Original Article | Published:

Cytoplasmic p27 is oncogenic and cooperates with Ras both in vivo and in vitro

Oncogene volume 30, pages 28462858 (23 June 2011) | Download Citation


p27Kip1 (p27) can have opposing roles during malignant transformation depending on cellular context: on one hand it functions as a tumor suppressor by inhibiting cyclin–cyclin-dependent kinase (CDK) activity in the nucleus and on the other it may adopt an oncogenic role that is less well understood. To gain further insight into the roles played by p27 during tumorigenesis, we compared the susceptibility with urethane-induced tumorigenesis of two p27 mouse models, p27−/− and p27CK− knockin, in which p27 cannot bind or inhibit cyclin–CDKs. In this K-Ras-driven tumorigenesis model, p27CK− mice had an increase in both tumor number and aggressiveness compared with p27−/−, indicating a cooperation between p27CK− and activated Ras. In the lung, increased tumorigenesis was associated with cytoplasmic localization of p27CK− and bronchiolaveolar stem cell amplification. The ability of p27CK− to cooperate with other oncogenes was not universal. When c-Myc was used as a transforming agent, p27 status became irrelevant and c-Myc was equally potent in transforming p27+/+, p27−/− and p27CK− cells. In fact, c-Myc induced the degradation of wild-type p27 via the Skp-Cullin-F-box (SCF)–Skp2 pathway. In contrast, p27CK− levels were not affected by c-Myc expression, as p27CK− is insensitive to Skp2-mediated degradation because of its inability to bind cyclin E/CDK2. However, in presence of c-Myc, p27CK− remained mostly nuclear, providing an explanation for its inability to cooperate with Myc during transformation. Thus, we propose that the p27CK− protein needs to be localized in the cytoplasm in order to function as an oncogene, otherwise it just behaves similar to a null allele.


Deregulation of cell cycle control mechanisms has a significant role in the process of malignant transformation. This is commonly driven by the inactivation of tumor suppressor genes that restrain proliferation and the activation of proto-oncogenes that provide growth-promoting signals (Hanahan and Weinberg, 2000). The cyclin/cyclin-dependent kinase (CDK) inhibitor p27Kip1 (p27) functions as a tumor suppressor by its ability to block cell proliferation (Sherr and Roberts, 1999). Knockout of the cdkn1b gene in mice has confirmed the tumor suppressor role of p27 in vivo, and p27 loss promotes both the formation of spontaneous and carcinogen or irradiation-induced tumors (Fero et al., 1996, 1998). Nevertheless, p27 is an atypical tumor suppressor, as mutations of its gene are extremely rare in human tumors (Besson et al., 2008; Chu et al., 2008) or in mouse models of cancer (Fero et al., 1998; Kelly-Spratt et al., 2009). In fact, tumor cells have evolved other mechanisms to inactivate p27, including increased proteolytic degradation and exclusion from the nucleus (Besson et al., 2008; Chu et al., 2008). Overall, the net loss of nuclear p27 expression is correlated with tumor aggressiveness and poor clinical outcome (reviewed in Besson et al., 2008; Chu et al., 2008).

Proteolysis of p27 is mediated via the ubiquitin–proteasome pathway and involves different ubiquitin ligases. In quiescent cells, p27 turnover is dependent on its dephosphorylation of Ser10 and its interaction with cyclin–CDK complexes but the enzyme mediating this degradation remains elusive (Besson et al., 2006). As cells enter the G1 phase, Ser10 phosphorylation become an active signal for export of p27 from the nucleus, and once in the cytoplasm, p27, if not bound to cyclin–CDKs, becomes the target of the RING-finger containing KPC1/2 complex for ubiquitination and degradation (Kamura et al., 2004). More recently, another RING-finger containing ubiquitin ligase that mediates p27 degradation during G1 has been described (Hattori et al., 2007). As cells progress through G1 and into S-phase, p27 degradation is mediated by the Skp2–SCF complex composed of Skp1, Cul1, Cks1, Rbx1 and Skp2, and requires the phosphorylation of p27 on Thr187 and a stable interaction with cyclin–CDK (Montagnoli et al., 1999; Malek et al., 2001; Besson et al., 2006; Frescas and Pagano 2008). An extensive body of literature has described the oncogenic role of Skp2 via the increased degradation of p27 and its deregulation in tumors (reviewed in Frescas and Pagano 2008).

Inactivation of p27 also occurs via relocalization in the cytoplasm where it cannot exert its cyclin–CDK inhibitory role. Multiple phosphorylation events on p27 regulate its shuttling between the nucleus and the cytoplasm. Phosphorylation on Ser10 induces its export from the nucleus by creating a binding site for the CRM1 exportin (Rodier et al., 2001; Connor et al., 2003; Besson et al., 2006). Phosphorylation of p27 on Thr157 (a site that is not conserved in the mouse protein) or Thr198 can also result in the cytoplasmic sequestration of p27 via its association with 14-3-3 proteins (Liang et al., 2002; Fujita et al., 2003; Sekimoto et al., 2004; Shin et al., 2005; Chu et al., 2008; Hong et al., 2008). A number of studies investigated the clinical significance of cytoplasmic localization of p27, and it is generally associated with poor prognosis, high tumor grade or metastasis in several human malignancies (Liang et al., 2002; Besson et al., 2004a, 2008; Denicourt et al., 2007; Chu et al., 2008). These studies suggest that cytoplasmic p27 could actively participate in the process of malignant transformation and the aggressiveness of these tumors. Indeed, while cytoplasmic localization results in the loss of the cyclin–CDK inhibitory function of p27, it would also enhance cyclin–CDK-independent functions of p27 that are executed in the cell cytoplasm (Besson et al., 2008). For instance, we and others found that p27 could bind to the GTPase RhoA, preventing its activation, thereby regulating the actin cytoskeleton and cell migration and other aspects of RhoA signaling (Besson et al., 2004a, 2004b, 2008; Li et al., 2006; Papakonstanti et al., 2007; Larrea et al., 2009).

