Retinoblastoma Protein-dependent Growth Signal Conflict and Caspase Activity Are Required for Protein Kinase C-signaled Apoptosis of Prostate Epithelial Cells*

Both protein kinase C and the retinoblastoma tumor suppressor protein have been linked to the regulation of cell growth and cell death, suggesting the differential roles these factors play in mediating cell fate. In some cells, protein kinase C-induced activation of the retinoblastoma protein results in G1arrest. However, inducible overexpression and activation of the protein kinase Cα isozyme or the addition of 12-O-tetradecanoylphorbol-13-acetate in the prostate epithelial cell line, LNCaP, resulted in apoptosis preceded by induction of p21 and dephosphorylation of the retinoblastoma protein. Consistent with a role for the retinoblastoma growth suppressor protein in protein kinase C-induced apoptosis, DU145 cells, which do not express functional retinoblastoma protein or LNCaP cells, which have been transfected with the retinoblastoma inhibitor, E1a, were resistant to apoptosis. LNCaP apoptosis was initiated by a unique conflict between the growth-suppressive activity of the retinoblastoma protein and growth-promoting mitogenic signals. Thus, when this conflict was prevented by serum depletion, apoptosis was suppressed. The caspase family of cysteine proteases is believed to encompass the execution machinery of mammalian apoptosis, and addition of the cell-permeable caspase inhibitor, Z-Val-Ala-Asp-fluoromethylketone, afforded nearly total protection from protein kinase C-signaled apoptosis. This protection correlated with the total loss of caspase activity as measured by the proteolytic cleavage of nuclear poly(ADP-ribose) polymerase. On the basis of these results, we propose that protein kinase C regulates a novel cell death pathway that is initiated by a cellular conflict between retinoblastoma growth-suppressive signals and serum mitogenic signals in proliferating prostate epithelial cells and that is executed by the caspase family of cysteine proteases.

tem, nervous system, and the epithelial component of various tissues (1,2). Soluble molecules, however, are not exclusive in the initiation of an apoptotic response in cells. Suppression of apoptosis in anchorage-dependent cells requires their attachment to the extracellular matrix, disruption of which is sufficient to induce apoptosis. Consistent with this, we have presented evidence, in vitro and in vivo, that the loss of ␤ 1 integrindependent cell anchorage resulted in apoptosis of several epithelial cell lines requiring the retinoblastoma (Rb) 1 cell cycle-regulatory pathway (3).
As a regulator of cell cycle, Rb is central to a pathway that is frequently disrupted in human cancer, and its inactivation during tumorigenesis demonstrates the dire consequences of depriving the cell of a means with which to undergo growth arrest (4). Mutational events that occur in the genesis of prostate cancer involve alterations in apoptotic pathways present in normal androgen-dependent epithelium. Two such alterations of the genetic machinery in prostate carcinogenesis include mutation of the cyclin-dependent kinase inhibitor, p21 (5), and allelic loss at the retinoblastoma locus (6 -8). Rb encodes a 110-kDa phosphoprotein that regulates the transition between G 1 and S phases in the cell cycle and also functions in transducing growth-inhibitory signals that regulate differentiation and growth (9). Functional regulation of Rb is cell cycle dependent, being strictly controlled by the activity of cyclin-dependent kinases (cdks) during G 1 . Rb is functionally active in an underphosphorylated conformation and is inactivated during late G 1 by cyclin D/cdk4/6-and cyclin E/cdk2-dependent phosphorylation, allowing the cell to proceed from G 1 to S (4). Additionally, cdk activity and Rb function are regulated by a family of inhibitors that block the catalytic activity of the cdk/cyclin complex. The cdk inhibitors p21 (SDI, WAF1, CIP1) and p27 (KIP1) are universal in their specificity and inhibit the activity of a number of different cdks. p16 (MTS1, INK4A), on the other hand, selectively blocks the activity of cdk4 and cdk6. Regardless of specificity, overexpression of either p21 or p16 have been shown to activate Rb, resulting in G 1 arrest of cells (10,11).
