725 Involvement of the Mitochondrion Respiratory Chain in the Synergy Achieved by Treatment of Human Ovarian Carcinoma Cell Lines with both Tumor Necrosis Factor-a and Cis-Diamminedichloroplatinum Ruchan Uslu, M.D.’ Benjamin Bonavida, Ph.D.’ Address for reprints: Benjamin Bonavida, Ph.D., Department of Microbiology and Immunology, UCLA School of Medicine, Center for Health Sciences, 10833 Le Conte Avenue, Los Angeles, CA 90095-1747. BACKGROUND. Previous studies have demonstrated that treatment of human tumor cell lines with a combination of cis-diamminedichloroplatinum (CDDP) and tumor necrosis factor-a (TNF-a) results in additivelsynergistic cytotoxic effects and reverses tumor cell resistance to TNF drugs. Free radical intermediates are induced by both TNF-a and CDDP; however, the role of free radicals in synergy is not known. This study investigated the effect of two inhibitors on synergy, phenoxan (Phe) and butylated hydroxyanisole (BHA), which inhibit Complex I and Complex 1 and 11 of the mitochondrion respiratory chain, respectively. METHODS. Three human ovarian carcinoma cell lines of different sensitivity to TNF-a andlor CDDP were selected for the study and consisted of 222, a TNFl CDDP-sensitive line, 222TR (TNF-resistant), a TNF-resistant, CDDP-sensitive line, and AD10, a TNF-sensitive, CDDP-resistant line. Cytotoxicity was determined by the microculture tetrazoliuni dye assay. RESULTS. Synergy in cytotoxicity was achieved in all three lines treated with a combination of TNF-a and CDDP. Cytotoxicity by either TNF-u or CDDP or by both TNF-a and CDDP was inhibited in the presence of either Phe or BHA. Pretreatment of tumor cells with either Phe or BHA for up to 4 hours, washed and followed by the addition of the cytotoxic agents (alone or combined), resulted in no inhibitory effect. Pretreatment of the cells with the cytotoxic agent for up to 2 hours, washed and then followed by the addition of Phe, resulted in significant inhibition of cytotoxicity. In contrast to Phe, the addition of BHA as late as 12 hours post pretreatment of the cells with the cytotoxic agent(s) still inhibited cytotoxicity. These results demonstrated that free radicals are involved in cytotoxicity mediated by a single agent, and in synergy with both agents. Further, the results demonstrated that Phe acts at an early stage of the cytotoxic pathway and that BHA acts at both an early and a late stage of the cytotoxic pathway. CONCLUSIONS. These results demonstrated that both TNF-a and CDDP rapidly stimulate the induction of free radicals but a lag of several hours was necessary to initiate the irreversible program of cell death. Further, the studies demonstrated that synergy and reversal of drug resistance in ovarian tumor cells by TNF-CY and CDDP, used in combination, share the same pathway of cytotoxicity as that mediated by TNF-a or CDDP used as a single agent. Cancer 1996; 7Z725-32. 0 1996 Ainericun Cancer Sociefy. Received February 16, 1995; revisions received August 14, 1995, and October 16, 1995; accepted October 16, 1995. KEYWORDS: cytotoxicity, ovarian carcinoma, CDDP, TNF-a, synergy, resistance, free radicals. Department of Microbiology and Immunology, UCLA Sc:hool of Medicine, University of California at Los Angeles, Los Angeles, California. Supported in part by a grant from the Boiron Research Foundation and a grant from Yuksek Ogretim Kurumu, lurkey (RU). Q 1996 American Cancer Society 726 CANCER February 15,1996 / Volume 77 / Number 4 R ecent studies from our laboratory have reported that resistance of human tumor cells to drugsltumor necrosis factor (TNF)' can be reversed by the synergistic cytotoxic activity of TNF-a and drugs.' ' In some studies, we reported that synergy by TNF-a and adriamycin (ADR)/cis-diamminedichloroplatinum(CDDP) is not influenced by the expression of gp170 in a multidrug resistance phenotype (MDR)' cell line but correlated with the expression of TNF messenger ribonucleic acid (mRNA).'