differential oxygen radical susceptibility of adriamycin ... · superoxide dismutase (3050 units/mg...

[CANCER RESEARCH 49. 8-15. January 1. 1989]

Differential Oxygen Radical Susceptibility of Adriamycin-sensitive and -resistantMCF-7 Human Breast Tumor CellsEdward G. Mimnaugh,1 Lata Dusre, John Atwell, and Charles E. Myers

Biochemical Pharmacology Section, Clinical Pharmacology Branch, Clinical Oncology Program, Division of Cancer Treatment, National Cancer Institute, NIH,Bethesda, Maryland 20892

ABSTRACT

Recent evidence supports the concept that Adriamycin cytotoxicitymay be mediated by drug semiquinone free radical and oxyradical generation. We tested this hypothesis further by exposing drug-sensitive(WT) and 500-fold Adriamycin-resistant MCF-7 human breast tumorcells (ADRR) to exogenous Superoxide- and hydrogen peroxide-generating

systems and subsequently monitored cell proliferation as a measure ofcytotoxicity. The ADR" tumor cells tolerated a 4-fold greater exposure

than sensitive cells to Superoxide generated by the xanthine/xanthineoxidase system. Likewise, exposure to hydrogen peroxide produced bythe action of glucose oxidase on glucose revealed a 4-fold diminishedsusceptibility of the drug-resistant cells to this reduced form of oxygen.

Similar results were obtained by the direct application of hydrogenperoxide to cells. For both cell lines, cytotoxicity was dependent uponthe magnitude and the duration of reactive oxygen exposure. When WTand ADR" cells were cultured under hyperoxia (95% O2:5% CO2), in

order to stimulate the intracellular production of oxyradicals, proliferation was inhibited to a greater extent in the drug-sensitive cell line.

Additionally, hyperoxia potentiated the cytotoxicity of Adriamycin toboth sensitive and drug-resistant cells, but the effect depended upon the

concentration of the drug. Under hyperoxic conditions, Adriamycincaused oxygen radical-dependent cytotoxicity to the WT tumor cells atclinically relevant drug concentrations as low as 2 to 3 IIM. With ADR"

tumor cells, hyperoxia increased the cytotoxicity of Adriamycin at concentrations above 5 MM-Paradoxically, both the WT and the ADR" tumor

cells were equally susceptible to the cytotoxic effects of y irradiation. Itis known that the Adriamycin-resistant MCF-7 cells greatly overexpress

glutathione peroxidase and glutathione transferase activities; however,other biochemical defenses against reactive drug intermediates and oxygen radicals have been reported to be similar in the two cell lines. Wehave reexamined those observations in this report. The resistance ofADR" breast tumor cells to Adriamycin appears to be associated with a

developed tolerance to Superoxide, most likely because of a twofoldincrease in Superoxide dismutase activity, and a decreased susceptibilityto hydrogen peroxide, most likely because of 12-fold augmented selenium-

dependent glutathione peroxidase activity. Acting in concert, these twoenzymes would decrease the formation of hydroxyl radical from reducedmolecular oxygen intermediates. Taken together, these observations support a role for oxygen radicals in the cytotoxic mechanism of Adriamycinat pharmacological levels of 10 IIM and below, and suggest further thatthe development of resistance to this drug by the MCF-7 WT tumor cells

may have a component linked to oxygen free radicals.

INTRODUCTION

A major impediment to the clear understanding of the phenomenon of anticancer drug resistance by tumors is the diversityof possible mechanisms that frequently and simultaneouslymanifest within a population of drug-resistant tumor cells (1,2). Compounding this situation, are the observations that certain tumors demonstrate a multidrug resistance or "pleiotropic"

resistance phenotype, and are cross-resistant to a variety ofagents which markedly differ in both chemical structure and

Received 4/21/88; revised 8/18/88; accepted 10/5/88.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 inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1To whom requests for reprints should be addressed, at the National Cancer

Institute. NIH. Building 10. Room 6N119. Bethesda, MD 20892.

putative mechanisms of action (2, 3). The clinically most important of these agents are the anthracyclines, Vinca alkaloids,and epidophyllotoxins.