To genetically dissect and study p27 functions in vivo, we generated two lines of knockin mice carrying mutated alleles of the p27 gene. p27S10A mice were partially resistant to urethane-induced tumorigenesis, and this was associated with the fact that export of the p27S10A protein into the cytoplasm is impaired, resulting in a largely nuclear localization even in presence of activated K-Ras, which is known to induce the cytoplasmic localization of p27 (Liu et al., 2000; Besson et al., 2006). In the p27CK− mice, the p27 allele has two mutations (R30A and L32A) in the cyclin-binding domain and two mutations (F62A and F64A) in the CDK-binding domain, making p27CK− unable to bind to or inhibit cyclin–CDK complexes (Besson et al., 2006, 2007). Owing to its inability to interact with cyclin–CDKs, p27CK− is insensitive to Skp2-mediated degradation and is degraded only during the G1 phase, resulting in an overall increase in p27CK− levels (Besson et al., 2006). p27CK− mice have an increased incidence of spontaneous tumors in multiple organs, even compared with p27-knockout mice, indicating an oncogenic role for p27 that is unrelated to its ability to regulate cyclins and CDKs (Besson et al., 2007). In the lung, spontaneous tumor development was associated with the amplification of a progenitor/stem cell population, the bronchioalveolar stem cells (BASCs; Besson et al., 2007).

Urethane treatment predominantly causes activating mutations in the K-Ras proto-oncogene and is a potent carcinogen in mice, especially in the lung, causing the formation of adenomas and adenocarcinomas (Meuwissen and Berns, 2005; Kelly-Spratt et al., 2009). In this study, the use of the urethane-induced tumorigenesis model in p27CK− mice revealed a potent cooperation between p27CK− and activated K-Ras, which was associated with cytoplasmic localization of p27CK− and BASC amplification. The ability of p27CK− to function as a cooperating oncogene was not universal, however, as when c-Myc was used as a transforming agent, the status of p27 became irrelevant and c-Myc was equally potent in transforming wild-type (WT), p27−/− and p27CK− cells. Indeed, c-Myc had no effect on p27 localization but induced the degradation of WT p27, c-Myc being known to upregulates components of the SCF–Skp2 ubiquitin ligase, but not of the Skp2-insensitive p27CK−. Thus, we propose that the p27CK− protein needs to be localized in the cytoplasm in order to function as a cooperating oncogene during transformation.


p27CK− mice are more susceptible to urethane-induced tumorigenesis than p27-null animals

Loss of p27 considerably increases the incidence and spectrum of tumors after urethane treatment (Besson et al., 2006; Kelly-Spratt et al., 2009). As p27CK− mice are highly susceptible to spontaneous development of tumors in multiple organs, we tested how this allele would affect K-Ras-driven tumorigenesis and its ability to function as a cooperating oncogene. Cohorts of p27+/+, p27+/−, p27−/−, p27+/CK− and p27CK−/CK− animals were injected with urethane and tumor incidence was analyzed after 20 weeks. We previously used mice in a 129S4 background and p27−/− animals developed harderian gland tumors with a 100% incidence, creating much discomfort to these mice and early lethality (Besson et al., 2006). To avoid this problem, we used mice in the mixed 129S4/C57BL6 genetic background for this study. p27CK−/CK− had an increased frequency of hyperplastic lesions in several tissues compared with the other genotypes (Supplementary Table 1), including the liver, ovaries, pituitary, harderian gland and adrenals. As expected, tumor penetrance in the lung was complete in all genotypes except for WT (93.5%). The average number of lung tumors per animal increased proportionally to the loss of p27 alleles (Figure 1a), consistent with p27 being a haploinsufficient tumor suppressor (Fero et al., 1998). Interestingly, the presence of a single p27CK− allele was sufficient to give a phenotype similar to the p27-null (14.13 tumors per animal in p27+/CK− and 14.55 in p27−/−), and p27CK−/CK− mice developed significantly more tumors (21.31 tumors per animal) than their p27−/− counterparts. Likewise, tumor size was increased in p27CK−/CK− compared with other genotypes (Figures 1b and c). Thus, the p27CK− seems to function as a cooperating oncogene with K-Ras activation during tumorigenesis.

Figure 1
Figure 1

Increased incidence and growth rate of urethane-induced lung tumors in p27CK− mice. (a) Total number of lung tumors per animal 20 weeks after a single urethane injection. Lungs were dissected and macroscopic tumors were counted by visual inspection. The results were analyzed using a one-way analysis of variance with the Tukey–Kramer multiple comparison test. (***) P<0.001; (**) P<0.01; (*) P<0.05. (b) Quantification of lung tumors <2 mm in diameter.(c) Quantification of lung tumors over 2 mm in diameter. Statistical analyses in (b, c) were performed as in (a). Error bars on the graphs represent the standard error of the mean (s.e.m.).

Tumors arising in p27CK−mice are more aggressive

Histological analyses of hematoxylin/eosin-stained lung sections indicated that in their large majority the tumors that developed were alveolar/bronchiolar adenomas with either solid, papillary or mixed growth patterns (Figure 2). However, mice carrying at least one p27CK− allele also developed alveolar/bronchiolar carcinomas with heterogenous growth patterns and pleiomorphic cytomorphological features and were characterized by the presence of infiltrating macrophages (Figure 2a, bottom picture). Moreover, tumors arising in p27CK− animals often infiltrated the surrounding parenchyma, whereas they remained mostly encapsulated in the other genotype (Figure 2b). Cell proliferation, evaluated by Ki67 immunohistochemistry (IHC), was largely circumscribed to the tumor mass in WT, less so in p27−/− and frequent hyperproliferative regions were observed in the bronchioles of p27CK−/CK− lungs (Supplementary Figure 1), suggesting the presence of numerous pre-neoplastic lesions and consistent with our previous observations (Besson et al., 2007). Although all tumors arising in WT, p27−/− (not shown) or p27S10A (Supplementary Figure 3a) mice were positive for cytokeratin by IHC, a fraction of the tumors in p27CK− animals were cytokeratin negative, indicating more heterogeneity in the type of tumors developing in these animals (Figure 3a). Cytokeratin staining also underscored the infiltrative nature of the tumors in the p27CK− mice (Figure 3a, top pictures), with diffuse staining extending from the tumor mass. Over half of the p27CK− lungs displayed signs of inflammatory response characterized by macrophage infiltration in the alveolar space and in the tumors, as seen by IHC for the macrophage marker F4/80 (Figure 3b). Lymphocytic infiltration was also observed in 40% of animals, some of these were metastases from thymic lymphomas. CD3 and B220 IHC revealed that the majority of these lymphocytic infiltrates were of T-cell origin; however, some of these infiltrates were predominantly of B-cell origin or of mixed lineages (Supplementary Figure 2).

