Introduction

Top Summary Introduction Materials and Methods Results Discussion References

The reversible phosphorylation of proteins, catalyzed by protein kinases and protein phosphatases (PP), is the key mechanism for the regulation of diverse cellular functions (Cohen, 1997). The serine/threonine PPs are encoded by the PPP and PPM gene families. The PPP family includes PP1, PP2A, and PP2B (calcineurin and Ca²⁺-calmodulin-regulated PP), whereas the PPM family consists of Mg²⁺-dependent PPs, which include PP2C and pyruvate dehydrogenase phosphatase. The catalytic subunit of PP1 (PP1c) is a 37-kDa protein that forms heterodimers with several regulatory units. PP2A is a family of trimeric holoenzymes composed of a 36-kDa catalytic subunit (PP2Ac) bound to a constant regulatory subunit of 65 kDa (PR654) and one of the number of variable subunits (52-130 kDa). The PPP family of PPs are involved in many cellular processes including glycogen metabolism, calcium transport, muscle contraction, protein synthesis, regulation of cell cycle, and RNA splicing (Cohen, 1997; Fischer, 1997). A number of naturally occurring agents such as okadaic acid (OA), microcystin, tautomycin, and calyculin A (Fujiki and Suganuma, 1993; Cohen, 1997), have been reported to inhibit the members of the PPP family (Table 1).

View larger version (36K): [in this window] [in a new window]   Fig. 1.   Effects of various concentrations of OA on the induction of apoptosis in NIH3T3 parental, -neo, and -ras cell lines. Cell lines were treated with either the vehicle DMSO, or with indicated concentration of OA dissolved in DMSO. Percentage of cells undergoing apoptosis was determined at 24 h (B) and 48 h (C) after treatment by flow cytometry after staining of PI-stained cells. Each point is an average of duplicate dishes (variation <5%). A, DNA histograms of cell lines treated with 20 and 40 nM OA for 48 h revealing the presence of the sub-G0-G1 peak and also the various phases of the cell cycle.

View larger version (99K): [in this window] [in a new window]   Fig. 2.   OA-induced apoptosis of NIH3T3 cell lines. A, C, and E are phase-contrast photomicrographs of DMSO-treated NIH3T3 parental, -ras, and -neo cell lines, respectively; whereas B, D, and F reveal the typical morphology of NIH3T3 parental, -ras, and -neo cell lines after treatment with 20 nM OA after 48 h. G, DNA fragmentation analysis at 48 h after treatment of NIH3T3 parental, -ras, and -neo cells with the indicated concentrations of OA.

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Correlation between OA-induced loss of cell viability and apoptosis. At 48 h after treatment of NIH3T3-ras and -neo cells with the indicated doses of OA, cell viability and apoptosis (sub-G0 cells) were determined by MTT assay and flow cytometric analysis, respectively. Each apoptosis value represents means ± S.E. from three separate dishes, whereas each viability value is an average of six readings from six wells.

View larger version (24K): [in this window] [in a new window]   Fig. 4.   Comparison of OA and TPA for their ability to affect cell viability and induce apoptosis in NIH3T3-ras cell line. NIH3T3-ras cells were treated with the tumor promoter TPA or OA (20-100 nM) for 48 h, and percentage of viable cells was assessed by MTT assay and a parallel experiment was set up to determine the sub-G0 population by flow cytometric analysis. Each point in the graph represents average values from duplicate plates (variation <5%) for flow cytometric analysis and means of six replicates for the MTT assay.

OA is the most commonly used inhibitor of PPs 1 and 2A. OA, a toxic polyether compound of a C₃₈ fatty acid, was first isolated from two sponges, Halicondria okadai and H. melanodocia (Fujiki and Suganuma, 1993). OA is a mouse skin tumor promoter that stimulates gene expression (Rosenberger and Bowden, 1996) and has the ability to induce a pseudomitotic effect. OA has also been shown to induce apoptosis in several cell lines such as human breast cancer (Kiguchi et al., 1994), retinoblastoma (Inomata et al. 1995), osteoblastic (Morimoto et al., 1997), rat kidney epithelial (Davis et al. 1996), neuroblastoma, and rat pituitary adenoma cell lines (Tergau et al., 1997). However, it is unclear whether OA-induced apoptosis is correlated to a common genetic abnormality present in various cell types.

