic Ras, but not the MEFs of other genotypes, under the 3% oxygen condition (Fig. 3C). Taken together, these results suggest that oncogenic Ras expression, in conjunction with the loss of p16INK4a and p21Waf1/Cip1, may have potential to counteract the pathway involved in contact inhibition in cultured MEFs.

Figure 2.

Induction of oncogene-induced cellular senescence by retrovirally transduced oncogenic Ras in wild-type, p16-KO, p21-KO, and DKO MEFs. MEFs of the indicated genotypes were infected with control retrovirus or retrovirus encoding oncogenic Ras, and a series of experiments were performed a week after infection. A, representative photographs of cells; B, percentage of senescence-associated β-galactosidase (SA-β-gal) positive MEFs; C, relative ROS levels. Columns, mean of three independent experiments; bars, SD. D, protein levels of p15INK4b, p19Arf, p53, Rb, and H-Ras in MEFs infected with retrovirus encoding oncogenic Ras (H-RasV12) or control empty vector (Vector). β-Actin was used as a loading control. Representative data of two independent experiments are shown. Uncropped images are shown in Supplementary Data.

Figure 2.

Induction of oncogene-induced cellular senescence by retrovirally transduced oncogenic Ras in wild-type, p16-KO, p21-KO, and DKO MEFs. MEFs of the indicated genotypes were infected with control retrovirus or retrovirus encoding oncogenic Ras, and a series of experiments were performed a week after infection. A, representative photographs of cells; B, percentage of senescence-associated β-galactosidase (SA-β-gal) positive MEFs; C, relative ROS levels. Columns, mean of three independent experiments; bars, SD. D, protein levels of p15INK4b, p19Arf, p53, Rb, and H-Ras in MEFs infected with retrovirus encoding oncogenic Ras (H-RasV12) or control empty vector (Vector). β-Actin was used as a loading control. Representative data of two independent experiments are shown. Uncropped images are shown in Supplementary Data.

Close modal
Figure 3.

DKO MEFs escape contact inhibition. A, the cell numbers of wild-type, p16-KO, p21-KO, and DKO MEFs expressing oncogenic Ras were counted 7 d after retrovirus infection (indicated as day 1). The cells were counted in triplicate and the relative cell numbers are shown. B, relative cell numbers of wild-type, p16-KO, p21-KO, and DKO MEFs expressing oncogenic Ras in the focus formation assay. The means ± SD of triplicate experiments are shown. C, Ras-expressing immortalized DKO MEFs, but not the MEFs of other genotypes, formed foci after 15 d of culture. Representative data of three independent experiments are shown.

Figure 3.

DKO MEFs escape contact inhibition. A, the cell numbers of wild-type, p16-KO, p21-KO, and DKO MEFs expressing oncogenic Ras were counted 7 d after retrovirus infection (indicated as day 1). The cells were counted in triplicate and the relative cell numbers are shown. B, relative cell numbers of wild-type, p16-KO, p21-KO, and DKO MEFs expressing oncogenic Ras in the focus formation assay. The means ± SD of triplicate experiments are shown. C, Ras-expressing immortalized DKO MEFs, but not the MEFs of other genotypes, formed foci after 15 d of culture. Representative data of three independent experiments are shown.

