S-Phase-specific Activation of PKCα Induces Senescence in Non-small Cell Lung Cancer Cells*

Protein kinase C (PKC) has been widely implicated in positive and negative control of cell proliferation. We have recently shown that treatment of non-small cell lung cancer (NSCLC) cells with phorbol 12-myristate 13-acetate (PMA) during G1 phase inhibits the progression into S phase, an effect mediated by PKCδ-induced up-regulation of the cell cycle inhibitor p21Cip1. However, PMA treatment in asynchronously growing NSCLC cells leads to accumulation of cells in G2/M. Studies in post-G1 phases revealed that PMA induced an irreversible G2/M cell cycle arrest in NSCLC cells and conferred morphological and biochemical features of senescence, including elevated SA-β-Gal activity and reduced telomerase activity. Remarkably, this effect was phase-specific, as it occurred only when PKC was activated in S, but not in G1, phase. Mechanistic analysis revealed a crucial role for the classical PKCα isozyme as mediator of the G2/M arrest and senescence, as well as for inducing p21Cip1 an obligatory event for conferring the senescence phenotype. In addition to the unappreciated role of PKC isozymes, and specifically PKCα, in senescence, our data introduce the paradigm that discrete PKCs trigger distinctive responses when activated in different phases of the cell cycle via a common mechanism that involves p21Cip1 up-regulation.

cellular effectors, of which PKC isozymes have been the most widely characterized. The PKC family comprises 3 subfamilies that include 10 structurally related phospholipid-dependent serine/threonine kinases (1,2). Members of the classical cPKC (␣, ␤I, ␤II, and ␥) and the novel nPKC (␦, ⑀, , and ) subfamilies are activated by phorbol esters and their cellular analogue, the second messenger diacylglycerol, which leads to redistribution (translocation) of the enzymes from cytosol to intracellular membrane compartments, where they phosphorylate specific substrates. The marked heterogeneity in the signaling events and cell type-specific responses triggered by phorbol esters could be explained by the distinctive pattern of expression and intracellular localization of PKC isozymes and their substrates, which ultimately results in selective pathway activation.
One of the paradigms that best exemplifies the functional versatility of PKC isozymes is the regulation of the cell cycle machinery. It became evident in the last years that PKCs can impact on the cell cycle both in positive and negative manners with a strict degree of cell type and isozyme specificity. PKC isozymes have been shown to regulate the progression of cells from G 1 to S phase as well as with the transition from G 2 to M phase (3) via transcriptional, translational, and post-translational mechanisms. PKCs control the activity of cyclin-Cdk complexes in G 1 by modulating the expression of cyclins and Cdk inhibitors (4,5). For example, early studies in vascular endothelial cells showed dual growth stimulatory or inhibitory roles for PKCs depending on which phase in the cell cycle PKC becomes activated. In HUVEC cells phorbol esters potentiate growth factor mitogenic activity when added in early G 1 phase, but they inhibit DNA synthesis when added in late G 1 phase (6). In NIH 3T3 cells, PKC␣ and PKC⑀ enhance cell cycle progression and proliferation by stimulating cyclin D1 transcription (7). On the other hand, PKC inhibits cdk2 activity in human keratinocytes, and overexpression of either PKC or PKC␦, but not PKC␣, leads to G 1 arrest and differentiation (8,9). In several cell types, the PKC activator phorbol 12-myristate 13-acetate (PMA) up-regulates Cdk inhibitors p21 Cip1 and/or p27 (10,11). These contrasting effects ultimately impact on the status of Rb phosphorylation, the expression of E2F-regulated genes, and the biological outcome. Several studies have also established key roles for PKC isozymes in G 2 . For example, PKC␤II activation was required for entry into mitosis in HL60 promyelocytic leukemia cells (12), whereas PMA was shown to induce G 2 arrest by suppression of cdc2 kinase activity in vascular endothelial cells (13). Based on this complexity, it is not surprising that both PKC activators (such as bryostatins or phorbol esters) and PKC inhibitors (such as the PKC␤ inhibitor enzastaurin) are in clinical trials for a number of neoplasias (14 -17). Because of the contrasting effects of PKC isozymes on cell cycle regulation, a thorough understanding of these mechanisms is essential for the rational design of PKC modulators with anti-cancer activity.