View larger version (19K): [in this window] [in a new window]   Fig. 5.   Induction of apoptosis in NIH3T3-ras by PP inhibitors. NIH3T3-ras cells were treated with the indicated dose of the phosphatase inhibitors (calyculin, dinophysitoxin, OTA, and OA) and percentage of sub-G0 cells were determined by flow cytometry after staining the cells with PI. Each point is an average of determinations from duplicate plates (variation <5%).

View larger version (19K): [in this window] [in a new window]   Fig. 6.   OA-induced cell death in an isogenic pair of human UEPCs. E/J-ras was transfected into SVHUC to yield SVHUC-ras cells. Cells were treated with either the vehicle, DMSO, or the indicated concentrations of OA in DMSO for 24 h. Percentage of cells undergoing apoptosis were analyzed by flow cytometry after staining with PI. Each point represents apoptosis represents means ± S.E from three separate dishes.

View larger version (59K): [in this window] [in a new window]   Fig. 7.   OA-induced apoptosis of SVHUC and SVHUC-ras cell lines as determined by dual-colored (PI and Annexin V) flow cytometry. Shown are the cytograms of annexin V binding (ordinate) versus PI uptake (abscissa) in DMSO- and OA-treated cells. Cell lines were treated with either the vehicle, DMSO (A and B)], or with 20 nM OA (C and D) for 24 h. The cells were harvested by trypsinization, washed with PBS, and subsequently stained with antibody to annexin V conjugated to FITC and with PI (10 µg/ml). Viable [annexin V/PI], preapoptotic [annexin V+/PI], apoptotic [annexin V+/PI+], and the residual damaged [annexin V/PI+] cells are shown in the respective quadrants. A, annexin V; P, PI.

View larger version (18K): [in this window] [in a new window]   Fig. 8.   Effect of OA-dose and time of treatment on the induction of apoptosis in T24 cells. T24 cells were treated with either the vehicle, DMSO, or with varying doses of OA (10-100 nM). Cells were harvested at 12, 24, and 48 h after treatment and were stained with PI to determine percentage of cells undergoing apoptosis. Each point in the graph is an average of duplicate dishes (variation <5%).

View larger version (39K): [in this window] [in a new window]   Fig. 9.   OA-induced apoptosis in cell lines with or without mutated ras. Semiconfluent cultures of all the indicated cell lines were exposed to 20 nM OA in DMSO for a period of 24 h except PC-3 and LNCaP, SK-LU-1, and Calu-6 cells, which were treated for 48 h, after which cells were harvested, fixed in ethanol, and then stained with PI. Percentage of cells undergoing apoptosis was analyzed by flow cytometry. Each value is mean ± S.E. of determinations from three separate plates for determining apoptosis. Similar results were observed in a repeat experiment. Arrow points to the cell lines with mutations in ras oncogene.

View larger version (28K): [in this window] [in a new window]   Fig. 10.   Analysis of cell cycle markers and DNA content by flow cytometry in OA-treated SVHUC and SVHUC-ras cells. Human UEPCs with (SVHUC-ras) and without (SVHUC) mutated H-ras, were treated with either DMSO or 20 nM OA for 24 h. Cells were fixed, permeabilized, incubated with the indicated anticyclin antibodies, and stained using a secondary antibody conjugated to FITC. Cells were also stained with PI for DNA content and analyzed on a FACScan flow cytometer.

View larger version (33K): [in this window] [in a new window]   Fig. 11.   Basal PP1 and PP2A activities in various cell lines. Each cell extract was prepared from untreated subconfluent cultures. Cell extract was treated either with 2 or 20 nM OA in vitro to distinguish between PP1 and PP2A activities. Phosphatase activity was measured using 32P-labeled phosphorylase a as a substrate as described in Materials and Methods.

