Mitochondrial Production of Reactive Oxygen Species Mediate Dicumarol-induced Cytotoxicity in Cancer Cells*

Dicumarol is a naturally occurring anticoagulant derived from coumarin that induces cytotoxicity and oxidative stress in human pancreatic cancer cells (Cullen, J. J., Hinkhouse, M. M., Grady, M., Gaut, A. W., Liu, J., Zhang, Y., Weydert, C. J. D., Domann, F. E., and Oberley, L. W. (2003) Cancer Res. 63, 5513–5520). Although dicumarol has been used as an inhibitor of the two-electron reductase NAD(P)H:quinone oxidoreductase (NQO1), dicumarol is also thought to affect quinone-mediated electron transfer reactions in the mitochondria, leading to the production of superoxide (\batchmode \documentclass[fleqn,10pt,legalpaper]{article} \usepackage{amssymb} \usepackage{amsfonts} \usepackage{amsmath} \pagestyle{empty} \begin{document} \(\mathrm{O}_{2}^{{\bar{{\cdot}}}}\) \end{document}) and hydrogen peroxide (H2O2). We hypothesized that mitochondrial production of reactive oxygen species mediates the increased susceptibility of pancreatic cancer cells to dicumarol-induced metabolic oxidative stress. Dicumarol decreased clonogenic survival equally in both MDA-MB-468 NQO1– and MDA-MB-468 NQO1+ breast cancer cells. Dicumarol decreased clonogenic survival in the transformed fibroblast cell line IMRSV-90 compared with the IMR-90 cell line. Dicumarol, with the addition of mitochondrial electron transport chain blockers, decreased clonogenic cell survival in human pancreatic cancer cells and increased superoxide levels. Dicumarol with the mitochondrial electron transport chain blocker antimycin A decreased clonogenic survival and increased superoxide levels in cells with functional mitochondria but had little effect on cancer cells without functional mitochondria. Overexpression of manganese superoxide dismutase and mitochondrial-targeted catalase with adenoviral vectors reversed the dicumarol-induced cytotoxicity and reversed fluorescence of the oxidation-sensitive probe. We conclude mitochondrial production of reactive oxygen species mediates the increased susceptibility of cancer cells to dicumarol-induced cytotoxicity.

We hypothesized that mitochondrial production of reactive oxygen species mediates the increased susceptibility of pancreatic cancer cells to dicumarol-induced metabolic oxidative stress. Dicumarol decreased clonogenic survival equally in both MDA-MB-468 NQO1 ؊ and MDA-MB-468 NQO1؉ breast cancer cells. Dicumarol decreased clonogenic survival in the transformed fibroblast cell line IMRSV-90 compared with the IMR-90 cell line. Dicumarol, with the addition of mitochondrial electron transport chain blockers, decreased clonogenic cell survival in human pancreatic cancer cells and increased superoxide levels. Dicumarol with the mitochondrial electron transport chain blocker antimycin A decreased clonogenic survival and increased superoxide levels in cells with functional mitochondria but had little effect on cancer cells without functional mitochondria. Overexpression of manganese superoxide dismutase and mitochondrial-targeted catalase with adenoviral vectors reversed the dicumarol-induced cytotoxicity and reversed fluorescence of the oxidation-sensitive probe. We conclude mitochondrial production of reactive oxygen species mediates the increased susceptibility of cancer cells to dicumarol-induced cytotoxicity.
Dicumarol (3,3Ј-methylenebis [4-hydroxycoumarin]) is a naturally occurring anticoagulant derived from coumarin that is obtained from the sweet clover (Melilotus alba) (1,2). Coumarin and its compounds have demonstrated numerous anti-tumor and antiproliferative effects. Studies have shown that coumarin compounds inhibit proliferation of human malignant cell lines in vitro (3,4) as well as affecting tumor activity in vivo (1,(5)(6)(7)(8). Dicumarol-induced cytotoxicity is not from interference with vitamin K epoxide oxidoreductase or a previously unrecognized aspect of vitamin K metabolism, since the addition of vitamin K does not impair the growth inhibiting effect of dicumarol. The growth inhibitory effect of dicumarol may also be relatively specific for tumor cells, since proliferation of normal human airway myocytes is not affected (9). Although such coumarin compounds as dicumarol have been utilized in cancer therapy, little is known about the mechanism of action of these drugs. NADPH:quinone oxidoreductase (NQO1) is responsible for the two-electron reduction of quinones using NADPH or NADH as an electron donor (10). Dicumarol inhibits NQO1 resulting in increased oxidative stress within the cell, leading to increased cell toxicity (11,12).