Lung cancer is one of the most common forms of cancers worldwide and the major cause of cancer-related mortality. Non-small cell lung carcinoma (NSCLC), the most frequent type of lung cancer, is treated in early stages mainly with surgery and radiotherapy, whereas advanced stages receive combination chemotherapy or radiation therapy (18). Unfortunately, the marked resistance to the various therapies accounts for the high lethality (19). Studies in the last years have proposed PKC isozymes, such as PKC␣, as targets for NSCLC therapy. However, a PKC␣-specific antisense oligonucleotide (ISIS 3521/LY900003) showed slight or no benefit in NSCLC patients, either alone or in combination with other chemotherapeutic agents (20). This is not unexpected since PKC␣, like other PKCs, has been shown to be either growth inhibitory or pro-apoptotic in various cancer cell models (21)(22)(23). One may speculate that activators of growth inhibitory PKCs rather than PKC inhibitors could have therapeutic benefits for lung cancer, as demonstrated for other types of cancers. It is evident that a deeper knowledge of the roles of individual PKC isozymes in lung cancer would be needed to rationalize the use of PKC as a therapeutic target.
Our recent studies have established that treatment of NSCLC cells with phorbol ester in early G 1 impairs the progression through S phase, and we have identified PKC␦ as the PKC responsible for this effect. PKC␦-induced G 1 arrest is mediated through the transcriptional up-regulation of the cell cycle inhibitor p21 Cip1 (24). However, as shown in the present article, the paradox seems to be more complex, because in asynchronously growing NSCLC cells phorbol esters lead to the accumulation of cells in G 2 /M, thus suggesting multiple points of regulation in the cell cycle by PKCs. A more detailed analysis revealed that activation of PKC in late G 1 -early S leads to a delay in the progression through S phase and irreversible arrest of cells in G 2 /M, followed by the appearance of a senescence phenotype. Both G 2 /M arrest and senescence depend on the upregulation of p21 Cip1 , but strikingly these effects are mediated by PKC␣ rather than PKC␦. Exploiting this irreversible induced growth arrest of lung cancer cells in response to PKC␣ activation may have significant therapeutic implications.

EXPERIMENTAL PROCEDURES
Materials-Cell culture media was purchased from Invitrogen (Carlsbad, CA). PMA was obtained from LC Laboratories (Woburn, MA). The pan-PKC inhibitor GF109203X (bisindolylmaleimide I) was from BIOMOL Research Laboratories Inc. (Plymouth Meeting, PA). Gö6976 and rottlerin were purchased from Alexis (San Diego, CA). Propidium iodide, DAPI, and hydroxyurea were from Sigma-Aldrich.
Cell Culture-H358 and H441 lung bronchoalveolar adenocarcinoma cells were obtained from ATCC (Manassas, VA). H322 lung bronchoalveolar carcinoma cells were kindly provided by Dr. Steven Albelda (University of Pennsylvania School of Medicine). Cells were cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum, penicillin (100 units/ ml), streptomycin (100 g/ml), and L-glutamine (2 mM) at 37°C in a humidified 5% CO 2 atmosphere. For synchronization at the G 1 /S boundary, cells were cultured in normal medium for 24 h, serum-starved for 24 h, and then treated with 1 mM hydroxyurea (HU) in complete medium. For synchronization in G 0 , cells were serum-starved for 48 h, and released from G 0 by the addition of serum.
Cell Proliferation and Cell Cycle Analysis-Cells (2 ϫ 10 5 ) were seeded in 60-mm dishes in triplicate and synchronized at the G 1 /S boundary with HU as described above. Upon release by extensive washing, cells were treated with PMA or vehicle for different times. After extensive washing to remove the PMA, complete growth medium was added. Cells were trypsinized at different intervals and counted with a hemocytometer. Proliferation was assayed by [ 3 H]thymidine incorporation. Briefly, 24, 48, or 72 h after PMA treatment cells were pulse-labeled with 3 Ci/ml of [methyl-3 H]thymidine (Amersham Biosciences) for 3 h, followed by trichloroacetic acid precipitation and scintillation counting. For determination of cell cycle profile, cells were stained with propidium iodide (0.1 mg/ml) and analyzed by flow cytometry, as previously described (24).