View larger version (37K): [in this window] [in a new window]   Fig. 12.   PP1 and PP2A activities of SVHUC, SVHUC-ras, T24, and DLD-1 cells treated with OA in vivo. Indicated cell lines were treated with 20 nM OA for 24 h and the level of PP1 and PP2A activities were measured as described in Materials and Methods.

The mutations in the ras genes are among the most widely observed genetic changes during tumorigenesis in humans (Bos, 1989; Vojtek and Der, 1998). Approximately 30 to 40% of all human cancers elicit mutations in the ras allele (Bos, 1989). Ras, a 21-kDa membrane-associated GTP-binding protein, regulates diverse extracellular signals controlling cellular processes (Vojtek and Der, 1998; Marshall, 1995). The mammalian ras gene family consists of four homologous members (H-ras, K-ras, N-ras, and R-ras) and each encodes a 21-kDa protein (McCormick, 1995). A series of proteins involving protein-protein interactions as upstream molecules are responsible for the control of ras activation. Ras activation, for its downstream signaling, involves its association with the plasma membrane. The association of ras with the outer plasma membrane is mediated by a 15-carbon farnesyl group that is covalently linked to ras (Lerner et al., 1997). The plasma membrane location of ras switches ras protein from the inactive GDP-bound to the active GTP-bound state. Mutations in ras genes localize in the N-terminal region controlling GTP-binding. Mutations lock ras in the GTP-bound state. Guanosine nucleotide exchange factors promote the formation of active GTP-bound forms of ras, whereas GTPase activating proteins promote the formation of the inactive GDP-bound form of ras. The mammalian ras gene family acquires oncogenic activity by somatic mutations predominantly localized to amino acids 12, 13, and 61 (Vojtek and Der, 1998). The role of ras oncogenes in tumorigenesis has been well documented in various animal models (Stanely, 1995). Mutations in one of the three ras genes has been associated with specific human cancers. Transitional cell carcinoma of human urinary bladder have predominately H-ras mutations, whereas 95% of pancreatic adenocarcinomas, and other human tumors including colon cancers contain mutation in the K-ras gene. N-ras mutations have been observed in some cases of acute myeloid leukemia (Boss, 1989).

We determined whether chronic activation of ras due to mutations contributes to the cell's sensitivity to the induction of apoptosis by OA. We present data indicating that 1) cell lines that have ras mutation (H or K), irrespective of p53 status (von Kleist et al., 1975; Taparowsky et al., 1982; Strickland et al., 1988; Shirasawa et al., 1993; Pergolizzi et al., 1993; Aoki et al., 1997), can be induced to undergo apoptosis by treatment with the inhibitors of PP1 and PP2A; 2) the H-ras mutation up-regulated the levels of PP2A activity; and 3) OA-induced apoptosis accompanied inhibition of ras-induced levels of PP2A activity.

	Materials and Methods

Top Summary Introduction Materials and Methods Results Discussion References

Materials

OA was purchased from Alexis Laboratories (San Diego, CA). 12-O-Tetradecanoylphorbol-13-acetate (TPA) was purchased from Life Systems (Newton, MA). All other chemicals, including propidium iodide (PI), RNase H, dimethyl sulfoxide (DMSO), and 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium (MTT) were purchased from Sigma Chemical Co. (St. Louis, MO). Phosphatase kit was purchased from Life Technologies/BRL (Gaithersburg, MD). Annexin V-staining kit was purchased from Clonetech (Palo Alto, CA). Anticyclin D1 was obtained from Santa Cruz Biotechnology, (Santa Cruz, CA). Anticyclin B1 was from Pharmingen (San Diego, CA).

