Effect of Anthracycl ine Antibiot ics on Oxygen Radical Formation

This investigation examined the effect of the anthracycline antitumor agents on reactive oxygen metabolism in rat heart. Oxygen radical production by doxorubicin, daunorubicin, and various anthracycline analogues was determined in heart homogenate, sarcoplasmic reticulum, mitochondria, and cytosol, the major sites of cardiac damage by the anthracycline drugs. Superoxide production in heart sarcosomes was significantly increased by anthracycline treatment; for doxorubicin, the reaction appeared to follow saturation kinetics with an apparent Kr~ of 1 12.62 #M, required NADPH as cofactor, was accompanied by the accumulation of hydrogen peroxide, and probably resulted from the transfer of electrons to molecular oxygen by the doxorubicin semiquinone after reduction of the drug by sarcosomal NADPH:cytochrome P-450 reductase (NADPH:ferricytochrome oxidoreductase, EC 1.6.2.4). Superoxide formation was also significantly enhanced by the anthracycline antibiotics in the mitochondrial fraction. Doxorubicin stimulated mitochondrial superoxide formation in a dose-dependent manner that also appeared to follow saturation kinetics (apparent Km of 454.55/~M); however, drug-related superoxide production by mitochondria required NADH rather than NADPH and was significantly increased in the presence of rotenone, which suggested that the proximal portion of the mitochondrial NADH dehydrogenase complex [NADH:(acceptor) oxidoreductase, EC 1.6.99.3] was responsible for the reduction of doxorubicin at this site. In heart cytosol, anthracycline-induced superoxide formation and oxygen consumption required NADH and were significantly reduced by allopurinol, a potent inhibitor of xanthine oxidase (xanthine:oxygen oxidoreductase, EC 1.2.3.2). Reactive oxygen production was detected in all of our studies despite the presence of both superoxide dismutase (superoxide:superoxide oxidoreductase, EC 1.1 5.1.1 ) and glutathione peroxidase (glutathione:hydrogen peroxide oxidoreductase, EC 1.1 1.1.9) in each cardiac fraction. These results suggest that free radical formation by the anthracycline antitumor agents, which occurs in the same myocardial compartments that are subject to drug-induced tissue injury, may damage the heart by exceeding the oxygen radical detoxifying capacity of cardiac mitochondria and sarcoplasmic reticulum. I N T R O D U C T I O N The anthracycline antitumor agents doxorubicin, daunorub1 This study was supported by USPHS Grant 31788-02 from the National Cancer Institute, Department of Health and Human Services. Presented in part on May 28, 1980, at the Annual Meeting of the American Association for Cancer Research, Inc., in San Diego, Calif. (13). 2 Recipient of a special fellowship from the Leukemia Society of America. To whom requests for reprints should be addressed, at the Pharmacology Section, Department of Medical Oncology, City of Hope National Medical Center, 1500 East Duarte Road, Duarte, Calif. 91010. Received March 11, 1982; accepted October 28, 1982. icin, rubidazone, and others are among the most effective antineoplastic drugs currently available for the treatment of several different human cancers (7). Unfortunately, the clinical usefulness of these drugs has been diminished by their association with a potentially life-threatening form of cardiac toxicity (28) that in most circumstances limits cumulative drug dosage to the equivalent of 500 to 550 mg of doxorubicin per sq m (28). Although the mechanism of anthracycline cardiac toxicity remains incompletely understood, recent studies have suggested that the cytotoxic effects of these agents may be related to the formation of semiquinone free radical drug intermediates in v ivo (45). Handa and Sato (18), Goodman and Hochstein (17), and Bachur et al. (3-5) have established that hepatic microsomes support the transfer of electrons from NADPH to the quinone moiety of doxorubicin. This is probably due to an interaction between the anthracycline ring and the flavin component of microsomal NADPH:cytochrome P-450 reductase (24). Under anaerobic conditions, this interaction leads to the enzymatic cleavage of the anthracycline glycosidic bond (26). Aerobically, the anthracycline semiquinone may donate unpaired electrons to molecular oxygen, forming reactive oxygen radicals (such as superoxide anion, hydrogen peroxide, or others) that can critically disrupt a wide range of essential intracellular macromolecules (1 6, 49). The significance of these observations is underscored by recent investigations suggesting that myocardial cells have a limited capacity to detoxify oxygen radicals enzymatically (12, 44, 49); hence, the heart may be particularly susceptible to injury from reactive oxygen species generated as a result of anthracycline administration. Conversely, enhancement of the endogenous defenses of the heart against oxygen radicals with free