Dissecting the Senescence-like Program in Tumor Cells Activated by Ras Signaling*

Activated Ras signaling can induce a permanent growth arrest in osteosarcoma cells. Here, we report that a senescence-like growth inhibition is also achieved in human carcinoma cells upon the transduction of H-RasV12. Ras-induced tumor senescence can be recapitulated by the transduction of activated, but not wild-type, MEK. The ability for H-RasV12 to suppress tumor cell growth is drastically compromised in cells that harbor endogenous activating ras mutations. Notably, growth inhibition of tumor cells containing ras mutations can be achieved through the introduction of activated MEK. Tumor senescence induced by Ras signaling can occur in the absence of p16 or Rb and is not interrupted by the inactivation of Rb, p107, or p130 via short hairpin RNA or the transduction with HPV16 E7. In contrast, inactivation of p21 via short hairpin RNA disrupts Ras-induced tumor senescence. In summary, this study uncovers a senescence-like program activated by Ras signaling to inhibit cancer cell growth. This program appears to be intact in cancer cells that do not harbor ras mutations. Moreover, cancer cells that carry ras mutations remain susceptible to tumor senescence induced by activated MEK. These novel findings can potentially lead to the development of innovative cancer intervention.

Activated Ras signaling can induce a permanent growth arrest in osteosarcoma cells. Here, we report that a senescence-like growth inhibition is also achieved in human carcinoma cells upon the transduction of H-Ras V12 . Ras-induced tumor senescence can be recapitulated by the transduction of activated, but not wild-type, MEK. The ability for H-Ras V12 to suppress tumor cell growth is drastically compromised in cells that harbor endogenous activating ras mutations. Notably, growth inhibition of tumor cells containing ras mutations can be achieved through the introduction of activated MEK. Tumor senescence induced by Ras signaling can occur in the absence of p16 or Rb and is not interrupted by the inactivation of Rb, p107, or p130 via short hairpin RNA or the transduction with HPV16 E7. In contrast, inactivation of p21 via short hairpin RNA disrupts Rasinduced tumor senescence. In summary, this study uncovers a senescence-like program activated by Ras signaling to inhibit cancer cell growth. This program appears to be intact in cancer cells that do not harbor ras mutations. Moreover, cancer cells that carry ras mutations remain susceptible to tumor senescence induced by activated MEK. These novel findings can potentially lead to the development of innovative cancer intervention.
Current strategies for treating cancer largely rely on inducing apoptosis to eradicate tumor cells (1). Indeed, many cancer therapeutic agents are useful due to their ability to induce cell death or apoptosis in tumor cells. Despite the promising outcomes, apoptosis-based therapeutic strategies can be limited by high toxicity and lack of selectivity (2). In addition, many cancer cells are refractory to chemotherapeutic agents intended to induce apoptosis (3,4). It is necessary, therefore, to explore alternative means of tumor suppression.
Senescence was originally defined as the limited replicative capacity of normal human diploid fibroblasts in culture (hence, termed replicative senescence) (5). It is now understood that replicative senescence occurs as a result of progressive telomere shortening, leading to permanent growth arrest of cells and a defined lifespan (6 -8). Emerging evidence, however, suggests that a senescence phenotype can also be triggered through mechanisms independent of telomere shortening, such as DNA damage, oxidative stress, ionizing radiation, and oncogene activation (9 -16). Importantly, several recent in vivo and in vitro studies have demonstrated that tumor cells also undergo senescence in response to various cancer treatments (17)(18)(19). These observations raise the possibility that senescence might be a useful therapeutic option to treat cancer. The molecular mechanisms involved in tumor senescence, however, remain poorly understood.