Cell Lines

NIH3T3 cells, stably transfected with either E/J-ras or the neoexpression vector, were obtained from Dr. Terry Oberley, Department of Pathology, Veterans Administration Hospital, University of Wisconsin (Madison, WI). Human uroepithelial cell (UEPC) lines (SVHUC and SVHUC-ras) were obtained from Dr. Catherine Reznikoff, Department of Human Oncology, University of Wisconsin (Madison, WI). Human pancreatic ductal adenocarcinomas cell lines [BxPC3 and MIAPaCa2; originally from the American Type Culture Collection (Rockville, MD)] were obtained from Dr. Pamela L. Crowell, Department of Biology, Indiana University-Purdue University at Indianapolis (Indianapolis, IN). BxPC3 cells were cultured in RPMI 1640 containing 10% fetal bovine serum (FBS). The MIAPaCa2 cells were grown in Dulbecco's modified Eagles' medium (DMEM) containing 10% FBS. Lung cancer cell lines (Calu-6 and SKLU-1) were obtained from Dr. Joan Schiller, University of Wisconsin (Madison, WI) and were grown in MEM medium supplemented with 8% FBS. Colon carcinoma cell lines DLD-1 and COLO 201 were maintained in RPMI 1640 supplemented with 10 and 20% FBS, respectively, whereas HCT116 and HT29 cell lines were maintained in DMEM medium supplemented with 10% FBS. Human prostate cancer cell line PC3 was cultured in RPMI 1640 medium supplemented with 5% FBS. T24 cells were maintained in MacCoys 5A medium supplemented with 10% FBS. The LNCaP and NIH3T3 cells were maintained in DMEM supplemented with 5% FBS. SVHUC and SVHUC-ras and mouse keratinocytes cell lines 308 and C50 ]provided by Dr. Jill Pelling, University of Kansas Medical Center (Kansas City, KS)], were cultured in Ham's F12 medium supplemented with 250 U/ml insulin, 5 µg/ml Transferrin, 15 mM dextrose, 0.1 mM nonessential amino acids, 1 µg/ml hydrocortisone, 2 mM L-glutamine, 100 U/ml penicillin, 100 µg/ml streptomycin, and 1% FBS. Mouse epidermal cell line JB6 ]provided by Dr. Nancy M. Colburn, Cell Biology Section, Laboratory of Viral Carcinogenesis, National Cancer Institute (Frederick, MD)] was maintained in MEM containing 5% FBS, whereas the normal human kidney cell line HS 715.K was grown in the above medium with 10% FBS.

Cell Treatments

The cells were allowed to grow in their respective media until they were 70% confluent. Stock solution of OA was prepared in ethanol and was further diluted in DMSO. Cells were treated with various concentrations of OA. Control dishes were treated with the vehicle DMSO (final concentration: <0.1%).

Cell Viability Assay

The viability of the cells were determined using the MTT assay (Mosmann, 1983).

Apoptosis Assays

The cells undergoing apoptosis were identified and quantitated as follows.

Phase-Contrast Microscopy. The cells exhibiting characteristic morphological features associated with apoptosis were viewed under a microscope with a 40× objective and then photographed.

Flow Cytometric Analysis. At appropriate times after treatment, the cells were harvested by trypsin-EDTA treatment, washed twice with ice-cold PBS and subsequently stained with antibody to Annexin V-fluorescein isothiocyanate (FITC) and PI. This dual staining technique revealed the viable (annexin V/PI], preapoptotic [annexin V⁺/PI], apoptotic [annexin V⁺/PI⁺], and damaged [annexin V/PI⁺] cells (Darzynkiewicz et al., 1996). For analysis of cell cycle markers, the harvested cell pellet was permeabilized using 0.25% Triton-X-100 solution in PBS for 5 min at 4°C. The cell pellet was washed and stained with anticyclin D1 and anticyclin B1 antibodies followed by incubation with the secondary antibody conjugated to FITC. The cell pellet was washed and stained with staining solution containing PI (5 µg/ml) and RNase H (200 µg/ml). This staining technique allowed the immunochemical detection of cyclins in relation to cell cycle position (DNA content) by mutiparameter flow cytometric analysis. The percentage of cells undergoing apoptosis were also quantitated by the method described by Darzynkiewicz et al. (1994). The cell pellet obtained after harvesting the treated and untreated cells was resuspended in ice-cold PBS and fixed in 95% ethanol. The fixed cells were pelleted, resuspended in a PI staining solution, filtered through a 40-µM pore nylon mesh, and analyzed using a Becton-Dickinson FACStar plus flow cytometer with excitation 488 nM. The DNA histograms elicited the sub-G₀-G₁ cells as the apoptotic peak in the various phases of the cell cycle.