Recent studies have demonstrated that dicumarol causes cytotoxicity against human pancreatic cancer cells that have overexpression and increased activity of NQO1 (13). Dicumarol increased intracellular levels of superoxide (O 2 . ), as measured by hydroethidine staining, and inhibited cell growth. Both of these effects (cell growth inhibition and hydroethidine staining) were blunted with infection of an adenoviral vector containing the cDNA for manganese superoxide dismutase (Mn-SOD), 2 a mitochondrial antioxidant enzyme that scavenges O 2 . . This study compliments the recent study from Li et al. (14) that similar compounds as dicumarol result in increased mitochondrial O 2 . production leading to apoptosis, whereas others have demonstrated that dicumarol acts intracellularly to uncouple mitochondrial oxidative phosphorylation at the NADH-cytochrome b5 site (15). These studies suggest that mitochondrial metabolism might be involved in the process that caused cytotoxicity with dicumarol treatment even though the majority of NQO1 is found in the cytoplasm (16,17). Metabolic oxidative stress results from an imbalance between prooxidants (i.e. superoxide, hydrogen peroxide, etc.) produced as byproducts of oxidative metabolism and the cellular antioxidants that eliminate these species. Studies to determine whether oxidative stress from mitochondrial metabolism * This work was supported by National Institutes of Health Grants CA115785, CA100045, CA66081, and HL07485-24, the Susan L. Bader Pancreatic Cancer Research Fund, and the Medical Research Service, Department of Veterans Affairs. 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. 1  was the mechanism responsible for dicumarol-induced cytotoxicity were accomplished. Dicumarol-induced cytotoxicity of MIA PaCa-2 cells was accompanied by a dose-dependent and time-dependent increase in apoptosis as measured by flow cytometry and diaminobenzidine staining, which was associated with cytochrome c release from mitochondria and poly-(ADP-ribose) polymerase cleavage (18). These results demonstrated that dicumarol treatment in these human tumor cells caused cytotoxicity as well as activation of mitochondrial signaling and pathways thought to be involved with cell death. We hypothesized that mitochondrial production of reactive oxygen species mediates the increased susceptibility of pancreatic cancer cells to dicumarol-induced metabolic oxidative stress. The current studies demonstrate that the presence or absence of NQO1 did not alter dicumarol-induced cytotoxicity but decreased clonogenic survival in transformed cell lines to a greater degree than normal cells. We also demonstrated that mitochondrial production of reactive oxygen species mediates the increased susceptibility of cancer cells to dicumarol-induced cytotoxicity in a series of three experiments. First, we demonstrated dicumarol-induced cytotoxicity, and increased superoxide levels were enhanced with the addition of mitochondrial electron transport chain (METC) blockers in cancer cell lines. Second, dicumarol with the addition of METC blockers induced cytotoxicity and increased superoxide levels in cells with functional mitochondria but had little effect on cancer cells without functional mitochondria. Finally, overexpression of Mn-SOD and mitochondrial-targeted catalase with adenoviral vectors reversed the dicumarol-induced cytotoxicity and decreased ROS levels.

