Oncogenic Ras mediates apoptosis in response to Protein Kinase C inhibition through the generation of reactive oxygen species

Ras is a well-established modulator of apoptosis. Suppression of PKC activity can selectively induce apoptosis in cells expressing a constitutively activated Ras protein. We wished to determine whether reactive oxygen species serve as an effector of Ras-mediated apoptosis. Ras-transformed NIH/3T3 cells contained higher basal levels of intracellular H 2 O 2 compared to normal NIH/3T3 cells, and PKC inhibition upregulated ROS to five-fold greater levels in Ras-transformed cells than in normal cells. Treatment with N-acetyl-L-cysteine reduced both the basal and inducible levels of intracellular H 2 O 2 in NIH/3T3-Ras cells and antagonized the induction of apoptosis by PKC inhibition. Culturing NIH/3T3-Ras cells in low-oxygen conditions, which prevents ROS generation, also inhibited the apoptotic response to PKC inhibition. These results suggest that reactive oxygen species are necessary as downstream effectors of the Ras-mediated apoptotic response to PKC inhibition. However, the generation of ROS alone is not sufficient to induce apoptosis in Ras-transformed cells because inhibition of cell cycle progression prevented the induction of apoptosis in NIH/3T3-Ras cells without inhibiting the generation of intracellular H 2 O 2 observed after PKC inhibition. These findings suggest that continued cell cycle progression of Ras-transformed cells during PKC inhibition is also necessary for the induction of apoptosis.


INTRODUCTION
The Ras proto-oncogene serves as a molecular switch controlling a variety of cellular processes including proliferation (1), differentiation (2), and senescence (3). Point mutations that cause single amino acid substitutions in the normal cellular Ras protein lead to its constitutive activation (4). This dominant mutant form of Ras plays a major role in the multi-step progression of tumorigenesis in many human tumors, with oncogenic ras mutations occurring in approximately 30% of all human tumors (5). Strikingly, over 90% of pancreatic tumors and 50% of colorectal tumors analyzed contain ras mutations. 6 2 hours and DNA profiles were analyzed by FACS using a FACScan flow cytometer (Becton Dickenson).

Measurement of ROS.
Cells were plated at 1 x 10 5 /plate in 60 mm dishes and treated for the indicated times. Cells were harvested with trypsin/EDTA, washed once in PBS, and resuspended in 5 µg 2'-7' dichlorodihydrofluorescein diacetate (DCF) (Molecular Probes) /ml in Hank's balanced salt solution (HBSS). Samples were incubated for 10 minutes at room temperature and analyzed immediately by FACS. For manipulation of ROS levels, cells were either treated with 20 mM NAC (Sigma) or placed in sealed chambers (Billups-Rothenberg), which had been flushed for 30 minutes with 95% N 2 /5% CO 2 as previously described (30) for 24 hours before treatment with HMG.
In vitro PKC assay. Cellular PKC activity was measured using a commercial assay, following the manufacturer's protocol (Upstate Biotechnology) and described previously (7).
Briefly, cells were treated with 150 µM HMG for 24 hours or left untreated in normal culture conditions before harvesting with trypsin/EDTA. Cells were lysed in 25 mM Tris-HCl (pH 7.5), 1% Triton X-100, 20 mM MgCl 2 , and 150 mM NaCl and extracts were normalized for protein concentration. Subsequently, 100 µg of extract was incubated with a PKC-specific peptide substrate, [ 32 P]ATP, and inhibitors of PKA and calmodulin kinase for 10 minutes at 30 o C. 32 P incorporated into the substrate was separated from residual 32 P using p81 phosphocellulose paper and quantified by scintillation counting.
Measurement of Ras activity. NIH/3T3-Ras cells were treated as indicated for 24 hours, either in low-oxygen culture conditions, or with sodium butyrate (5 mM) or TsA (100ng/ml). For control, NIH/3T3 cells were serum-starved in medium containing 0.5% DCS for 24 hours and endogenous Ras was activated by PDGF (30ng/ml) stimulation for 15 minutes. Ras activity was 7 measured using the technique developed by de Rooij and Bos (31

RESULTS
Our laboratory has previously shown that downregulation of PKC by chronic, high-dose PMA treatment (32) could selectively induce apoptosis in Jurkat human T lymphoblastoid cells stably-expressing v-Ha-ras (PH1 cells) when compared to normal Jurkat cells (7,8).