Figure 2
Figure 2

p27CK− mice develop more aggressive lung tumors than wild-type or p27−/− animals. (a) Lung tumor histology. Sections (5 μm) of paraffin-embedded lungs tissues from p27+/+, p27−/− and p27CK−/CK− mice were stained with eosin and hematoxylin. Arrowheads indicate sites of macrophage infiltration in p27CK−/CK− tumor. (b) Local invasion of lung tumors in the various genotypes. Tumor boundaries are indicated with the hatched lines on eosin and hematoxylin-stained sections. Magnification bars=100 μm.

Figure 3
Figure 3

IHC of urethane-induced lung tumors. (a) Pan-keratin IHC of urethane-treated p27+/+ and p27CK−/CK− lungs. (b) Macrophage infiltration in the lung and tumors of p27CK−/CK− mice was visualized by IHC for the macrophage marker F4/80. Magnification bars=100 μm.

p27CK− localizes in the cytoplasm of lung tumor cells

Mutations in the cdkn1b gene encoding p27 are exceedingly rare in tumors (Besson et al., 2008), including in lung tumors induced by urethane (Besson et al., 2006; Kelly-Spratt et al., 2009). In fact, p27 expression is usually maintained in these tumors but the protein is partially relocated to the cytoplasm of tumor cells (Besson et al., 2006; Kelly-Spratt et al., 2009). Ras activation is known to cause cytoplasmic translocation of p27, including in lung tumors, notably via the downstream activation of the Ral and phosphatidylinositol-3 kinase pathways (Liu et al., 2000; Kfir et al., 2005; Besson et al., 2006; Kelly-Spratt et al., 2009). Monitoring of p27 protein levels in urethane-induced lung tumors confirmed the maintenance of both WT p27 and p27CK− expression in these tumors at levels similar to that of the normal tissue (Figure 4a). Phosphorylation of p27 (and p27CK−) on Ser10, which we previously found to be important for cytoplasmic translocation of p27 in urethane-induced tumors (Besson et al., 2006), was abundant and proportional to total p27 levels. Tumors had increased proliferating cell nuclear antigen levels compared with normal lung, indicating increased cell proliferation in the tumors (Figure 4a). Interestingly, ‘normal’ (not treated with urethane) lung of p27CK− mice had intermediate proliferating cell nuclear antigen levels between the WT normal lung and tumors, consistent with our previous observation that p27CK− lungs exhibit aberrant bronchiolar proliferation, BASC amplification and spontaneously develop tumors (Besson et al., 2007). We used extracellular signal-regulated kinase (ERK) activation as a marker of Ras activation, and almost all tumors had elevated levels of phospho-ERK, thus confirming the urethane-induced K-Ras activation in these tumors (Figure 4a).

Figure 4
Figure 4

Mislocalization of p27 in lung tumors. (a) Immunoblot analysis of normal lung tissue and individual lung tumors from p27+/+, p27+/CK− and p27CK−/CK− urethane-treated mice. Membranes were probed successively with antibodies against phospho-Ser10 p27, p27, proliferating cell nuclear antigen (PCNA), phospho-ERK (active) and Grb2 for loading control. (b) p27 IHC of normal lung tissue and tumors. Sections (5 μm) of paraffin-embedded lung tissues from p27−/−, p27+/+ and p27CK−/CK− mice were stained using a polyclonal p27 antibody (C19, Santa Cruz Biotechnology). Magnification bars=50 μm. (c) Nuclear/cytoplasm ratio in bronchiolar cells and in tumors. p27 signal intensity in the cytoplasm and nucleus was quantified on images of IHC-stained tissues (20 cells per genotype and per treatment) and the nuclear to cytoplasm ratio calculated. The red dot indicates the average value of the plotted data points in each condition. The nuclear to cytoplasm ratio was significantly decreased in WT tumors versus WT bronchioles (**) P<0.01 and in CK tumors versus CK bronchioles (*) P<0.05. Statistical analyses were conducted using a non-parametric Kruskal–Wallis test with Dunn's multiple comparisons test.

p27 IHC on lung sections revealed that WT p27 and p27CK− behaved similarly: in the normal lung, the majority of p27 is found in the nucleus of bronchiolar cells, whereas it was partially translocated to the cytoplasm in tumor cells (Figure 4b). In contrast, p27S10A, for which nuclear export is impaired, was nearly exclusively present in the nucleus in both bronchiolar and lung tumor cells (Supplementary Figure 3b), as observed previously (Besson et al., 2006). Calculation of the nuclear to cytoplasm ratio of p27 in bronchiolar and tumor cells confirmed that a higher fraction of p27 localized in the cytoplasm of tumor cells compared with bronchiolar cells both in p27+/+ (P<0.01) and p27CK− tissues (P<0.05), although a slightly higher fraction of p27CK− was present in the cytoplasm even in bronchiolar cells (Figure 4c). Thus, p27CK− seems to function as a cooperating oncogene with K-Ras and this correlates with the cytoplasmic localization of the protein in tumors, whereas p27S10A remains nuclear and these mice are partially resistant to urethane-induced tumorigenesis (Besson et al., 2006).

Urethane causes bronchioalveolar stem cell amplification

K-Ras activation was previously reported to induce the amplification of BASCs, the lung stem cells that reside at the bronchioalveolar duct junction and are positive for both Clara cell-specific protein (CCSP/CC10) and the type II pneumocyte marker surfactant protein C (SP-C) (Kim et al., 2005; Dovey et al., 2008). We also found that this stem cell population was amplified in p27CK− mice and were likely to be the cells from which lung tumors spontaneously develop in these mice (Besson et al., 2007). We, therefore, examined the status of this stem cell population by double CCSP/SP-C immunostaining in urethane-treated lungs (Figure 5). Consistent with urethane causing mutations in K-Ras, BASC amplification was observed in all genotypes (BASCs are indicated with asterisks or circles, Figure 5). However, these were infrequent events in WT and p27−/− mice, with approximately one to three patches of BASC amplification per section, whereas they were much more frequent in p27+/CK− and p27CK−/CK− lungs with over 20% bronchioles affected. Regions of BASC amplification feeding into the tumors could be observed (p27+/CK− and p27CK−/CK− right images, tumors are brightly labeled with SP-C in red). BASC amplification was more extensive in urethane-treated lungs than previously observed in untreated mice (Besson et al., 2007). As seen with tumorigenesis, this data suggests that activated K-Ras and p27CK− cooperate to cause BASC amplification.