Recent investigations into the mechanisms of inherent andacquired multidrug resistance have revealed several markedlydifferent biochemical and pharmacological characteristics ofdrug-resistant tumors compared to their drug-sensitive counterparts. Among these, the most conspicuous appear to be (a)a decreased net cellular accumulation of anticancer drugs, resulting from an efflux pump function of overexpressed membrane P-glycoproteins (4), (b) the elevation of reduced glutathi

one (5, 6), (c) diminished drug metabolic activation resultingfrom the down-regulation of monooxygenase and other enzymeactivities (7), and (d) increased levels of drug- and reactiveoxygen-detoxifying enzymes such as glutathione S-aryltransfer-ases and selenium-dependent glutathione peroxidase (8, 9).Currently, it is not known if any one of these postulatedmechanisms predominates in the resistance of tumor cells toAdriamycin or other anticancer agents, and it is conceivablethat each of them contributes to some degree to the development of resistance.

Although still controversial, the mechanism of Adriamycincytotoxicity, at least in some tumors, appears to be linked toits enzymatic reductive activation to a semiquinone free radicalmetabolite with subsequent generation of extremely toxic reactive oxygen species (10,11). These activated forms of oxygen:Superoxide aniÃ³n radical, hydrogen peroxide and ultimatelyhydroxyl free radical, have been implicated in the cytotoxicityof Adriamycin in several in vitro tumor model systems (12-15),including MCF-7 human breast tumor cells (14, 15). Recently,an acquired subline ADRR 2 was derived from the MCF-7parental wild-type tumor cell line by slowly escalating Adriamycin exposure (8). This resistant subline is 500-fold resistantto Adriamycin, shows diminished ability to generate hydroxylradical in the presence of Adriamycin (15, 16), exhibits three-to 9-fold decreased anthracycline accumulation depending upon

the drug concentration, and has concurrently developed thephenotype of multidrug resistance, to include resistance to VP-16, actinomycin D, and vincristine (8,9). In addition, the ADRR

subline shows enhanced expression of DNA sequences homologous to the P-glycoprotein gene (17), enhanced glycolysis andATP production (18), and stimulated pentose-shunt activity inthe presence of exogenous peroxides (19).

We have designed the present study in order to examine thepossibility that the diminished cytotoxicity of Adriamycin toADRR drug-resistant tumor cells may be related to their abilityto tolerate exposure to Adriamycin-mediated reactive oxygen

species. To this end, we have compared the susceptibility ofAdriamycin-sensitive (WT) and Adriamycin-resistant MCF-7breast tumor cells (ADRR) to several reactive oxygen systems.

Additionally, we have investigated the cytotoxic interaction ofAdriamycin and hyperoxia to both sensitive and resistant tumor

2The abbreviations used are: ADR", Adriamycin-resistant MCF-7 cell line;WT. wild-type MCF-7 cell line; IC50, concentration of toxin that inhibited cellproliferation by 50% compared to the untreated controls.

Research. on October 6, 2020. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

OXYGEN RADICAL RESISTANCE BY BREAST TUMOR CELLS

cells. It appears that higher levels of enzymatic defenses againsttoxic reactive oxygen species in the resistant tumor cells providesignificant protection against oxyradical exposure. This findingsuggests that elevated levels of oxygen free radical defenses,likewise, can at least partially modulate the cytotoxicity ofAdriamycin in the multidrug-resistant MCF-7 human breast

tumor model.

MATERIALS AND METHODS

Chemicals and Drugs. Xanthine. xanthine oxidase (SO units/ml),Superoxide dismutase (3050 units/mg protein), catalase (10,000 units/mg protein), glucose oxidase (242 units/mg protein), NADPH, reducedglutathione, glutathione reducÃase (500 units/mg protein), bovineserum albumin, EDTA, cytochrome c, diethylenetriamine-pentaaceticacid, l-chloro-2,4-dinitrobenzene, eumene hydroperoxide, and 5,5'-

dithiobis-nitrobenzoic acid were from Sigma Chemical Co. (St. Louis,MO); hydrogen peroxide and sulfosalicylic acid were from FisherChemical Co. (Fair Lawn, NJ). Desierai mesylate was from BenVenueLabs. (Bedford, OH), ultra-pure Tris was purchased from BethesdaResearch Labs. (Gaithersburg, MD), and all other chemicals were ofthe highest purity available. Adriamycin hydrochloride (NSC 123127)was provided by the Drug Development Branch. National CancerInstitute, NIH, Bethesda, MD). The water used in this study was firstdeionized then distilled in glass.