We have previously reported that apoptosis of prostate epithelial cells is regulated by activators and inhibitors of the protein kinase C (PKC) family (12,13). PKC is a synonym for a multimember family of serine/threonine kinases that has traditionally been associated with signal transduction regulating cell growth and differentiation (14) but has recently been as-sociated with the regulation of cell death as well (12,13). Most PKC isozymes require the physiological activator diacylglycerol, which is derived from membrane phospholipids; for full activity, however, the phorbol ester, 12-O-tetradecanoylphorbol-13-acetate (TPA), or the macrocyclic lactone, bryostatin 1, will also activate PKC. Additionally, PKC activity also requires association with cellular membranes and/or cytoskeletal components to execute many of its physiological functions (15,16). PKC modulates signal transduction pathways that have been linked to both positive and negative regulation of the cell cycle and the initiation of apoptosis; however, the specialized roles of the individual isozymes and their physiological targets that modulate these events are not known. The response of a particular type of cell to PKC activators appears to depend not only on specific PKC isozyme expression but also on the status of the functional targets of PKC. Thus, growth regulation by PKC is likely mediated by the cell cycle machinery, but the regulation of cell death may require other, undefined components of the PKC signaling pathway.
By using an inducible overexpression system of PKC␣ in the present study, we examined the role of Rb in PKC-regulated apoptosis of LNCaP cells. Results from this study are the first to demonstrate that PKC-induced apoptosis was preceded by the rapid induction of p21 and Rb dephosphorylation and required Rb activity for cell death. Moreover, Rb could only induce apoptosis in opposition to serum mitogenic signals in proliferating LNCaP cells and that the execution of apoptosis was dependent on the activation of the caspase family of cysteine proteases. Taken together, the experimental evidence indicates that PKC and Rb are linked in a novel signal transduction pathway which in the presence of mitogenic signals culminates in caspase-regulated apoptosis.
Transfection of LNCaP Cells-The previously described cell lines LNGK9 (vector control) and LN␣17 are subclones of LNCaP cells in which expression of exogenous human PKC␣ is regulated by the presence or absence of tetracycline in the medium. 2 These clones have stably incorporated plasmid pUHD 15-1, encoding a tetracycline-responsive transactivator protein, and plasmid pUHD 10-3/PKC␣, from which human PKC␣ is expressed (clone LN␣17) under control of a tetracycline-repressible promoter. For stable clones expressing E1a or E1a 928, the pCMV-E1a-12S or pCMV-E1a-12S-928 plasmids (3) were transfected into LNCaP cells and clones were isolated in 500 g/ml of G418.

Regulation of p21 Expression and Rb Activation by PKC
Regulators-We have previously demonstrated that depriving anchorage-dependent cells of matrix contact induced a novel apoptotic pathway dependent on Rb activity (3). We have also developed a model of TPA-induced apoptosis in Rb ϩ/ϩ prostate cells that exhibits induction of PKC␣ expression and membrane translocation (12,13). Taken together, these studies implicated an apoptotic mechanism uniquely regulated by Rb in epithelial cells that may be initiated by PKC signal transduction. To ascertain the functional significance of Rb in PKCmediated apoptosis, we began by examining the response of Rb ϩ/ϩ and Rb Ϫ/Ϫ prostate epithelial cells to the PKC activator, TPA. The human prostate epithelial cell line, DU145, which does not express functional Rb (18), failed to undergo TPAinduced apoptosis as measured by the exclusion of trypan blue, terminal deoxynucleotidyltransferase-mediated dUTP-biotin nick end labeling assay, and nuclear condensation ( Fig. 1A and results not shown). At moderate concentrations of TPA, which induced apoptosis in 70% of the Rb ϩ LNCaP culture by 48 h, the DU145 cultures remained viable.