Figure 5
Figure 5

p27CK− and K-Ras activation cooperate to cause BASC amplification. Paraffin-embedded lung sections from urethane-treated p27+/+, p27−/−, p27+/CK− and p27CK−/CK− mice were stained for the Clara cell marker CCSP/CC10 (goat anti-CC10, green) and the type II pneumocyte marker surfactant protein C (SP-C; rabbit anti-SPC, red) and merged images are shown. BASCs are double positive for these proteins. Individual BASCs are indicated on the images using asterisks (*) and clusters of BASCs are shown with white circle. Tumor cells are SP-C positive (p27+/CK− and p27CK−/CK−, right pictures). Magnification bars=50 μm.

Oncogenic cooperation by p27CK− requires its cytoplasmic localization

We next wanted to determine whether p27CK− would cooperate with another oncogene during transformation. For this we chose to use c-Myc as it does not transform cells by the same mechanism as Ras. In colony formation in soft agar assays, activated K-RasV12 only inefficiently transformed human papillomavirus E6 immortalized WT, p27S10A and p27−/− mouse embryonic fibroblasts (MEFs). In contrast, p27CK− MEFs were transformed by K-Ras (Figure 6) with a much greater frequency (P<0.001 compared with the other genotypes), consistent with the cooperation we observed in vivo. In sharp contrast, c-Myc was equally potent at inducing growth in soft agar in all genotypes. Thus, it seems that c-Myc-driven transformation is independent of p27 status. As K-Ras drives p27, at least in part, into the cytoplasm in lung tumors, we hypothesized that the ability of p27CK− to function as an oncogene was linked to its cytoplasmic localization. To test this, we compared the ability of activated K-Ras to transform cells expressing mutants of p27 that are predominantly sequestered in the nucleus (Besson et al., 2006). In these soft agar assays, the oncogenic activity of the CK mutation was greatly diminished when this mutation was present in the S10A background (p27S10A/CK−; P<0.001; Supplementary Figure 4). Thus, our data indicates that p27CK− needs to relocalize to the cytoplasm in order to function as a cooperating oncogene.

Figure 6
Figure 6

c-Myc-mediated transformation is independent of p27 status. Human papillomavirus-E6-immortalized MEFs were infected or not with retroviruses encoding K-RasV12 or c-Myc and used in soft agar colony formation assays. Experiments (n=4) were performed in triplicates and at two seeding densities (see Materials and methods). Representative results from experiments with 5 × 103 cells seeded per well are shown. After 4 weeks, colonies were stained with 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) and images were acquired at a × 12 magnification. The graph shows the average number of colonies (counted on the images from each experiment; error bars represent the s.e.m.). p27CK− and activated K-Ras cooperated for transformation compared with the other genotypes expressing K-Ras V12 (***, P<0.001). On the other hand, c-Myc was equally potent in transforming all genotypes. Also, c-Myc was significantly more potent at inducing growth in soft agar than E6 alone or E6+K-Ras V12 (***, P<0.001 for all other conditions except *, P<0.05 for p27−/− c-Myc versus p27CK− K-Ras V12 and ns, P>0.05 for p27S10A c-Myc vs p27CK− K-Ras V12). Statistical analyses were conducted using a one-way analysis of variance with a Tukey–Kramer multiple comparisons test.

The best-characterized function of p27 is the inhibition of cyclin–CDK complexes. To ascertain that differences in CDK-associated activities were not responsible for the different phenotypes observed between p27−/− and p27CK− cells, we performed in vitro kinase assays on immunoprecipitated cyclin E, CDK2 and CDK1 from exponentially growing MEFs of the different genotypes expressing or not K-Ras or c-Myc (Supplementary Figure 5). As previously reported (Besson et al., 2007), no significant difference of CDK activity was observed between p27−/− and p27CK− cells, whereas CDK activity was slightly increased in p27+/+ cells in presence of c-Myc. Similarly, no significant difference in CDK2 activity was detected in lung tumors from p27+/+, p27−/− and p27CK− mice (Supplementary Figure 5). Thus, we conclude that differences in CDK activity do not account for the different susceptibility to transformation by K-Ras activation.

To understand the basis for the differences between K-RasV12 and c-Myc's abilities to drive cellular transformation in vitro, we examined p27's localization and expression levels by immunostaining and immunoblotting in MEFs expressing c-Myc or K-RasV12. In human papillomavirus E6-immortalized MEFs, WT p27 and p27CK− were abundant in the nucleus and displayed only very weak cytoplasmic staining as expected (Figures 7a and b). However, in the presence of activated K-Ras, both WT p27 and p27CK− were partially relocalized to the cytoplasm as observed previously for WT p27 (increased cytoplasmic p27 in WT E6 K-Ras versus WTE6, P<0.001; Figures 7a and b; Besson et al., 2006). In contrast, WT p27 and p27CK− had completely different behaviors in the presence of c-Myc: WT p27 nuclear levels were reduced in WT E6 c-Myc compared with WT E6 cells (P<0.001), whereas neither p27CK− levels nor its subcellular localization were affected by c-Myc and p27CK− remained largely nuclear in c-Myc-expressing cells (Figures 7a and b). The cytoplasmic localization of WT p27 induced by Ras and its downregulation by c-Myc were confirmed in fractionation experiments (Figure 7c). K-Ras did not significantly affect p27 levels by immunoblotting (Figure 7d). On the other hand, WT p27 was decreased in presence of c-Myc and p27CK− levels were unaffected (Figure 7e).