Tumor Cell Culture and Cytotoxicity Assays. MCF-7 human breasttumor cells (WT) were grown in improved modified essential medium,supplemented with 2 mM L-glutamine, 2 mg/liter L-proline, 50 Â¿ig/mlgentamicin and 5% fetal bovine serum (Grand Island Biologicals Co.,New York), at 37Â°Cin a humidified incubator under an atmosphere of

5% carbon dioxide in air. The Adriamycin-resistant MCF-7 tumor

cells, were derived from the parental line by stepwise selection inincreasing concentrations of Adriamycin until the cells were capable ofpropagating in 10 MMdrug, as described previously (8). Typically, theresistant cells were grown for a minimum of four passages in mediumlacking Adriamycin before they were used in experiments. Tumor cellcytotoxicity assays were conducted by seeding 50,000 cells/well in 6-chambered Linbro dishes (Costar) in complete medium, then 24 to 30h later, fresh medium containing the appropriate concentrations ofAdriamycin, oxyradical-generating system, or other agents was added,the cells were incubated for various times (minutes to 24 h-see figurelegends), replenished with fresh medium, and grown for 72 h longerunder normal conditions before the cells were trypsinized and countedwith a model ZM Coulter counter (Hialeah, FL). In the experimentswith hydrogen peroxide, this agent was diluted in glass-distilled water,then added directly to the cells in wells containing freshly replacedcomplete medium; an equal volume of water was added to the controls.In some experiments, the clonogenic assay was used to estimate cytotoxicity by inoculating 300-500 cells/well after exposure to the toxins,growing the cells for 12 days and counting the resulting colonies afterwashing with ethanol and staining with mÃ©thylÃ¨neblue. Cytotoxicitywas usually determined by using a range of concentrations of drug ortoxic oxygen species and inhibitor concentration 50% (ICso)values werederived from computer plots (Macintosh).

Superoxide- and Hydrogen Peroxide-Generating Systems. To studythe relative susceptibility of the WT and ADR" breast tumor cells to

an extracellular exposure of reactive oxygen species, two enzymaticoxyradical generating systems were used. The first was the xanthine/xanthine oxidase reaction that produces Superoxide and hydrogenperoxide during the enzymatic conversion of xanthine to urate. Tumorcells growing in monolayers were exposed to various activities of thisenzyme in the presence of 1 mM xanthine in complete medium for afixed time of 60 min, or to 10 mU/ml of xanthine oxidase and 1 mMxanthine for various times. The reaction mixtures were then aspirated,fresh medium was added and the inhibition of cell growth was assessed3 days later. The second enzyme system, the glucose oxidase-mediatedoxidation of glucose to gluconic acid, reduces molecular oxygen directlyto hydrogen peroxide. We investigated the effects of glucose oxidaseactivities and the duration of exposure to this hydrogen peroxide-

generating system on the survival of sensitive and drug-resistant tumorscells. The concentration of glucose was initially 11 mM in all experiments. These two enzymatic oxyradical-generating systems provided

the capability to carefully control both the rate and the amount ofSuperoxide and hydrogen peroxide exposed to the cells. In some experiments, we also studied the toxicity of hydrogen peroxide directly addedto the medium.

Exposure of the Tumor Cells to Hyperoxic Conditions. Hyperoxia(>21% oxygen) can cause leakage of reactive oxygen species fromoxidoreductase enzymes such as NADPH-cytochrome P-450 reducÃaseand mitochondrial respiratory electron transport chain enzymes (20-23) and from the autoxidation of reduced cellular biochemicals. WTand ADR" cells were exposed to an atmosphere of 95% oxygen-5%

carbon dioxide using a Forma CH/P incubator with oxygen control forvarious lengths of time ( 1 to 4 days) and the effects of that >4-foldincreased oxygen level on cell proliferation were examined. In theseexperiments the cells were inoculated into 6-well plates and allowed toattach and begin to proliferate for 24 to 30 h prior to initiating theexposure to 95% oxygen. Adriamycin was included in the mediumduring the hyperoxic exposure in separate experiments.