EXPERIMENTAL PROCEDURES
Cell Culture-MIA PaCa-2 (undifferentiated) and AsPC-1 (poorly to moderately differentiated) cells, human pancreatic adenocarcinoma cells, were obtained from American Type Culture Collection (Manassas, VA). MIA PaCa-2 was maintained in Dulbecco's modified Eagle's medium supplemented with 10% heat-inactivated fetal bovine serum and 2.5% horse serum. AsPC-1 was maintained in RPMI 1640 with 20% heat-inactivated bovine serum and 1% sodium pyruvate. Control MDA-MB-468 breast cancer cells and the isogenic cell line MDA-MB-468 stably transfected with constitutive NQO1 expression vector under the control of a CMV promoter were generous gifts from Dr. David Boothman, University of Texas Southwestern and were maintained in RPMI 1640 supplemented with 10% calf serum, 2 mM L-glutamine, 100 units/ml penicillin, and 100 mg/ml streptomycin (19). Dr. Michael King (Thomas Jefferson University) kindly provided the human osteosarcoma 143BTKrho(ϩ) and rho(0) cells (20). IMR90 human lung fibroblasts and their SV40-transformed counterparts (IMR90-SV40) were gifts from Dr. Douglas Spitz, University of Iowa, and were maintained in minimal essential medium (MEM) supplemented with 1ϫ MEM vitamins, 1ϫ MEM essential amino acids, and 2ϫ nonessential amino acids, with 20% fetal calf serum and buffered with 2.2 g/liter sodium bicarbonate (21). GM00637G SV-40-transformed human fibroblasts were also a gift from Dr. Douglas Spitz and were maintained in minimum Eagle's medium with 10% fetal bovine serum. All media was obtained from Invitrogen, and all cell lines were maintained at 37°C. All cells were routinely tested for mycoplasma and only utilized when found to be negative.
Generation of rho(0) Cells-To determine the role of mitochondrial electron transport chain activity in dicumarol-induced cytotoxicity and ROS production, human osteosarcoma cells (143BTK-rho(ϩ) and rho(0) cells) (20) and MIA PaCa-2 rho(0) cells depleted of mitochondria DNA were generated by incubating wild type cells (rhoϩ) for 6 -8 weeks with 100 ng/ml ethidium bromide. The medium was supplemented with 4.5 mg/ml glucose, 50 g/ml uridine, and 100 g/ml pyruvate to compensate for the respiratory metabolism deficit as described (22). After selection, the MIA PaCa-2 rho(0) cells were cultured in the above-specified medium without ethidium bromide. To verify the mitochondria DNA depletion, PCR demonstrated an absence of mitochondria DNA in rho(0) cells as described (22).
Clonogenic Cell Survival Experiments-At the beginning of each experiment, cells were rinsed with phosphate-buffered saline and placed in media supplemented with 2% serum as previously described (19). Control and treatment cultures were treated identically. Dicumarol (50 -100 M) was added to cell cultures after a 50 mM concentration of dicumarol was dissolved in water by the dropwise addition of 0.1 N NaOH. The addition of up to 2.5 l of this solution/ml (250 M highest final concentration) did not change the pH of complete medium. Cultures were then placed in an incubator, and at each time point cells were trypsinized, counted, diluted, and plated for clonogenic cell survival assay as previously described (13). Surviving colonies were fixed and stained after 7-10 days and counted under a dissecting microscope.
Cell Viability-As an indicator of cell metabolic viability, the 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide (MTT) assay was used. Cells were seeded at 1 ϫ 10 4 in a 96-wellplate in full media. 5 mg/ml MTT was added to the wells and incubated at 37°C for 3 h. Lysing buffer consisting of 20% SDS in a 1:1 solution of N,N-dimethyl formamide was added and incubated at 37°C for 16 h. The plate was read at 590 nm on an Ultramark microplate imaging system (Bio-Rad).
Drug Treatment-Drugs were added to cells at final concentrations of 10 M antimycin A (AntA), 50 M myxothiazol (Myx), 50 M rotenone, and 2 M dinitrophenol. Stock solutions of all drugs were obtained from Sigma, dissolved in Me 2 SO, and used without further purification. All cells were incubated in media containing 2% serum with different drugs for the specified times.