--!TNF-a is a cytokine with a wide range of biologic activities. It was first described in serum of endotoxin-treated mice that can induce tumor necrosis in vivo and selectively kills transformed and neoplastic cell lines in vitro.' TNF-a has two types of receptors and its multiple activities are mediated by these two receptors. The surface expression of TNF-a receptors is necessary but not sufficient for the biologic activities of TNF-CX.~ Among the well-known effects of TNF-a are its oxidative damageG and DNA fragmentation.' Moreover, anaerobic conditions," glucocorticoids," and antioxidants' can block TNFa-mediated killing. These studies suggested that the generation of reactive oxygen intermediates by TNF-a may be responsible for damaging cellular macromolecules such as unsaturated lipids, DNA, RNA, and proteins and further, free radicals may enhance enzymes that are involved in TNF-cu-mediated cytotoxicity pathway."' One of the most important sources of reactive-oxygen intermediates is the mitochondria." Recently, it has been shown that a functional mitochondrial respiratory chain is crucial to the events leading to TNF-a-cytotoxicity in the classical TNF-a-sensitive fibrosarcoma cell line L929 and in the WEHI164 clone 13." Also, mitochondrial intermediates control both the cytotoxic and gene regulatory effects of TNF-a.' ' Other investigators have shown that a high level of manganese-superoxide dismutase (Mn-SOD), an important mitochondrial antioxidant enzyme, is up-regulated in TNF resistant lines and/or it can be induced by TNF-cy in several tumor cell lines.'4 Conversely, inhibition of Mn-SOD and reduction of glutathione levels lead to augmentation of TNF-a cytotoxicitY.''.l4 Cisplatin is a widely used chemotherapeutic agent in the treatment of several tumors such as testicular, ovarian, head and neck, and small cell lung cancer. However, drug resistance is a major obstacle in cancer treatmentt5 Oxidative damage has been implicated in cytotoxicity by cisplatin."' It has been shown that cellular resistance to oxidative stress is accompanied by resistance to cisplatin." It has also been shown that the enzymatic activities of catalase, glutathione peroxidase and glutathioneS-transferase were significantly decreased whereas Mn-SOD activity increased gradually after CDDP adniinistration." Studies on the subcellular distribution of cis- FIGURE 1. Schematic diagram of the mitochondrion electron transport system. platin revealed that the mitochondrion and the cytosol accuniulate the highest concentration of platinum in the cell.'" Also, cisplatin inhibits enzymes of the mitochondrial respiratory chain." CDDP increases the accumulation of mitochondrial calcium,'' and can cause damage to the mitochondrial DNA." Since cytotoxicity by both TNF-cy a n d cisplatin involve the mitochondria, we hypothesized that both synergy a n d reversal of drug resistance of ovarian carcinoma cell lines by TNF-a a n d CDDP might b e sharing a common pathway(s) of reactive-oxygen-mediated cytotoxicity. This hypothesis was tested by examining the role of mitochondrial dysfunction in the generation of reactive-oxygen intermediates by two inhibitors, namely, phenoxan (Phe)'"'' a n d butylated hydroxyanisole (BHA), antioxidant'5 a n d potent mitochondrial electron transport inhibitors at site of Complex I and 11, respectively (Fig. 1 ) . MATERIALS AN0 METHOOS Tumor Cells Human tumor cells were obtained from malignant ascites of a patient with a n ovarian tumor and were established as a cell line named 222. Then, a TNF-resistant subline of 222, named 222TR. was established by culturing it in increasing concentrations of TNF-a.' The 222TR cell line is TNF-resistant but remains sensitive to cisplatin. The ADlO ovarian carcinoma cell line was obtained from Dr. Robert Ozols, Philadelphia, PA. The A D l O cell line is an adriamycin resistant variant developed from the ovarian line A2780, and is an MDR' line, and sensitive to TNF-a cytotoxicity. The three tumor cell lines are adherent in culture and are maintained in 10% heat inactivated bovine calf serum (Gibco, Grand Island, NY) added to RPMI1640 medium (Gibco) with 1% pyruvate (Gibco), 1% nonessential amino acids (Gibco) and 1% Fungi-bact solution (Irvine Scientific, Santa Ana, CA) which contains 10,000 Ulml penicillin G , 10 ing/ml streptomycin and 25 Role of the Mitochondrion in Resistance/Uslu et al. pglml Fungizone. When the tumor cell lines were used as target cells, they were first detached with TrypsinEDTA (ICN Biomedicals, Costa Mesa, CA), washed, and resuspended in complete medium. 