Cellular transformation is a complex process involving genetic alterations associated with multiple signaling pathways. Development of a transformation model using defined genetic elements has provided an opportunity to elucidate the role of oncogenes and tumor suppressor genes in the initiation and development of ovarian cancer. To study the cellular and molecular mechanisms of Ras-mediated oncogenic transformation of ovarian epithelial cells, we used a proteomic approach involving two-dimensional electrophoresis and mass spectrometry to profile two ovarian epithelial cell lines, one immortalized with SV40 T/t antigens and the human catalytic subunit of telomerase and the other transformed with an additional oncogenic rasV12 allele. Of ∼2200 observed protein spots, we have identified >30 protein targets that showed significant changes between the immortalized and transformed cell lines using peptide mass fingerprinting. Among these identified targets, one most notable group of proteins altered significantly consists of enzymes involved in cellular redox balance. Detailed analysis of these protein targets suggests that activation of Ras-signaling pathways increases the threshold of reactive oxidative species (ROS) tolerance by up-regulating the overall antioxidant capacity of cells, especially in mitochondria. This enhanced antioxidant capacity protects the transformed cells from high levels of ROS associated with the uncontrolled growth potential of tumor cells. It is conceivable that an enhanced antioxidation capability may constitute a common mechanism for tumor cells to evade apoptosis induced by oxidative stresses at high ROS levels.
Oncogenic transformation involves multiple biochemical and cellular changes, which include self-sufficient production of growth signals, suppression of antigrowth signals, unlimited replication potential, evasion of apoptosis, and capabilities of tissue invasion and angiogenic nourishment of a tumor mass (1) . These seemingly divergent processes can be propagated by changes in a few discrete genes in human cells. This has been demonstrated through the induction of malignant transformation in human epithelia and fibroblast cells by SV40 T/t antigens, human telomerase (hTERT), and oncogenic rasV12 (2) . Since then, oncogenic transformation using defined genetic elements has been achieved in primary human cells from several different tissue origins (3, 4, 5, 6) . The ability to generate human cancer models using defined genetic elements not only suggests that certain genes play a critical role in tumorigenesis but also provides a powerful tool for elucidating the molecular mechanism of oncogenic transformation at different organ sites.
Ovarian carcinoma is a major cancer burden among women, accounting for 5% of all female cancer deaths in the United States (American Cancer Society Statistics for 2003). Although the exact etiology of ovarian carcinoma is not clear at present, extensive studies clearly suggest the involvement of various genetic alterations (7) . Two important tumor suppressor genes, p53 and the retinoblastoma (Rb), are inactivated frequently in human ovarian cancers. Mutation of p53 is observed in >50% of advanced and early stage ovarian carcinomas (8) . Whereas allelic loss at the Rb locus has been shown in only 30% of ovarian carcinomas (9) , detailed analysis of signaling molecules up-stream of Rb suggests that at least 80% of ovarian tumors exhibit disruptions in the general pathway of p16-CDK4/cyclin D1-pRb (10) . Activation of the catalytic subunit of hTERT is observed in all ovarian carcinomas, whereas telomerase activity is rarely found in benign ovarian tumors (11 , 12) , suggesting that hTERT plays an essential role in ovarian carcinogenesis. Mutation and amplification of proto-oncogenes such as c-myc, pkb/akt, and ras are also often observed in ovarian carcinoma cells. For example, whereas amplification of c-myc has been reported in 30% of ovarian cancers (13 , 14) , activation of the PKB/Akt pathway has also been detected in >30% of ovarian cancer tumors (15 , 16) . The small GTPase Ras is known to play a critical role in tumor growth and development. Constitutively active Ras mutations are found in >30% of all human cancers (17) . Alterations in the Ras-signaling pathway are also common in ovarian carcinomas and can occur by several mechanisms, which include direct activating mutations in Ras (18, 19, 20) , mutation of its downstream mediator, B-Raf (21 , 22) , and amplification/overexpression of its up-stream regulator HER-2/neu (23) . In addition, elevated activation of Ras pathways in ovarian cancer cells has been shown even in the absence of apparent activating mutations (24) .