Transduction of Adenoviral Vectors Containing Antioxidant Enzymes-To determine whether O 2 . or H 2 O 2 were responsible for the cytotoxic effects of dicumarol, we transduced MIA PaCa-2 pancreatic cancer cells with adenovirus constructs containing antioxidant enzymes. The adenovirus constructs used were replication-defective, E1-and partial E3-deleted recombinant adenovirus (23). Inserted into the E1 region of the adenovirus genome are the human antioxidant genes (Mn-SOD, CuZn-SOD, mitCAT, CAT ) or BglII, all of which are driven by a cytomegalovirus promoter. Except for AdmitCAT, the adenovirus constructs were originally prepared by John Engelhardt, University of Iowa (23). A plasmid with the full-length catalase cDNA with the Mn-SOD mitochondrial leader sequence added to the construct was originally prepared by Dr. Andres Melendez (24). Approximately 10 6 MIA PaCa-2 cells were plated in 10 ml of complete media in a 100-mm 2 plastic dish and allowed to attach for 24 h. Cells were then washed 3 times in serum-and antibioticfree media. The adenovirus construct(s), suspended in 3% sucrose, was then applied to cells suspended in 4 ml of serum-and antibiotic-free media. Control cells are treated with the adenovirus-BglII construct. Cells were incubated with the adenovirus constructs for 24 h. Media was then replaced with 4 ml of complete media for an additional 24 h before cells were harvested. Measurement of Prooxidant Levels-Prooxidant levels (presumably H 2 O 2 ) were measured at various time intervals by labeling cells for 15 min at 37°C using the oxidation sensitive (C-400; 5-(and -6)-carboxy-2Ј,7Ј-dichlorodihydrofluorescein diacetate; Molecular Probes, Eugene, OR) and insensitive (C-369; 5-(and -6)carboxy-2Ј,7Ј-dichlorofluorescein diacetate) fluorescent dyes dissolved in Me 2 SO with flow cytometric analysis as described (13). The oxidation insensitive probe was utilized to control for changes in uptake, ester cleavage, and efflux so that differences in fluorescence could definitively be attributed to changes in oxidation of the probe. Intracellular generation of the O 2 . was assessed using hydroethidine fluorescence. O 2 . reacts with hydroethidine to form oxidized products, which bind to nuclear DNA and form red fluorescent complexes. The fluorescence of hydroethidine-labeled cells was quantitated using flow cytometry. Cells were grown to subconfluence in 60-mm dishes and initially treated with or without drugs for the specified time, washed, and incubated with hydroethidine (10 M) for 40 min. The cells were removed by trypsinization, which was neutralized with phosphate-buffered saline containing 2% fetal calf serum and then analyzed by flow cytometry (BD Bioscience FACScan). Intracellular hydroperoxide production was estimated by adding 10 M DCFDA (C-400) to cells. After incubation at 37°C for 20 min, cells were harvested using a cell scraper and transferred to a collection tube along with the floating cells in the incubation buffer. After centrifugation, supernatant was removed by aspiration; cells were lysed with 200 l of 0.5% Nonidet P-40. After centrifugation, the fluorescence levels in cell lysates were measured with a TECAN Spectrafluor Plus (excitation 485 nm, emission 535 nm). Extracellular H 2 O 2 production was determined using a fluorometric assay (25). Incubation of H 2 O 2 with horseradish peroxidase in the presence of para-hydroxyphenyl acetic acid (pHPA) (Sigma) results in the formation of a fluorescent pHPA dimer. Cells were treated with 50 M dicumarol at 37°C in Dulbecco's modified Eagle's medium supplemented with 2% fetal bovine serum. After the treatment the cells were washed with phenol red-free Hanks' balanced salt solution. To the cells were added 1 ml of Hanks' balanced salt solution supplemented with 6.5 mM glucose, 1 mM HEPES, 6 mM sodium bicarbonate, 1.6 mM pHPA, and 0.95 g/ml horseradish peroxidase. Cells were incubated at 37°C for 60 min. H 2 O 2 -induced pHPA dimer formation was measured with a spectrofluorometer (PerkinElmer Life Sciences LS50B) (excitation 323 nm, emission 400 nm). H 2 O 2 production was estimated by constructing a standard curve using known H 2 O 2 concentrations.
Cell Homogenization and Protein Determination-Cells were washed 3 times in phosphate-buffered saline (pH 7.0), scraped from the dishes using a rubber policeman, and then collected in phosphate buffer (pH 7.8). This was followed by sonic disruption on ice for 30 s in 10-s bursts using a VibraCell sonicator (Sonics and Materials Inc., Danbury, CT) at 100% power. Protein concentration was determined using the Bio-Rad Bradford dye-binding protein assay kit according to the manufacturer's instructions.