radical scavenging agents (such as N-acetylcysteine) may reduce the cardiac membrane damage that follows treatment with anthracycline drugs (1 1,37, 41, 51 ). To evaluate the role of oxygen radical formation in anthracycline cardiac toxicity further, we have attempted to quantitate the effect of these drugs on reactive oxygen metabolism in various subcellular fractions from rat heart. Our results indicate that drugs of the anthracycline class significantly increase both superoxide anion and hydrogen peroxide production in cardiac mitochondria, cytosol, and sarcoplasmic reticulum. Thus, the pathological picture of anthracycline cardiac toxicity characterized by disruption of heart mitochondrial and sarcoplasmic reticular membranes (15) may be explained by drug-induced free radical formation in specific myocardial compartments. M A T E R I A L S A N D M E T H O D S Experimental Animals. Male Sprague-Dawley rats weighing 180 to 200 g were obtained from Mission Laboratory Supply Co., Rosemead, Calif. From the time of weaning, these animals were maintained on a 460 CANCER RESEARCH VOL. 43 Research. on September 23, 2017. © 1983 American Association for Cancer cancerres.aacrjournals.org Downloaded from diet of Wayne Lab-Blox rat pellets with water available ad libitum. Rat heart was chosen for study because the rat appears to develop both acute and chronic cardiac toxici ty after anthracycline treatment that is similar in many respects to anthracycline cardiomyopathy in humans (35). Materials. Doxorubicin hydrochloride of clinical grade, as well as chromatographically pure drug, was obtained from Adria Laboratories, Inc., Wilmington, Del. Daunorubicin, rubidazone, and 5-iminodaunorubicin were supplied by the Drug Synthesis and Chemistry Branch, Division of Cancer Treatment, National Cancer Institute, Bethesda, Md. Aclacinomycin A was generously supplied by Dr. W. T. Bradner, Bristol Laboratories, Syracuse, N. Y. Actinomycin D was obtained from Merck Sharp and Dohm, West Point, Pa. The drugs were reconstituted in sterile water unless indicated otherwise and protected from light until used. Glutathione (reduced form), glutathione reductase type III, sodium azide, ATP, MgCI2, bovine erythrocyte SOD, 3 [2900 uni ts/mg as assayed by the method of McCord and Fridovich (33)], bovine albumin Fraction V, xanthine, xanthine oxidase (Grade 1), cytochrome c (type VI from horse heart), EDTA, NADPH type III, NADH Grade III, NADP § NAD + Grade V, flavin adenine dinucleotide Grade III, flavin mononucleotide, sodium succinate, sucrose, D-mannitol, dimethyl sulfoxide Grade I, L-histidine, EGTA, sodium HEPES, DTNB, rotenone, dicumarol, D-o~-tocopherol acid succinate type IV, D-~x-tocopherol type I, AMP type III, and adenosine were purchased from Sigma Chemical Co., St. Louis, Mo. Deferoxamine mesylate was purchased from Ciba Pharmaceutical Co., Summit, N. J. Methanol (spectral grade), hydrogen peroxide (30% solution), ethyl alcohol (99% pure), potassium cyanide, sodium acetate, and acetic anhydride were obtained from Fisher Scientific Co., Fair Lawn, N. J. Chelex 100 resin (100 to 200 mesh, sodium salt) was purchased from Bio-Rad Laboratories, Richmond, Calif. Catalase (EC 1.11.1.6) of analytical grade (65,000 uni ts/mg protein) was obtained from Boehringer Mannheim Biochemicals, Indianapolis, Ind., and was devoid of SOD activity when assayed by the method of McCord and Fridovich (33). Only glass-distilled, deionized water was used in these studies. Preparation of Rat Heart Subcellular Fractions. Experimental animals were killed by cervical dislocation; the cardiac ventricles were excised, blotted dry, trimmed of connective tissue, and then minced into 20 to 30-mg replicates while being kept on melted ice. To prepare rat heart homogenate, the minced cardiac ventricles were vigorously washed free of erythrocytes with an iced solution of 225 mM mannitol and 75 mM sucrose, pH 7.4, containing 1 mM EGTA, 100 /~M D-~tocopherol, and 1 /~g SOD per ml. The mannitol:sucrose:EGTA was treated with Chelex 100 resin before use to remove trace quantities of iron salts in the buffer. The ventricles were homogenized at 4 ~ for 10 sec in 4 volumes of the mannitol:sucrose:EGTA buffer that contained tocopherol and SOD with a Brinkmann Model PCU-2-110 Polytron (Brinkmann Instruments, Inc., Westbury, N. Y.). We have shown previously that this procedure leads to minimal contamination of the homogenate by erythrocyte-derived proteins (12). The rat heart homogenate was centrifuged at 4 ~ and 480 x g for 5 min to remove nuclear and myofibrillar debris, and the resulting supernatant was used without further modification as the crude homogenate fraction. To prepare heart sarcosomes, the minced cardiac ventricles were vigorously washed with an iced solution of 100 mM KCI containing 5 mM histidine, pH 7.3, and were then homogenized with the Polytron at 4 ~ for 2 min in 4 volumes of the KCl:histidine buffer. For