Biochemical and Enzymatic Assays. Prior to determining the activities of oxidoreductase enzymes known to activate Adriamycin andenzymes known to play important roles in oxyradical detoxification,the WT and ADRR tumor cells were trypsinized and washed severaltimes in ice-cold phosphate buffered saline, or scraped off the mono-layers directly into the appropriate chilled buffers, centrifuged andresuspended in buffer prior to sonication with a microprobe at setting5, for three 10-s bursts (model W-225; Heat Systems Ultrasonics, Inc.,New York). Cytosolic enzymes were then assayed in the supernatantsfollowing centrifugation in an Eppendorf microfuge for 20 min. Forsome measurements, microsomal (24) and mitochondrial (25) subcellular fractions were isolated by differential centrifugation techniques.NADPH-cytochrome P-450 reducÃaseactivities were assayed in themicrosomal fractions from tumor cells using cylochrome c as theelectron acceptor (26), and NADH-dehydrogenase activities were esti-maled by moniloring ihe reduction of cytochrome c by tumor mitochondria in the presence of NADH. Glutathione peroxidase activitywas determined by the method of Paglia and Valentine (27) with slightmodifications, using both HiOi and eumene hydroperoxide, glutathione5-aryl transferase activity was assayed by the reaction of reducedglutathione (1 mM) with chlorodinitrobenzoate (28), and reduced glutathione levels were measured in tumor cell sonicates deproteinatedwith sulfosalicylic acid by the method of Ellman (29). Superoxidedismutase activities were assayed in tumor cell sonicates by the methodof McCord and Fridovich (30), using the acetylcytochrome c modification as described by Azzi et al. (31). Protein determinations were bythe method of Lowry et al. (32).

T Irradiation of Cells. To determine whether there were differencesin the relative toxic effects of intracellular oxygen radicals generatedby irradiation, WT and ADR" tumor cells were trypsinized, suspended

in chilled complete medium, and exposed to various doses of y radiationwith a Shepherd Mark I model 68 l37Cs irradiator with a dose rate of

5.74 Gy/min. Tumor cells were then inoculated in triplicate at 50,000cells/well for proliferation inhibition assays and at 500 cells/well forthe clonogenic assay of cytotoxicity.

Statistics. Data were analyzed by Student's f test (33), and differences

between mean values at P < 0.05 were considered to be significant.

RESULTS

Adriamycin Cytotoxicity to MCF-7 WT and ADRR Tumor

Cells. It is known that there are marked differences in the abilityof Adriamycin to inhibit the proliferation of sensitive and drug-resistant MCF-7 breast tumor cells (8, 9, 16). As shown in Fig.1, the ADRR subline was found to be more than 500-fold

resistant to Adriamycin than the parental WT cells. The ICsovalues for Adriamycin by continuous exposure were 5 to 7 nMand 2 to 3 /Â¿Mfor the WT and ADRR cells, respectively. TheADRR-resistant phenotype has been previously reported to be




Sensitive 5io7nM Kesislanl 2to3nM

?Bn-o.'~Z 00AHL\!VÃ•Ih.vâ€žIO"2

10' * 10Â° IO1 10

Adriamycin ( nM )

O0c

60'

oS

20-I10Bâ€¢3

10-2 10Aik^V550%'is,-'10Â° io1 10

Adriamycin (^M)

Fig. 1. Comparative cytotoxicity of Adriamycin to MCF-7 WT (A) and ADR"

(B) human breast tumor cells determined by the inhibition of tumor cell proliferation resulting from the continuous exposure to Adriamycin for 3 days. TheICso values were S to 7 nM and 2 to 3 uM Adriamycin for the sensitive andresistant cells, respectively. Dual curves, results from two separate experimentswith three to four samples/data point.

about 200-fold resistant to Adriamycin, and to maintain stableresistance over extended subculture for at least 12 months (8).However, in the present study, although the ICso value forAdriamycin toward the ADRR cells was almost identical to thatvalue published previously (8), the IC50for the WT MCF-7 cellswas considerably lower. This apparent change in susceptibilityto Adriamycin may result from temporal alterations in thebiochemistry of the WT cells or be due to slight differences inthe methods for the determination of the ICso values, such asthe inoculation density or the number of drug concentrationsused in the experiments. In the experiments conducted in thisstudy, we used cells showing a relative 500-fold resistance toAdriamycin by the continuous drug exposure, inhibition ofproliferation assay.