Western Analysis-Immunoreactive protein corresponding to antioxidant enzymes was identified and quantified from total cell protein by the specific reaction of the immobilized protein with its antibody. Total protein was electrophoresed in a 12.5% SDS-polyacrylamide running gel and a 5% stacking gel. The proteins were then electrotransferred to nitrocellulose sheets. After blocking in 20% fetal bovine serum for 1 h, the sheets were washed and then treated with antisera to either Mn-SOD (1:1000), CuZn-SOD (1:500), or catalase (1:1000). Polyclonal rabbit-anti-human antibodies to Mn-SOD and CuZn-SOD have been prepared and characterized in our laboratory (26). The antibody for catalase was purchased from Athens, Inc. (Athens, GA). The blots were incubated with horseradish peroxidase-conjugated goat-anti-rabbit (Sigma) IgG (1:10,000) for 1 h at room temperature. The washed blot was then treated with ECL Western blot detection solution (Amersham Biosciences) and exposed to x-ray film. All Western blots were performed in duplicate.
Enzyme Activity-Superoxide dismutase activity was measured using an indirect competition assay between SOD and an indicator molecule, nitro blue tetrazolium (26). Sodium cyanide (5 mM) inhibits CuZn-SOD; therefore, activity measured in the crude homogenate in the presence of sodium cyanide indicates only Mn-SOD activity. Specific activity was reported as units/mg of protein. Catalase activity was measured by the method of Beers and Sizer (27) with the analysis of Aebi (28). All measurements were normalized to protein content using the method of Lowry et al. (29). Briefly, the catalase assay is a spectrophotometric procedure that measures peroxide removal.
Statistical Analysis-Statistical analysis was performed by means of Systat (Systat inc., Evanston Ill). A single factor ANOVA, followed by post-hoc Tukey test, was used to determine statistical differences between means. All means were calculated from three experiments, and error bars represent standard error of mean (S.E.). All Western blots were repeated at least twice. All data are expressed as means Ϯ S.E.

RESULTS
The Effects of Dicumarol Are Independent of NQO1 Status-NAD(P)H:quinone oxidoreductase (NQO1), a homodimeric, ubiquitous, flavoprotein, catalyzes the two-electron reduction of quinones to hydroquinones. This reaction prevents the oneelectron reduction of quinones by cytochrome P450 reductase and other flavoproteins that results in oxidative cycling with generation of ROS. Previous studies have suggested that selective inhibition of NQO1 with dicumarol alters the malignant phenotype of tumor cells with high NQO1 levels (13). To determine the role of NQO1 in dicumarol-induced cytotoxicity in  (Fig. 1C). Consistent with the data in the clonogenic survival, hydroethidine fluorescence was somewhat higher in the NQO1Ϫ cell lines and lower in the NQO1ϩ cell line, which may be due to the antioxidant properties of NQO1.
Dicumarol-induced Cytotoxicity Is More Profound in Human Transformed Cells than Untransformed Cells-Transformed cells have been found to be more susceptible to oxidative stress   DECEMBER 8, 2006 • VOLUME 281 • NUMBER 49 than untransformed cells (31). To determine the effect of dicumarol on transformed versus untransformed cells, IMR90 human lung fibroblasts and their SV40-transformed counterparts (IMR90-SV40) were treated with dicumarol (50 and 100 M) for 24 h (Fig. 2). Although dicumarol had little effect on clonogenic survival in the IMR90 fibroblast cell line, dicumarol had a profound effect in decreasing clonogenic survival in the IMR90-SV40-transformed cell line, resulting in no colony formation when 100 M was given for 24 h.

Dicumarol and Mitochondrial Reactive Oxygen Species
Mitochondrial Electron Transport Chain Blockers Enhance Dicumarol-induced Cytotoxicity-Previous investigations have demonstrated that dicumarol acts intracellularly to uncouple mitochondrial oxidative phosphorylation at the NADH-cytochrome b5 site (15). The effect of mitochondrial electron transport chain blockers on dicumarol-induced cytotoxicity was determined in MIA PaCa-2 human pancreatic cancer cells and in GM00637G SV40-transformed human fibroblasts (Fig. 3).