Rb is maintained in its active state through cdk inactivation achieved either by loss of the regulatory cyclins or by induction of the p21/p27 and p16 family of cdk inhibitors. To determine whether Rb is activated through increased cdk inhibitor expression or reduced cyclin expression following TPA treatment, we examined p21, p27, and p16 protein levels following the addition of TPA to LNCaP cells. The cdk inhibitor p21, which inhibits both cyclin E-cdk2 and cyclin D1-cdk4/6, was shown to be strongly induced within 12 h of TPA addition in LNCaP cells and remained elevated throughout the 48-h time course (Fig.  1A). In contrast, no change in the level of the cdk inhibitors p27 or p16 was evident, nor were there changes in the levels of cyclin D1 or E (data not shown). Increased p21 expression might be expected to lead to dephosphorylation and activation of Rb. In Fig. 1A, we show that activation of Rb does indeed follow the induction of p21 with significant conversion from hyper-to hypophosphorylated Rb by 12 h following TPA, coincident with the onset of apoptosis. By 24 h, Rb had been completely converted to the hypophosphorylated form correlating to 40% reduction in cell viability. The protein remained dephosphorylated at 48 h at which time cell viability had been reduced to 31%.
Since TPA is a potent activator of the PKC family (14), we wanted to confirm that PKC activity was associated with the induction of p21 and Rb activation. Staurosporine, a potent but nonspecific catalytic inhibitor of protein kinase C (19), and bryostatin 1, a specific activator and down regulator of PKC, have both been shown to prevent TPA-induced apoptosis (12,13). Pretreatment of LNCaP cultures with staurosporine or bryostatin 1 prior to TPA addition not only inhibited apoptosis but also completely prevented p21 induction and Rb activation ( Fig. 1B and data not shown).
E1a Inhibition of Rb Function Suppressed PKC-mediated Apoptosis-Our results using activators and inhibitors of PKC suggested that PKC recruits the Rb cell cycle control pathway to initiate LNCaP apoptosis. Since we had previously demon-strated that expression of exogenous Rb was sufficient to induce apoptosis in LNCaP cells (3), we next wanted to determine whether functional inhibition of endogenous Rb could suppress apoptosis in these cells. To address this question, the viral oncogene E1a, which binds to and functionally inhibits Rb (20), was stably integrated into LNCaP cells, and the resulting transfectants were challenged with TPA. As shown in Fig. 2, LNCaP cells stably expressing E1a did not undergo apoptosis after addition of TPA. Western blot analysis of Rb in the E1a-LNCaP line demonstrated that Rb was dephosphorylated with normal kinetics following TPA treatment (data not shown), indicating that Rb function was not disrupted upstream in its regulatory pathway but following dephosphorylation in the presence of E1a as expected. Because E1a has cellular targets in addition to Rb, we used a control E1a expression vector with a mutation at nucleotide 928 that blocks interaction with Rb without disrupting interactions with the other proteins (21). In contrast to wild-type E1a, the Rb-specific E1a point mutant (E1a-928) did not block apoptosis in TPA-treated LNCaP cells, suggesting that the anti-apoptotic activity of E1a results from inhibition of Rb function alone.
Apoptosis-resistant LNCaP Cells Do Not Exhibit p21 Induction or Rb Dephosphorylation-In many cells, TPA-induced membrane translocation of PKC is followed by degradation and depletion of PKC cellular pools (22). In LNCaP cells, membrane translocation of PKC␣ following TPA addition persists for at least 36 h before down-regulation is observed, and TPA exposure longer than 48 h results in an apoptotic-resistant phenotype with undetectable levels of PKC␣ protein (13). If p21 and Rb reside in a PKC-regulated pathway, it is likely then that long term exposure of LNCaP to TPA should result in the emergence of apoptotic-resistant cells incapable of p21 induction or Rb activation. When LNCaP cells were cultured in TPA for 6 days, we observed an initial apoptotic response of 57% by day 2 (Fig. 3). Interestingly, by day 4 post-TPA treatment, the surviving cells had grown back to the original plating number and continued to proliferate out through day 6. Apoptosis of LNCaP cells correlated with the induction of p21 and Rb dephosphorylation at day 1. However, by day 2 the levels of p21 had begun to decline and Rb was just beginning to convert to the hyperphosphorylated form. By day 4, the cells were TPA insensitive and have begun to proliferate and p21 expression and Rb phosphorylation had returned to normal (Fig. 3). Thus, TPA sensitivity and the emergence of an apoptotic-resistant subpopulation correlated precisely with p21 levels and the activation state of Rb.