Downregulation of PKC by high-dose PMA treatment also selectively induced cell death in both v-Ki-ras transformed Balb fibroblasts (KBalb) and v-Ha-ras transformed NIH/3T3 fibroblasts (NIH/3T3-Ras) when compared to normal Balb and NIH/3T3 cells (Fig. 1a). Treatment with 100 nM PMA, a concentration that activates PKC but is insufficient to cause downregulation of PKC, consistently caused less than 10% loss in cell viability of the Ras-transformed fibroblasts. In contrast, chronic high-dose treatment of Ras-transformed fibroblasts with 500 nM PMA, a concentration known to cause downregulation of PKC after prolonged exposure (32), caused between 85%-95% cell death after 72 hours. No loss in cell viability was seen in normal Balb or NIH/3T3 fibroblasts with either 100 nM or 500 nM PMA treatment. These findings demonstrate that the suppression of PKC activity, but not activation of PKC, can trigger cell death in the presence of oncogenic Ras activity.
To determine whether a pharmacological inhibitor of PKC could also selectively induce cell death in Ras-transformed cells, the DAG antagonist HMG was used to treat both normal and Ras-transformed fibroblasts. Treatment of KBalb and NIH/3T3-Ras fibroblasts with 150 µM HMG caused Ras-specific cell death, similar to that achieved using chronic, high-dose PMA downregulation of PKC. Ras-transformed cells treated with 150 µM HMG underwent 80%-90% loss in cell viability after 48 hours, while less than 10% loss of viability was observed after treatment of normal Balb and NIH/3T3 cells (Fig. 1b). Other selective PKC inhibitors, such as 9 transformed cells 2 . Overexpression of Bcl-2 in KBalb cells by transfection with a retroviral expression vector inhibited the apoptosis induced by HMG treatment, consistent with our previous findings that Bcl-2 overexpression effectively inhibited death induced by chronic, highdose PMA treatment in cells expressing activated Ras (8). To verify that the dose of HMG that selectively induced cell death was also inhibiting PKC activity, an in vitro PKC activity assay was performed. Treatment with 150 µM HMG suppressed the cellular PKC activity of cells growing logarithmically in 10% serum-containing medium by greater than four-fold (Fig. 1c).
Photomicroscopy of NIH/3T3-Ras cells treated with HMG showed changes in morphology consistent with apoptosis ( Fig. 1d). Quantification of apoptosis by propidium iodide staining and FACS analysis showed that NIH/3T3-Ras cells underwent increasing amounts of DNA fragmentation over a 72 hour period, with over 70% of the population containing a hypodiploid DNA content at 72 hours (Fig. 2a). In contrast, no increase in DNA fragmentation was observed in parental (normal) NIH/3T3 after 72 hours of HMG treatment when compared to basal levels ( Fig. 2b). Concurrent analysis of cell cycle indicated that over 98% of parental NIH/3T3 cells were arrested in the G 0 /G 1 phase after PKC inhibition (Fig. 2c). This cell cycle arrest was reversible, and NIH/3T3 cells resumed cell cycle progression once HMG was washed out of culture 2 .
To determine if reactive oxygen species served as effectors of Ras-mediated apoptosis, intracellular ROS levels were measured by FACS using the peroxide-sensitive fluorescent indicator, DCF. NIH/3T3-Ras cells showed higher basal levels of ROS compared to parental NIH/3T3 cells (Fig. 3a). KBalb cells also showed higher basal ROS levels compared to parental Balb cells 2 . HMG treatment for 24 hours caused an upregulation of ROS levels in both NIH/3T3 and NIH/3T3-Ras (Fig. 3b). However, while ROS levels were upregulated by only two-fold in by guest on March 23, 2020 http://www.jbc.org/ Downloaded from 10 NIH/3T3 cells in response to HMG, ROS levels increased by more than seven-fold in NIH/3T3-Ras cells. Treatment with the structurally-unrelated PKC inhibitors Ro-31-8220 and GF109203X also resulted in similar upregulation of ROS levels 2 , suggesting that ROS generation by HMG correlated with PKC inhibition.