MIA PaCa-2 cell survival decreased after 4 h of dicumarol (50 M) in the presence of antimycin A, myxothiazol, and rotenone, relative to control, but these drugs were not toxic by themselves during this time frame (Fig. 3A). No toxicity was seen in cells treated with dinitrophenol in the presence of dicumarol. In the GM00637G cells there was a more dramatic decrease in clonogenic survival after treatment of dicumarol (50 M) in the presence of antimycin A, myxothiazol, and rotenone (Fig. 3B). Once again, dinitrophenol in the presence of dicumarol had little effect on clonogenic survival.  ence of AntA. These data support the hypothesis that blocking the mitochondrial electron transport chain will contribute to dicumarol-induced cytotoxicity.

Mitochondrial Electron Transport Chain Blockers Enhance Dicumarol-induced Cytotoxicity and Prooxidant Production-
To test the hypothesis that mitochondrial electron transport chain blockers enhance levels of intracellular prooxidants and associated cytotoxicity, cells were labeled with hydroethidine, and fluorescence was quantitated using flow cytometry. The human pancreatic cancer cell lines AsPC-1 and MIA PaCa-2 were treated with dicumarol (50 M) alone, antimycin A (10 M), and the combination of both drugs, and clonogenic survival and hydroethidine were measured (Fig. 4). Cells treated with dicumarol for 4 h in the presence of AntA demonstrated enhanced cytotoxicity as well as increases in mean fluorescence intensity when labeled with hydroethidine. Fig.  4A demonstrates clonogenic survival in the MIA PaCa-2 pancreatic cancer cell line after treatment with dicumarol, antimycin, or the combination of drugs for 4 h. As previously demonstrated, dicumarol (50 M) alone decreased clono-genic survival, whereas the addition of antimycin A further decreased clonogenic survival. As demonstrated in Fig. 4B there was a concomitant increase in hydroethidine fluorescence as quantitated by flow cytometry corresponding to the decreased clonogenic survival. This was also seen in the AsPC-1 pancreatic cancer cell line whereupon dicumarol decreased clonogenic survival (Fig. 4C), which was enhanced with antimycin A, which corresponded to increased prooxidants production as measured by hydroethidine fluorescence (Fig. 4D).
Estimation of H 2 O 2 Levels during Dicumarol Treatment-To test the hypothesis that dicumarol increases the levels of prooxidants, intracellular H 2 O 2 was determined by a DCF plate reader assay in which DCFDA was oxidized to fluorophore DCF. H 2 O 2 can easily diffuse from the intracellular site of production to the extracellular milieu so we also measured extracellular H 2 O 2 , which was determined using a fluorometric assay (25) that measures a fluorescent pHPA dimer. Fig. 5A demonstrates that intracellular DCF fluorescence increases after 1 h of dicumarol treatment and remains elevated for up to 5 h. In addition, there was an  DECEMBER 8, 2006 • VOLUME 281 • NUMBER 49 JOURNAL OF BIOLOGICAL CHEMISTRY 37421 incremental increase in extracellular H 2 O 2 , which peaked at 4 h with dicumarol treatment (Fig. 5B).

Dicumarol-induced Cytotoxicity and Prooxidant Production in Cells Lacking Functional Mitochondrial Electron Transport
Chains-To provide further data demonstrating the role of mitochondrial electron transport chain activity in dicumarolinduced cytotoxicity and ROS production, human osteosarcoma cells and human pancreatic cancer cells deficient in fully functional mitochondrial DNA were utilized. Fig. 6 shows clonogenic survival in mitochondrial deficient rho(0) cells and the parental rho(ϩ) cells after dicumarol treatment for up to 24 h in the presence of AntA. In the human osteosarcoma rho(ϩ) cells, dicumarol 50 M resulted in 83 and 90% cell killing for 8 and 24 h of dicumarol treatment, respectively, in the presence of AntA (Fig. 6A). In contrast, rho(0) cells exhibited 10 and 0% cell killing during the same time frame in the presence of AntA (Fig.  6A). Furthermore, after 2 h of dicumarol treatment in the presence of AntA, hydroethidine fluorescence was greater in rho(ϩ) cells relative to rho(0) cells (Fig. 6B). In human pancreatic cancer cells deficient in fully functional mitochondria, a similar pattern was seen. Fig. 6C demonstrates clonogenic survival in mitochondrial-deficient pancreatic cancer rho(0) cells and the parental MIA PaCa-2 rho(ϩ) cells after dicumarol treatment for up to 24 h in the presence of AntA. In the rho(ϩ) cells, dicumarol 50 M resulted in 23 and 65% cell killing during 4 and 24 h of dicumarol treatment, respectively, in the presence of AntA (Fig. 6C). Once