222 I CDDPluM) 222 losssl TNF(pM) 727 1033 Reagents TNF-a w,as purchased from Peprotech (Nutley, NJ). TNFa was diluted in PBS at 10 pg/ml and stored at 4" C. CDDP (Lot. No. 103H3451) and butylated hydroxyanisole (BHA) (Lot. No. 12H0147) were purchased from Sigma, St. Louis, MO. The phenoxan (Phe) was a gift from Dr. Elena Burlakova (Moscow, Russia). Stock solutions of CDDP, BHA, and phenoxan were prepared in DMSO (Sigma, Lot. No. 93H4650) and the DMSO concentration in the assay did not exceed 0. I % and was not cytotoxic to the tumor cells. Cytotoxicity Assay The microculture tetrazolium dye (MTT) assay was used to determine cytotoxicity as described elsewhere.' Briefly, target tumor cells were resuspended in medium at lo5 cells/ml after verifying the cell viability by trypan blue dye (Sigma) exclusion. One hundred pl of cell suspension was distributed into each well of a 96-well flat bottom microtiter plate (Costar, Cambridge, MA) and each plate was incubated for 24 hours at 37" C in a humidified 5% CO, atmlosphere to allow cells to proliferate. The next day, 100 ,ul of reagent solutions at the desired concentrations, or medium alone used for positive control, were distributed into each well. 'Two hundred p1 of the medium alone without cells and reagents were also distributed for use as negative control. Then, the microtiter plate was incubated for the desired periods of time. Thereafter, 20 p1 of MTT dye (5 mg/ml) (Sigma) was added into each well. The unreactive supernatants in the wells were carefully aspirated and replaced with 100 p1 of isopropanol (Sigma) supplemented with 0.05N HC1 to solubilize the reactive dye. The OD values of each well at 540 nm were read using an automatic multiwell spectrophotometer (Titerek Multiscan MCC/340). The negative control well was used for zeroing the OD. The percent cytotoxicity was calculated using the background-corrected OD as follows: Percent of Cytotoxicity 11 ~ = (OD of experimental wells/OD of positive control wells)] x 100. Statistical Analysis All assays were set up in triplicates, and the results were expressed as the mean plus or minus the standard deviation (SD:I (mean -C SD). Statistical analysis was determined b y the Student's t test. Synergy was calculated 0 I 4 180 400 0 I 222TR 4 160 400 222TR 80 70 -10 -10 0 4 1 0 400 AD10 0 4 180 400 AD10 .0 160 4 400 PHE FIGURE 2. Inhibitory effect of Phe on TNF-a and CDDP-mediated cytotoxicity against 222, 222TR, and AD10 cell lines. Cytotoxicity was determined in a 24-hour MTT assay. The results are derived from three different experiments (standard deviation < 5). and illustrated by a modified isobole analysis based on Berenbaum's methods.26 RESULTS Inhibition by Phe and BHA of TNF-a and CDDP-Mediated Cytotoxicity on Human Ovarian Carcinoma Tumor Cell Lines Three human ovarian carcinoma tumor cell lines were selected for the study, namely 222 (TNF-a, CDDP-sensitive), 222TR (TNF-resistant, CDDP-sensitive), and AD10 (TNF-a-sensitive, CDDP-resistant). The addition of Phe inhibited TNF-a cytotoxicity of all three lines tested, with maximal inhibition obtained at concentrations of Phe 2 160 pM and TNF-a < 588 pM (Fig. 2). The addition of Phe also inhibited CDDP-mediated cytotoxicity at concentrations 2 160 pM (Fig. 2). Optimal inhibition of TNFa and CDDP cytotoxicity was achieved at BHA concentrations > 100 pM for all three lines (Fig. 3 ) . These results demonstrate that inhibition of free radicals by Phe and BHA abrogates both TNF-a and CDDP cytotoxicity. Based CANCER February 15,1996 / Volume 77 / Number 4 728 222 -10 222 TNF(pH 0 0 033 33 0 ~. ~ 02 04 06 08 ~~ 1 ~ 222TR -10 1w Zen 10 1 222 CDDP(uMJ J 0 10 1w 200 222TR 222TR 80 70 0 40 E* -10 0 33 AD10 30 70 33 0 02 04 06 08 1 - t AD10 - 20 10 0 -10 I 0 10 1w 200 AD10 0 10 100 200 AD10 0 0 33 CDDP ["MI 0 02 04 06 F I C of CDDP 08 I FIGURE 4. Synergy of cytotoxicity of AD10 by combination of TNF and CDDP. Synergy is demonstrated in the right side by isobolograph analysis. 