Figure 7
Figure 7

K-Ras causes cytoplasmic exclusion of p27, whereas c-Myc causes downregulation of WT p27. (a) Immunofluorescence analysis of p27 stained (left column) p27−/−, p27+/+ and p27CK−/CK− MEFs infected or not with retroviruses encoding K-RasV12 or c-Myc counterstained with Hoechst H33342 (merge column, right). Exposure time was determined on the p27−/− MEFs to account for background in the p27 channel and the same camera settings were used to acquire all images in this series. Magnification bars=50 μm. (b) Quantification of p27 signal in the nucleus and cytoplasm. Graph shows the average p27 signal intensity in the nucleus and cytoplasm determined in 30 cells per condition, nuclei position was determined by Hoechst H33342 staining. Errors bars represent the s.e.m. p27 signal in the cytoplasm of WT E6+K-Ras V12 cells was significantly higher than in WT E6 cells (***, P<0.001). Also, nuclear p27 signal in WT E6+c-Myc was significantly decreased compared with WT E6 (***, P<0.001). Statistical analyses were conducted using a one-way analysis of variance with a Tukey–Kramer multiple comparisons test. (c) Biochemical fractionation of MEFs. Nuclear and cytoplasmic fractions were prepared as described in (Kelly-Spratt et al., 2009) and 80 μg of proteins were resolved on SDS–polyacrylamide gel electrophoresis, transferred on polyvinylidene difluoride and probed for p27, stripped and reprobed for lamin A. (d) Immunoblot of MEFs expressing or not activated K-Ras probed successively for p27, K-Ras and Grb2 for protein loading. (e) Immunoblotting for p27 levels in c-Myc-infected cells. MEFs-expressing c-Myc or not were probed successively for p27, c-Myc and Grb2 for protein loading.

Thus, it seems that c-Myc induces the downregulation of WT p27, thereby making WT cells similar to p27-null cells. On the other hand, c-Myc is unable to induce p27CK− downregulation, but the protein is maintained in the nucleus, where it behaves similar to a null because of its inability to bind or inhibit cyclin–CDK complexes. Altogether, our data suggests that in order to function as an oncogene, p27CK− needs to localize in the cytoplasm, otherwise it just behaves similar to a null allele.


In this study, we provide further insight into the roles played by p27 during tumorigenesis and cellular transformation. Our evidence indicates that the p27CK− allele function as a cooperating oncogene with activated K-Ras, but not with c-Myc. Moreover, our data suggests that the basis for p27CK− oncogenic cooperation is its localization in the cytoplasm where it exerts cyclin–CDK-independent functions, such as the regulation of Rho GTPase signaling.

p27 seems to be exquisitely sensitive to genetic background, as there are striking differences in tumor incidence and spectrum in the p27 knockout between our previous study carried out in pure 129S4 genetic background (Besson et al., 2006) and the present one carried out in a mixed 129S4/C57BL6 background. Indeed, these differences suggest the existence of one or more modifier gene in 129S4 that accentuate the phenotypes caused by the loss of p27 since in that background, urethane-treated mice developed more tumors in the pituitary (100% in 129S4 versus 40% in 129/B6), harderian gland (100 versus 0%), ovaries (82 versus 37.5%) and uterus (73 versus 0%). One exception is observed in the case of lymphomas for which the susceptibility is reversed: none were observed in 129S4 p27−/− animals compared with 40% in the 129/B6 p27−/− mice. Identifying this modifier gene would certainly provide invaluable insight into p27 function and its role in tumor suppression.

Urethane treatment also increased the phenotype of macrophage infiltration that is observed in p27CK− mice. A current research direction is to investigate what causes this inflammatory phenotype in these animals and whether the presence of these immune cells may participate in the development of tumors in the lung. Indeed substantial evidence indicate a prominent role for inflammation during tumor progression (reviewed in Condeelis and Pollard, 2006).

Our major finding is that p27CK− does not universally cooperate with other oncogenes during transformation, and that in order to function as an oncogene it needs to localize in the cytoplasm. Ras activation indirectly causes cytoplasmic localization of p27 via the activation of its effectors Ral-GEF/Ral and phosphatidylinositol-3 kinase/Akt and Raf1/MEK/ERK pathways (Liu et al., 2000; Kfir et al., 2005; Besson et al., 2006). Phosphorylation of p27 on three different sites has been shown to result in cytoplasmic localization: phosphorylation on Ser10 by Akt, ERK2, CDK5 or hKIS promotes CRM1-mediated nuclear export; whereas phosphorylation on Thr157 by Akt, SGK1 or Pim, or on Thr198 by Akt, RSK1, Pim, or LKB1/AMPK were shown to result in cytoplasmic retention of the protein via binding with 14-3-3 proteins and also for Thr157 by inactivating the nuclear localization sequence of p27 (Boehm et al., 2002; Ishida et al., 2002; Liang et al., 2002; Fujita et al., 2002, 2003; Shin et al., 2005; Kawauchi et al., 2006; Chu et al., 2008; Hong et al., 2008; Morishita et al., 2008; Short et al., 2008; Larrea et al., 2009). In fact, several laboratories found that cytoplasmic localization of p27 promotes its oncogenic activity or that its nuclear sequestration inhibits tumorigenesis (Besson et al., 2006, 2007; Wu et al., 2006; Denicourt et al., 2007; Kelly-Spratt et al., 2009). Moreover, the cytoplasmic localization of p27 in subsets of different tumor types has been associated with aggressive tumors, poor prognosis and metastasis (reviewed in Besson et al., 2004a; Chu et al., 2008).

In our experiments, p27+/+ and p27−/− MEFs were equally susceptible to c-Myc-driven transformation. This was surprising given that in a lymphomagenesis assay in vivo, Myc activation by Moloney murine leukemia virus insertion occurred more frequently in p27−/− than in WT mice (Hwang et al., 2002; Martins and Berns, 2002). p27 deficiency also increased lymphomagenesis in Eμ-Myc mice (Martins and Berns, 2002). However, when Eμ-Myc transgenics were crossed in a Skp2-null background, resulting in elevated p27 levels compared with WT, lymphomagenesis was not significantly reduced (Old et al., 2010). Interestingly, in MEFs, the ability of c-Myc to induce proliferation was unaffected by the loss of p27 (Berns et al., 2000). One possibility is that loss of p27 may accelerate Myc-driven tumorigenesis in the absence of other genetic lesion, and that in our model in which p53 is inactivated by human papillomavirus E6, the loss of p27 is no longer critical for Myc-mediated transformation.