Although the involvement and importance of various tumor suppressors and oncogenes in ovarian carcinomas have been clearly established, the molecular mechanisms by which these genes affect the development of ovarian carcinomas remain elusive. This is in part due to the lack of appropriate human ovarian cancer models. A human ovarian cancer cell model has been established recently using defined genetic elements to transform human ovarian surface epithelia (HOSE) cells by disrupting the p53, Rb, ectopic expression of the telomerase catalytic subunit, and activation of Ras signaling pathways. These cells produce tumors morphologically similar to high-grade, undifferentiated human ovarian carcinomas (25) . This novel model system provides an opportunity to elucidate the molecular mechanisms by which specific tumor suppressors or oncogenes play in the initiation and development of human ovarian cancer. In this study, we applied a two-dimensional electrophoresis proteomic approach to study the protein changes associated with Ras-mediated oncogenic transformation in human ovarian epithelium. We have identified several major cellular pathways activated upon Ras transformation of ovarian cells and have additionally studied how these pathways may contribute to the initiation and development of oncogenic transformation in ovarian epithelium.
MATERIALS AND METHODS
Cell Line Generation and Culture Methods.
T29 and T29H cells were generated and propagated in 1:1 MCDB105 and Media 199 (Sigma-Aldrich) supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin (Life Technologies, Inc.) as described previously (25) .
T29 and T29H cells grown in 75-cm2 T-flasks at 70–80% confluence were trypsinized, pelleted, and subsequently lysed with 500 μl of lysis buffer containing 7 m urea, 2 m thiourea, 4% 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonic acid, 1 mm EDTA, 1 mm EGTA, 60 mm DTT, 1 mm phenylmethylsulfonyl fluoride, 25 μg/ml leupeptin, 10 μg/ml aprotinin, 1 mm benzamidine, 1 mm sodium orthovanadate, and 1 mm microscystin at room temperature for 1 h. Whole cell extracts were obtained by centrifugation of the cell lysate at 180,000 × g for 1 h at 22°C. Protein concentration was determined using a detergent and reducing agent compatible Bradford RC/DC assay (Bio-Rad). Protein aliquots in 200 μg or 500 μg were used fresh for analytical or preparative gels, respectively, or frozen at −80°C.
Protein samples in 350 μl of lysis buffer containing 0.5% IPG buffer (pH 4–7) were applied onto Immobilin dry strips [(pH 4.0–7.0) 18 cm; Pharmacia Biosciences)] and allowed to rehydrate for 12 h at 20°C. Isoelectric focusing was performed at 20°C on an IPGphor unit (Pharmacia Biosciences) for a total of 54,000 Vh. After focusing, strips were either immediately loaded onto polyacrylamide gels or frozen at −80°C. The second dimension was performed on 10% Tricine-SDS gels (20 × 22 cm) with a Bio-Rad Protein IIxi two-dimensional unit at 12°C overnight for 2,800 Vh. Analytical gels were silver stained as described (26) , whereas preparative gels were silver stained using a low fixation protocol (27) .
Digital images of two-dimensional gels were recorded using an Alpha Innotech Imager 5500. Silver stained proteins were analyzed and quantified using Phoretix two-dimensional analysis software (Nonlinear). Intensity of each protein spot was measured as percentage of volume, corresponding with pixel intensity integrated over the area of each spot and divided by the sum of all of the spots in the gel. Protein spot changes between the treatment and control pair were scored when the increase or decrease was >50% in magnitude.
Tryptic In-Gel Digestion and Mass Spectrometry.
Protein spots were excised from multiple wet preparative gels with end-removed pipette tip. Gel pieces from 1–4 or 8–10 gels were pooled for high and low abundance spots, respectively. Excised protein spots were destained (28) and digested in-gel according to Lewis et al. (29) . Digested peptide samples were cocrystallized with an equal volume of α-cyano-4-hydroxy-trans-cinnamic acid matrix (Hewlett Packard) on a gold-coated sample plate and analyzed by matrix-assisted desorption ionization-time of flight mass spectrometry (Voyager-DE STR; Applied Biosystems) to obtain peptide mass. Peptide mass data were analyzed by searching against the National Center for Biotechnology Information nonredundant database using the programs Protein Prospector (University of California, San Francisco, San Francisco, CA) 3 and Profound (Rockefeller University) 4 as described (30) .
Cell Viability Analysis and 3-(4,5-Dimethylthiazol-2-yl)-2,5-Diphenyltetrazolium Bromide Assay.