0 10 100 2w BHA FIGURE 3. Inhibitory effect of BHA on TNF-cr and CDDP-mediated cytotoxicity against 222, 222TR, and AD10 cell lines. Cytotoxicity was determined in a 24-hour MTT assay. The results are derived from three different experiments (standard deviation > 5). on the optimal concentration of inhibition, a constant concentration of Phe at 400 p,M and of BHA at 200 pM were used in subsequent experiments. All three lines tested showed synergy when TNF-a and CDDP were used in combination (Fig. 4). Synergy was achieved with low concentrations (58.8 pM TNF-a and 0.33 pM CDDP) and higher concentrations (294 pM TNF and 3.3 ph4 CDDP) (data not shown) of the agents. The addition of Phe or BHA significantly inhibited the synergistic activity of TNF-a and CDDP used in combination (Fig. 5). There were no noticeable differences in the degree of inhibition obtained in all three lines tested. Mechanism of Phe and BHA-Mediated Inhibition of Cytotoxicity Phe inhibits the mitochondria1 electron transport at the site of Complex I and B H A acts at site of Complex I and I1 (Fig. 1 ) . We wished to examine how these two inhibitors, which act at different steps in the metabolic pathway involved in free radical formation and detoxification, block cytotoxicity. Initially, experiments were performed whereby sequential treatments were done for various periods of time. Tumor cells were treated with Phe for a period of 0.5-4 hours, washed, and TNF-a, CDDP, or a combination of TNF-a and CDDP were added for the duration of the MTT cytotoxicity assay. There was no detectable inhibitory effect of pretreatment of tumor cells with Phe on cytotoxicity by TNF-a, CDDP, or a conibination on all three cell lines tested. Also, there were no differences when different concentrations of the agents were used. Similar findings to those obtained with Phe were also observed with BHA. The time kinetics of addition of inhibitors subsequent to pretreatment with the cytotoxic agent(s) was examined. Pretreatment with TNF-a, CDDP, or a combination for up to two hours, and wash, and addition of Phe subsequently for the remaining of the assay, resulted in inhibition of cytotoxicity (Table 1). However, a fourhour pretreatment with TNF-a, or a CDDP four-hour pretreatment when Phe was added four hours after, the inhibitory effect was no longer detected (Table 1, Fig. 5). These results demonstrate that Phe acts at an early step in the pathway of cytotoxicity after initiation of the cytotoxic pathway. After two hours, the cells are not inhibitable by Phe and thus, the pathway for cytotoxicity could not be reversed. Unlike Phe, BHA reversed cytotoxicity up to 12 hours following pretreatment of tumor cells with the cytotoxic agents. Pretreatment of tumor cells for four hours with Role of the Mitochondrion in Resistance/Uslu et al. ADIO AD10 ADIO 45 1 r-- 0 45 EM(+) 5 -5 -5 0.5h ih 2h 4h 4h 6h 8h 12h CDDP+BHA CDDP+PHE AD10 AD10 50 H y -- - .Phe(+) 10 0 CISh ih 2h 4h Time _ _ _ ~ 4h 6h 8h Time 'TNF+CDOP+PHE way by which synergy is achieved by TNF-a and CDDP involves the mitochondrion as two anti-oxidant inhibitors abrogated cytotoxicity. Furthermore, the data suggest that induction of cytotoxicity is as late as 12 hours following treatment and, prior to that time, cytotoxicity can be blocked. Two sites of the mitochondrial electron transport chains have been identified as sources of superoxide radicals. The first is dependent on the auto-oxidation of the flavin mononucleotide from reduced nicotinamide adenine dinucleotide (NADH)-dehydrogenase (Complex I), and the second is dependent o n the autooxidation of the unstable ubisemiquinone (Complex III)." The generation of oxygen radicals at the ubiquinone site is supported by NADH-linked substrates and is decreased by inhibitors of Complex 120,29(Fig. 1). Schulze-Osthoff et al".'' have shown that inhibition of the production of oxygen radicals at the site of Complex I inhibits TNF-@-mediated cytotoxicity. Other investigators have shown that antioxidants and free radical scavengers protect cells against TNF-a-induced cytolysis.R.'2.'"Likewise, oxidative damage has also been implicated in cytotoxicity by CDDP.'"'" Also, CDDP has been shown to have a n effect on the mitochondrial membrane, the respiratory chain, and the mitochondrial DNA."