Senescence is a signaling program that can be activated through various stimuli to induce permanent growth arrest of cells (18). Ras proteins are vital signaling molecules that mediate various cellular responses (20). The ras family comprises three genes that encode highly homologous protein isoforms: H-, N-, and K-Ras (21)(22)(23)(24). A single mutation on codons 12, 13, or 61 renders the proteins constitutively active, leading to the aberrant activation of their downstream effector pathways, including Raf, phosphatidylinositol 3-kinase, and Ral.GDS pathways (24,25). Activated Ras proteins can transform many immortal cells to a tumorigenic state and are thus known to have oncogenic potential (26,27). Indeed, activating mutations in ras are found in Ͼ30% of human cancers (24). Despite its role as an oncogene, transduction of H-Ras V12 induces a senescence program in normal human and rodent cells (9,28,29). H-Ras V12 -induced senescent cells share many features with cells that have undergone replicative senescence. These include flat and enlarged morphology, the induction of specific markers associated with senescence, including elevated senescence-associated (SA) 2 -␤-galactosidase and senescence-associated heterochromatic foci (SAHF) formation, as well as the induction of cell cycle regulators, such as p16, p53, and p21 (9, 28 -31). Studies done using genetic and pharmacological approaches reveal that Ras-induced senescence is mediated through the activation of the Raf/MEK/MAPK pathway (15,28,32). Furthermore, normal human cells lacking functional p16 are resistant to Ras-induced senescence (33)(34)(35), and inactivation of Rb via shRNA also allows human fibroblasts to bypass Ras-induced senescence (31). These studies indicate that the Ras/Raf/MEK/ MAPK signaling pathway is linked to a senescence program in normal human cells. Moreover, the p16/Rb tumor suppressor pathway is crucial for Ras-mediated senescence in normal human fibroblasts.
We have recently shown that transduction of H-Ras V12 induces a permanent growth inhibition in human osteosarcoma U2OS cells (36). In contrast to normal human fibroblasts, U2OS cells undergo H-Ras V12 -induced growth inhibition in the absence of p16 (34 -37). These results suggest that an alternative senescence-like program can be activated by Ras signaling to suppress tumor cell growth (36). In the current study, the potential for activated Ras signaling to induce tumor senescence is investigated in various human carcinoma cell lines. To gain insight into the molecular basis of Ras-induced tumor senescence, an RNA interference approach was used to determine the roles of important cell cycle regulators involved. Our findings uncover a potentially novel tumor suppression program induced by activated Ras signaling.
Retroviral Gene Transfer-The following retroviral vectors were used: pBabe-puro and its derivatives expressing H-Ras V12 (9), MEK Q56P , or wild-type MEK. Retroviral gene transfer was performed as previously described (28). Retroviral constructs (pLXSN.neo) expressing E7 or an empty vector control were kindly provided by Dr. G. Ferbeyre at the University of Montreal (38). Briefly, helper virus-free retroviruses were produced by Phoenix retrovirus packaging cells and were used to infect target cells. Stably transduced cell populations were established by selection in puromycin (2.5 g/ml) or G418 (400 g/ml).
Growth Curve Analyses-Growth curves were generated by plating cells (in duplicate) in 12-well dishes at 2 ϫ 10 4 cells/well on day 0 post-selection. Total viable cell number was counted at 2, 4, and 8 days post-plating using a Coulter Counter (Beckman). To obtain a value for time zero, cells were trypsinized 6 h after plating and counted. Each experiment was repeated at least three times, and results are displayed as a line graph (with standard deviation) showing the -fold increase in total cell number as compared with values obtained at time zero.
BrdUrd Incorporation Assays-BrdUrd incorporation was carried out as previously described with minor modifications (29). Briefly, cells from each cell type were plated into chamber culture slides (Falcon). The next day, cells were pulsed with bromodeoxyuridine (BrdUrd, Sigma) and 5-fluoro-2Ј-deoxyuridine (Sigma) for 30 min. After pulsing, cells were fixed, permeabilized, and denatured. BrdUrd incorporation was measured by immunofluorescence using an anti-BrdUrd antibody (BD Pharmingen). Alexa Fluor Conjugate 488 (Molecular Probes) was used as a secondary antibody, and the nuclei were stained with 4Ј,6-diamidino-2-phenylindole (DAPI, Sigma). The stained cells were visualized under a fluorescence microscope (Zeiss), and at least 200 cells were scored for BrdUrd incorporation. Each experiment was repeated at least three times, and the results are displayed as percentage of BrdUrdpositive cells (with standard deviation) in a histogram.