T29 and T29H cells were grown to 50–70% confluence then treated with increasing amounts of H2O2. Overall cell viability at various time points was monitored by phase-contrast microscopy. Cell proliferation was measured by reduction of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide, which corresponds with living cell number and metabolic activity (31) . Cells were seeded in 96-well plates at a density of 2000 cells/well in 50 μl of medium. After 8 h of incubation at 37°C, an additional 50 μl of medium with (treated) or without (control) 200 μm of H2O2 was added to each well. After treatment with H2O2 (1–4 days), 50 μl of 0.5 mg/ml 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide was added to each well for 3 h, followed by 200 μl of DMSO for 1 h. The absorbance of each well was measured at 595 nm using a microplate reader.
Nuclear DNA Condensation Staining.
Two to 3 × 105 T29 or T29H cells were plated in 3 cm of tissue culture plates containing polylysine coated coverslips. Cells were grown for 6 h, then treated with 100 μm H2O2 and incubated at 37°C for 12 h. Cells on coverslips were stained with Hoescht 22658 for 30 min, rinsed with PBS, and sealed onto glass slides. Nuclear DNA condensation was examined using a fluorescence microscope (Olympus BX51) equipped with a Hamamatsu digital camera (C4742–95).
Caspase Activity Assays.
T29 and T29H cells were grown in 100-mm plates to 50–70% confluence, then treated for 12 h with 5 ml of regular medium or medium containing 100 μm of H2O2. After treatment, cells were harvested, washed twice in PBS, once in wash buffer [100 mm HEPES (pH 7.4), 0.5 mm EDTA, 1 mm DTT, and protease inhibitors (0.1 mm phenylmethylsulfonyl fluoride, 2 μg/ml leupeptin, and 2 μg/ml pepstatin)] then lysed on ice for 30 min with 150 μl of wash buffer with 1% of Triton-X100. Cell debris was removed by centrifugation at 16,000 × g for 20 min at 4°C. Lysate protein concentration was determined by Bradford assay (Bio-Rad). Fifty μg of total protein was incubated with the fluorogenic substrate DEVD-AFC, and caspase 3 activity was measured by monitoring the release of free AFC using a Fluoroskan fluorescence microplate reader with excitation and emission set at 405 and 510 nm, respectively.
Ras Activity Assay and Immunoblotting.
Cells at ∼70% confluence were harvested from 100-mm tissue culture plates. After three washes in cold PBS, cells were lysed in buffer containing 25 mm HEPES (pH 7.5), 150 mm NaCl, 1% NP40, 10% glycerol, 25 mm NaF, 10 mm MgCl2, 0.25% sodium deoxycholate, 1 mm EDTA, 1 mm Na3VO4, 10 μg/ml aprotinin, and 10 μg/ml leupeptin. The cell lysate was incubated with 40 μl of glutathione-Sepharose beads with 30 μg of glutathione S-transferase-Raf-RBD bound at 4°C for 1 h. Beads were washed three times with lysis buffer. The GTP-bound Ras was eluted by the addition of 30 μl of SDS sample buffer. Samples were loaded onto 12% SDS polyacrylamide mini-gels (Bio-Rad), transferred to polyvinylidene difluoride membranes by wet transfer, and blocked overnight in 5% milk in TBS-Tween. Membranes were incubated with anti-Ras primary antibody (1:3000) for 1.5 h, followed by secondary antibody (antimouse, 1:2000) for 45 min. Antigen-antibody complexes were detected by enhanced chemiluminescence (Cell Signaling). Blots for total cellular Ras were done in a similar fashion, except that whole cell lysates after protein quantification were loaded straight onto SDS-PAGE gels.
Inhibition of H-Ras Gene Expression by Small Interfering RNA (siRNA).
T29H cells were transfected with two retrovirus-mediated H-Ras siRNA vectors (designated H1 and H2) that have been described previously (32) . Briefly, H1 selectively silences mutant H-RasV12, whereas H2 suppresses both V12 mutant and wild-type Ras expression. T29H cells were transfected with H1 and H2 retrovirus generated from Phoenix cells and selected for 7–10 days in 0.7 mg/ml of G418 to establish stable cell lines. The expression and activation levels of Ras proteins in these siRNA cell lines were assayed as mentioned above.
Immortalization and Transformation of HOSE 29 Cell Line by hTERT and Oncogenic ras.