-" The phenolic compound 2,6-di-tert-butyl+methoxy-phenol (butylated hydroxyanisole1BHA) is a widely used antioxidant to prevent foods from oxidative deterioration and rancidity. In addition to its preventive effect on lipids and proteins from peroxidation,25BHA has been shown to inhibit the mitochondrial electron transport system3"."' at the site of Complex I and Complex 11. Further, BHA has been shown to increase glutathione and glutathione-s-transferase levels,'* to increase gammaglutamylcysteine synthetase activity,"3 to increase cytosolic free calcium levels,34 and to induce certain oncogenes.35 These studies demonstrated that BHA acts at both the production of free radicals and at the detoxification of free radicals. In contrast, Phenoxan is a new oxazole-pyrone from myxobacteria and inhibits the mitochondria] electron transport system at the site of Complex I and acts primarily on the production of free radicals.""?' The present study demonstrates that both TNF-a and CDDP-mediated cytotoxicity and synergy by TNF-a and CDDP are inhibited by both Phenoxan and BHA, suggesting a common initial trigger of cytotoxicity. The results also demonstrate that induction by the cytotoxic agents must precede the block by the inhibitors suggesting d e novo induction of free radicals. Thus, modulation of the mitochondria1 respiratory chain by TNF-a and CDDP is a n early event in the cytotoxic pathway. Phenoxan acts at this early stage of the cytotoxic pathway, and is effective only for 2 hours following which the cells ;VTri AD10 i2h TNF+CDDP+BH FIGURE 5. Effect of sequence of treatment of AD10 tumor cells with Phe or BHA and TNF-a. CDDP, or TNF and CDDP on cytotoxic activity. The tumor cells were pretreated with TNF-a (58.8 pM), CDDP (0.33p M ) , or both for desired periods of time. The medium was aspirated and the tumor cell:; were washed twice with RPMl medium. The cells were then incubated with Phe (400 p M ) or BHA (200 pM) for desired periods of time. Cytotoxicity was assessed in a 24-hour MTT assay. The results are derived from three different experiments (SD < 5). the cytotoxic agents and wash was sufficient to trigger cytotoxicity to the same extent as cells pretreated for 12 hours or longer (Table 1 and Fig. 5). These results demonstrate that BHA acts at an early and at a late stage of the pathway of cytotoxicity (i.e. both Phe and BHA block the inductive phase while BHA also blocks a post inductive phase of cytotoxicity). DISCUSSION Studies have reported that TNF-a and CDDP used in combination can exert a synergistic cytotoxic activity against 1°F-a andlor CDDP-resistant human tumor cell lines and can also reverse tumor cell resistance to TNFN andlor CDDP. The mechanisms of synergy and overriding tumor cell resistance, however, have not been elucidated. The present study provides evidence that the path- 729 730 CANCER February 15, 1996 / Volume 77 / Number 4 TABLE 1 Effect of Sequential Treatment of TNF-a or CDDP with Wash on the Inhibitory Effect of PHE and BHA Firstlsecond treatment Tumor 222 Time of 1st treatment (h) 0.5 1 7 222TK AD10 4 0.5 1 2 4 0.5 Tie of 1st TNFl medium TNFlPhe Phe'lTNF CDDPl medium CDDPlPhe Phe"l CDDP TNF t CDDPl Med TNF t CDDPl Phe Phea TNF t CDDP 4.1 i 2.1 5.2 ? 2.5 9.2 2 3.5 11.0 2 1.8 1.1 t 2.8 1.2 ? 1.7 1.8 2.9 3.1 i 1.3 5.8 i 3.2 6.2 t 2.3 8.8 i 2.7 12.1 ? 2.7 0.3 i 2.1 0.9 ? 2.1 5.5 i 1.7 11.1 i 2.4 1.0 2 3.1 0.8 1 2.1 0.2 i- 1.1 3.2 i- 2.0 -1.2 2 3.1 2.3 2 1.2 4.2 i 0.8 12.3 i 2.4 13.5 ? 2.7 13.7 t- 1.9 12.5 2 1.7 12.7 2 1.3 4.5 i 2.5 4.2 t 2.4 3.7 2 1.8 4.3 i 0.7 14.3 i 1.7 14.0 t- 3.3 13.5 2 2.2 13.3 i 1.2 6.1 1 1.3 7.8 i 2.9 10.9 i- 2.4 12.1 I 1.5 7.0 1 0.7 8.2 I1.9 9.5 i 1.9 13.2 i 1.0 O . i 5 1.2 1.0 i- 0.3 1.5 t 2.7 2.2 i- 1.8 1.8 t 1.0 3.2 ? 