Adenoviral Infections-Cells were infected with replicationdeficient adenoviruses (AdV) for either PKC␣ or LacZ as a control (23, 25) 4 h (MOI, 100 pfu/cell) in serum-free RPMI 1640 medium. After removal of the AdV by extensive washing, cells were incubated in complete medium for 20 h. Expression of PKC␣ was readily detected 24 h after infection and remained stable for several days (data not shown). Amplification of AdVs was carried out in HEK293 cells. Titers of viral stocks were normally higher than 1 ϫ 10 9 pfu/cell.
Plasmid Transfections and Promoter Analyses-H358 cells in 12-well plates (5 ϫ 10 4 cells/well) were transiently transfected with 0.5 g of a p21 Cip1 Firefly luciferase reporter vector (26) using Fugene 6 (Roche Applied Science, Indianapolis, IN). A Renilla luciferase expression vector (50 ng, pRL-TK, Promega, Madison, WI) was co-transfected for normalization of transfection efficiency. After transfection, cells were grown overnight in complete medium, synchronized at the G 1 /S boundary, and lysed. Cells extracts were subject to luciferase determination using the Dual-Luciferase Reporter Assay System (Promega). Results were expressed as the ratio between Firefly and Renilla luciferase.
Immunofluorescence-Cells were plated on coverslides placed on 35-mm dishes. After synchronization, cells were stimulated with either PMA or vehicle, and at the indicated times washed twice with PBS, fixed for 10 min with methanol, washed three times for 5 min with PBS, and permeabilized for 15 min with 0.25% Triton X-100 in PBS, followed by a 10-min incubation in 100 mM glycine in PBS. After blocking for 30 min with 3% fetal bovine serum in PBS, a mouse anti-p21 Cip1 monoclonal antibody (1:250) was added (1 h), followed by washing with 0.1% Tween-20 in PBS, and incubation with a CY3-conjugated anti-mouse antibody (1:1000; Jackson Immunoresearch Laboratories, Inc.) was added. After additional washings, DNA was stained using DAPI (0.1 g/ml, 10 min). Coverslides were washed three times with PBS, mounted with Vectashield, and visualized with a Nikon Eclipse TE2000 inverted microscope equipped with a q-imaging Exi digital cooled camera (1360 ϫ 1036 pixels, Burnaby, Canada). Recordings were done using Northern Eclipse 6.0 software (Empix Imaging Inc., Cheektowaga, NY). A 40ϫ planar objective (Nikon) was used for all recordings.
Subcellular Fractionation-Cells were washed, collected in ice-cold PBS, and then pelleted and fractionated into cytosolic and nuclear fractions, as described elsewhere (27). Separation of cytosolic and particulate fractions was performed by ultracentrifugation, as described previously (28).
Analysis of Telomerase Activity-Non-isotopic telomerase activity was determined by the telomeric repeat amplification protocol (TRAP) using the TRAPEZE telomerase detection kit (Chemicon International, Temecula, CA) according to the manufacturer's instructions.
Statistical Analysis-Data are presented as mean Ϯ S.D. and were analyzed using a Student's t test. A p value of Ͻ0.05 was considered statistically significant.

PMA Causes Irreversible G 2 /M Arrest in NSCLC Cells-PKC
activation with phorbol esters causes a profound inhibition of cell proliferation in lung cancer cells, and our previous work has identified a PKC␦-dependent inhibitory mechanism that limits cell progression from G 1 into S phase (24). However, when asynchronously growing H358 NSCLC cells were treated for 30 min with the PKC activator PMA (100 nM), a significant accumulation of cells in G 2 /M was observed, as determined by flow cytometry analysis (Fig. 1A, left panel). At 48 and 72 h the population of G 2 /M cells in response to PMA treatment doubled relative to vehicle-treated cells. A significant accumulation of cells in G 2 /M in response to PMA was also observed in H441 and H322 cells (Fig. 1A, right panel). PKC activation by PMA induces apoptosis in several cellular models (1,5,23), but no evidence of apoptosis in response to PMA was observed in H358, H441, or H322 cells either by flow cytometry (absence of a sub-G 0 /G 1 cell population) or by microscopy (absence of fragmented nuclei after DAPI staining, data not shown).