The immortalized T29 HOSE cell line, generated from IOSE29 cells (33) using hTERT, can bypass senescence in culture but form few colonies on soft agar. T29 cells also fail to form tumors upon injection into nude mice. T29 cells were subsequently infected with retrovirus containing the oncogenic H-RasV12 to generate the T29H cell line. As shown in Fig. 1 ⇓ , levels of active Ras, measured by selective pull-down of GTP-bound Ras protein, are significantly higher in T29H cells. This higher level of Ras activation in T29H cells led to markedly increased anchorage-independent growth in soft agar and the ability to form tumors when injected into nude mice (25) .
Proteomic Changes Associated with Ras-Mediated Oncogenic Transformation.
To explore the biochemical and cellular basis for H-Ras mediated oncogenic transformation, comparative proteomic analyses between T29 and T29H cells were performed. More than 2200 distinct protein spots could be resolved within the ranges of pH 4–7 and molecular weight 6,000–200,000 on silver-stained two-dimensional electrophoresis gel images obtained using whole cell lysates. To determine significant protein changes between H-Ras-transformed and control cells, four pairs of well-resolved gels from two independent experiments were analyzed using Phoretix 2-D analysis software (Nonlinear). Spots with average normalized volume increased or decreased by at least 50% between the T29 and T29H cell lines were selected as candidates with significant changes. Results from these automated analyses were additionally confirmed by direct visual inspection of all of the gel images by two independent observers. These spots were then processed by in-gel digestion, analyzed by peptide mass fingerprinting, and matched to entries in the human nonredundant database. Thirty proteins associated with Ras-mediated transformation were positively identified by peptide mass fingerprinting. These proteins include enzymes involved in cellular metabolism, DNA/protein methylation, proteolysis/apoptosis, redox balance, and calcium signaling pathways. 5 To confirm our proteomic results, we independently verified the expression of 4 randomly selected proteins from the 30 identified targets, to which commercial antibodies were available, using immunoblot analysis.
Up-Regulation of Multiple Proteins Involved in Major Cellular Redox Pathways.
A large portion of the proteomic changes associated with Ras-mediated transformation are proteins important in cellular metabolism and redox balance. Five of the 30 proteins identified are enzymes involved either directly in metabolizing ROS or in maintaining the redox balance of the cell (Table 1) ⇓ . These proteins include thioredoxin peroxidase (peroxiredoxin 4), a thiol-specific antioxidant enzyme using thioredoxin as a source of reducing equivalents to scavenge hydrogen peroxide; peroxiredoxin 3, a mitochondrial member of the antioxidant family of thioredoxin peroxidases; NADH dehydrogenase ubiquinone Fe/S protein, a 30 kDa subunit of mitochondrial complex I of the electron transport chain; glyoxyalase I, a key enzyme in detoxification of the reactive dicarbonyl compound, methylglyoxal, a side product of glycolysis; and selenophosphate synthetase, a rate-limiting enzyme for incorporating selenium into key redox metabolizing enzymes, such as glutathione peroxidase and thioredoxin reductase (Fig. 2) ⇓ .
Resistance of H-RasV12-Transformed HOSE Cells to H2O2-Mediated Oxidative Stress.
The up-regulation of multiple antioxidant proteins after Ras-mediated transformation led us to hypothesize that Ras-mediated transformation in human ovarian cells activates resistance mechanisms to oxidative stress. Therefore, we conducted experiments to determine if T29H cells would be resistant to oxidative stress. T29 and T29H cells were treated with increasing concentrations of H2O2 (25–400 μm). T29 cells responded to the treatment of H2O2 at much lower doses than did T29H cells. At 100 μm of H2O2, T29 cells showed marked cell death after 6 h, whereas T29H cells were largely unharmed under the same conditions (Fig. 3A) ⇓ . The effect of H2O2 treatment on cell survival/proliferation was additionally assessed using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay. Whereas the growth of T29 cell was dramatically suppressed after H2O2 (100 μm) treatment, the growth of T29H cells was not significantly affected (Fig. 3B) ⇓ .
Resistance of T29H Cells to Apoptosis Induced by Oxidative Stress.