2.6 8.1 i 3.2 12.3 1' 2.1 0.8 ? 2.7 1.1 i 2.2 6.9 i 2.8 13.1 2 3.4 -0.7 2 1.4 -0.3 t 1.4 0.9 2 2.8 2.4 2 3.0 14.0 i- 1.4 13.9 ? 1.7 13.7 ? 1.9 13.0 i 2.6 14.9 i- 1.5 14.5 2 2.3 13.4 i 1.7 14.0 t 2.4 3.8 t- 0.4 3.7 2 1.8 2.9 i 1.0 2.9 i 3.4 26.9 i 3.1 28.1 ? 3.3 39.8 2 3.0 40.9 ? 2.7 18.2 2 3.0 20.9 2.1 26,8 t 0.8 36.1 ? 1.8 15.5 2 2.1 17.5 1.3 24.9 t- 1.3 34.8 2 2.1 6.8 I 2.4 10.5 2 2.7 2 0.5 2 3.2 41.1 i 1.9 3.2 2 1.0 4.1 2 3.0 15.9 i 2.7 36.5 i 1.8 0.6 5 1.9 3.5 2 2.8 13.7 i 1.8 34.5 i 2.3 45.0 i- 3.2 44.3 t- 1.2 43.3 2 1.1 42.4 t 1.9 38.9 t- 2.2 38.3 i- 1.7 36.3 ? 3.2 36.9 i 1.7 38.4 t 2.0 36.8 i- 2.0 35.1 i 2.4 36.2 i 2.4 +_ treatment (h) TNFl medium TNF/BHAa BHA/TNF CDDPl medium CDDPl BHA BHA/ Tumor CDDP TNF t CDDPl med TNF t CDDPl BHA BHA/TNF t CDDP 222 4 8 12 4 8 12 4 8 12 12.0 i- 2.1 12.3 2 2.3 11.9 2 3.7 2.9 i 1.7 3.1 1' 1.6 2.9 2 0.4 13.9 t 1.8 13.1 2 1.3 12.1 t 1.7 -2.8 2 1.7 -2.2 i 0.7 -0.9 t 0.8 -2.7 ? 1.1 -2.0 % 1.3 -1.0 2 3.0 -2.3 i- 0.7 -1.9i 1.8 -0.6 2 2.2 12.3 2 2.8 12.6 ? 2.7 11.5 t- 0.5 3.2 i 1.5 2.9 i- 0.9 2.9 i 2.1 14.3 t 2.2 13.8 2 1.8 13.2 i 1.1 12.7 i 1.0 11.6 i 0.7 11.7 t 2.6 12.5 i 2.0 12.9 i- 1.0 12.9 i 2.4 2.4 t 1.6 2.6 i- 0.6 2.2 i 1.6 -2.7 -2.1 -1.1 -2.4 -1.7 -0.5 -2.2 -1.2 -0.8 13.1 2 1.0 12.2 ? 1.1 11.2 5 0.9 14.0 i- 0.7 13.0 ? 1.7 11.8 t 1.7 3.1 t- 1.9 1.2 ? 0.7 2.8 1' 1.7 45.0 i 2.3 41.0 i 2.4 40.9 t 1.7 33.7 ? 0.9 14.8 ? 1.3 33.6 2 1.1 32.0 I 2.1 31.9 i 1.7 31.9 t 2.8 5.4 2 2.2 8.1 i 1.0 9.6 5 2.6 -1.9 i- 2.9 1.0 i 3.4 5.3 2 2.2 -1.4 2 0.8 -1.0 ? 2.3 6.9 2 1.2 46.0 i- 3.2 41.8 2 0.9 40.3 i 3.1 37.9 ? 3.5 35.9 i 3.0 32.7 2 2.8 33.1 2 3.4 33.8 2 1.4 32.3 2 1.8 222TR AD10 2 2.1 t- 2.1 i 2.0 2 2.4 2 1.6 ? 1.3 i 1.8 t- 1.9 ? 1.1 - "First treaunent with Ilhe or Btl4 showed the same findings as first treatment with medium alone i.e. Phe and BHA are not toxic to the cells. The bold nuinbers represent the inhibition mediated by the addition of Phe. The cells w e pretreated with medium only, TNF-n 158.8 pM), CDDP 10.33 pM),Phe 1400 pM1 or BHA 1200 phl) or TNF-a 158.8 pM) t CDDP 10.33 pMJ for desired periods of time (first treatment). The medium was aspirated and the cells were washed twice with RPMl medium. I h e cells were then incubated with medium only, TNF-n (58.8 pMI, CDDP 10.33 !MI, Phe (400 phl] or BHA (200 pMi or TNF-a (58.8 pM1 t CDDP 10.33 pMI for desired periods of time [second treatment). Cytoroxicity was assessed in one day MTT assay. The results are erpressed as the mean S.D. are programmed for cell death and cannot be reversed. However, in addition to what Phe does, BHA also has a free radical scavenger effect which prevents peroxidation of lipids and proteins. This preoxidation stage appears to be a late stage of the cytotoxic pathway initiated by the free radicals. Hence, BHA can block the synergistic cytotoxicity as late as 12 hours post treatment of the cells with the cytotoxic agent(s). Clement et a13' have shown that DNA fragmentation by anti-fas antibody takes place in less than 2 hours, but that TNF-mediated DNA fragmentation can occur as late as 24 hours. It is generally accepted that DNA fragmentation is irreversible and cannot be b10cked.~'In our experiments, no significant DNA fragmentation was detected up to 12 hours following treatment with a concentration of CDDP or TNF-a (data not shown), and accordingly, BHA could reverse CDDP and TNF-a-mediated cytotoxicity up to 12 hours prior to irreversible DNA damage. Tumor cells develop resistance to chemotherapy. Ovarian cancer is a model disease to investigate chemotherapeutic resistance because of both its intrinsic and its acquired resistance to drugs. Approximately 75% of patients have surgically-incurable disease at time of diagnosis, thus chemotherapy plays a major role in the management of ovarian cancer.38 Aggressive treatment of these patients with platinum-based combination chemotherapy usually results in severe adverse effects and it is rarely curative. Therefore, effective subtoxic therapeutic modalities are needed. Such modalities include biological response modifiers, cytokine based chemotherapy, passive administration of tumor infiltrating