To begin dissecting the mechanisms that lead to the accumulation of H358 cells in G 2 /M, we decided to synchronize cells with hydroxyurea (HU) and analyze the effects of PMA on S 3 G 2 progression. Approximately 70% of H358 cells were synchronized in late G 1 after HU treatment. Upon removal of HU by extensive washing, cells began progressing into S phase at ϳ2 h, and into G 2 phase at ϳ6 h. HU treatment did not seem to cause significant replicative stress, as judged by the ability of cells to complete the cell cycle (supplemental Fig. S1) and reach confluence, the absence of apoptosis, as well as the possibility of subculturing cells for several passages after HU treatment (data not shown). To analyze the effect of phorbol ester treatment on S 3 G 2 progression, H358 cells were treated with PMA (100 nM, 30 min) or vehicle 2 h after HU release, and cell cycle distribution was determined at different times. Fig. 1B shows that PMA caused a significant delay in S 3 G 2 transition. Indeed, while it takes 8 h to reach the maximum % of control cells in G 2 , this effect is achieved at 11 h in PMA-treated cells. The effect was dependent on the PMA concentration (Fig. 1C, upper panel), as well as the length of incubation, with maximum response at 30 min (Fig. 1C, center panel). The S 3 G 2 delay was observed only when PMA was added in late G 1 -early S phase (t ϭ 2-5 h) but not when the phorbol ester was added in late S phase (t ϭ 6 h) (Fig. 1C, lower panel). Notably, despite the short duration of the incubation with PMA, the majority of the cells still remained in G 2 /M 72 h after treatment (Fig. 1B, right  panel). Therefore, the delay in S 3 G 2 transition was accompanied by an irreversible arrest that prevented cells to complete the cycle and progress into G 1 . Indeed, treatment of cells synchronized in the G 1 /S phase boundary (t ϭ 2 h after HU release) with 100 nM PMA (30 min) abolished proliferation, as determined by cell counting (Fig. 1D, left panel). In agreement with this data, PMA treatment of HU-synchronized cells resulted in a marked inhibition of DNA synthesis, as judged by analysis of [ 3 H]thymidine incorporation (Fig. 1D, right panel). Irreversible arrest by PMA was also observed in asynchronous cultures of H358 cells, as determined both by cell counting and [ 3 H]thymidine incorporation (Fig. 1E). Thus, PMA treatment of H358 cells in S phase leads to irreversible accumulation of cells in G 2 , arguing for multiple points of regulation of the cell cycle by PKC activation. PMA Induces Senescence in H358 Cells in a Cell Cycle Phasespecific Manner-Morphological analysis of H358, H441, and H322 cells 72 h after PMA treatment, either in asynchronous or HU-synchronized cultures, showed that a significant number of cells became large and flat, and exhibited enlarged nuclei. These features, together with irreversible growth, are hallmarks of senescence (30). Remarkably, cells remain attached for at least 10 days post-PMA treatment. At that time the phenotypic changes were more pronounced ( Fig. 2A), with cells even larger, multinucleated, and with a characteristic vacuolization. We then determined the effect of PMA on telomerase activity, using a telomeric repeat amplification protocol (TRAP). PMA treatment in S phase caused a significant reduction of telomerase activity in H358 cells (Fig. 2B). To further establish the presence of a senescence phenotype, we measured the expression of SA-␤-Gal, a well-established marker of senescence (29). Notably, Ͼ50% of H358 cells in asynchronous cultures became SA-␤-Gal positive after PMA treatment, compared with Ͻ10% in response to vehicle. Similarly, a significant fraction of H441 and H322 become SA-␤-Gal positive (Fig. 2C). The proportion of SA-␤-Gal positive H358 cells was even higher (Ͼ75%) when PMA treatment was carried out in HU-synchronized cells. However, when PMA was added to H358 cells synchronized in early G 1 phase, there was only a slight increase in the number of ␤-Gal positive cells. Therefore, the PMA effect is phase specific, as senescence does not occur when PKC activation was triggered in early G 1 . Altogether, these data strongly suggest that the irreversible G 2/ M arrest induced by PMA during S phase leads to senescence. The senescence phenotype was also observed in H460 lung cancer cells, as well as in HT-29 and HCT-116 colon cancer cells. 