Colony Formation Assay-Cells of each cell type were plated in 6-well dishes (5 ϫ 10 3 cells/well for MCF7 cells and MDA-MB231 cells, 1 ϫ 10 3 cells/well for MDA-MB468, HCT116, and T24 cells) and cultured for at least 2 weeks prior to fixation with 4% paraformaldehyde and stained with 0.1% crystal violet (Sigma). Each cell type was plated in duplicate, and each experiment was repeated at least three times independently. Representative data are presented.
SAHF Formation-SAHF formation was detected as previously described (31). Briefly, cells were fixed in chamber slides with 4% paraformaldehyde, stained with DAPI, and documented using a charge-coupled device camera attached to a fluorescence microscope.

Human Breast Carcinoma Cell Lines Undergo a Permanent Growth Arrest upon the Activation of the Ras/MEK/MAPK
Pathway-Our recent studies demonstrated that human U2OS osteosarcoma cells undergo permanent growth arrest upon the transduction of H-Ras V12 , raising the possibility that Ras signaling can activate a cell cycle control program to inhibit the growth of tumor cells (36). Considering that Ͼ90% of human cancers are carcinomas, it is important to determine whether this growth inhibition can also be achieved in human carcinoma cells. To address this question, human breast carcinoma cell lines, MDA-MB468 and MCF7, were transduced with H-Ras V12 or an empty vector via retroviral gene transfer, as previously described (9,28,29). Strikingly, in contrast to the vector controls, both MDA-MB468 and MCF7 cells transduced with H-Ras V12 exhibit senescence-like flat and enlarged cell morphology, 4 days after selection (Fig. 1A). Cells transduced with H-Ras V12 also exhibit reduced BrdUrd incorporation when compared with the vector controls, suggesting that DNA synthesis is inhibited in these cells (Fig. 1B). Colony formation assays show that both MDA-MB468 and MCF7 cells transduced with H-Ras V12 fail to form colonies in comparison with cells transduced with the vector controls (Fig. 1C). All of the cells remain viable in culture for several weeks, without significant cell death (data not shown). Furthermore, both MDA-MB468 cells and MCF7 cells transduced with H-Ras V12 exhibit elevated SA-␤-galactosidase activity (Fig. 1D). Taken together, these data demonstrate that both human breast carcinoma cell lines undergo a senescence-like permanent growth arrest upon the transduction of H-Ras V12 .
To gain insight into the molecular basis of H-Ras V12 -induced permanent growth arrest in human breast carcinoma cells, we performed Western blot analyses to examine the expression pattern of important cell cycle regulators in these cells (Fig. 1E). The Rb tumor suppressor plays an essential role in regulating G 1 -S transition during cell cycle progression (43) and is crucial for H-Ras V12 -mediated senescence in normal human diploid fibroblasts (31,44). Interestingly, MDA-MB468 cells do not express Rb, while the expression of Rb proteins is dramatically reduced in MCF7 cells transduced with H-Ras V12 , when compared with the vector control (Fig. 1E). The p16 CDK inhibitor regulates cell cycle progression through the Rb pathway (45,46). In normal human cells, p16 plays an essential role in senescence, including replicative senescence and H-Ras V12 -induced senescence (34,35,47,48). Interestingly, MCF7 cells do not express p16, whereas the expression levels of wild-type p16 in MDA-MB468 cells is not up-regulated but, in fact, slightly down-regulated upon the transduction of H-Ras V12 (Fig. 1E).