DNA Fragmentation. To detect the DNA ladder typical of the apoptotic cells, DNA fragmentation analysis was performed. At approximate times after treatment, the genomic DNA was isolated (Gentra Systems, Minneapolis, MN) and subjected to agarose gel electrophoresis (1.8%). The DNA in the gel was visualized with the Fluoroimager SI (Molecular Dynamics, Sunnyvale, CA) after staining for 1 h with SYBR Green I nucleic acid gel stain (Molecular Probes, Eugene, OR).

PP Assay. PP activity was assayed using ³²P-labeled phosphorylase a as a substrate which detects both PP1 and PP2A activities. We used the PP kit supplied by Life Technologies/BRL. At appropriate times after treatment, the cell pellet was homogenized in the extraction buffer containing 20 mM Tris-HCI, pH 7.5, 5 mM EDTA, 10 mM EGTA, 15 mM -mercaptoethanol, 0.25 M sucrose, 0.3% Triton X-100, 5 µg/ml leupeptin, and 5 µg/ml aprotinin and centrifuged (40,000g) to give a soluble supernatant. The PP activity in the clear supernatant was determined by measuring the trichloroacetic acid-soluble counts released after incubation of the ³²P-labeled phosphorylase a in the cell extract. The PP activity was linear up to assay times of 10 min and 5 µg protein of the cell extract. Routinely, incubation for PP activity was carried out for 10 min with an extract containing 5 µg of protein as determined by the Bio-Rad assay (Bio-Rad, Hercules, CA).

	Results

Top Summary Introduction Materials and Methods Results Discussion References

Ras Mutation Imparts Sensitivity to Cells for Induction of Apoptosis by OA. To determine whether ras mutation is the key genetic event that is associated with the susceptibility of cells to undergo apoptosis in response to treatment with the PP inhibitor OA, we used genetically identical cell lines that differed only in the expression of mutated ras. In our initial experiments, we used NIH3T3 and human UEPC stably transfected with either pSV2-neo or PREJ ras (mutated c-H-ras with 12th codon mutation).

OA Induces Apoptosis in a Variety of Cell Types with Mutations in Either H- or K-ras. To further obtain evidence that ras mutations determine a cell's sensitivity to OA-induced apoptosis, we used several established cell lines with or without ras mutations. As shown in Fig. 9, we observed a distinct difference in the percentage of cells undergoing apoptosis between cell lines with and without mutations in the ras oncogene. Cell lines harboring mutations in either H-ras [e.g., T24 (Taparowsky et al., 1982) and 308 (Strickland et al., 1988) cells or in K-ras (DLD1 and HCT116; Shirasawa et al., 1993); PC3 and LNCaP (Pergolizzi et al., 1993); Calu-6 and SKLU-1 (Lehman et al., 1991); and MIAPaCa2 (Aoki et al., 1997)] were more sensitive to OA-induced apoptosis (Fig. 9) than cell lines with wild-type ras oncogene [mouse epidermal cell lines C50, and JB6 (Sun et al., 1993); murine fibroblast cell line NIH3T3, human kidney epithelial cell line Hs715.K, human UEPC line SVHUC human pancreatic cancer cell line BxPC3 (Aoki et al., 1997); and human colon cancer cell line HT29 (von Kleist et al., 1975)].