5 p21 Cip1 Is Required for PMA-induced G 2 /M Arrest and Senescence Induction-Analysis of relevant cell cycle markers in NSCLC cells upon HU release revealed a marked elevation in the levels of the Cdk inhibitor p21 Cip1 in response to PMA, which was sustained even 72 h after treatment (Fig. 3, A and B). p21 Cip1 up-regulation is a characteristic feature of senescent cells (31). Additional studies in H358 cells revealed no significant changes in p27 levels. PMA treatment led to a marked reduction in Rb phosphorylation and in the levels of the transcription factor E2F-1. A sustained reduction in the levels of cyclin E, cyclin A, and cyclin B, as well as induction of cyclin D1, were also observed (Fig. 3B). While the levels of p21 Cip1 protein did not change significantly across S phase in H358 cells after HU release (Fig. 3C, upper panel, ϪPMA), a progressive p21 Cip1 up-regulation was observed in response to PMA (Fig. 3C, upper  panel, ϩPMA). Similar results were observed at the mRNA level, as determined by RT-PCR (Fig. 3C, lower panel). In addition, PMA treatment of HU-synchronized cells promoted a sig- nificant activation of a p21 Cip1 luciferase reporter (Fig. 3D). As newly synthesized p21 Cip1 protein translocates to the nucleus to exert its inhibitory activity on cyclin-Cdk complexes (32), we carried out a subcellular fractionation analysis. ATF-2 and vinculin were used as controls for the nuclear and cytosolic fractions, respectively. A marked elevation of nuclear p21 Cip1 in response to PMA was observed (Fig. 3E). Elevated p21 Cip1 levels were also detected in the cytosolic fraction, which probably reflects protein recently synthesized. Likewise, immunofluorescence staining revealed a prominent nuclear p21 Cip1 staining in PMA-treated compared with vehicle-treated H358 cells (Fig. 3F).
To determine whether a causal relationship exists between p21 Cip1 up-regulation, G 2 /M arrest and the senescent phenotype, we used a RNAi approach. Two different RNAi target sequences were used in order to minimize the chances of off-target effects. H358 cells were transfected with either p21 Cip1 RNAi or control duplexes, and 24 h later cells were HU-synchronized and treated with either PMA or vehicle. A Ͼ90% reduction in the induction of p21 Cip1 was achieved with either RNAi duplex (Fig. 4A), which lasted for at least 4 days (data not shown). p21 Cip1 depletion did not affect the synchronization protocol (supplemental Fig. S2). Notably, p21 Cip1 -depleted H358 cells failed to arrest in G 2 /M in response to PMA (Fig. 4B). Moreover, p21 Cip1 RNAi markedly reduced the induction of the senescence marker SA-␤-Gal by the phorbol ester (Fig.  4C). Therefore, p21 Cip1 up-regulation is required for G 2 /M arrest and the induction of a senescence phenotype by PMA in H358 cells.
PMA-induced p21 Cip1 Up-regulation, G 2 /M Arrest, and Senescence Are Mediated by PKC␣-H358, H441, and H322 cells express three phorbol ester-responsive PKCs: the classical PKC␣, and the novel PKCs ␦ and ⑀ (24, and data not shown). While PKC⑀ is mostly a mitogenic/ survival kinase, growth inhibitory roles have been ascribed to PKC␣ and/or PKC␦ in several cell models. We have previously described that PKC␦ mediates PMA-induced G 1 arrest in H358 cells (24). Expression levels of PKC␣, PKC␦, and PKC⑀ in H358 cells remained constant across the S phase (supplemental Fig. S3A). To determine the involvement of PKC isozymes in p21 Cip1 induction by PMA we first used a pharmacological approach. As we have found that translocation of PKC␣, PKC␦, and PKC⑀ in response to PMA (a hallmark of PKC activation) remained for the whole S phase (Fig. S3B), PKC inhibitors were added before PMA treatment (Ϫ50 min) and left until the end of the S phase. As shown in Fig. 5A, the pan-PKC inhibitor GF109203X (bisindolylmaleimide I) completely blocked p21 Cip1 up-regulation by PMA in H358, H441, and H322 cells. Gö6976, a specific inhibitor of cPKCs (33), also blocked p21 Cip1 up-regulation. On the other hand, the PKC␦ inhibitor rottlerin was ineffective or less effective. As PKC␣ is the only cPKC pres-  (n ϭ 3). E, H358 cell extracts were collected 4 h after PMA treatment (t ϭ 6 h after HU release) and subject to fractionation as described under "Experimental Procedures." p21 Cip1 protein was determined by Western blot in cytosolic and nuclear fractions. ATF-2 and vinculin were used as markers for the nuclear and cytosolic fractions, respectively. F, localization of p21 Cip1 by immunofluorescence in H358 