We hypothesize that the inability of p27CK− to cooperate with c-Myc in cellular transformation is because of the fact that c-Myc does not induce the cytoplasmic localization of p27CK−. On the other hand, c-Myc caused the downregulation of WT p27. In fact c-Myc is known to cause p27 downregulation by inducing the expression of several components of the SCF-Skp2 pathway, such as Cul1, Cks1 and Skp2; as well as by inducing the expression of cyclin E, which promotes p27 phosphorylation on Thr187 and its degradation, and the expression of cyclin D2 and CDK4, which can sequester p27 away from cyclin E/CDK2 complexes (Leone et al., 1997; Muller et al., 1997; Bouchard et al., 1999; Hermeking et al., 2000; O’Hagan et al., 2000; Keller et al., 2007; Old et al., 2010). In all likelihood, the lack of effect of c-Myc on the p27CK− protein is because of the inability of p27CK− to be phosphorylated on Thr187 by cyclin E/CDK2 and targeted for degradation by the Skp2 pathway (Besson et al., 2006, 2007). Determining the status of p27 in tumors (both levels and localization) could represent a useful indicator for treatment choices. Moreover, the targeted inhibition of the pathways causing the cytoplasmic localization of p27 would both serve to re-activate the cyclin–CDK inhibitory function of the protein and inactivate its cytoplasmic, oncogenic function.

Materials and methods

Urethane-induced tumorigenesis

p27+/− and p27+/CK− mice in the mixed 129S4/C57BL6 genetic background were bred to generate the p27+/+, p27+/−, p27−/−, p27+/CK− and p27CK−/CK− animals used in the study. Mouse genotyping was performed as previously (Fero et al., 1996; Besson et al., 2006). The mice were maintained and procedure performed in accordance with federal and institutional regulations. Mice between 21 and 28 days old received a single intraperitoneal injection of urethane (1 g/kg). Mice were killed after 20 weeks, dissected and examined for gross lesions of internal organs. Lungs were dissected and lung tumors were counted and their size measured. Tissues were fixed overnight in formalin, transferred in 70% ethanol for 24 h and embedded in paraffin. Paraffin blocks were sectioned at 5 μm thickness for histochemistry or immunostaining.

Histology and IHC

Paraffin sections were deparaffinized and either stained with hematoxylin and eosin, or used for immunostaining. IHC using anti-p27 (C19, Santa-Cruz, Biotechnology Inc, Santa Cruz, CA, USA), pan-keratin multi Ab-1 (clone C-11, Thermo Scientific, Waltham, MA, USA) or F4/80 pan macrophage marker (BM8, AdB Serotec, Oxford, UK) antibodies were performed using the VectaStain ABC Kit (Vector Laboratories, Burlingame, CA, USA) following the manufacturer's instructions. Briefly, hydrated tissue sections were steamed for 30 min in low pH unmasking solution (Vector Laboratories) and blocked in serum for 20 min at room temperature. Samples were incubated with the indicated antibody for 1 h and then with biotinylated secondary antibody for 30 min and visualized using the chromogen 3′3′-diaminobenzidine (ImmPACT-DAB, Vector Laboratories). Except for p27 IHC, slides were counterstained with Mayer's hematoxylin for 20 s and rinsed abundantly in H20 before dehydration and mounting.

BASC immunostaining was performed as previously (Besson et al., 2007). Images were acquired on a Nikon Eclipse 90i microscope equipped with CoolSnap HQ2 and Nikon DS-Fi1 cameras using the NIS-Br software (Nikon Instruments Europe, Evry, France).

Western blotting

Cells and tumor samples were lysed in lysis buffer (50 mM hepes pH 7.5, 150 mM NaCl, 1 mM EDTA, 2.5 mM EGTA, 0.1% Tween-20, 10% glycerol and 1% NP-40 complemented with 1 mM dithiothreitol, 10 mM β-glycerophosphate, 10 mM NaF, 10 mM sodium orthovanadate, 10 μg/ml each of leupeptin, aprotinin and pepstatin A). Tissue samples were crushed with a pestle. After sonication for 10 s, extracts were centrifuged for 5 min at 12 000 r.p.m. to retrieve supernatants. Proteins were separated on SDS–polyacrylamide gel electrophoresis 12% gels and transferred onto polyvinylidene difluoride membranes (Millipore, Billerica, MA, USA). After blocking for 1 h in phosphate-buffered saline (PBS) 0.1% Tween 10% milk, membranes were incubated overnight at 4 °C with the indicated primary antibodies diluted in PBS containing 0.1% Tween-20 and 5% milk. Membranes were rinsed three times in PBS 0.1% Tween and incubated with secondary horseradish peroxidase-conjugated antibodies (Jackson ImmunoResearch Laboratory, West Grove, PA, USA) for 4 h at room temperature. Immunodetection was performed with the ECL kit (GE Healthcare, Piscataway, NJ, USA). The antibodies used were: polyclonals against p27 (C19), phospho-Ser10 p27 and lamin A (H102) and monoclonals against proliferating cell nuclear antigen (PC10), Myc (9E10) and K-Ras (F234) were from Santa Cruz Biotechnology. Rabbit anti phospho-ERK1/2 (Thr202/Tyr204) was from Cell Signaling Technology Inc (Danvers, MA, USA) and mouse anti Grb2 was from BD-Transduction Laboratories (Franklin Lakes, NJ, USA). Nuclear/cytoplasmic fractionation was performed as described in Kelly-Spratt et al. (2009).


Cells were seeded on coverslips and grown overnight before being fixed with 2% paraformaldehyde in PBS for 20 min at 37 °C. Following permeabilization for 3 min with PBS 0.2% Triton X-100, cells were rinsed three times in PBS and incubated for 1 h at 37 °C with primary antibodies (p27, C19) diluted 1/100 in PBS 3% bovine serum albumin 0.05% Tween and 0.08% sodium azide. After three 5 min washes in PBS, cells were incubated for 30 min at 37 °C with Cy2-conjugated secondary antibodies at a 1/400 dilution. Coverslips were then incubated in PBS containing 0.1 μg/ml Hoechst H33342, washed twice in PBS and mounted on glass slides. Image analysis and signal quantification on immunofluorescence and IHC pictures was performed using the NIS-Br software (Nikon) and statistical analyses using the GraphPad Instat 3 software (GraphPad Software Inc, La Jolla, CA, USA).