To establish whether the upregulation of ROS levels is necessary for Ras-mediated apoptosis, the antioxidant N-acetyl-cysteine (NAC) was used to pre-treat Ras-transformed cells.
Treatment of NIH/3T3-Ras cells with 20 mM NAC for 24 hours reduced the basal ROS to levels similar to those in parental NIH/3T3 (Fig. 4a). Moreover, pre-treatment of NIH/3T3-Ras cells with NAC completely inhibited the upregulation of ROS levels induced by HMG, reducing ROS to levels comparable to those found in untreated NIH/3T3-Ras cells (Fig. 4b). Pre-treatment of NIH/3T3-Ras cells with NAC also effectively inhibited the DNA fragmentation induced by HMG. DNA fragmentation was reduced from a thirteen-fold increase in cells unprotected by NAC treatment, to two-fold in cells pre-treated with NAC ( Fig. 4c). Culturing cells under a low oxygen environment is another method for inhibiting intracellular ROS generation (33).
NIH/3T3-Ras cells were cultured in either normoxic or hypoxic conditions for 24 hours before treatment with HMG. After PKC inhibition, cells were maintained in normoxic or hypoxic conditions for another 48 hours before analysis of DNA fragmentation by FACS. While To determine a mechanism by which PKC inhibition generates ROS, studies were performed using diphenylene iodonium (DPI), an inhibitor of NAD(P)H oxidase (35). Pretreatment of NIH/3T3-Ras cells with DPI did not inhibit the upregulation of ROS by HMG, suggesting the lack of a role for NAD(P)H oxidase in generating ROS downstream of PKC inhibition (data not shown). Since mitochondria play an important role in many forms of apoptosis and are also a major source for ROS production (36), we investigated the involvement of mitochondria as the source of ROS downstream of PKC inhibition. To explore the possibility that the inhibition of PKC generates ROS as a consequence of the opening of the mitochondrial permeability transition pore (PTP), studies were performed using cyclosporine A (CsA), an inhibitor of the mitochondrial PTP (37). Pretreatment of NIH/3T3-Ras cells with CsA did not affect the basal level of ROS in the cells, but effectively blocked the upregulation of ROS by HMG treatment (Fig. 6a). Moreover, the inhibition of HMG-induced ROS generation by CsA correlated with the inhibition of apoptosis. NIH/3T3-Ras cells pretreated with CsA exhibited nearly a ten-fold reduction in DNA fragmentation induced by HMG (Fig. 6b).
Previous results from our laboratory suggested that cell cycle progression may also be necessary for Ras-mediated apoptosis (7). Attempts to use the cell cycle-arresting agents hydroxyurea or aphidicolin to inhibit the cell cycle progression of Ras-transformed cells resulted in high levels of cytotoxicity. However, the cell cycle-arresting agents sodium butyrate and trichostatin A (TsA) were relatively nontoxic to NIH/3T3-Ras cells and were therefore tested for their ability to inhibit Ras-mediated apoptosis induced by PKC inhibition. Pre-treatment of NIH/3T3-Ras cells with either sodium butyrate (5 mM) (Fig. 7a) or TsA (100ng/ml) (Fig. 7b) reduced the induction of DNA fragmentation in cells by HMG treatment from over thirteen-fold to less than two-fold. Concurrent cell cycle analysis showed inhibition of cell cycle progression 12 and accumulation of cells in G 1 induced by both sodium butyrate and TsA treatments (Fig. 8a).