cells 4 h after PMA treatment (t ϭ 6 h). Green, p21 Cip1 ; blue, DAPI. ent in H358 cells, these experiments suggest that this PKC mediates p21 Cip1 induction, and presumably also mediates G 2 /M arrest. To determine the involvement of PKC␣ in mediating the G 2 /M arrest we used RNAi. Fig. 5B shows that Ͼ90% reduction in PKC␣ levels was achieved upon delivery of two different specific PKC␣ RNAi duplexes into H358 cells, without affecting the expression of other phorbol ester-responsive PKC isozymes (Fig. 5B). In agreement with the results using inhibitors, PKC␣ depletion impaired PMA-induced up-regulation of p21 Cip1 (Fig. 5C), suggesting a crucial role for this cPKC in mediating p21 Cip1 up-regulation in S phase. Moreover, in PKC␣-depleted cells PMA failed to stimulate p21 Cip1 reporter luciferase activity (Fig. 5D). Because p21 Cip1 up-regulation is required for G 2 /M arrest and senescence induced by PMA, we reasoned that PKC␣ is involved in mediating the senescence phenotype. Analysis of cell cycle distribution revealed that PKC␣-depletion impaired PMA-induced G 2 /M arrest in H358 cells (Fig. 6A). Interestingly, in PKC␣-depleted cells the morphological alterations induced by PMA could not be observed (Fig. 6B). Moreover, the percentage of SA-␤-Gal-positive cells as a consequence of PMA stimulation was drastically reduced in H358 cells subject to PKC␣ RNAi (Fig. 6C). Furthermore, when PKC␣ was knocked down, the inhibitory effect of PMA on telomerase activity was impaired (Fig. 6D).
In the last set of experiments we used an adenoviral approach to overexpress PKC␣. H358 cells were infected with either control LacZ or PKC␣ AdVs (24). Cells were synchronized with HU and subject to PMA treatment, as described above. Neither LacZ nor PKC␣ overexpression affected the synchronization protocol (supplemental Fig. S4). Interestingly, the induction of p21 Cip1 expression by PMA was significantly enhanced in PKC␣-overexpressing H358 cells (Fig. 7A). While overexpression of PKC␣ using 100 MOI of PKC␣ AdV did not cause significant p21 Cip1 up-regulation in the absence of PMA stimula- tion, higher MOIs (500 pfu/cell) slightly increase the basal levels of p21 Cip1 (data not shown). The percentage of cells accumulated in G2/M in response to PMA was higher in PKC␣ AdV-infected cells (100 MOI) compared with control cells (Fig.  7B). Lastly, we took advantage of a specific PKC␣ agonist, the DAG-lactone HK-434 (34). Treatment of H358 cells with HK-434 induced p21 Cip1 levels to a similar extent than PMA (Fig. 7C). Furthermore, HK-434 caused a marked accumulation of cells in G 2 /M (Fig. 7D), as well as characteristic senescence features, including cell flattening and enlargement (Fig. 7E), and induction of SA-␤-Gal activity (Fig. 7F). Taken together, these data strongly suggest that activation of PKC␣ by PMA during S phase leads to p21 Cip1 up-regulation, irreversible G 2 /M arrest, and the induction of a senescence phenotype in H358 lung cancer cells.

DISCUSSION
A considerable body of evidence supports the notion that the differential proliferative versus antiproliferative effects of phorbol esters is conferred by the distinct ability of PKC isozymes to regulate the cell cycle machinery. This study introduces three novel paradigms. The first is that activation of PKC␣ can lead to a senescence phenotype, an effect mediated by the induction of p21 Cip1 . Second, senescence is strictly dependent upon the phase of the cell cycle in which PKC activation occurs. Third, this is the first study to our knowledge showing that in a defined cellular model phorbol esters trigger cell arrest in different phases of the cell cycle via a common mechanism (p21 Cip1 induction) but through distinct PKC isozymes. Our previous studies have shown that phorbol ester treatment of lung cancer cells in early G 1 impairs their progression into S phase. This G 1 arrest is mediated by PKC␦ (24). In contrast, when cells are treated with PMA in late G 1 or S phases, they undergo permanent arrest in G 2 /M accompanied by senescence, and in this case the effect is mediated by PKC␣ rather than PKC␦. Despite the differential involvement of PKC isozymes, both effects occur via the up-regulation of the cell cycle inhibitor p21 Cip1 . These cell cycle phase-specific effects not only attest to the functional complexity of PKC isozymes, but also suggest that PKC regulation of cell cycle via p21 Cip1 involves multiple mechanisms.