Cell culture and colony formation in soft agar assays

Cells were grown at 37 °C and 5% CO2 in Dulbecco's modified Eagle medium, 2 mM glutamax, 4.5 g/l glucose supplemented with 10% fetal calf serum, 0.1 mM non-essential amino acids, 2 μg/ml penicillin–streptomycin and 1 mM sodium pyruvate. All tissue culture reagents were from Life Technologies (Carlsbad, CA, USA). Primary MEFs were prepared as described previously (Besson et al., 2006). MEFs were immortalized by infection with retroviruses encoding the human papilloma virus E6 protein and selected with hygromycin. MEFs were infected with pWZL-c-Myc or a pBabe-Puro K-RasV12 vector and were selected with 2 μg/ml blasticydin or with 4 μg/ml puromycin, respectively.

Soft agar assays were performed in six-well plates, in triplicate for each cell type and condition with either 5 × 103 or 5 × 104 cells. A volume of 3 ml of a 0.8% SeaPlaque low melting agarose (Lonza, Basel, Switzerland)–Dulbecco's modified Eagle medium 10% fetal calf serum formed the lower layer. After solidification of the first layer, 3 ml of 0.5% agarose–Dulbecco's modified Eagle medium 10% fetal calf serum containing the cells were added and let a few minutes a room temperature to solidify. Finally, the cell layer was overlaid with 3 ml of 0.8% agarose–Dulbecco's modified Eagle medium 10% fetal calf serum. Soft agar cultures were incubated for 4 weeks at 37 °C and 5% CO2. Live cells were then stained with 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) for 4 h at 37 °C and 5% CO2. Pictures of the wells were taken on a Nikon SMZ800 dissection microscope (Nikon) at a magnification of × 12 with a Nikon DS-Fi1 camera (Nikon) using the NIS-Br software. Colonies were counted on picture using Image J and statistical analyses using the GraphPad Instat 3 software.


  1. , , , , , . (2000). p27kip1-independent cell cycle regulation by MYC. Oncogene 19: 4822–4827.

  2. , , . (2004a). Regulation of the cytoskeleton: an oncogenic function for CDK inhibitors? Nat Rev Cancer 4: 948–955.

  3. , , . (2008). CDK inhibitors: cell cycle regulators and beyond. Dev Cell 14: 159–169.

  4. , , , , , . (2006). A pathway in quiescent cells that controls p27Kip1 stability, subcellular localization and tumor suppression. Genes Dev 20: 47–64.

  5. , , , , . (2004b). p27Kip1 modulates cell migration through the regulation of RhoA activation. Genes Dev 18: 862–876.

  6. , , , , , et al. (2007). Discovery of an oncogenic activity in p27Kip1 that causes stem cell expansion and a multiple tumor phenotype. Genes Dev 21: 1731–1746.

  7. , , , , , et al. (2002). A growth factor-dependent nuclear kinase phosphorylates p27Kip1 and regulates cell cycle progression. EMBO J 21: 3390–3401.

  8. , , , , , et al. (1999). Direct induction of cyclin D2 by Myc contributes to cell cycle progression and sequestration of p27. EMBO J 18: 5321–5333.

  9. , , . (2008). The Cdk inhibitor p27 in human cancer: prognostic potential and relevance to anticancer therapy. Nat Rev Cancer 8: 253–267.

  10. , , . (2010). Role of BCR-ABL-Y177-mediated p27kip1 phosphorylation and cytoplasmic localization in enhanced proliferation of chronic myeloid leukemia progenitors. Leukemia 24: 779–787.

  11. , . (2006). Macrophages: obligate partners for tumor cell migration, invasion, and metastasis. Cell 124: 263–266.

  12. , , , , , et al. (2003). CRM1/Ran-mediated nuclear export of p27(Kip1) involves a nuclear export signal and links p27 export and proteolysis. Mol Biol Cell 14: 201–213.

  13. , , , , . (2007). Relocalized p27Kip1 tumor suppressor functions as a cytoplasmic metastatic oncogene in melanoma. Cancer Res 67: 9238–9243.

  14. , , , . (2008). Bmi1 is critical for lung tumorigenesis and bronchioalveolar stem cell expansion. Proc Natl Acad Sci USA 105: 11857–11862.

  15. , , , , . (1998). The murine gene p27Kip1 is haplo-insufficient for tumour suppression. Nature 396: 177–180.

  16. , , , , , et al. (1996). A syndrome of multiorgan hyperplasia with features of gigantism, tumorigenesis, and female sterility in p27(Kip1)-deficient mice. Cell 85: 733–744.

  17. , . (2008). Deregulated proteolysis by the F-box proteins SKP2 and beta-TrCP: tipping the scales of cancer. Nat Rev Cancer 8: 438–449.

  18. , , , . (2002). Akt-dependent phosphorylation of p27Kip1 promotes binding to 14-3-3 and cytoplasmic localization. J Biol Chem 277: 28706–28713.

  19. , , . (2003). Phosphorylation of p27Kip1 at threonine 198 by p90 ribosomal protein S6 kinases promotes its binding to 14-3-3 and cytoplasmic localization. J Biol Chem 278: 49254–49260.

  20. , . (2000). The hallmarks of cancer. Cell 100: 57–70.

  21. , , , , , et al. (2007). Pirh2 promotes ubiquitin-dependent degradation of the cyclin-dependent kinase inhibitor p27Kip1. Cancer Res 67: 10789–10795.

  22. , , , , , et al. (2000). Identification of CDK4 as a target of c-MYC. Proc Natl Acad Sci USA 97: 2229–2234.

  23. , , , , , . (2008). mTOR-raptor binds and activates SGK1 to regulate p27 phosphorylation. Mol Cell 30: 701–711.

  24. , , , , , et al. (2002). Identification of oncogenes collaborating with p27Kip1 loss by insertional mutagenesis and high-throughput insertion site analysis. Proc Natl Acad Sci USA 99: 11293–11298.

  25. , , , , , . (2002). Phosphorylation of p27Kip1on serine 10 is required for its binding to CRM1 and nuclear export. J Biol Chem 277: 14355–14358.