OA-Induced Apoptosis Is Independent of p53 Status. The tumor suppressor p53, a nuclear phosphoprotein, plays a key role in the induction of apoptosis. Post-translational modification by phosphorylation of p53, mediated by several kinases, is essential for its transcriptional activity (Mowat, 1998). p53 is dephosphorylated by PP1 and PP2A. The findings that OA has been shown to induce hyperphosphorylation of p53 (Yatsunami et al., 1993) prompted us to determine whether OA-induced apoptosis involves a p53-dependent pathway. In these experiments, we determined the ability of OA to induce apoptosis in various cell lines with altered p53 status. At first we determined the effects of OA on the induction of apoptosis in SVHUC cells. SV40 encodes two proteins, large "T" antigen and "t" antigen. The large "T" antigen complexes with p53 and pRb family of proteins (Kao et al., 1993). As shown in Fig. 6, OA-induced apoptosis in SVHUC in the absence of active p53 due to complex formulation with the large T antigen. We also found that cell lines with or without mutations in p53 were susceptible to OA-mediated apoptosis (Fig. 9). Cell lines sensitive to OA-induced apoptosis included the prostate cancer lines PC3 and LNCaP. PC3 is androgen independent and is defective for both its alleles for the p53 gene, whereas LNCaP is androgen sensitive and expresses the wild-type p53 gene (Israel et al., 1995). Human bladder transitional carcinoma T24 cells carry an inframe deletion of tyrosine 126 in the p53 gene (Cooper et al., 1994). The colon cancer cell line HCT 116 expresses wild-type p53 gene (Take et al., 1996), whereas DLD-1 cells harbor a mutation at position 241 (resulting a change from serine to phenylalanine) in p53. The lung cancer cell line Calu-6 has mutations at positions 196 and 213 whereas SKLU-1 cells exhibits mutation at position 193 in the p53 gene. These results imply that OA-induced apoptosis may involve a p53-independent pathway.

OA-Induced Apoptosis Accompanies Changes in Cell Cycle Progression. Because OA has been reported to induce the hyperphosphorylation of cell cycle regulator proteins pRb and p53 (Yatsunami et al., 1993), we determined whether OA-induced apoptosis accompanies alterations in the cell cycle profile. Asynchronously growing NIH3T3 parental, NIH3T3-ras, and NIH3T3-neo cells were treated with 20 nM OA for 48 h, whereas SVHUC and SVHUC-ras cells were treated with 20 nM OA for 24 h. Both cell lines were stained with PI to determine the percentage of cells in various phases of the cell cycle. The NIH3T3-ras and SVHUC-ras cells treated with OA for 48 h displayed a decrease in the G₀-G₁ phase of the cell cycle accompanied with an increase in the sub-G₂-M phase of the cell cycle as well as in the sub-G₀-G₁ phase of the cell cycle. The parental NIH3T3, the neotransfected NIH3T3 cells and SVHUC cells did not reveal a distinct G₂-M block in their cell cycle progression (Table 3).

H-Ras Mutation Up-Regulates PP2A Activity. A possibility was explored to determine whether ras mutations result in alterations in the basal PP1 and PP2A activities. In this experiment (Fig. 11), the indicated cell lines were grown to subconfluency and then were used to assay PP activity. PP2A and PP1 activities were distinguished by their sensitivity to 2 and 20 nM OA, respectively (Huang et al., 1995; Murata et al., 1996). The PP1 and PP2A activities were detected, although at different levels, in all the cell lines evaluated. A novel observation is the finding that H-ras mutation leads to an increase in the PP2A activity. As shown in Fig. 11, NIH3T3 and human UEPCs after transfection with the mutated H-ras displayed an increase in PP2A activity. It is noteworthy that SVHUC, which are immortalized after transfection with the wild-type SV40, which has both large T and small t antigens (Pratt et al., 1992), had reduced or no PP2A activity. This is because of the fact that small t antigen specifically inhibits PP2A activity (Yang et al., 1991; Howe et al., 1998). As shown in Fig. 10, SVHUC-ras up-regulated PP2A activity. Similar results were observed in other H-ras transfected cell lines (Fig. 11).