These data are in stark contrast to H-Ras V12 -induced senescence in normal human diploid fibroblasts, and demonstrate that H-Ras V12 -mediated tumor senescence can occur in the absence of p16 or Rb. Topoisomerase II␣ is essential for DNA replication and cell proliferation (49,50), and a marked reduction of topoisomerase II␣ is highly specific in senescent cells (51). Consistent with the observed growth inhibition, both MDA-MB468 and MCF7 cells transduced with H-Ras V12 express reduced levels of topoisomerase II␣ when compared with the vector controls (Fig. 1E). The p53 protein plays a crucial role in transcription and cell cycle control (52) and is the most frequently inactivated tumor suppressor in human cancers (53,54). MDA-MB468 cells harbor a hemizygous "gain-of-function" mutant p53 (GOFp53), which is highly expressed and is considered to be an oncoprotein associated with malignancy (55)(56)(57). Interestingly, transduction of H-Ras V12 in MDA-MB468 cells results in a significant down-regulation of the GOFp53 (Fig. 1E). In contrast, MCF7 cells express wild-type p53, which is up-regulated upon the transduction of H-Ras V12 (Fig. 1E). The p21 CDK inhibitor is a transcriptional target of p53 and is best known for its important role in mediating cell cycle arrest (58,59). Regardless of the p53 status, the p21 CDK inhibitor is significantly up-regulated in both MDA-MB468 and MCF7 cells transduced with H-Ras V12 , when compared with the vector controls (Fig. 1E).
Previous studies have shown that H-Ras V12 -induced senescence in normal human fibroblasts depends on signal intensity and threshold (60). Interestingly, Ras proteins are consistently expressed at a much lower level in MCF7 cells than MDA-MB468 cells upon gene transfer. In fact, the ectopic expression level of Ras proteins in MCF7 cells is comparable to levels found in cancer cells that harbor endogenous activating ras mutations; specifically, it is similar to that in human bladder T24 carcinoma cells (Fig. 1E), which carry an activating mutation on codon 12 in the H-ras gene (H-Ras V12 ) (61). These data suggest that the observed growth inhibition in tumor cells can be achieved by the expression of H-Ras V12 at relatively low levels.
Collectively, these data demonstrate that transduction of H-Ras V12 induces a senescence-like permanent growth arrest in human breast carcinoma MDA-MB468 and MCF7 cells, independent of the status of the p16 or Rb tumor suppressors. Moreover, growth inhibition of breast carcinoma cells induced by H-Ras V12 is associated with the up-regulation of wild-type p53 and p21, as well as down-regulation of GOFp53.

Transduction of Activated MEK, but Not Wild-type MEK, Is Sufficient to Induce Growth Inhibition in Breast Carcinoma
Cells-Our recent studies show that H-Ras V12 -induced growth inhibition in osteosarcoma U2OS cells can be interrupted by treatment with inhibitors of MEK or p38MAPK, suggesting that activation of the MEK/MAPK pathway is required for H-Ras V12 -induced tumor suppression (36). Here, we further investigate the ability of activated MEK and wild-type MEK to induce growth inhibition in breast carcinoma cells. MDA-MB468 cells were transduced with either an activated MEK allele (MEK Q56P ) or wild-type MEK (WT MEK) via retroviral gene transfer. Transduction of H-Ras V12 or an empty vector served as controls. The results show that MDA-MB468 cells transduced with activated MEK exhibit a senescence-like flat and enlarged cell morphology similar to that of cells transduced with H-Ras V12 . In contrast, cells transduced with WT MEK exhibit cell morphology comparable to cells transduced with an empty vector control ( Fig. 2A). Growth curve analyses further demonstrate that cells transduced with H-Ras V12 or activated MEK fail to accumulate but maintain constant cell numbers in culture over time (Fig. 2B). Consistent with this observation, cells transduced with either activated MEK or H-Ras V12 fail to form colonies in the colony formation assays (Fig. 2C), and exhibit elevated SA-␤-galactosidase activity (data not shown).
Western blot analyses further show that MDA-MB468 cells transduced with MEK Q56P exhibit similar changes in cell cycle regulators as cells transduced with H-Ras V12 , including the down-regulation of GOFp53 (Fig. 2D). Although the expression of MEK proteins are elevated in both cells transduced with MEK Q56P and WT MEK, phosphorylation of MAPK is seen in MEK Q56P -transduced cells but not in WT MEK-transduced cell (Fig. 2D). Similar findings are also observed in MCF7 cells (data not shown). These studies demonstrate that transduction of MEK Q56P , but not WT MEK, is sufficient to activate MAPK, leading to tumor senescence.