Close modal

DKO mice are extremely susceptible to DMBA/TPA–induced skin cancer

To explore the role of p16INK4a and p21Waf1/Cip1 in oncogenic Ras expression in a more physiologic setting, wild-type, p16-KO, p21-KO, and DKO mice were subjected to a conventional chemically induced skin carcinogenesis protocol, with a single dose of DMBA followed by biweekly treatment of TPA for 20 weeks. Because this protocol is known to cause an oncogenic mutation in the endogenous H-ras gene, it seemed to be an ideal system for studying the physiologic response to oncogenic Ras expression in living animals (37). In p16-KO mice, the frequency of benign skin papilloma formation was not increased, and the maximum number of papillomas was 6.8 per mouse on average, whereas 6.0 papillomas developed in wild-type mice (Fig. 4A and B). Interestingly, however, 23.08% of the p16-KO mice developed at least one malignant skin tumor (Fig. 4B; Table 1), whereas 5.56% of the wild-type mice developed skin cancer (Fig. 4A; Table 1) at 30 weeks after DMBA treatment. Note that a malignant tumor that appeared in the wild-type mice was carcinoma in situ, whereas two thirds of the malignant tumors in the p16-KO mice were more aggressive and invasive cancers (Table 1). The malignant conversion ratio from benign papillomas in p16-KO mice was approximately five times higher than that in wild-type mice (0.93% in wild-type mice versus 4.55% in p16-KO mice; see Table 1), indicating that p16INK4a plays a role in preventing the malignant conversion of benign skin tumors, but not benign skin tumor formation itself, at least in this setting. This is consistent with the observation that p16INK4a expression is upregulated in the late stage (30 weeks after DMBA treatment), but not in the early stage (10 weeks after DMBA treatment), of DMBA/TPA–induced skin papillomas (Supplementary Fig. S3; ref. 29). In addition, the level of p19Arf expression was also increased in the late papillomas (Supplementary Fig. S3), as seen in Ras-induced senescence in cultured MEFs (Fig. 2D). These results suggest that the upregulation of p19Arf, at least in part, contributes to the induction of Ras-induced senescence in vitro and in vivo in mice.

Figure 4.

Analysis of DMBA/TPA–induced skin tumors. A to D, wild-type (n = 18; A), p16-KO (n = 13; B), p21-KO (n = 13; C), and DKO (n = 19; D) mice. Left, average number of papillomas at the indicated times and their relative size distribution. Middle, representative images of skin tumors at 30 wk after DMBA treatment. Right, histology of the representative tumors (indicated by red arrows in the tumor images; H&E staining; magnification, ×100). A, a low-grade papilloma showing thickening of epidermis with slight cellular atypia mainly in the parabasal layer and with hyperkeratosis. B, a large tumor with ulceration in the center of the tumor. Histologically, the tumor is diagnosed as squamous cell carcinoma (SCC) showing papillary proliferation of atypical epidermal cells with dyskeratosis. C, a high-grade papilloma showing papillary proliferation of the epidermis with cellular atypia, including high nucleus/cytoplasm (N/C) ratio, increase of nuclear chromatin, and increase of mitotic figures, although the polarity of maturation is still preserved. D, a spindle cell tumor (SCT), not otherwise specified, which showed diffuse proliferation of relatively uniform short spindle cells. Yellow arrows, muscle layers. Representative results of two independent experiments are shown.

Figure 4.

Analysis of DMBA/TPA–induced skin tumors. A to D, wild-type (n = 18; A), p16-KO (n = 13; B), p21-KO (n = 13; C), and DKO (n = 19; D) mice. Left, average number of papillomas at the indicated times and their relative size distribution. Middle, representative images of skin tumors at 30 wk after DMBA treatment. Right, histology of the representative tumors (indicated by red arrows in the tumor images; H&E staining; magnification, ×100). A, a low-grade papilloma showing thickening of epidermis with slight cellular atypia mainly in the parabasal layer and with hyperkeratosis. B, a large tumor with ulceration in the center of the tumor. Histologically, the tumor is diagnosed as squamous cell carcinoma (SCC) showing papillary proliferation of atypical epidermal cells with dyskeratosis. C, a high-grade papilloma showing papillary proliferation of the epidermis with cellular atypia, including high nucleus/cytoplasm (N/C) ratio, increase of nuclear chromatin, and increase of mitotic figures, although the polarity of maturation is still preserved. D, a spindle cell tumor (SCT), not otherwise specified, which showed diffuse proliferation of relatively uniform short spindle cells. Yellow arrows, muscle layers. Representative results of two independent experiments are shown.

Close modal
Table 1.