Inducible Overexpression and Activation of PKC␣ Resulted in p21 Induction, Rb Activation, and Apoptosis-Data from the PKC activator/inhibitor studies coupled to the observation that depletion of PKC resulted in a nonresponsive phenotype suggested that PKC has a regulatory role in cell cycle events leading to apoptosis. Because we previously identified PKC␣ as the predominant isotype associated with TPA-induced apoptosis (13), we next examined the effects of inducible PKC␣ overexpression in LNCaP cells. In two separate clones (LN␣1, LN␣17), which express exogenous PKC␣ under control of the tetracycline-responsive promoter, removal of tetracycline resulted in a 30-and 50-fold induction of PKC␣ protein (Fig. 4).
In the presence of tetracycline, expression of PKC␣ was not detectable at higher levels in the clones than in parental LN-CaP cells. Inducible overexpression of PKC␣ by itself did not result in the induction of p21, Rb dephosphorylation, or apoptosis. One explanation for the lack of apoptosis in PKC␣-overexpressing cells may be the absence of an activation signal for PKC␣. Bryostatin 1, which activates and down-regulates certain PKC isozymes with potencies different than those of TPA, will down-regulate the low endogenous levels of PKC␣ in parental LNCaP cells between 12 and 24 h. Bryostatin 1 treatment of LNCaP cells prior to TPA treatment prevented p21 induction, Rb activation, and apoptosis (data not shown); however, addition of bryostatin 1 to LN␣17 could not down-regulate the extremely high levels of exogenous PKC␣ and resulted in p21 induction, Rb activation and apoptosis ( Fig. 4 and data not shown). Raf-1, one of the immediate targets of PKC, was hyperphosphorylated (activated) following bryostatin 1 treatment in the PKC␣-overexpressing cells, indicating that bryostatin 1 functions as an activator of PKC␣ in these cells 3 and suggests that Raf-1 represents a functional link between PKC␣ and the induction of p21. Bryostatin 1 by itself had no effect on cell viability, p21 expression, or Rb phosphorylation in parental LNCaP cells, nonexpressing (Tetϩ) PKC␣ cells, or on the empty vector control, LNGK9 ( Fig. 4 and data not shown).

A Cellular Conflict between Rb Growth-inhibitory Signals and Serum Mitogenic Signals Triggers Apoptosis in LNCaP
Cells-Cytokines and growth factors in the extracellular environment are critically important in determining cell survival (23). However, opposing growth-regulatory signals that arise in some cells, such as forced c-myc expression (growth promoting) in fibroblasts in the absence of serum growth factors (growth inhibiting), precipitates a conflict resulting in a dysfunctional cell cycle and apoptosis (23). Because Rb transfection, which is growth inhibitory, induced apoptosis of proliferating epithelial cells (3), we predicted the opposite scenario in our model: that PKC␣-mediated Rb activation provides a negative signal for growth arrest in direct opposition to a positive mitogenic signal in proliferating LNCaP cells; and that this unique conflict results in apoptosis. If this conflict initiated apoptosis in PKC␣-overexpressing cells, then removal of the mitogenic signal by depletion of serum growth factors might prevent apoptosis. Indeed, serum-starved PKC␣-expressing cells were resistant to apoptosis induced by the PKC activator, bryostatin 1 (Table I). In addition, parental LNCaP cells under the same conditions were also completely resistant to TPA-induced apoptosis (data not shown). Serum withdrawal had no effect on the endogenous or exogenous levels of PKC␣ expression in LNCaP parental cells or in the PKC␣-expressing clones (data not shown). Thus, it appears that an apoptotic conflict arises when an Rb growth-suppressive signal opposes growth-promoting signals from serum growth factors.