To determine whether T29 cells were succumbing to apoptosis in response to H2O2-mediated stress, nuclear DNA status and caspase 3 activity were assessed. After T29 and T29H cells were treated for 12 h with 100 μm of H2O2, cells were examined by Hoescht 22658 staining to observe nuclear DNA condensation and fragmentation. The T29 cells exhibited a significant increase in chromosomal fragmentation and condensation compared with T29H cells after H2O2 treatment. More than 25% of the T29 cells showed chromosomal condensation, whereas only 6% of the T29H cells were affected (Fig. 4A) ⇓ . There were also significantly less T29 cells after H2O2 treatment, indicating that some of the T29 cells might have already died. To additionally confirm that Ras-mediated oncogenic transformation protects cells from H2O2 induced apoptosis, we measured caspase 3 activity in T29 and T29H cell lines after treatment with H2O2. Caspase 3 is one of the major downstream effectors of the apoptotic cascade and is frequently used as a marker of apoptosis. T29 cells displayed significantly higher caspase 3 activity than T29H cells when treated with 100 μmof H2O2, indicating that T29 cells were more susceptible to apoptosis under oxidative stress than Ras-transformed T29H cells (Fig. 4B) ⇓ . Taken together, these data suggest that Ras-transformed HOSE cells are resistant to oxidative stress-mediated apoptosis.
Ras Activation Is Responsible for Resistance to H2O2-Mediated Apoptosis in T29H Cells.
To verify that H-RasV12 was responsible for the resistance to H2O2 seen in the transformed ovarian cells, we used the farnesyl transferase inhibitor FTI-277 to block Ras activity. For Ras to be functionally active in vivo, it must be incorporated into the plasma membrane. The farnesylation of the extreme COOH-terminal CAAX motif of Ras is necessary for its membrane incorporation (34) . FTI-277 significantly inhibited Ras activation in T29H cells as measured by Ras activation assay using glutathione S-transferase-tagged Raf-RBD protein (Fig. 5A) ⇓ . T29H cells that were pretreated for 24 h with FTI-277 showed a marked decrease in resistance to H2O2-induced cell death when compared with the same cells treated with DMSO vehicle (Fig. 5B) ⇓ . Similar results were obtained when we used a natural ovarian cancer cell line SKOV-3 with a high level of Ras activation. This effect of FTI-277 to resensitize T29H cells to H2O2-mediated cell killing is likely mediated by caspase 3 activation. FTI-277 pretreatment increased the level of caspase 3 activity significantly in T29H cells when subjected to H2O2-mediated oxidative stress (Fig. 5C) ⇓ . These results suggest that Ras-activation is essential to protect T29H cells from oxidative stress-mediated apoptosis.
H-Ras siRNA Expression Resensitizes H-RasV12-Transformed HOSE Cells to H2O2-Mediated Oxidative Stress.
Although FTI-277 has been used as a common reagent to inhibit Ras activation in vivo, this compound can potentially inhibit other cellular processes that are also dependent upon farnesylation. To ensure that resistance to H2O2-mediated oxidative stress observed in T29H cells is indeed mediated by Ras activation, we examined the levels of resistance to H2O2 treatment in T29H cells with cellular Ras levels selectively suppressed by retroviral-mediated siRNA, which has been applied successfully to suppress the expression levels of Ras and thereby diminish the transformation efficiency and tumorigenicity in a similar ovarian cancer cell model as used in this study (32) . Expression of two retroviral-mediated siRNA vectors, H1-siRNA, which targets the H-RasV12 mutation, and H2-siRNA, which targets sequences in the wild-type H-Ras, significantly reduced the levels of active Ras, Ras-GTP (Fig. 6A) ⇓ . As shown in Fig. 6B ⇓ , stable expression of H1 and H2 in T29H cells led to a significant increase in sensitivity to H2O2-mediated killing as compared with the parental cells. The H1 and H2 siRNA cells are equally or more sensitive to H2O2-mediated killing as compared with untransformed T29 cells. Furthermore, suppressing Ras activation using H1 or H2 siRNA increased the level of caspase 3 activity significantly in T29H cells when subjected to H2O2-mediated oxidative stress (Fig. 6C) ⇓ . These results clearly demonstrate that Ras activation is responsible for the observed resistance of T29H cells to H2O2-mediated cell death.