Skin tumor susceptibility by DMBA/TPA in p16 and/or p21 knockout mice

Mouse genotype
Wild-type (n = 18)p16−/− (n = 13)p21−/− (n = 13)p16−/−;p21−/− (n = 19)
Tumor type, n* 
    Low-grade papilloma 16 
    High-grade papilloma 
    High-grade papilloma with microinvasion 
    Carcinoma in situ 
    Squamous cell carcinoma 
    Malignant spindle cell tumor 
Mice with carcinomas, % 5.56 23.08 15.38 84.21 
Maximum no. of papillomas/mouse 6.0 6.8 12.5 16.0 
Malignant conversion ratio at 30 wk after DMBA (no. of carcinomas/total no. of papillomas), % 0.93 4.55 1.23 5.92 
Mouse genotype
Wild-type (n = 18)p16−/− (n = 13)p21−/− (n = 13)p16−/−;p21−/− (n = 19)
Tumor type, n* 
    Low-grade papilloma 16 
    High-grade papilloma 
    High-grade papilloma with microinvasion 
    Carcinoma in situ 
    Squamous cell carcinoma 
    Malignant spindle cell tumor 
Mice with carcinomas, % 5.56 23.08 15.38 84.21 
Maximum no. of papillomas/mouse 6.0 6.8 12.5 16.0 
Malignant conversion ratio at 30 wk after DMBA (no. of carcinomas/total no. of papillomas), % 0.93 4.55 1.23 5.92 

NOTE: Tumors were analyzed 30 wk after DMBA treatment.

*The number of mice with the indicated tumors is shown.

One p16−/− mouse harbored two SCCs.

Two p16−/−;p21−/− mice harbored two SCCs.

In stark contrast with p16-KO mice, the frequency of benign skin papilloma formation was strikingly increased in p21-KO mice (the maximum number was 12.5 papillomas per mouse; Fig. 4C). However, surprisingly, the malignant conversion rate to advanced cancers was not increased in these p21-KO mice (1.23%) at 30 weeks after DMBA treatment in this setting (Table 1). These results are consistent with the previous observation by Weinberg and colleagues (38), but not with the report by Topley and colleagues (39). However, because each group used different protocols and mouse strains, these seemingly contradictory data are, at least partly, due to the differences in the experimental conditions between our study and the study of Topley and colleagues (39). Nevertheless, a substantial level of p21Waf1/Cip1 expression, but not p16INK4a expression, was induced in the early stage of the benign skin papillomas (10 weeks after DMBA treatment; Supplementary Fig. S3). These data imply that a major role of p21Waf1/Cip1 is likely to be preventing benign skin papilloma formation rather than the malignant conversion of benign skin tumors, at least in this setting, although it is also possible that p21Waf1/Cip1 may have the ability to block progression to higher-grade benign papillomas.

It is interesting that the frequencies of both the benign papilloma formation (the maximum number was 16.0 papillomas per mouse) and the incidence of cancer-bearing mice were dramatically enhanced when DKO mice were subjected to the DMBA/TPA–induced skin carcinogenesis protocol (Fig. 4D; Table 1). By 30 weeks after the DMBA/TPA treatment, 84.21% of the DKO mice developed at least one carcinoma (Fig. 4D; Table 1). Notably, however, the malignant conversion rate of benign papillomas in DKO mice (5.92%) was not increased compared with that in p16-KO mice (4.55%; Table 1). These results indicate that the high incidence of cancer in the mice with the DKO genotype is likely to be due to a cooperative effect of increased benign skin papilloma formation due to p21Waf1/Cip1 loss and increased malignant conversion of benign skin tumors due to p16INK4a loss. It is also worth emphasizing that 31.6% (6 of 19) of the DKO mice developed malignant spindle cell tumors, which were more aggressively invasive through the muscle layers compared with squamous cell carcinomas (Fig. 4D; Table 1).

An intact DNA damage response prevents the progression of DMBA/TPA–induced skin tumors

It was reported that oncogenic Ras induces DNA damage signaling and the activation of the cell cycle checkpoint, which are critical for the cellular senescence and tumor suppression phenotypes (4042). Therefore, we have examined γ-H2AX foci formation, a DNA damage response marker, in tumors in all the genotypes. γ-H2AX foci were detectable in a significant percentage of cells in the low-grade papillomas in the wild-type, p21-KO, and p16-KO mice. In contrast, a very small percentage of cells were positive for γ-H2AX foci in the low-grade papillomas that developed in DKO mice (Supplementary Fig. S4). These results suggest that p21Waf1/Cip1 and p16INK4a play important roles in conducting the DNA damage signaling pathway. Note that γ-H2AX foci were hardly detectable in most of the high-grade papillomas and malignant tumors, such as squamous cell carcinomas and spindle cell tumors, regardless of the genotype, suggesting that the higher-grade papillomas and the malignant tumors seem to have a defect in responding to DNA damage signals.