again, the corresponding rho(0) cells exhibited 7 and 20% cell killing during the same time in the presence of AntA (Fig. 6D). determine whether O 2 . or H 2 O 2 were responsible for the cytotoxic effects and increased levels of prooxidants with dicumarol treatment, we transduced MIA PaCa-2 pancreatic cancer cells with adenovirus constructs containing antioxidant enzymes. As mentioned previously, NQO1 is primarily a cytosolic protein (16). Our studies in the NQO1Ϫ and NQO1ϩ MDA-MB-468 cell lines demonstrated that dicumarol-induced cytotoxicity was not dependent on NQO1 (Fig. 1). To provide further evidence that dicumarol-induced cytotoxicity was not dependent on NQO1, we initially overexpressed AdBglII  Table 1 demonstrates the increase in both Mn-SOD activity and catalase activity induced by treatment with the adenoviral vectors; thus, both immunoreactive protein and enzymatic activity increased as expected after adenovirus transduction. MIA PaCa-2 human pancreatic cancer cells that overexpressed both Mn-SOD and mitCAT had significant reductions in dicumarol-induced cytoxicity when compared with either controls, AdBglII-treated cells, or cells treated with either the AdMn-SOD or AdmitCAT vectors alone (Fig. 7B). In pancreatic cancer cells treated with the single adenoviral vectors, there was approximately a 40% reduction in clonogenic survival when compared with controls from each group. In contrast, cells infected with both AdMn-SOD and AdmitCAT had clonogenic survival that was similar to controls suggesting a protective effect of these antioxidants on dicumarol-induced cytotoxicity.

The Effect of Overexpression of Antioxidant Enzymes on Dicumarol-induced Cytotoxicity and Prooxidant Production-To
To provide further evidence to test the hypothesis that mitochondrial reactive oxygen species mediates dicumarol-induced cytotoxicity, the same MIA PaCa-2 cells were labeled with an oxidation-sensitive dye during dicumarol treatment. Dicumarol significantly increased mean fluorescence intensity when labeled with the oxidation sensitive probe (C-400, Fig. 7C). The addition of dicumarol (50 M) in cells infected with the AdBglII (200 m.o.i.) vector control increased mean fluorescence inten- Ϫdicumarol/ϩAntA; means Ϯ S.E., n ϭ 3. B, dicumarol and antimycin A increased superoxide production in human osteosarcoma cells with functional mitochondria (rho(ϩ)) but had little affect on mitochondrial functionally deficient (rho(0)) cells. Cells were incubated in various treatment groups and then stained with hydroethidine. Mean fluorescence intensity was measured via flow cytometry, corrected for background fluorescence levels, and normalized to controls; means Ϯ S.E., n ϭ 3. C, the combination of dicumarol and antimycin A decreased cell survival in human pancreatic cancer cells with functional mitochondria (rho(ϩ)) in a time-dependent manner but had little effect on cells lacking functional mitochondria (rho(o)). There are significant differences between in rho(ϩ) cells in the dicumarol ϩ antimycin A compared with antimycin alone at 4 and 24 h. There are no differences in the rho(0) cells in the 4-and 24-h time points when compared with the 0-h time point. Clonogenic survival is presented as the surviving fraction at various time intervals after plating 500 cells per plate. *, p Ͻ 0.05 versus rho(ϩ)/Ϫdicumarol/ϩAntA; means Ϯ S.E., n ϭ 3. D, dicumarol and antimycin A increased superoxide production in human pancreatic cancer cells with functional mitochondria (rho(ϩ)) but had little affect on mitochondria functionally deficient (rho(o)) cells. Cells were incubated in various treatment groups and then stained with hydroethidine. Mean fluorescence intensity was measured via flow cytometry, corrected for background fluorescence levels, and normalized to controls; means Ϯ S.E., n ϭ 3. DECEMBER 8, 2006 • VOLUME 281 • NUMBER 49  , n ϭ 3). The mean fluorescence intensity of cells labeled with the oxidation-insensitive probe (C-369) was unchanged in the presence or absence of dicumarol (Fig. 7D). These results demonstrate that changes in mean fluorescence intensity seen with dicumarol treatment when the  cells were labeled with the oxidation-sensitive probe were indicative of changes in quantities of the dye being oxidized and are not due to changes in uptake, ester cleavage, or efflux of the probe. Thus, increases in mean fluorescence intensity seen in Fig. 7C and the subsequent decrease in mean fluorescence intensity with antioxidants can be interpreted as indicative of increases in steady-state levels of intracellular prooxidants during dicumarol treatment.