Using functional proteomic approaches, we examined the proteomic changes associated with Ras-mediated oncogenic transformation in a genetically defined human ovarian cancer model. On the basis of >30 protein targets identified, we have unveiled a few cellular pathways that are potentially involved in the malignant transformation mediated by the ras oncogene in ovarian epithelial cells. The two most notable groups of protein targets altered significantly are enzymes involved in cellular metabolism and redox balance. It is understandable that many of the metabolic enzymes are up-regulated in T29H cells, because cancer cells usually exhibit growth advantage over normal cells. The increased protein levels for a large number of enzymes responsible for metabolizing ROS were unexpected. We have identified 5 redox-related proteins that are up-regulated upon Ras-mediated transformation of the ovarian epithelial cell. These proteins include thioredoxin peroxidase, peroxiredoxin 3, selenophosphate synthetase, and NADH dehydrogenase ubiquinone Fe/S protein.
As shown in Fig. 7 ⇓ , the majority of these proteins are involved directly in the major cellular redox-balancing pathways. Under normal conditions, the mitochondrial respiration chain is a major site of ROS production in vivo. ROS are generated as a byproduct of the electron-transfer reaction. Leakage of electrons to oxygen at the coenzyme Q/ubiquinone complex I and complex III leads to the formation of superoxide (35) , which can induce serious oxidative damage in cells when converted to H2O2 and eventually to the highly reactive hydroxyl radical. To protect against oxidative stress induced by ROS, cells use several antioxidants or reductants to maintain the intracellular redox environment in a highly reduced state. For example, superoxide is first converted to H2O2 by superoxide dismutase. Cellular H2O2 and other peroxides are then metabolized by glutathione- and thioredox-based peroxidase/reductase systems (36 , 37) . In addition, cells also possess an enzyme, catalase, that catalyzes the degradation of H2O2 to H2O and O2 (Fig. 7) ⇓ . Partial inhibition of mitochondrial complex I has been reported to enhance electron leakage from the electron transport chain, leading to an increase in superoxide generation and sensitizing of leukemia cells to apoptosis induced by the anticancer agents doxorubicin and ionizing radiation (38) . Therefore, up-regulation of NADH dehydrogenase ubiquinone Fe/S protein, a 30 kDa subunit of mitochondrial complex I of the electron transport chain, can potentially increase the efficiency of the mitochondrial electron transport chain and subsequently decrease the production of superoxide. Peroxiredoxin 3, a mitochondrial member of the antioxidant family of thioredoxin peroxidases, is essential for neoplastic transformation in R1a-myc and MCF7/ADR cells (39) . It has also been shown that peroxiredoxin 3 is overexpressed in hepatocellular carcinomas (40) . Increased expression of peroxiredoxin 3 protects cancer cells against hypoxia and drug-induced hydrogen peroxide-dependent apoptosis (41) . Another redox-related target protein that is up-regulated in T29H cells is selenophosphate synthetase, a key enzyme of the selenoprotein biosynthesis pathway. Although selenophosphate synthetase is not involved directly in metabolizing ROS, this enzyme is critical for maintaining the redox balance of the cell (42 , 43) , because selenium is one of the important components of antioxidant enzymes such as glutathione peroxidase and thioredoxin reductase. Both glutathione peroxidase and thioredoxin reductase contain selenium in their active site, and are therefore dependent on selenophosphate synthetase for their function. In addition to the 5 antioxidant proteins identified by our functional proteomic study, another important ROS-metabolizing enzyme, the mitochondrial (Mn, Cu) superoxide dismutase, was also found up-regulated at the transcriptional level in T29H cells by gene array study (25) . Taken together, these results depict a picture in which the overall cellular pathway of ROS metabolism, especially in the mitochondria, is enhanced in a coordinated fashion in T29H cells (Fig. 7) ⇓ .
Increased expression of major antioxidant enzymes in T29H cells raises an intriguing question: what is the functional role of up-regulation of these antioxidant enzymes in Ras-mediated oncogenic transformation? Oncogenically transformed cells exhibit an increased growth rate, and thus produce a higher level of ROS associated with this increased cellular respiration (44) . Numerous studies have shown that the level of oxidants and peroxides is higher in tumors than in surrounding normal tissue (45) . In addition, a previous study shows that Ras-transformed human ovarian epithelial cells, including T29H, expressed higher levels of proinflammatory cytokines, which place additional oxidative stress upon these cells. High levels of free radicals are thought to be an important factor in causing the genetic alterations that lead to tumor development (45) . However, excessive ROS stress can cause extensive cellular damage and eventually lead to the killing of the cancer cell by activating apoptotic programs. For example, direct damage of the mitochondria by ROS will lead to the leakage of the mitochondrial membrane, and subsequent release of cytochrome c and other apoptotic factors, which, in turn, activate the downstream apoptotic cascade. How cancer cells defend themselves against cell damage/destruction caused by excessive free radicals associated with oncogenic transformation is not clear. The up-regulation of antioxidant enzymes, especially in mitochondria, the primary site for ROS production, would give the tumor cell certain survival advantages under its intrinsically high oxidative conditions. Increased antioxidant capacity would allow cancer cells to avoid oxidative damage that would lead to cell death under normal circumstances.