In this study, we have generated for the first time the p16INK4a;p21Waf1/Cip1 double mutant mice (DKO mice) on a C57BL/6 background, a carcinogenesis-resistant strain, and analyzed the effects of p16INK4a and p21Waf1/Cip1 deficiency on the susceptibility to cancer in cooperation with oncogenic Ras expression. MEFs derived from DKO mice easily escaped Ras-induced senescence and overrode contact inhibition in culture (Fig. 3). Moreover, DKO-mice on the C57BL/6 background were extremely susceptible to DMBA/TPA–induced skin carcinogenesis (Fig. 4). Our analyses clearly show a significant increase of cancer incidence in the DKO background because p21Waf1/Cip1 deletion promotes benign skin tumor development and p16INK4a deletion promotes the incidence of malignant conversion of benign skin tumors (Table 1).

There are reports showing that compound mice expressing a mutant form of CDK4 (CDK4-R24C), which is resistant to all of the p16INK4a family members, on a p21Waf1/Cip1-null background are more susceptible to malignant tumors compared with mice either lacking p21Waf1/Cip1 or expressing CDK4-R24C (27, 43). However, because CDK4-R24C has the ability to counteract all of the p16INK4a family members, it has been unclear whether p16INK4a or other p16INK4a family members play a role in cooperating with p21Waf1/Cip1 in the induction of cellular senescence and/or tumor suppression in vivo. Our study has clarified that the increased tumor incidence and immortalization of the previously described cdk4r24c mutant mice and MEFs with the p21Waf1/Cip1-null background are most likely to be due to the absence of p16INK4a from among the INK4 family both in vitro and in vivo.

It is also noteworthy that 31.6% of the DKO mice developed malignant spindle cell tumors, which were more aggressive than squamous cell carcinomas (Fig. 4D; Table 1). It is interesting that all of these spindle cell tumors exhibited a substantial increase in the expression of the Oct4 gene, which is known to play an important role in stem cell maintenance and reprogramming (Supplementary Fig. S5), although the molecular mechanism behind this upregulation remains unclear. This is consistent in part with recent reports showing that both the p16INK4a and p53-p21Waf1/Cip1 pathways play an inhibitory role in the reprogramming of differentiated cells toward becoming induced pluripotent stem cells (4449). It is therefore tempting to speculate that the inactivation of both p16Ink4a and p21Waf1/Cip1, in conjunction with Ras activation, increases the chance of reprogramming differentiated cells or stimulates the proliferation of undifferentiated cells in vivo. Further analysis is required to understand how the p16Ink4a and p21Waf1/Cip1 pathways are linked to Oct4 expression.

It is still possible that the phenotypic effects seen in p16INK4a knockout mice and DKO mice are compromised, at least to some extent, by developmental or somatic compensation by upregulation of the remaining p16INK4a family members (3033). Nonetheless, our results provide direct evidence that p16Ink4a and p21Waf1/Cip1 play cooperative roles in the onset and/or the maintenance of Ras-induced senescence in vitro and show for the first time that p16INK4a per se plays a crucial role in preventing the malignant conversion of benign skin tumors in vivo.

No potential conflicts of interest were disclosed.

We thank Dr. P. Leder for p21Waf1/Cip1-knockout mice, Dr. N.E. Sharpless for p16INK4a-knockout mice, and Dr. M. Serrano [Spanish National Cancer Research Center (CNIO), Madrid, Spain] for retrovirus vector encoding H-RasV12. We also thank C. Sugita and S. Ohtomi for their assistance in mouse experiments.