H-Ras V12 -mediated Tumor Senescence Is Compromised in Cancer Cells Harboring Endogenous Activating Mutations in ras Genes-Recent studies suggest that Ras-induced senescence occurs in vivo during carcinogenesis, and disruption of this senescence program is crucial for tumor development and progression (62,63). Our observation that transduction of H-Ras V12 induces growth inhibition in cancer cells suggests that cancer cells may maintain a cell cycle control program that can be activated by Ras signaling, leading to tumor senescence. Given that Ͼ30% of human cancers have activating mutations in ras genes (24), we next examined whether the tumor sup- pression pathway that can be activated by H-Ras V12 is disrupted in cancer cells harboring endogenous activating mutations in ras genes. To achieve this goal, we ectopically expressed H-Ras V12 in three cancer cell lines that have endogenous activating mutations in ras genes: human bladder carcinoma T24 cells (H-Ras V12 ), human breast carcinoma MDA-MB231 cells (K-Ras D13 ), and human colon carcinoma HCT116 cells (K-Ras D13 ).
Breast carcinoma MDA-MB468 cells harbor wildtype ras and, therefore, were used as a control. The results show that, in contrast to MDA-MB468 cells, none of the cancer cell lines carrying activating ras mutations exhibit senescence-like flat and enlarged cell morphology after ectopically transduced with H-Ras V12 (Fig. 3A). In fact, these cells continue to accumulate in culture at a growth rate comparable with their vector controls (Fig. 3B). Colony formation assays reveal that these cancer cells transduced with H-Ras V12 reproducibly form colonies similar in size to those of vector controls (Fig. 3C). Consistent with their ability to proliferate, Western blot analyses show that the protein expression levels of topoisomerase II␣ and Rb in these three cancer cell lines remain relatively unchanged upon H-Ras V12 transduction (Fig. 3D). Together, these results show that the ability for H-Ras V12 to induce tumor senescence is dramatically compromised in cancer cells that harbor endogenous activating mutations in ras genes, suggesting that these cancer cells may have additional genetic alterations that disrupt the tumor senescence program that can be activated by H-Ras V12 .

Transduction of MEK Q56P Induces Growth Inhibition in Cancer Cells
Harboring Endogenous Activating Mutations in ras-We next investigated whether cancer cells harboring endogenous activating mutations in ras remain susceptible to activated MEK-induced growth inhibition. Breast carcinoma MDA-MB231 or bladder carcinoma T24 cells were transduced with H-Ras V12 , MEK Q56P , or an empty vector. Interestingly, in contrast to cells transduced with H-Ras V12 or the vector controls, MDA-MB231 cells transduced with MEK Q56P exhibit reduced BrdUrd incorporation (Fig. 4A). Western blot analysis shows that Rb protein expression levels are reduced, whereas p21 is significantly elevated in cells transduced with MEK Q56P (Fig. 4B). Moreover, MDA-MB231 and T24 cells transduced with MEK Q56P exhibited senescence-like flat and enlarged cell morphology (Fig. 4C) and their ability to form colonies in the colony formation assays was significantly reduced, when compared with the vector controls or cells transduced with H-Ras V12 (Fig. 4D). Similar results were seen in human pancreatic carcinoma MIA-PaCa2 cells, which harbor a mutation in codon 12 of K-ras (supplemental Fig. S1). Lastly, MDA-MB231 cells transduced with MEK Q56P exhibit elevated SA-␤-galactosidase activity and SAHF formation (Fig. 4E). These studies suggest that cancer cells with activating mutations in ras genes may remain susceptible to MEK Q56P -mediated tumor senescence.