We postulate that Ras up-regulates antioxidant proteins to mitigate the detrimental effects of excessive oxidative stresses that the tumor cell incurs. This hypothesis is confirmed by our observation that H-Ras-transformed tumorigenic T29H cells are much more resistant to H2O2-mediated cell killing compared with the immortalized but nontumorigenic parental T29 cells. Activation of caspase 3 upon oxidative stress is abrogated in H-Ras-transformed cells, suggesting that up-regulation of these antioxidant enzymes protects T29H cells from caspase-3 mediated apoptosis when the cell undergoes oxidative stress. In addition, inhibition of Ras activation by siRNA and FTI-277, a general inhibitor of Ras signaling, resensitizes T29H cells to H2O2-mediated cell killing. FTI-277 and Ras siRNA-treated T29H cells also behave in a similar manner to parental T29 cells with regard to caspase 3 activation when under oxidative stress, indicating that Ras is involved directly in the resistance to ROS-mediated apoptosis in transformed cells. Therefore, activation of Ras signaling pathways is responsible for the enhancement of the overall antioxidant capacity in the mitochondria, thus protecting tumor cells from apoptosis that would normally occur with excessive ROS. This is likely a general mechanism related to the Ras signaling pathways, given that a natural ovarian carcinoma cell line SKOV-3, which displays high levels of Ras activity, is also highly resistant to H2O2-mediated cell killing.
In summary, by examination of a genetically defined human ovarian cancer cell model using functional proteomic approaches, we identified an important cellular pathway that is potentially involved in Ras-mediated oncogenic transformation. To our knowledge, it is the first such attempt combining a human ovarian oncogenic transformation cancer model with functional proteomic approaches to elucidate novel mechanisms involved in transformation. Our study suggests that activation of Ras signaling pathways increases the threshold of ROS tolerance by up-regulating the overall antioxidation pathways in cells, especially in the mitochondria. This enhanced antioxidant capacity protects the transformed cells from high levels of ROS associated with the uncontrolled growth potential of tumor cells. It is conceivable that an enhanced antioxidation capability may constitute a common mechanism for tumor cells to evade apoptosis induced by oxidative stresses at high ROS loads. Furthermore, data presented in this study suggest that inhibitors of mitochondrial antioxidant enzymes might act as effective cancer therapeutic agents through preferential accumulation of ROS in malignant cells that usually produce high levels of ROS and are under intrinsic oxidative stress.
We thank Dr. Natalie Ahn (University of Colorado, Boulder, CO) for continuous support and insightful discussions. We thank Drs. Alex Kurosky and Tony Haag in the Biomedical Resource Facility (supported by NCI Grant R24CA88317 and NIEHS Center Grant ES06676) for mass spectrometry analysis.
Grant support: American Cancer Society Research Scholar Award RGS-01-035-01-TBE and NIH Grant GM060170 (X. Cheng), and RGS-04-028-1-CCE (J. Liu).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Requests for reprints: Xiaodong Cheng, Department of Pharmacology and Toxicology, The University of Texas Medical Branch, Galveston, TX 77555-1031. Phone: (409) 772-9656; Fax: (409) 772-9642; E-mail: firstname.lastname@example.org; or Jinsong Liu, Department of Pathology, The University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Blvd., Houston, TX 77030-4095. Phone: (713) 745-1102; Fax: (713) 792-5529; E-mail:
↵3 Internet address: http://prospector.ucsf.edu/.
↵4 Internet address: http://prowl.rockefeller.edu/profound_bin/webprofound.exe.
↵5 T. W. Young, C. F. Mei, J. Liu, and X. Cheng. Probing Ras-mediated transformation of human ovarian epithelia cell by functional proteomics, manuscript in preparation.
- Received January 23, 2004.
- Revision received April 5, 2004.
- Accepted April 28, 2004.
- ©2004 American Association for Cancer Research.