Grant Support: Ministry of Education, Science, Sports and Technology of Japan, the Astellas Foundation for Research on Metabolic Disorders, the Vehicle Racing Commemorative Foundation, the Takeda Science Foundation, the Uehara Memorial Foundation, and the Mishima Kaiun Memorial Foundation.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1
Noda
A
,
Ning
Y
,
Venable
SF
,
Pereira-Smith
OM
,
Smith
JR
. 
Cloning of senescent cell-derived inhibitors of DNA synthesis using an expression screen
.
Exp Cell Res
1994
;
211
:
90
8
.
2
Hara
E
,
Smith
R
,
Parry
D
,
Tahara
H
,
Stone
S
,
Peters
G
. 
Regulation of p16CDKN2 expression and its implications for cell immortalization and senescence
.
Mol Cell Biol
1996
;
16
:
859
67
.
3
Alcorta
DA
,
Xiong
Y
,
Phelps
D
,
Hannon
G
,
Beach
D
,
Barrett
JC
. 
Involvement of the cyclin-dependent kinase inhibitor p16INK4a in replicative senescence of normal human fibroblasts
.
Proc Natl Acad Sci U S A
1996
;
93
:
13742
7
.
4
Serrano
M
,
Lin
AW
,
McCurrach
ME
,
Beach
D
,
Lowe
SW
. 
Oncogenic ras provokes premature cell senescence associated with accumulation of p53 and p16INK4a
.
Cell
1997
;
88
:
593
602
.
5
Randle
DH
,
Zindy
F
,
Sherr
CJ
,
Roussel
MF
. 
Differential effects of p19Arf and p16Ink4a loss on senescence of murine bone marrow-derived preB cells and macrophages
.
Proc Natl Acad Sci U S A
2001
;
98
:
9654
9
.
6
Takahashi
A
,
Ohtani
N
,
Yamakoshi
K
, et al
. 
Mitogenic signalling and the p16INK4a-Rb pathway cooperate to enforce irreversible cellular senescence
.
Nat Cell Biol
2006
;
8
:
1291
7
.
7
McConnell
BB
,
Gregory
FJ
,
Stott
FJ
,
Hara
E
,
Peters
G
. 
Induced expression of p16INK4a inhibits both CDK4- and CDK2-associated kinase activity by reassortment of cyclin-CDK-inhibitor complexes
.
Mol Cell Biol
1999
;
19
:
1981
9
.
8
Mitra
J
,
Dai
CY
,
Somasundaram
K
, et al
. 
Induction of p21WAF1/CIP1 and inhibition of Cdk2 mediated by the tumor suppressor p16INK4a
.
Mol Cell Biol
1999
;
19
:
3916
28
.
9
Sherr
CJ
,
Roberts
JM
. 
CDK inhibitors: positive and negative regulators of G1-phase progression
.
Genes Dev
1999
;
13
:
1501
12
.
10
Li
W
,
Kotoshiba
S
,
Berthet
C
,
Hilton
MB
,
Kaldis
P
. 
Rb/Cdk2/Cdk4 triple mutant mice elicit an alternative mechanism for regulation of the G1-S transition
.
Proc Natl Acad Sci U S A
2009
;
106
:
486
91
.
11
Collado
M
,
Blasco
MA
,
Serrano
M
. 
Cellular senescence in cancer and aging
.
Cell
2007
;
130
:
223
33
.
12
Jacobs
JJ
,
de Lange
T
. 
p16INK4a as a second effector of the telomere damage pathway
.
Cell Cycle
2005
;
4
:
1364
8
.
13
Ben-Porath
I
,
Weinberg
RA
. 
When cells get stressed: an integrative view of cellular senescence
.
J Clin Invest
2004
;
113
:
8
13
.
14
Okada
H
,
Mak
TW
. 
Pathways of apoptotic and non-apoptotic death in tumor cells
.
Nat Rev Cancer
2004
;
4
:
592
603
.
15
Sharpless
NE
,
DePinho
RA
. 
Cancer: crime and punishment
.
Nature
2005
;
436
:
636
7
.
16
Campisi
J
. 