	Discussion

Top Summary Introduction Materials and Methods Results Discussion References

Our results that show that the cell lines that harbor ras mutations are more sensitive than cell lines with wild-type ras to OA-induced apoptosis, were unexpected. OA is a mouse skin tumor promoter (Fujiki and Suganuma, 1993). A majority of the mouse skin papillomas and carcinomas, elicited by initiation with 7,12-dimethylbenz[a]anthracene and by promotion with TPA, contain a point mutation (A to T) in the Ha-ras oncogene at condon 61 (Pelling et al., 1988; Fujiki et al., 1989; Kim et al., 1997). Similarly, Fujiki et al. (1989) reported similar codon 61 mutations in the Ha-ras gene in mouse skin tumors developed by initiation with 7,12-dimethylbenz[a]anthracene and promotion with OA class tumor promoters (OA, dinophysistoxin-1, and calyculin A) and TPA-type tumor promoters (teleocidin and aplysiatoxin). These results led to the current hypothesis that tumor promoters selectively cause the clonal expansion of a population of initiated cells, which harbor Ha-ras mutations, to form skin papillomas and carcinomas. Our cell lines, which express Ha-ras mutations, were insensitive to TPA for the induction of apoptosis (Fig. 4). However, the question of how OA, which selectively may expand cells with Ha-ras mutations in intact mouse skin in vivo, induces apoptosis in vitro of cell lines that contain ras mutations, is critical to the interpretation of present results (Hennings et al., 1992; Fujiki and Suganuma, 1993). The role of ras mutations in mouse skin tumor formation by the inhibitors of PP 1 and 2A needs reevaluation. We hypothesize that the initiated cells, which escape OA-induced apoptosis and proliferate to form papillomas and carcinomas, have other genetic lesions in addition to Ha-ras mutations. All epidermal cells, which have only Ha-ras mutations, undergo apoptosis during tumor promotion with OA. We now present results indicating that the proliferative signals of ras-activated cells can be diverted toward apoptosis by inhibition of PP1 and PP2A by OA.

OA-induced cell death was due to apoptosis and not the results of necrosis as confirmed by several criteria such as morphological characteristics, analysis of preapoptotic cells entering apoptosis by dual-colored (PI and Annexin V) flow cytometry, and DNA fragmentation by agarose gel electrophoresis (Figs. 1-3 and 6-8). OA-induced apoptosis appears to involve mitochondrial changes and DNA fragmentation (Fig. 2). It is clear from the results, using isogenic cell lines, that ras mutation imparts sensitivity to cells to OA-induced apoptosis (Figs. 1-3 and 6 and7). To obtain additional evidence to support the conclusion that ras mutation imparts sensitivity to cells to OA-induced apoptosis, we screened several established cell lines harboring mutations in either H-ras (T24 and 308) and K-ras genes (DLD-1, HCT116, Calu-6, SK-LU-1, and MIAPaCa2). The results illustrated in Fig. 9 clearly indicate that the sensitivity of several murine and human cell lines to OA-induced apoptosis correlated with the presence of mutation in the ras oncogene H or K. Because the concentrations of OA used in the present study inhibited both PP1 and PP2A (Figs. 11 and 12), the onset of apoptosis appears to link to the sustained phosphorylation of PP1 and PP2A substrates.

The role of functional p53 in determining cell's sensitivity to OA-induced apoptosis has been inconsistent. In this context the reports by Sheikh et al. (1996), and Boe et al. (1991) are noteworthy. They observed the induction of apoptosis by OA in breast cancer cell lines with disrupted p53 function. However, others have proposed both p53-independent and p53-dependent pathways in OA-induced apoptosis in NIH3T3 (p53/) (Yan et al., 1997) and rabbit lens epithelial cells (Li et al., 1998). Interestingly, we found that cell lines with or without mutations in p53 were equally susceptible to OA-induced apoptosis (Figs. 6 and 9), implying that OA-induced G₂-M arrest and apoptosis involves p53-independent pathways.