Treatment with either sodium butyrate or TsA reduced the percentage of cells in S phase by more than five-fold. In contrast, treatment with isobutyramide, an analog of sodium butyrate (38), had no effect on the cell cycle distribution of NIH/3T3-Ras cells (Fig. 8a) and also did not prevent induction of apoptosis by HMG 2 . Thus, inhibition of apoptosis by these compounds correlated with inhibition of cell cycle progression. Our laboratory has shown that continued Ras activity is necessary to mediate apoptosis induced by PKC inhibition, as inhibition of Ras activity with a farnesyltransferase inhibitor can inhibit apoptosis 3 . To rule out the possibility that the inhibition of apoptosis by hypoxic culture conditions, sodium butyrate treatment, or trichostatin A treatment were due to effects on Ras activity, Ras activity was analyzed by affinity-binding to a Raf Ras-binding-domain (RBD) peptide which only binds the activated (GTP-bound) form of

Ras. Stimulation of serum-starved NIH/3T3 cells with PDGF (30 ng/ml) resulted in an increase
in Ras activity as measured by the Raf-RBD-GST pulldown assay (Fig. 8b1). NIH/3T3-Ras cells constitutively expressed a relatively high level of activated Ras and culturing NIH/3T3-Ras cells under hypoxic conditions, or treatment with either sodium butyrate or TsA, had no inhibitory effect on the levels of Ras activity (Fig. 8b2).
Interestingly, despite the strong protection against apoptosis, pre-treatment of NIH/3T3-Ras cells with sodium butyrate also had no significant effect on either basal ROS levels, or the upregulation of ROS levels by HMG (Fig. 8c). These results, together with the protective effects of antioxidants agents like NAC, suggest that the observed increase in ROS levels after PKC inhibition are necessary for Ras-mediated apoptosis, but that increased ROS levels alone are not sufficient, and that enforced cell cycle progression may be required as well. To determine whether the transition between specific phases of cell cycle might be necessary for Ras-mediated Pre-treatment of NIH/3T3-Ras cells with nocodazole (400 ng/ml) did not prevent the DNA fragmentation induced by PKC inhibition (Fig. 9a), suggesting that transition through M phase is not necessary for Ras-mediated apoptosis. Previous reports have shown that nocodazole treatment can induce endoreduplication, the process in which cells bypass mitosis but continue to pass through the other cell cycle phases and replicate DNA (39), and we also observed a large fraction of NIH/3T3-Ras cells treated with nocodazole undergoing endoreduplication. FACS analysis showed that after 24 hours of nocodazole treatment, more than 80% of NIH/3T3-Ras cells were arrested in the M phase of cell cycle and contained a 4n quantity of DNA (Fig. 9b).
By 72 hours though, only 20% of the population remained at 4n, while 52% of the cells contained an 8n quantity of DNA, and thus had passed at least twice through S phase. We conclude from these experiments that mitosis is not necessary for Ras-mediated apoptosis induced by PKC inhibition, but the results using sodium butyrate or TsA suggest that transition through other phases of cell cycle may be important.

DISCUSSION
The concept of using oncogenically-mutated Ras as a molecular target for cancer therapy is an appealing one, but it has not been clear how Ras is able to induce apoptosis or sensitize cells to apoptosis in response to various apoptotic stimuli. We have shown that downregulation of PKC by chronic, high-dose PMA treatment selectively induces apoptosis in cells expressing oncogenic Ras. Since suppression of PKC activity by chronic high-dose PMA exposure is not an approach that can be translated into in vivo studies, we wished to establish a model for studying Ras-mediated apoptosis using a pharmacological inhibitor of PKC that could potentially be developed as a clinical therapeutic against human tumors containing oncogenically-mutated Ras. Our studies show that HMG can be used to inhibit cellular PKC activity, and to selectively induce apoptosis in Ras-transformed cells while sparing normal cells. Interestingly, HMG-PC was first described to have anti-neoplastic activity against the promyelocytic leukemia line HL-60, but not against the erythroblastic leukemia line K562 (44). HMG-PC has also been reported to be non-cytotoxic to either normal human neutrophils or skin fibroblasts (45). The mechanism for the differential cell line-specific cytotoxicity of HMG-PC observed in these previous studies is unclear. No correlation was found between cell-specific sensitivity to the effects of HMG-PC and either alkyl cleavage enzyme activity (as a mechanism for drug by guest on March 23, 2020 http://www.jbc.org/ Downloaded from activation) (46), or incorporation of drug into the plasma membrane and changes in membrane fluidity (47). However, HL-60 cells contain an activating mutation of N-ras while K562 cells do not contain any ras mutations (48). This raises the possibility that when HMG-PC accumulates in the plasma membrane and is metabolized into HMG (40), the resulting inhibition of PKC activity triggers cell death in HL-60 cells by a Ras-mediated mechanism. Furthermore, the ether phospholipid ET-18-OCH3 also inhibits PKC (49), and has been reported to both increase ROS levels (50) and induce apoptosis in HL-60 cells (51), consistent with our studies showing that Ras-mediated apoptosis in response to PKC inhibition by HMG is accompanied by ROS generation.