  26. , , , , , et al. (2004). Cytoplasmic ubiquitin ligase KPC regulates proteolysis of p27(Kip1) at G1 phase. Nat Cell Biol 6: 1229–1235.

  27. , , , . (2006). Cdk5 phosphorylates and stabilizes p27kip1 contributing to actin organization and cortical neuronal migration. Nat Cell Biol 8: 17–26.

  28. , , , , , et al. (2007). Myc targets Cks1 to provoke the suppression of p27Kip1, proliferation and lymphomagenesis. EMBO J 26: 2562–2574.

  29. , , , , , . (2009). Inhibition of PI-3K restores nuclear p27Kip1 expression in a mouse model of Kras-driven lung cancer. Oncogene 28: 3652–3662.

  30. , , , , , . (2005). Pathway- and expression level -dependent effects of oncogenic N-Ras: p27Kip1 mislocalization by the Ral-GEF pathway and Erk-mediated interference with Smad signaling. Mol Cell Biol 25: 8239–8250.

  31. , , , , , et al. (2005). Identification of bronchioalveolar stem cells in normal lung and lung cancer. Cell 121: 823–835.

  32. , , , , , et al. (2009). RSK1 drives p27Kip1 phosphorylation at T198 to promote RhoA inhibition and increase cell motility. Proc Natl Acad Sci USA 106: 9268–9273.

  33. , , , , . (1997). Myc and Ras collaborate in inducing accumulation of active cyclin E/Cdk2 and E2F. Nature 387: 422–426.

  34. , , , , , et al. (2006). Cyclin D1 Induction of Cellular Migration Requires p27KIP1. Cancer Res 66: 9986–9994.

  35. , , , , , et al. (2002). PKB/Akt phosphorylates p27, impairs nuclear import of p27 and opposes p27-mediated G1 arrest. Nat Med 8: 1153–1160.

  36. , , , , , et al. (2000). Disruption of TGF-beta growth inhibition by oncogenic ras is linked to p27Kip1 mislocalization. Oncogene 19: 5926–5935.

  37. , , , , , . (2001). A mouse knock-in model exposes sequential proteolytic pathways that regulate p27Kip1 in G1 and S phase. Nature 413: 323–327.

  38. , . (2002). Loss of p27(Kip1) but not p21(Cip1) decreases survival and synergizes with MYC in murine lymphomagenesis. EMBO J 21: 3739–3748.

  39. , . (2005). Mouse models for human lung cancer. Genes Dev 19: 643–664.

  40. , , , , , et al. (1999). Ubiquitination of p27 is regulated by Cdk-dependent phosphorylation and trimeric complex formation. Genes Dev 13: 1181–1189.

  41. , , , , . (2008). Pim kinases promote cell cycle progression by phosphorylating and down-regulating p27Kip1 at the transcriptional and posttranscriptional levels. Cancer Res 68: 5076–5085.

  42. , , , , , et al. (1997). Cdk2-dependent phosphorylation of p27 facilitates its Myc-induced release from cyclin E/cdk2 complexes. Oncogene 15: 2561–2576.

  43. , , , , , et al. (2000). Myc-enhanced expression of Cul1 promotes ubiquitin-dependent proteolysis and cell cycle progression. Genes Dev 14: 2185–2191.

  44. , , , , , et al. (2010). Skp2 directs Myc-mediated suppression of p27Kip1 yet has modest effects on Myc-driven lymphomagenesis. Mol Cancer Res 8: 353–362.

  45. , , . (2007). The p110delta isoform of PI 3-kinase negatively controls RhoA and PTEN. EMBO J 26: 3050–3061.

  46. , , , , , et al. (2001). p27 cytoplasmic localization is regulated by phosphorylation on Ser 10 and is not a prerequisite for its proteolysis. EMBO J 20: 6672–6682.

  47. , , . (2004). 14-3-3 suppresses the nuclear localization of threonine 157-phosphorylated p27(Kip1). Embo J 23: 1934–1942.

  48. , . (1999). CDK inhibitors: positive and negative regulators of G1-phase progression. Genes Dev 13: 1501–1512.

  49. , , , . (2005). Phosphorylation of p27Kip1 at Thr-157 interferes with its association with importin {alpha} during G1 and prevents nuclear re-entry. J Biol Chem 280: 6055–6063.

  50. , , , , , et al. (2008). AMP-activated protein kinase signaling results in cytoplasmic sequestration of p27. Cancer Res 68: 6496–6506.

  51. , , , , , et al. (2006). Reduction of cytosolic p27(Kip1) inhibits cancer cell motility, survival, and tumorigenicity. Cancer Res 66: 2162–2172.

Download references


MPS is supported by a studentship from the Ministère de l’Enseignement Supérieur et de la Recherche. CC is supported by a fellowship from the Fondation pour la Recherche Médicale. This work was supported by NIH grant #1R01CA118043 to JMR. AB is supported by grants from the Association pour la Recherche sur le Cancer and Ligue Nationale Contre le Cancer.

Author information


  1. INSERM UMR1037-Cancer Research Center of Toulouse, Toulouse, France

    • M P Serres
    • , E Zlotek-Zlotkiewicz
    • , C Concha
    • , V Daburon
    •  & A Besson
  2. Université de Toulouse, Toulouse, France

    • M P Serres
    • , E Zlotek-Zlotkiewicz
    • , C Concha
    • , V Daburon
    •  & A Besson
  3. CNRS ERL5294, Toulouse, France

    • M P Serres
    • , E Zlotek-Zlotkiewicz
    • , C Concha
    • , V Daburon
    •  & A Besson
  4. Fred Hutchinson Cancer Research Center, Division of Basic Sciences, Seattle, WA, USA

    • M Gurian-West
    •  & J M Roberts


  1. Search for M P Serres in:

  2. Search for E Zlotek-Zlotkiewicz in:

  3. Search for C Concha in:

  4. Search for M Gurian-West in:

  5. Search for V Daburon in:

  6. Search for J M Roberts in:

  7. Search for A Besson in:

Competing interests

The authors declare no conflict of interest.

Corresponding author

Correspondence to A Besson.

Supplementary information

About this article

Publication history







Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc).

Further reading