While phorbol esters are mitogenic in some cellular models (35,36), they can also inhibit proliferation or promote apoptosis in a large number of cell types. Dual G 0 /G 1 and G 2 /M arrest by phorbol esters has been reported in endothelial (13), breast cancer (37) and melanoma cells (38). PKC␣ was shown to play negatives roles in G 1 3 S transition as well as in G 2 3 M transition (4,5,38,39). However, there has been little information on the role of PKC activation in S phase. It has been reported that in G 1 /S-synchronized Demel and MCF-7 cells, phorbol ester stimulation induces a transitory G 2 arrest (37,38). This fits well with our observations that in H358 lung cancer cells PKC activation in late G 1 or early S phases causes a significant delay in S 3 G 2 transition. In addition to this effect, our data provide strong evidence that PMA activation promotes irreversible G 2 /M arrest and inhibition of proliferation by inducing cellular senescence. The cellular senescent phenotype is defined by the appearance of characteristic morphological changes and biochemical markers, including the induction of SA-␤-Gal activity and p21 cip1 , and inhibition of telomerase activity, leading ultimately to irreversible proliferation arrest. Cellular senescence has been originally described in primary cells as a consequence of telomere shortening (40). However, senescence can be also triggered by a variety of signals such as DNA damage, loss of tumor suppressors, and oncogenic hyperproliferative signals (30). The fact that PMA triggers cellular senescence in lung cancer cells argues for the existence of senescence mechanisms dependent on PKC. In support of our data, a very recent study has shown that long-term (24 h) treatment with the diterpene ester PEP005 induces irreversible cell cycle arrest in G 1 and G 2 /M phases and senescence in melanoma cells via PKC (41). However, long-term PKC activation leads to PKC down-regulation, and therefore it is unclear from that study whether PKC activation or PKC depletion was responsible for the phenotype in melanoma cells, as well as which PKC isozymes were responsible for the effect. In our studies in H358 cells, a short incubation with PMA is capable of inducing senescence, suggesting that PKC␣ activation rather than down-regulation is triggering a senescence program. Indeed, we do not detect PKC down-regulation as a consequence of the PMA treatment (supplemental Fig. S3B). Interestingly, H358 lung cancer cells go into senescence when treated with PMA in late G 1 -S, but only a small percentage of cells undergo a senescence phenotype when treated with PMA in early G 1 . This small fraction of senescent cells may arise from cells in late G 1 present as a consequence of incomplete G 0 synchronization by serum starvation (24). Altogether, these results argue that the senescent population originates from cells arrested in G 2 /M. Although the molecular mechanisms underlying the induction of senescence by PKC␣ still require investigation, our studies unambiguously show the involvement of the cell cycle inhibitor p21 Cip1 . PMA promotes a robust elevation in p21 Cip1 mRNA and protein levels during S phase. Furthermore, p21 Cip1 up-regulation represents an obligated event, as p21 Cip1 RNAi significantly impaired the induction of G 2 /M arrest and senescence by the phorbol ester. PKC␣ RNAi depletion inhibited the induction of p21 Cip1 by PMA, both at the level of mRNA (data not shown) and protein, and consequently blocked the induction of G 2 /M arrest and senescence by PMA, thus providing strong evidence for a PKC␣-p21 Cip1 -G 2 arrest-senescence link. The unique phenotypes observed in response to PKC activation at different cell cycle phases probably reflect distinct roles for p21 Cip1 in G 1 and G 2 . Indeed, while p21 Cip1 is an inhibitor of G 1 cyclin-Cdk complexes, it also inhibits the activation of Cdc2 in G 2 (42,43). PKC␣ was shown to play a role in G 1 /S arrest via p21 induction and to mediate cell cycle exit and a differentiation program in intestinal cells (4), suggesting that this cPKC can  exert multiple effects depending on the cell type. p21 Cip1 mRNA stability can be regulated by diverse factors such as the RNA-binding proteins HuR and CP1 (44). A very recent study showed that PKC␣ modulates HuR nuclear-cytosolic shuttling, thus favoring mRNA stabilization (45). One attractive hypothesis is that the specific PKC␣-mediated p21 Cip1 up-regulation in late G 1 /S phase could be mediated by HuR-dependent p21 Cip1 mRNA stabilization. However, preliminary studies have not revealed any noticeable changes in HuR total levels or relocalization during G 1 , S, or G 2 , either in response to PMA treatment or after PKC␣ overexpression (data not shown). Further studies would be required to address the mechanistic basis of the differential regulation of p21 Cip1 by PKCs.