The source of the increased basal and inducible H 2 O 2 levels we observed in Rastransformed cells has not yet been defined. Our findings that the intracellular peroxide levels in
Ras-transformed mouse fibroblasts are higher than in normal fibroblasts are consistent with the report by Irani, et al. that Ras-transformed NIH/3T3 cells produced larger amounts of superoxide than normal NIH/3T3 under basal conditions (52). This superoxide production is thought to be mediated by a Rac-dependent activation of NADPH oxidase localized to the cellular membrane in non-phagocytic cells (52)(53)(54) (57). Elevated ROS levels may contribute to a transformed phenotype by increasing proliferation in a Raf-MAPK independent manner (52) and inducing the transformed morphology by Rac-dependent changes in the actin cytoskeleton (58,59). Moreover, ROS may contribute to the oncogenic process by damaging DNA, increasing genetic instability, and causing the loss of tumor suppressors (60,61). Indeed, our laboratory has shown that Rastransformed fibroblasts demonstrate increased genetic instability and decreased p53 expression (62). Finally, the induction of senescence by Ras was shown to be dependent on the generation of H 2 O 2 , as treatment with NAC or culturing under low oxygen rescues these cells from Rasinduced senescence (55).
We have shown that the inhibition of PKC can cause an increase in intracellular peroxide levels in both normal and Ras-transformed cells, with much greater ROS induction in Rastransformed cells. The level of ROS induced by PKC inhibition in normal cells was not sufficient to induce apoptosis, while the higher level of ROS production induced by PKC inhibition in Rastransformed cells was necessary for apoptosis, as inhibition of intracellular H 2 O 2 generation by NAC treatment or low-oxygen culture conditions inhibited apoptosis. This finding is significant, as ROS generation is not necessary for many forms of apoptosis (63). The mechanism by which PKC inhibition leads to increased ROS generation is unclear. We have previously reported that Bcl-2 can inhibit Ras-mediated apoptosis (8), and that the ability of Bcl-2 to inhibit Rasmediated apoptosis depends on its phosphorylation (64). Others have reported that phosphorylation of Bcl-2 by mitochondrial PKCα may be necessary for suppression of apoptosis (65). Many mechanisms for how Bcl-2 inhibits apoptosis have been reported, one mechanism being the prevention of ROS generation (66), and another being the prevention of the opening of the mitochondrial permeability transition pore (67). Our results with CsA suggest that the opening of the mitochondrial permeability transition pore is responsible for ROS generation downstream of PKC inhibition, potentially due to reduced Bcl-2 phosphorylation and function at the mitochondria. The cell cycle is thought to have an interdependent connection with the cellular apoptotic program as a mechanism to restrict uncontrolled cell growth and oncogenesis (70,71). Though the connection between cell cycle and apoptosis is not well understood, the cell death program appears to be induced when growth regulatory signals are inappropriate in amplitude or temporal sequence, or otherwise conflicting. We have shown here and in previous studies that suppression of PKC activity induces a reversible G 1 arrest in normal cells. Therefore, we believe apoptosis is triggered when a strong growth-stimulating signal, such as oncogenic Ras activity, induces cell cycle progression while the cells are simultaneously receiving a growth-inhibitory signal such as the suppression of PKC activity. Agents that were effective in inducing cell cycle arrest in Rastransformed cells, such as sodium butyrate and TsA, protected those Ras-transformed cells from apoptosis upon PKC inhibition. In contrast, nocodazole, which blocked mitosis but still allowed cells to re-enter S phase, was ineffective at preventing Ras-induced apoptosis, suggesting that progression through S phase may be the critical cell cycle event in initiating apoptosis. The mechanism for the induction of apoptosis by Ras appears to be p53-independent (15,62,68), a finding of potential clinical importance as p53 is frequently lost during tumorigenesis (72).