Senescent cells, tumor suppression, and organismal aging: good citizens, bad neighbors
.
Cell
2005
;
F120
:
513
22
.
17
Pei
XH
,
Xiong
Y
. 
Biochemical and cellular mechanisms of mammalian CDK inhibitors: a few unresolved issues
.
Oncogene
2005
;
24
:
2787
95
.
18
Krimpenfort
P
,
Quon
KC
,
Mooi
WJ
,
Loonstra
A
,
Berns
A
. 
Loss of p16INK4a confers susceptibility to metastatic melanoma in mice
.
Nature
2001
;
413
:
83
6
.
19
Sharpless
NE
,
Bardeesy
N
,
Lee
KH
, et al
. 
Loss of p16INK4a with retention of p19Arf predisposes mice to tumorigenesis
.
Nature
2001
;
413
:
86
91
.
20
Deng
C
,
Zhang
P
,
Harper
JW
,
Elledge
SJ
,
Leder
P
. 
Mice lacking p21Waf1/Cip1 undergo normal development, but are defective in G1 checkpoint control
.
Cell
1995
;
82
:
675
84
.
21
Brugarolas
J
,
Chandrasekaran
C
,
Gordon
JI
,
Beach
D
,
Jacks
T
,
Hannon
GJ
. 
Radiation-induced cell cycle arrest compromised by p21 deficiency
.
Nature
1995
;
377
:
552
7
.
22
Martín-Caballero
J
,
Flores
JM
,
García-Palencia
P
,
Serrano
M
. 
Tumor susceptibility of p21Waf1/Cip1-deficient mice
.
Cancer Res
2001
;
61
:
6234
8
.
23
DiGiovanni
J
,
Prichett
WP
,
Decina
PC
,
Diamond
L
. 
DBA/2 mice are as sensitive as SENCAR mice to skin tumor promotion by 12-O-tetradecanoylphorbol-13-acetate
.
Carcinogenesis
1984
;
5
:
1493
8
.
24
Hennings
H
,
Glick
AB
,
Lowry
DT
,
Krsmanovic
LS
,
Sly
LM
,
Yuspa
SH
. 
FVB/N mice: an inbred strain sensitive to the chemical induction of squamous cell carcinomas in the skin
.
Carcinogenesis
1993
;
14
:
2353
8
.
25
Woodworth
CD
,
Michael
E
,
Smith
L
, et al
. 
Strain-dependent differences in malignant conversion of mouse skin tumors is an inherent property of the epidermal keratinocyte
.
Carcinogenesis
2004
;
25
:
1771
8
.
26
Pantoja
C
,
Serrano
M
. 
Murine fibroblasts lacking p21 undergo senescence and are resistant to transformation by oncogenic Ras
.
Oncogene
1999
;
18
:
4974
82
.
27
Carbone
CJ
,
Grana
X
,
Reddy
EP
,
Haines
DS
. 
p21 loss cooperates with INK4 inactivation facilitating immortalization and bcl-2-mediated anchorage-independent growth of oncogene-transduced primary mouse fibroblasts
.
Cancer Res
2007
;
67
:
4130
7
.
28
Dimri
GP
,
Lee
X
,
Basile
G
, et al
. 
A biomarker that identifies senescent human cells in culture and in aging skin in vivo
.
Proc Natl Acad Sci U S A
1995
;
92
:
9363
7
.
29
Yamakoshi
K
,
Takahashi
A
,
Hirota
F
, et al
. 
Real-time in vivo imaging of p16Ink4a reveals cross talk with p53
.
J Cell Biol
2009
;
186
:
393
407
.
30
Krimpenfort
P
,
Ijpenberg
A
,
Song
JY
, et al
. 
p15INK4b is a critical tumor suppressor in the absences of p16INK4a
.
Nature
2007
;
448
:
943
6
.
31
Ramsey
MR
,
Krishnamurthy
J
,
Pei
XH
, et al
. 
Expression of p16Ink4a compensates for p18Ink4c loss in cyclin-dependent kinase 4/6-dependent tumors and tissues
.
Cancer Res
2007
;
67
:
4732
41
.
32
Wiedemeyer
R
,
Brennan
C
,
Heffernan
TP
, et al
. 
Feedback circuit among INK4 tumor suppressors constrains human glioblastoma development