lymphocytes, genes therapy, and targeting of tumor cells with immunotoxins. In addition, the development of drugs that inhibit cellular anti-oxidant pathways should reverse tumor cell sensitivity to cytotoxic agents. Recent studies have shown that treatment of tumor Role of the Mitochondrion in Resistance/Uslu et al. cells with combinations of biologic response modifiers such as IFN-a, IL-2, and TNF-a and drugs or toxins results in reverse drug resistance and often, These combination treatments were also effective at concentrations that were subtoxic. Such treatment modalities may also boost the immune system for a more effective and complimentary antitumor immune response. REFERENCES I. 2. 3. 4. 5. 6. 7 I . 8. 9. 10 11 12. 13. 14. 15. 16. Mizirtani Y. Bonavida B. Overcoming CDDP resistance of human ovarian tumor cells by combination treatment with CDDP and TNF-u. Cnrzr-er 1993:72:809-18. Ronavida B, Tsuchitani T. Zighelboim 1, Berek J . Synergy is documented in vitro with low-dose recombinant human tumor necrosis, factor, cisplatin, and doxorubucin in ovarian cancer cells. Gyriecol Oiicol 1990;38:333-9. Mizutani Y. Bonavida B. Overcoming TNF-a and CDDP resistance of B htiman ovarian cancer cell line (C30) by treatment with buthionine sulfoximine in combination with TNF-a andlor CDDP. h i t 1 Oncol 1993;3:229-35. Carswell EA, Old I J , Kassel RL, Green S, Fiore N , Williamson B. Am endotoxin-induced serum factor that causes necrosis of tumors. Proc Nntl Acntl Sci U S A 1975;72:3666-70. Shalaby MR, F'alladino MA lr., Hirabayashi SE, Essalu TE, Lewiis (XI, Shepard HM, et al. Receptor binding a n d activation of polymorphonuclear neutrophils by tumor necrosis factor-alpha. 1 I.etrkac Biol 1987;41:196-204. Zimmerman RJ,Marafino BJ Jr., Chan A, Landre P. Winhelke 11.. The role of oxidant injury in tumor cell sensitivity to recombinant human tunior necrosis factor in vivo. J f r n r n t r rzml 1989; 142:1405-9. Schmid DS, Hornung R, McGrath KM, Paul N , Ruddle NH. 'l'arget cell DNA fragmentation is mediated by lymphotoxin and TNI:. Lyinphokine Cyrokine Res 1987;6:195-202. M at t h ew N , Neal ML, Jackson SK, Stark JM. Tumor cell killing by tumor necrosis factor: inhibition of anaerobic conditions, free radical scavengers, and inhibitors of arachidonatr nietabolism. Ininiiirrology 1987;62: 153- 5. Tsujimoto M , [Dkamura N. Adachi H. Dexamethasone inhibits tlhe cytotoxic activity of TNF. Biocliem Biophys Kes Cornn ~ ~ i 1988; r? 153:104-15. Jones CRN. Free radicals in immunological killing: the case of TNF. Merl l-lypotheses 1986:21:267-71. Boveris A. Mitochondria1 production of superoxide radical and hydrogen peroxide. In: Reivich M, Coburn R, Lahiri S, Chance H. editors. Tissue liypoxia and ischemia. New York: l'leiiuni Press, 1977:67-82. Schltilze-Osthoff K, B a k e r AC, Vanhaesbroeck B, Beyaert R, Jacob WA, Fiers W. Cytotoxic activity of TNF is mediated by early damage of mitochondria1 functions. I Biol Chem 1992;267(8):5317-23. Schulze-Osthoff K. Beyaert R , VandervoordeV, Haegernan G, Fiers W. Depletion of the mitochondria1 electron transport abrogates the cytotoxir and gene-inductive effects of TNF. EM.RO I 1993; 12(8):3095-3104. Woing GHW, Elwell JH, Oberley WL, Goeddel DV. Manganese superoxide dismutase is essential for cellular resistance to cytotoxicity of 'INF. Cell 1989;58:922-31. Timmer-Hosscha H,Mulder NH. deVries EGE. Modulation of CDDP resistance: a review. Br I Cnncer 1992;66:227-38. Hannernan I, Baumann K. Cisplatin-induced lipid peroxidation and decrease of gluconeogenesis in rat kidney cortex: 731 different effects of antioxidants and radical scavengers. Toxicology 1988;51:119-32. A d a m DT, Sherman CM, Deen DF, 17. Spitz DR, Phillips JW, Li GC. Cellular resistance to oxidative stress is accompanied by resistance to cisplatin. The significance of increased catalase activity and total glutathione in H202 resistant fibroblasts. J Cell Physiol 1993; 156:72-9. 18. Sadzuko Y,Shoji T, Takino Y. Effect of cisplatin on the activities enzymes which protect against lipid peroxidation. Biothem Plinrmacol 1992;43:1872-5. 