Rb Family Proteins Are Dispensable for H-Ras V12 -mediated Tumor Senescence-Rb is a bona fide tumor suppressor, and disruption of Rb alone is sufficient to disrupt the senescence phenotype in normal human fibroblasts (31,64). Our studies demonstrate, however, that transduction of H-Ras V12 induces a permanent growth arrest in the Rb-null human breast carcinoma MDA-MB468 cells and osteosarcoma SAOS2 cells ( Fig. 1  and data not shown). The Rb family of pocket proteins is comprised of three members, Rb, p107, and p130, which are known to regulate the cell cycle through their interaction with E2F transcription factors (65). Therefore, we further investigated the requirement of pocket proteins in H-Ras V12 -mediated growth inhibition of tumor cells. To achieve this goal, we utilized a sequence-specific RNA interference approach in MCF7 cells, which harbor wild-type Rb, p107, and p130. MCF7 cells were transduced with RbshRNA, p107shRNA, p130shRNA, or a scrambled control, respectively. After selection, cells were sequentially transduced with H-Ras V12 or a vector control, selected, and subjected to various analyses. The results show that transduction of RbshRNA, p107 shRNA, p130shRNA, or the scrambled control does not change the senescence-like cell morphology or the reduced BrdUrd incorporation upon the transduction of H-Ras V12 (Fig. 5, A and B). Moreover, these cells remained unable to form colonies in the colony formation assays (Fig. 5C). Consistent with these observations, Western blot analyses show that the expression levels of topoisomerase II␣ is markedly reduced in these cells, when compared with vector controls (Fig. 5D). Importantly, each shRNA efficiently knocked down the expression of the target molecule in MCF7 cells (Fig. 5D).
Rb pocket proteins have redundant functions (65). Therefore, inactivation of each pocket protein alone may not be sufficient to disrupt Ras-induced tumor senescence. To address this concern, we transduced human papilloma virus-16 E7 to simultaneously inactivate all three pocket proteins in human osteosarcoma U2OS cells (38,66). Notably, transduction of H-Ras V12 in E7-expressing cells still results in a senescence-like growth arrest as demonstrated by the inability of these cells to form colonies in the colony formation assay (Fig. 5E). Taken together, these findings suggest that H-Ras V12 -induced tumor senescence can occur in the absence of Rb pocket proteins.
Knockdown of p21 Disrupts H-Ras V12 -induced Growth Inhibition in MCF7 Cells-The p21 CDK inhibitor plays an important role in cell cycle control (67). Indeed, p21 is induced in cells that have undergone senescence in response to various chemotherapeutic agents (68). Our studies show that H-Ras V12 -induced and MEK Q56P -induced growth inhibition of cancer cells is consistently associated with the induction of p21. Therefore, we next investigated the requirement of p21 in H-Ras V12 -induced growth inhibition of MCF7 cells. MCF7 cells were transduced with p21shRNA or a scrambled control. After selection, cells were sequentially transduced with H-Ras V12 or an empty vector, selected, and subjected to various assays. Interestingly, transduction of p21shRNA blocks the senescence-like cell morphology normally seen after the transduction of H-Ras V12 (Fig.  6A). BrdUrd incorporation and colony formation assays further indicate that p21shRNA disrupts H-Ras V12 -mediated growth inhibition in MCF7 cells (Fig. 6, B and C). Western blot analyses confirm the efficient knockdown of p21 expression in MCF7 cells by p21shRNA and the resulting recovered expression of Rb and topoisomerase II␣ in these cells (Fig. 6D). Similar results were also seen in MDA-MB231 cells transduced with p21shRNA and MEK Q56P (supplemental Fig. S2). These results were further validated using two additional sequence-specific p21shRNAs in MCF7 cells, MDA-MB468 cells, and U2OS cells (supplemental Fig. S3). Knockdown with either p21shRNA allows cells transduced with H-Ras V12 to incorporate BrdUrd in a similar manner to cells transduced with empty vector controls (supplemental Fig. S3A). Furthermore, Western blot analyses of these cells confirm the knockdown of p21 in these cells and the recovery of topoisomerase II␣ protein expression (supplemental Fig. S3B). Together, these results establish a crucial role for p21 in H-Ras V12 -induced tumor senescence.

DISCUSSION
In recent years, encouraging evidence suggests that senescence provides a less toxic means to manage tumor growth and is potentially an alternative approach to apoptosis in treating cancer cells that are refractory to cell death (18,69). Indeed, tumor senescence can be provoked by various stimuli, including differentiating agents (17,18,69,70). Despite its increasing significance, the underlying signaling mechanism involved in tumor senescence remains elusive. Here, we report the identification of a signaling pathway that induces tumor senescence in human carcinoma cells and the characterization of the crucial determinants involved.