DISCUSSION
Our data support the hypothesis that mitochondrial production of reactive oxygen species mediates dicumarol-induced cytotoxicity in cancer cells. Mitochondrial electron transport chain blockers enhance dicumarol-induced cytotoxicity and superoxide production, whereas cancer cells with fully functional mitochondrial electron transport chains significantly contribute to dicumarol-induced cytotoxicity and increased levels of reactive oxygen species. Overexpression of antioxidant enzymes targeted to the mitochondria ameliorates dicumarolinduced cytotoxicity and production of reactive oxygen species.
Our present study correlates with previous studies demonstrating the increased susceptibility of cancer cells to dicumarol. Cross et al. (32) demonstrated simultaneous inactivation of stress-activated protein kinase and NFB activation while enhancing apoptotic cell death in dicumarol-treated cells. Previous studies from our laboratory also demonstrated in increases in cancer cell death both in vitro and in vivo after dicumarol treatment (13,18). Additionally, Cross et al. (32) demonstrated that expression of the NQO1 gene reduced sensitivity to stress-activated protein kinase inhibition after dicumarol treatment. Using lower doses of dicumarol (50 M), our study also demonstrated a slightly greater, but not significant, prevention of dicumarol-induced cytotoxicity in MDA-MB-468 cells that overexpress NQO1. This may be explained by recent reports from Siegel et al. (30) that demonstrate the antioxidant properties of NQO1.
Oberley et al. (33) have suggested that tumor cells have increased steady-state levels of intracellular superoxide and hydrogen peroxide associated with aberrant respiration, which could trigger immortalization, uncontrolled cellular proliferation, and the development of the malignant phenotype through the activation of signaling pathways. Mitochondrial metabolism appears to be integrally related to the metabolic production of intracellular hydroperoxides formed as by products of mitochondrial electron transport chain activity (31,34). Given that mitochondrial metabolism appears to be involved with the production of intracellular hydroperoxides and other studies have suggested that cancer cells demonstrate increased intracellular hydroperoxide production (35), tumor cells may increase mitochondrial intracellular hydroperoxide production as a consequence of mitochondrial respiration (31,34). Our study suggests that dicumarol has potential as a therapeutic intervention designed to manipulate mitochondrial electron transport chainmediated oxidative energy metabolism, and mitochondrial hydroperoxide production would be predicted to kill tumor cells versus normal cells via metabolic oxidative stress (31,34).
Mitochondrial electron transport chain blockers have been utilized to determine the various complexes in mitochondria that produce superoxide and hydrogen peroxide (34,36,37). Our present study demonstrates that all mitochondrial electron transport chain blockers tested enhanced dicumarol-induced cytotoxicity and superoxide production. Furthermore, cancer cells with fully functional mitochondrial electron transport chains are significantly more sensitive to dicumarol-induced cytotoxicity and increased levels of reactive oxygen species. In general, mitochondria have been shown to increase their production of O 2 . when exposed to different mitochondrial electron transport chain blockers such as antimycin A, myxothiazol, and rotenone (36,37). This is believed to occur because blocking the mitochondrial electron transport chains effectively inhibits the four-electron reduction of O 2 at complex IV, and the only places for electrons that enter the electron transport chains to exit are sites where one-electron reductions of O 2 can occur.
Our results suggest that when superoxide production is increased from any of the METCs, that dicumarol cytotoxicity is increased to a greater extent in cancer versus normal human cell. These data also provide support for the speculation that fundamental differences between cancer versus normal cell mitochondrial superoxide production may contribute to differential susceptibility to dicumarol toxicity that could be used to gain a therapeutic advantage when treating human cancers with cytotoxic therapies.