Which of the major senescence pathway(s) play a role in PMA-induced senescence? Most senescence responses converge on either p53 (involving DNA damage) and/or pRB (mostly through stress or oncogene activation of p16) (30). It is well established that DNA lesions can trigger a checkpoint, which delays cell cycle progression and activate DNA-repair mechanisms (reviewed in Ref. 46). Preliminary data show no detectable Chk-1 or histone H2A.X phosphorylation during S-phase in response to PMA (data not shown), suggesting the absence of checkpoint activation by PKC␣. It has been postulated that G 2 arrest via p21 Cip1 induction may involve the suppression of cdk2/cyclin A and cdc2/cyclin B activities, changes in cdc2 phosphorylation status, and inhibition of cdc25 expression (4). PKC activation can inhibit Cdc2 and entry into mitosis by reducing the activity of Cdc25 phosphatases (which dephosphorylate and activate Cdc2), a mechanism that bypasses Chk1 (38). H358 cells do not express p16 (47), suggesting that senescence is independent of this cell cycle inhibitor. Also, the cell lines used in this study are p53-null or mutant (48,49), arguing for a p53independent mechanism in the induction of p21 Cip1 by PKC␣. Recent studies have implicated ERK activation as a mediator of p53-independent p21 Cip1 induction and G 2 /M arrest in response to PMA (50). p21 Cip1 up-regulation in HeLa and SKOV-3 cells in response to growth factors, oxidative stress, or PMA is dependent upon the ERK MAPK cascade (10,51,52). Regulation of p21 Cip1 via ERK may involve transcriptional or post-transcriptional mechanisms that extend p21 Cip1 half-life (26,53). Pharmacological inhibition of MEK blocks the senescence phenotype triggered by the diterpene PEP005 in melanoma cells (41). We have observed a strong and sustained activation of ERK in H358 cells upon PMA stimulation that preceded p21 Cip1 up-regulation in S phase, and we found that MEK inhibition markedly reduced p21 Cip1 mRNA and protein up-regulation (data not shown), suggesting that a similar mechanism might take place in lung cancer cells.
NSCLC cells are very resistant to conventional anti-cancer treatment. Induction of cellular senescence has been proposed as a strategy for cancer therapy (reviewed in Ref. 54). Studies in breast cancer have shown that induction of senescence following adjuvant therapy correlates with favorable outcome (55). One attractive scenario is that targeting PKC␣ in NSCLC with specific activators may promote senescence in tumor cells. Our studies strongly suggest that activation rather than inhibition of PKC␣ function/expression may serve for therapeutic purposes, and may help to explain the failure of PKC␣ antisense approaches for NSCLC treatment (20). Indeed, PKC agonists have been examined as antitumor agents in vivo. Bryostatin 1, a potent PKC activator that induces only a subset of those responses induced by PMA, has been in clinical trials for a number of malignancies, including lung cancer (1,17). Moreover, PMA has been administered to patients for the treatment of hematological malignancies (15)(16)(17). DAG lactones that selectively activate PKC␣ have anti-proliferative activity not only in lung cancer cells but also in other cancer cell lines (34). The fact that senescence induction in normal cells may facilitate tumor progression (56) argues for the need of additional studies to better assess the validity of PKC␣ or downstream effectors as a therapeutic targets.
In summary, our studies established that activation of PKC␣ in lung cancer cells during late G 1 -early S phase leads to cellular senescence, an effect that involves p21 Cip1 up-regulation and irreversible inhibition of cell proliferation. In addition, our results revealed important differences with regards to the role of individual PKC isozymes in different phases of the cell cycle.
In our previous study we demonstrated that PMA-induced activation of PKC␦ in early-mid G 1 inhibits H358 cell progression into S phase, inducing cell accumulation in G 1 (24), whereas PKC␣ was dispensable for this effect. The unique growth inhibitory effect of PKC isozymes in specific phases of the cell cycle opens a window of opportunity for targeting individual PKCs for therapeutic purposes.