Ras-induced senescence in normal human diploid fibroblasts is dependent on signal intensity (60). In contrast, tumor senescence induced by the activation of the Ras/ MEK/MAPK signaling pathway can be repeatedly achieved when the transgenes are expressed at levels comparable with that of the endogenous activated Ras or MEK (Fig. 1E and data not shown), suggesting that the observed endpoint is not solely a consequence of expressing the transgenes at supraphysiological levels. In fact, overexpression of H-Ras V12 failed to induce growth arrest in cells that carry ras mutations (e.g. T24), indicating that the ability of activated Ras or MEK to induce tumor senescence is strictly dependent on cellular context, a notion that is consistent with previous reports (9,29,71).
In addition to the ample in vitro evidence that links activated Ras signaling to senescence, recent in vivo studies strongly indicate that Ras-induced senescence occurs as a barrier to carcinogenesis (62,63,72). In this view, cells that harbor endogenous FIGURE 5. Knockdown of Rb pocket protein family members does not prevent H-Ras V12 -induced senescence in tumor cells. MCF7 cells were transduced with RbshRNA, p107shRNA, p130shRNA, or a scrambled shRNA control (Sc shRNA). U2OS cells were transduced with HPV16 E7 or an empty vector control. The cells were sequentially transduced with H-Ras V12 (Ras V12 or R) or an empty vector control (Vector or V) and analyzed in the following assays. A, representative photomicrographs illustrating morphology of the indicated cell types taken on day 6 post-selection. Photos were taken using a phase-contrast microscope at 100ϫ magnification. B, BrdUrd incorporation of the indicated cell types. At least 200 cells were scored, and quantification of BrdUrd incorporation is presented as a percentage of total cells counted (with standard deviation) in the histogram. DAPI, a DNA dye, was used to counterstain nuclei. C, colony formation assay of the indicated cell types. Cells were plated at day 6 post-selection and maintained in culture for at least 2 weeks before fixation and staining with crystal violet. D, Western blot analysis of the specified proteins in the indicated cell types. Cells were harvested on day 6 post-selection. Expression of tubulin was used as a protein loading control (S.E., short exposure; L.E., long exposure). E, U2OS cells were transduced with HPV16 E7 (E7) or an empty vector control (Vector) and subsequently transduced with H-Ras V12 (Ras V12 or R) or an empty vector control (Vector or V). Left panel, for the colony formation assay cells were plated on day 6 post-selection and maintained in culture for at least 2 weeks before fixation and staining with crystal violet. Right panel, Western blot analysis of the indicated proteins. Cells were harvested on day 6 post-selection. Expression of tubulin was used as a protein loading control.
activating ras mutations must encounter disruptions in the senescence program before carcinogenesis can progress. In support of this notion, we show that the ability of activated Ras to induce senescence is severely compromised in cancer cells that harbor endogenous activating mutations in ras genes. Given that activating mutations in ras genes occur in 30% of all human cancers, it is imperative to further identify signaling mechanisms that will lead to the growth inhibition of these cancer cells. Due to the genetic heterogeneity in cancer cells, it may be challenging to pinpoint a common mechanism for tumor senescence. However, our observation that some of these cancer cells remain sensitive to senescence induced by activated MEK provides a promising perspective. Future studies should explore whether activated MEK can also suppress growth of tumor cells harboring mutations in RAF, such as human melanoma cells (73). Potentially, our findings are clinically relevant, given that cancer cells that have ras mutations are surprisingly resistant to pharmacological inhibitors, such as farnesyl transferase inhibitors (74 -76).