The mechanisms by which the expression of H-ras up-regulates PP2A activity (Fig. 11) are unknown. The native form of PP2A consists of A and C subunits that are considered to form the core of the enzyme. The C unit binds stably to the carboxyl terminal region of the A subunit. The B subunit binds to the amino terminal region of the A subunit. The B family has three members (B, B, and B) (~55 kDa). The B' family consists of several isoforms (~54-68 kDa), whereas the B" family has two members (72 and 130 kDa). The combination of B subunits to the core PP2A heterodimeric complex gives rise to the several PP2A variants that exhibit substrate specificity. The B subunits act as targeting subunits that direct PP2A to specific subcellular locations (Strack et al., 1998). Ras-induced up-regulation of PP2A activity may be the result of increased synthesis of PP2A subunits (A, C, and B) due to increased transcription or translation. PP2A activity is also regulated by post-translational modification (methylation) of the catalytic subunit of PP2A (Turowski et al., 1995; Kloeker et al., 1997). Ras activation may up-regulate PP2A activity by increasing the production of ceramide catalyzed by sphingomyelinase. The ceramide 2 has been shown to activate PP2A activity (Galadari et al., 1998).

The cell lines, treated in vivo with OA, inhibited both PP1 and PP2A activity (Fig. 12). The mechanisms by which inhibition of PP1 and PP2A activity lead to the induction of apoptosis in ras-activated cells remains to be defined. OA-induced apoptosis preceded arrests cells in the progression of G₂-M phase (Table 3). Entry into mitosis is triggered by the maturation promoting factor, which is composed of cdk2 and cyclin B. In early G₂, the human cdc2 protein is phosphorylated on tyrosine 15 and threonine 14 by Weel and Mik1 to inhibit its activation. G₂ to M progression requires activation of cdc2 by dephosphorylation of cdc2 by the dual specificity PP cdc25 (Draetta and Eckstein, 1997). Cdc25C activity in G₂ phase is negatively regulated by PP2A. In human cells, there are three cdk activating phosphatases (cdc25A, cdc25B, and cdc25C). Cdc25A is expressed early in the G₁ phase of the cell cycle, whereas cdc25B and cdc25C are expressed at the G₂-M boundary (Draetta and Eckstein, 1997). It is likely that OA-induced G₂-M arrest may be the result of inhibition of cyclin B1 accumulation by OA (Table 4). However, the roles of other cell cycle regulatory proteins such as other cyclins and cyclin-dependent kinase inhibitors in OA-induced G₂-M arrest remain to be explored.

OA treatment leads to the production of tumor necrosis factor- (TNF) in BALB/c 3T3 cells (Komori et al., 1993). TNF is a cytokine (molecular mass 17 kDa) and is produced mainly by monocytes and/or macrophages. TNF is implicated in various biological processes such as immunoregulation, inflammation, cachexia and mitogenesis. OA and TNF have been shown to share similar effects such as the induction of protein phosphorylation and the expression of early response genes. TNF, like OA, was found to be a tumor promoter (Guy et al., 1992; Fujiki et al., 1997). We also found that OA-induced apoptosis of human cell lines accompanied release of TNF in the medium (D.R. and A.K.V., unpublished observations). The role of TNF in OA-induced apoptosis is currently being investigated.

In summary, we have shown, using isogenic cell pairs (Figs. 1-3, 6, and 7) and other established cell lines (Fig. 9), that OA and other inhibitors of PP1 and PP2A (e.g., calyculin, tautomycin, dinophysitoxin, and microcystin) induce apoptosis in cell lines that harbor either H- or K-ras mutations. OA-mediated apoptosis does not appear to require the presence of wild-type p53 (Fig. 9). An overall survey of the basal levels of PP1 and PP2A indicates the presence of both enzymes at different levels in different cell lines (Fig. 11). However, H-ras mutation appears to affect the level of PP2A activity (Fig. 11). The results presented indicated that the cell survival signals imparted by the oncogenic ras can be diverted to cell death signals by inhibition of PP1 and PP2A activities. Thus, the PP1 and PP2A and their associated signaling molecules (e.g., hyperphosphorylated proteins and genes) may be potential targets for human cancer treatments.