Rb-mediated Apoptosis Is Carried Out by the Caspase Family of Cysteine Proteases-Our results have demonstrated that PKC-signaled apoptosis of LNCaP cells is triggered by an
Rbdependent conflict with mitogenic signals. Because a deathpromoting role for Rb contradicts its traditional role in cell cycle regulation, we were compelled to identify the mechanism by which cell death was executed. Since the caspase family of cysteine proteases are thought to encompass the execution phase of programmed cell death, we next examined the role of caspases in Rb-mediated apoptosis. As shown in Table I, the cell-permeable caspase inhibitor ZVAD-fmk afforded nearly total protection from apoptosis in the bryostatin 1-treated PKC␣-expressing line, LN␣17. Due to the high level of protection by ZVAD-fmk pretreatment, we expected to see a strong correlation with the loss of caspase proteolytic activity. To this end we examined the efficacy with which ZVAD-fmk might inhibit caspase function by analyzing proteolytic cleavage of the caspase substrate, PARP, by Western blot analysis using extracts from cycling LN␣17 cells and serum-starved LN␣17 cells. We also examined the phosphorylation status of Rb in these extracts to determine the sequence of events between Rb activation and PARP cleavage in this system. As shown in Fig.  5, PARP is completely cleaved in cycling LN␣17 cells treated with bryostatin 1 by 24 h. In stark contrast is the lack of any detectable cleavage of PARP in serum-deprived LN␣17 cells. When we examined extracts from LN␣17 cells that were pretreated with the PKC inhibitor, staurosporine, which inhibits PKC-mediated initiation of this pathway, or ZVAD-fmk, which inhibits the terminal caspase-mediated execution phase of this pathway, no PARP cleavage was detectable through 72 h of bryostatin 1 treatment (Fig. 6). Additionally, the Rb Ϫ/Ϫ DU145 cells, which cannot undergo apoptosis, do not exhibit PARP cleavage following TPA treatment (Fig. 6). Coupled to the viability data, these results indicate that the execution of apoptosis activated by this unique growth signal conflict is achieved through the caspase family of cysteine proteases. DISCUSSION We and others have previously demonstrated that castration-induced apoptosis of prostate epithelial cells coincides with the loss of G 1 cdk activity and increased synthesis of Rb mRNA and protein (3,24,25). Interestingly, the Rb that had accumulated in the glandular epithelium following castration was in the hypophosphorylated (active) form. 3 Similar to this observation was our current findings in LNCaP cells, where TPA treatment or overexpression and activation of PKC␣ led to the loss of G 1 cdk activity, resulting in the accumulation of hypophosphorylated Rb. The resulting growth signal conflict initiated by Rb growth-suppressive activity in this system did not lead to growth arrest, but in opposition to mitogenic signals in proliferating cells, resulted in apoptosis. This conflict-mediated apoptosis did not occur in Rb Ϫ/Ϫ prostate epithelial cells and was prevented by functional inhibition of Rb or removal of mitogenic signals. From our use of caspase inhibitors and examination of their activity, it is apparent that execution of this apoptotic pathway is mediated through this proteolytic pathway. Taken together, these results suggest that PKC signals a unique apoptotic pathway that is initiated by an Rb-dependent growth signal conflict in proliferating LNCaP cells and executed by the caspase proteases.
Programmed cell death is regulated by diverse cellular factors, some of which possess differential functions that may include the maintenance of cell growth and survival. Two such regulatory molecules, PKC and Rb, which traditionally are implicated in the transduction and execution of growth-regulatory and differentiation signals (26 -29), have been linked to the regulation of apoptosis as well (3,12,13). Although much of the molecular function of PKC and Rb are known, specific PKC isozymes and the mechanism of Rb-mediated apoptosis that it signals are completely unknown. The current study was initiated based on the observation that the Rb ϩ/ϩ LNCaP cells undergo TPA-induced apoptosis associated with p21 induction and Rb activation and that DU145 cells, which exhibit p21 induction but lack functional Rb, do not. Therefore, it appears that the signal to decrease cdk activity through p21 is sent in TPA-treated Rb Ϫ/Ϫ DU145 cells, but they do not undergo apoptosis, perhaps because they lack Rb. In total, these results suggested that Rb activity is required for TPA-induced apoptosis. To investigate this further, we focused on this pathway in the Rb ϩ/ϩ LNCaP line. TPA-treated LNCaP cells did not exhibit loss of cyclin D1 or cyclin E expression; however, there was a dramatic increase in the level of p21. In contrast to p21 induction, there was little change in the level of the related inhibitor p27 or the cdk4/6 inhibitor p16. Therefore, it appeared that a PKC-mediated increase in p21 expression leads to a reduction in G 1 cdk activity, resulting in Rb hypophosphorylation and cell death.