The p53 tumor suppressor is the most frequently mutated gene in human cancers (53,77). Mutations in p53 can result in either loss-of-function, a dominant-negative protein, or a gainof-function protein (54,56,57,78). Codon 273 is among the most frequently mutated codons in the p53 gene, and certain nucleotide substitutions in this codon are known to result in a gain-of-function mutant p53 (GOFp53) (55)(56)(57). These GOFp53 proteins are considered to be "oncoproteins" that foster tumorigenicity (55)(56)(57)78). It is conceivable that understanding the regulation of GOFp53 will aid in identifying novel anti-cancer strategies. This study reports for the first time that a GOFp53 can be negatively regulated by Ras/MEK/MAPK signaling. The notion that p53 can be regulated by a signaling pathway is not unprecedented. For example, treatment with epidermal growth factor induces growth arrest in MDA-MB468 cells, which is accompanied by the down-regulation of GOFp53 mRNA levels (79). Given the link between epidermal growth factor and Ras signaling (80), it is possible that activation of the Ras/MEK/MAPK pathway contributes to the downregulation of the GOFp53 in epidermal growth factor-treated cells. Together, these results uncover a novel signaling mechanism for specifically regulating an oncoprotein, namely GOFp53. Future studies should aim to determine whether similar events also occur in other GOFp53 proteins. Understanding the molecular basis of this regulation will potentially lead to the development of cancer therapeutic strategies targeting GOFp53. As a major transcriptional target of p53, p21 plays a crucial role in p53-mediated cell cycle arrest (58). However, p21 can also be activated independently of p53 to halt the progression of the cell cycle (81). Induction of p21 is closely associated with tumor senescence induced by cancer therapeutic agents (68,82). Our study further FIGURE 6. Knockdown of p21 prevents H-Ras V12 -induced senescence in MCF7 cells. MCF7 cells were transduced with p21shRNA or a scrambled shRNA control. The cells were sequentially transduced with H-Ras V12 (Ras V12 or R) or an empty vector control (Vector or V) and analyzed in the following assays. A, representative photomicrographs illustrating morphology of the indicated cell types taken on day 6 post-selection. Photos were taken using a phase-contrast microscope at 100ϫ magnification. B, BrdUrd incorporation of the indicated cell types. At least 200 cells were scored, and quantification of BrdUrd incorporation is presented as a percentage of total cells counted (with standard deviation) in the histogram. DAPI, a DNA dye, was used to counterstain nuclei. C, colony formation assay of the indicated cell types. Cells were plated at day 10 post-selection and maintained in culture for at least 2 weeks before fixation and staining with crystal violet. D, Western blot analysis of the specified proteins in the indicated cell types. Cells were harvested on day 6 post-selection. Expression of tubulin was used as a protein loading control. establishes a crucial role for p21 in tumor senescence induced by activation of Ras signaling. Given the important roles of p16 and Rb in Ras-induced senescence in normal human diploid fibroblasts, it is surprising that Ras-induced tumor senescence can occur in the absence of p16 and Rb pocket proteins. Furthermore, the finding that inactivation of p21, but not Rb, alone is sufficient to disrupt Ras-induced tumor senescence, raises the possibility that p21 might act in a Rb-independent manner to mediate the growth inhibition of cancer cells. In support of this notion, studies in human diploid fibroblasts show that p21 is crucial for senescence induced by p400 inactivation, whereas Rb is dispensable (83). Although the mechanisms by which p21 induces growth arrest in response to activated Ras signaling needs to be further investigated, it is clear that the up-regulation of p21 alone is insufficient to induce growth arrest. Specifically, T24 cells that are resistant to H-Ras V12 -induced senescence exhibit elevated p21 protein expression (data not shown). Indeed, several reports, including ours, have demonstrated that mere up-regulation of a protein is not sufficient for it to exert its function (36,84). These studies suggest that additional posttranslational events may also play a role in determining p21 activity. Tumor suppressors p53, p16, and Rb are frequently inactivated in cancer cells (53,85,86); therefore, it is important to identify alternative tumor suppression mechanisms that can be activated. Given that p21 is rarely inactivated in cancer cells, it is expected that a further understanding of its activity in mediating tumor senescence will impact the development of novel cancer interventions.
In conclusion, this study reports the identification and characterization of a potentially novel tumor suppression program induced by activation of the Ras/MEK/MAPK signaling pathway (Fig. 7A). Cancer cells that do not harbor endogenous activating mutations in ras genes might maintain a senescence program that can be activated through the Ras/MEK/MAPK pathway (Fig. 7, A and B). Moreover, cancer cells that harbor endogenous activating ras mutations might remain sensitive to senescence induced by activated MEK (Fig. 7C). These findings provide a basis for understanding the senescence signaling pathway in tumor cells and will potentially lead to the design of novel strategies, alternate to apoptosis, to effectively control cancer progression.