.
Cancer Cell
2008
;
13
:
355
64
.
33
Ciemerych
MA
,
Sicinski
P
. 
Cell cycle in mouse development
.
Oncogene
2005
;
24
:
2877
98
.
34
Parrinello
S
,
Samper
E
,
Krtolica
A
,
Goldstein
J
,
Melov
S
,
Campisi
J
. 
Oxygen sensitivity severely limits the replicative lifespan of murine fibroblasts
.
Nat Cell Biol
2003
;
5
:
741
7
.
35
Kamijo
T
,
Zindy
F
,
Roussel
MF
, et al
. 
Tumor suppression at the mouse INK4a locus mediated by the alternative reading frame product p19ARF
.
Cell
1997
;
91
:
649
59
.
36
Stott
FJ
,
Bates
S
,
James
MC
, et al
. 
The alternative product from the human CDKN2A locus, p14ARF, participates in a regulatory feedback loop with p53 and MDM2
.
EMBO J
1998
;
17
:
5001
14
.
37
Kemp
CJ
. 
Multistep skin cancer in mice as a model to study the evolution of cancer cells
.
Semin Cancer Biol
2005
;
15
:
460
73
.
38
Weinberg
WC
,
Fernandez-Salas
E
,
Morgan
DL
, et al
. 
Genetic deletion of p21WAF1 enhances papilloma formation but not malignant conversion in experimental mouse skin carcinogenesis
.
Cancer Res
1999
;
59
:
2050
4
.
39
Topley
GI
,
Okuyama
R
,
Gonzales
JG
,
Conti
C
,
Dotto
GP
. 
p21WAF1/Cip1 functions as a suppressor of malignant skin tumor formation and a determinant of keratinocyte stem-cell potential
.
Proc Natl Acad Sci U S A
1999
;
96
:
9089
94
.
40
Di Micco
R
,
Fumagalli
M
,
Cicalese
A
, et al
. 
Oncogene-induced senescence is a DNA damage response triggered by DNA hyper-replication v
.
Nature
2006
;
444
:
638
42
.
41
Bartkova
J
,
Rezaei
N
,
Liontos
M
, et al
. 
Oncogene-induced senescence is part of the tumorigenesis barrier imposed by DNA damage checkpoints
.
Nature
2006
;
444
:
633
7
.
42
Halazonetis
TD
,
Gorgoulis
VG
,
Bartek
J
. 
An oncogene-induced DNA damage model for cancer development
.
Science
2008
;
319
:
1352
5
.
43
Quereda
V
,
Martinalbo
J
,
Dubus
P
,
Carnero
A
,
Malumbres
M
. 
Genetic cooperation between p21Cip1 and INK4 inhibitors in cellular senescence and tumor suppression
.
Oncogene
2007
;
26
:
7665
74
.
44
Hong
H
,
Takahashi
K
,
Ichisaka
T
, et al
. 
Suppression of induced pluripotent stem cell generation by the p53–21 pathway
.
Nature
2009
;
460
:
1132
5
.
45
Utikal
J
,
Polo
JM
,
Stadtfeld
M
, et al
. 
Immortalization eliminates a roadblock during cellular reprogramming into iPS cells
.
Nature
2009
;
460
:
1145
8
.
46
Marión
RM
,
Strati
K
,
Li
H
, et al
. 
A p53-mediated DNA damage response limits reprogramming to ensure iPS cell genomic integrity
.
Nature
2009
;
460
:
1149
53
.
47
Li
H
,
Collado
M
,
Villasante
A
, et al
. 
The Ink4/Arf locus is a barrier for iPS cell reprogramming
.
Nature
2009
;
460
:
1136
9
.
48
Kawamura
T
,
Suzuki
J
,
Wang
YV
, et al
. 
Linking the p53 tumour suppressor pathway to somatic cell reprogramming
.
Nature
2009
;
460
:
1140
4
.
49
Banito
A
,
Rashid
ST
,
Acosta
JC
, et al
. 
Senescence impairs successful reprogramming to pluripotent stem cells
.
Genes Dev
2009
;
23
:
2134
9
.

Supplementary data