The response of a particular type of cell to PKC activators appears to depend not only on the combination of PKC isozymes expressed in the cell but also on the expression and status of the functional targets of PKC. As an "early " functional target of PKC, Raf-1 becomes phosphorylated (activated) by TPA in a variety of cell types, which exhibit various responses, such as proliferation, transformation, growth arrest, and differentiation. However, in LNCaP cells overexpressing activated PKC␣, Raf-1 is phosphorylated very early in a pathway that terminates in cell death. This finding suggests that Raf-1 phosphorylation by PKC␣ represents a putative signaling cascade leading to Rb hypophosphorylation and apoptosis. The association between PKC signal transduction and activation of Rb was further established by staurosporine-mediated inhibition of p21 induction and Rb activation. Coupled to the observation that this pathway was inhibited by the depletion of PKC, suggested a signal-transducing role for PKC in G 1 regulation that in LNCaP cells results in apoptosis.
Because these studies provided correlative evidence but not definitive proof of PKC function in this pathway, we next examined this mechanism in inducible PKC␣-overexpressing lines. The p21/Rb pathway was indeed activated following PKC␣ induction in the presence of the PKC activator, bryostatin 1, implicating this isoform in cell cycle regulation and apoptosis. The effects of inducible overexpression of PKC␣ have not been examined previously, and our results contrast with those of Ways et al., who, employing constitutive overexpression of PKC␣ in MCF-7 cells, showed that TPA and bryostatin 1-induced cell death was associated with G 2 -M accumulation, minimal p21 induction, and lack of Rb activation (30). Additionally, parental MCF-7 cells treated with TPA demonstrated G 1 arrest associated with p21 induction and Rb activation and were mostly unaffected by bryostatin 1. The results in MCF-7 cells were complicated by the finding that constitutive overexpression of PKC␣, in the absence of PKC activators, resulted in increased expression of PKC␤ and decreased expression of PKC and p53 (31). None of these changes were observed in LNCaP cells following inducible overexpression of PKC␣. Interestingly, Livneh et al. reported that ectopic expression of the epithelial-specific PKC isoform regulated a similar program in NIH 3T3 cells, resulting in p21 induction and Rb dephosphorylation (32). However, these PKC-expressing fibroblasts did not undergo apoptosis as do inducible PKC␣-expressing LNCaP epithelial cells but remain G 1 arrested.
Our observations along with those of other laboratories suggest that accumulation of hypophosphorylated Rb is a general response of PKC activity leading to G 1 arrest in a variety of cell types (33). It has also been demonstrated in fibroblasts that disruption of integrin-mediated contact results in the induction of p21 and G 1 cyclin inactivation leading to dephosphorylation of Rb and growth arrest (34,35). As in fibroblasts, however, we have found that the same integrin-regulated pathway induced in several different epithelial cell lines, including prostate, resulted in p21 induction and Rb activation, but in contrast to growth arrest of fibroblasts, the epithelial cells underwent apoptosis (3). The present results using ectopic expression of PKC␣ demonstrate that PKC activation, like ␤ 1 integrin, initiates a signaling pathway that results in Rb hypophosphorylation and apoptosis of prostate epithelial cells. The evidence that Rb is central to this PKC-induced apoptotic pathway was confirmed by the finding that LNCaP clones, stably expressing the adenovirus E1a protein, prevented TPA-induced apoptosis. It therefore appears that activation of Rb, which leads to growth suppression in some cell types, uniquely regulates an apoptotic pathway in prostate epithelial cells and possibly other epithelial cells (3).