certain studies, the homogenization was performed in KCl:histidine buffer that contained 100 #M D-~-tocopherol and 1 /~g SOD per ml and that had been treated with Chelex to decrease the potential for membrane peroxidation during the homogenization process. The sarcosomal fraction was obtained by differential ultracentrifugation of the tissue homogenate as described by Martonosi (32) and was resuspended before use in 150 3 The abbreviations used are: SOD, superoxide dismutase; EGTA, ethyleneglycolbis(tg-arninoethylether)-N,N'-tetraacetic acid; HEPES, N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid; DTNB, 5,5'-dithiobis(2-nitrobenzoic acid). Cardiac Oxygen Radical Formation by Doxorubicin mM potassium phosphate buffer, pH 7.4, containing 100 /~M EDTA. Where indicated, heart sarcoplasmic reticulum was also prepared by the method described by Harigaya and Schwartz (19). To produce the rat cardiac mitochondrial fraction, the minced heart muscle was vigorously washed 4 times with an iced solution of 225 mM mannitol and 75 mM sucrose containing 1 mM EGTA and then homogenized for 10 sec with the Polytron on ice in 10 volumes of iced mannitol:sucrose:EGTA. Where specified, the cardiac tissue was homogenized in iced mannitol:sucrose:EGTA that had been treated with Chelex and that contained 100 /J,M n-~-tocopherol and 1 /~g SOD per ml. The heart mitochondrial fraction was prepared from the homogenate by a technique described previously (34) and was resuspended before use in 250 mM sucrose:20 mM HEPES buffer, pH 7.4. Where indicated, the mitochondrial fraction was frozen and thawed 3 times in a dry icemethanol mixture to ensure membrane disruption prior to assays for oxygen radical formation. To prepare rat heart cytosol, the minced ventricles were washed 4 times in 250 mM sucrose containing 1 mM EDTA and then homogenized on ice with the Polytron for 2 rain in 4 volumes of iced sucrose:EDTA. The homogenate was centrifuged at 4 ~ and 1000 x g for 20 min to eliminate nuclei and membranous debris; the resulting supernatant was centrifuged subsequently at 8000 x g for 20 min to remove cardiac mitochondria. The postmitochondrial supernatant was then centrifuged for 1 hr at 105,000 x g and 4 ~ in a Beckman Model L5-50 ultracentrifuge (Beckman Instruments, Inc., Palo Alto, Calif). The final supernatant was decanted at the end of ultracentrifugation and was used directly as the cardiac cytosol fraction. The experimental heart homogenate, as well as the sarcosomal, mitochondrial, or cytosolic fractions were used on the day of preparation. Superoxide Assay. Superoxide anion production in experimental samples was determined by the rate of SOD-inhibitable acetylated cytochrome c reduction. The cytochrome c utilized was acetylated before use, as described by Azzi et al. (2), to eliminate interfering reactions by cytochrome c oxidases or reductases in the cardiac sarcosomal, mitochondrial, or cytosolic fractions. The initial, linear rate of acetylated cytochrome c reduction was determined spectrophotometr ical ly at 550 nm and 37 ~ in a Gilford Model 250 recording spectrophotometer (Gilford Instrument Laboratories, Inc., Oberlin, Ohio) equipped with a circulating water bath. Superoxide formation was calculated from the rate of acetylated cytochrome c reduction that was inhibited by SOD using an extinction coefficient for cytochrome c (reduced minus oxidized) of 19.6 mM -1 cm -1 (55). For experiments assessing the effect of DTNB on superoxide production, the sulfhydryl reagent was added to the paired reaction mixtures which were then preincubated for 2 min at 37 ~ before the addition of NADPH. Preincubation was not performed in experiments examining the effect of other agents on the rate of superoxide formation. Superoxide formation in the rat heart mitochondrial fraction was examined in a fashion similar to that for heart sarcosomes as described in Table 7. Chemotherapeutic agents were added, where indicated, before initiation with NADH. Superoxide production in rat cardiac cytosol was determined as described in Table 11. Measurement of NADPH Consumption. The oxidation of NADPH by heart sarcosomes was determined in triplicate at 37 ~ by the initial, linear change in optical density at 340 nm using the Gilford spectrophotometer. The 1-ml reaction mixture contained 150 #mol potassium phosphate buffer, pH 7.4, 1 O0 nmol EDTA, 200/~g sarcosomal protein, 1 O0 nmol NADPH, and the indicated amount of anthracycline. NADPH consumption was initiated by the addition of the sarcosomal protein and was calculated using an extinction coefficient of 6.22 mM -1 cm -~ (20). Oxygen Consumption Measurements in Cardiac Subcellular Fract ions. Oxygen consumption was measured at 37 ~ with a YSI Model 53 oxygen-monitoring system (Yellow Springs Instrument Co., Yellow Springs, Ohio). Oxygen consumption by the crude rat heart homogenate was determined in a 3-ml reaction mixture that contained 750