19. Levi J, Jacobs C, Kalman SR, Mctigue M, Weiner MW. Mechariisrn of cisplatinum nephrotoxicity: 1. Effect of sulthydryl groups in rat kidneys. J P1inrnzncol Exp Tlzer 1988:213:545- -- 50. 20. Rosen M, Figliorneni M, Simpkins H. The interaction of platinum antitumor drugs with mouse liver mitochondria. h i t J EX^ P ~ t l ~ 1992;73:61-74. ol 21. Gordon JA, Gattone VH 11. Mitochondria1 alterations in cisplatin-induced acute renal failure. An? J Physiol 1986; 250:F991-8. 22. Singh G, Maniceia-Bozzo E. Evidence for lack of mitochondrial DNA repair following cis-dichlorodiammine platinum treatment. Cancer Cliemotlier Phnrrnncol 1990;26:97- 100. 23. Kunze B, Jansen R, Pridzun L, Jurkiewicz E, Hunsmann G, Hofle G, et al. Phenoxan: a new oxazole-pyrone from myxobacteria: production, antimicrobial activity and its inhibition of the electron transport in Complex I of the respiratory chain. J Antihior 1992;45(9):1549-52. 24. Friedrich T, Van Heek P, Leif H, Ohnislii T, Forche E, Kunze B, et al. Two binding sites of inhibitors in NADH: ubiquinone oddoreductase (Complex 1). Relationship of o n e site with the ubiquinone-binding site of bacterial glucose: ubiquinone oxidoreductase. Ezir J Biochenz 1994;219:691-8. 25. Verhagen H, Schilderman PAEL, Kleinjans-10s CS. Butylated Hydroxyanisole: in perspective. Bzern Biol fnternct 199 1; 80:109-34. 26. Berenbaum MC. Criteria for analyzing interactions between biologically active agents. Aclv Cnricer Res 1981;35:269-77. 27. Turrens JF, Alexandre A, Lehninger AL. Ubisemiquinone is the electron donor for superoxide forination by Complex 111 of heart mitochondria. Arclz Biocheni Biophys 1985;237:40814. 28. Cino M. Del Maestro RF. Generation of hydrogen peroxide by brain mitochondria: the effect of reoxygenation following postdecapitative ischemia. Arch Bioclienz Bioplzys 1989; 2691623-38. 29. Cadenas E, Boveris A. Enhancement of hydrogen peroxide formation by protophores and ionophores in antirnycinsupplemented mitochondria. Bioclzern J 1980; 188:431-7. 30. Ferreira J. Effect of RHA o n electron transport in rat liver mitochondria. Bioclzern Pharnzacol 1990;40(4):677-84. 31. Aldunate J , Ferreira J, Letelier ME, Repetto Y, Morello A. tbutyl-4-hydroxyanisole. a novel respiratory chain inhibitor. FEBS Lert 1986; 195:295-7. 32. Sharma R. Ahmad H, Singhal SS, Saxema M, Srivastava SK, Awasthi YC. Comparative studies o n effect of BHA on glutathione and glutathione-s-transferases in the tissues of male 1993; and female CD-I mice. Corizp Biochern Physiol 105(1):31-7. 33. Eaton DL, Hamel DM. Increase in gamma-glutamylcysteine synthetase activity as a mechanism for butylated hydroxyanisole-mediated elevation of hepatic glutathione. 7oxicol AppI Pharnincol 1994; 126(1):145-9. [a 732 CANCER February 15, 1996 I Volume 77 I Number 4 34. David M, Horvath G, Schimke 1, Mueller MM, Nagy I. Effects 38. Perez PR, Hamilton TC, Ozols RF, Young RC. Mechanism of the antioxidant BHA on cytosolic free calcium concentration. Toxicology 1993; 77(1-2): 115-21. 35. Choi HS, Moore DD. Induction of c-fos and c-jun gene expression by phenolic antioxidants. Mol Endocrinol 1993; and modulation of resistance to chemotherapy in ovarian cancer. Cancer 1993;71(Supp1):1571-80. 39. Krosnick ]A, Mule JJ, McIntosh JK, Rosenberg SA. Augmentation of antitumor efficacy by the combination of recombinant TNF and chemotherapeutic agents in vivo. Cancer Res 7( 12):1596- 1602. 36. Clement MV, Stamenkovic I. Fas and tumor necrosis factor- mediated cell death: similarities and distinctions. J Exp Med 1994; 180:557-67. 37. Mogil RJ, Shi Y, Bissonnette RP. Bromley P, Yamaguchi I, Green DR. Role of DNA fragmentation in T cell activationinduced apoptosis in vitro and in vivo. J Immunol 1994; 152:1674-83. 1989;49:3729-33. 40. Morimoto H, Safrit JT, Bonavida B. Synergistic effect of TNF- and diphtheria-toxin-mediated cytotoxicity in sensitive and resistant human ovarian tumor cell lines. J lmmunol u 1991; 147:2609-16. 41. Mulch DG, Powell CB, Kao MS, Collins JL. In vitro analysis of the anti-cancer potential of tumor necrosis factor in combination with cisplatinum. Gynecol Oncol 1989;34:328-33.