Because LNCaP cells respond to Rb activation by undergoing apoptosis, Rb itself must encompass some aspect of the cell death signal; however, the nature of such a signal is unknown. Apoptosis is thought to be a default pathway in cells receiving opposing cell proliferation signals. Evan et al. have published extensively on growth signal conflict-induced apoptosis using c-myc overexpression in fibroblasts deprived of serum growth factors (23). Although our findings demonstrate that conflicting growth signals induce apoptosis, it is due to the opposite scenario, where negative Rb growth-inhibitory signals precipitate an apoptotic conflict with positive mitogenic signals in proliferating prostate epithelial cells. To confirm this possibility, we found that if such a conflict was prevented in LNCaP cells, by removing the mitogenic signal (serum depletion) or eliminating the growth-inhibitory signal (Rb deletion in the DU145 cells or Rb inactivation by E1a transfection in LNCaP cells), then apoptosis does not occur.
We next set out to determine the mechanism through which apoptosis is executed by this Rb-mediated conflict. We began by investigating the interleukin-1␤-converting enzyme-related proteases, which have been implicated in apoptosis of many diverse systems. When we used the cell-permeable caspase inhibitor, ZVAD-fmk, we observed total protection from apoptosis in the bryostatin 1-treated PKC␣-expressing line, LN␣17. When we examined cleavage of the caspase substrate, PARP, we found it unaltered by 12 h of PKC␣ activation, which is the time when Rb is becoming activated, but that complete cleavage had occurred by 24 h, at which time Rb is fully activated, suggesting that Rb activation preceded activation of the caspases. In determining the order of events in this apoptotic pathway, we observed that inhibition of PKC by staurosporine (initiation phase) not only inhibits Rb activation but also inhibits apoptosis and the associated cleavage of PARP. Coupled to the findings that ZVAD-fmk inhibited both apoptosis and PARP cleavage (execution phase) in LNCaP cells but did not inhibit Rb hypophosphorylation places Rb downstream of PKC yet upstream of caspase activation. Therefore, we postulate that the cellular conflict signal provided by Rb in proliferating cells is initiated by PKC and terminates in the recruitment of caspases to execute apoptosis.
Tumor suppression is manifest not only by cell cycle arrest but also through the initiation of cell death programs as well. The role of the p53 tumor suppressor gene product in apoptosis has been extensively documented in cells that are responding to DNA-damaging agents or chemotherapeutic agents or in cells that have a deregulated cell cycle (reviewed in Ref. 1). Although Rb has been shown to inhibit multiple p53-dependent apoptotic pathways (1,23), accumulation of hypophosphorylated Rb leading to G 1 arrest and apoptosis occurs in multiple p53-independent pathways as well (36,37). Two such p53independent pathways are induced by release of the lipid second messenger ceramide (38,39) or by DNA-damaging agents (40), both of which result in the accumulation of hypophosphorylated Rb, G 1 arrest, and apoptosis. Apoptosis also occurs normally in the androgen-dependent prostate epithelium of p53 Ϫ/Ϫ mice following castration (41), indicating that this is a p53-independent process. p53 transcriptionally activates a number of different genes including p21 in response to DNA damage; however, p21 expression can occur independently of p53 function (42).
PKC and Rb modulate a variety of physiologic responses in cells ranging from the regulation of cell growth and differentiation to cell death. Therefore, it is likely that cell survival is dictated by a complex intracellular environment through which PKC signaling pathways are responding to common extracellular cues. Some of these PKC-regulated pathways terminate in the accumulation of hypophosphorylated Rb; however, depending on the type of cell and its environment, PKC and Rb activities may encompass differential roles, directing the cell toward quiescence, differentiation, or apoptosis. We maintain that the primary role of Rb in regulating cell cycle is to signal growth arrest; however, we suggest that it is the extracellular and intracellular mitogenic influences that dictates whether the cell will survive or die in response to that signal.