Sensitivity of WT and ADR" Cells to Superoxide Free Radical.

The xanthine oxidase/xanthine system was used as a means toexpose tumor cells to an external source of Superoxide. Thissystem also produces hydrogen peroxide and most likely, hy-droxyl radical in the presence of trace quantities of transitionmetal ions found in the cell culture medium, but the initial andpredominating enzymatically reduced form of oxygen is super-oxide. When the WT and the ADRR cells were exposed to the

xanthine oxidase system, we observed a differential cytotoxicityof Superoxide to the sensitive and drug-resistant cells. In thepresence of 1 HIMxanthine, the amounts of xanthine oxidaserequired to diminish cell numbers by 50% following a 60 minexposure were calculated to be 3 and 9 m units/ml, for the WTand the ADRR cells, respectively (Fig. 2A). Incubating the cells

with either xanthine oxidase alone at 100 m units/ml or withxanthine alone up to 3 mM resulted in no substantial cytotoxicity (<5%) to either cell line (data not presented). In Fig. 2B,it is shown that the duration of exposure to the superoxide-generating system also resulted in differential cytotoxicity; 50%killing of the WT cells required an 8- to 10-min exposure, incontrast, 50% killing of the ADRR cells required about 45 min.

The xanthine oxidase reaction at 10 m units/ml was capable ofgenerating 1.8 //mol of Superoxide during the 60-min exposureperiod. Thus, the Adriamycin-resistant cells were comparativelythree- to 4-fold more resistant to Superoxide exposure in theextracellular environment than the Adriamycin-sensitive cells.

Cytotoxicity to Hydrogen Peroxide by Glucose Oxidase. Asmentioned above, the xanthine oxidase system forms bothSuperoxide aniÃ³n and hydrogen peroxide as products of xanthine oxidation. There is considerable evidence that Superoxideitself is neither reactive nor especially toxic, and that damageto cells or tissues exposed to Superoxide results from the formation of the hydroxyl radical through the intermediary of

1 MI 40â€¢

| 20

O

As<f"\nsltl\'f4T\VR\^Â»lit\ntS

=!

o 40e

Â» 20

O

BSens\\Uve

\^NIK\^Resi\\~~^,tant

40 60 80

Xanthine oxidase (mil ml) Minutes of exposure

Fig. 2. Cytotoxicity of Superoxide aniÃ³nfree radical to WT and ADR" tumor

cells. In A, the activity of xanthine oxidase was varied to determine the relativesusceptibility of the sensitive and drug-resistant tumors cells to an extracellularexposure to Superoxide. In B, both cell lines were exposed to 10 m units ofxanthine oxidase/ml in the presence of I mM xanthine for various lengths oftime. Tumor cells were counted 3 days after replacing the medium containing thexanthine oxidase/xanthine with fresh complete medium. Xanthine oxidase alonewas not toxic to either cell type. Data are expressed as mean Â±SD.

AI^*=!nsitivr1\\1e^SRes^\\\istant\

:\II

0.0 10 20 5.0 100 20.0Glucose Oxidase ( mu ml )

0 20 40 60 80 100120140Exposure time (minutes)

Fig. 3. A comparison of the cytotoxicity of hydrogen peroxide, generatedextracellularly by the action of glucose oxidase, to WT and ADRRs tumor cells.

A, effect of increasing activities of the enzyme and the cytotoxic response to bothcell lines; /(. time-dependent effects of exposure to 10 m units of glucose oxidase/ml for up to 2 h. In both experiments, the tumor cell proliferation was measured3 days after exposure to glucose oxidase. Data are expressed as mean Â±SD.

hydrogen peroxide (34,35). We therefore studied the inhibitionof tumor cell proliferation by hydrogen peroxide generated bythe glucose oxidase-catalyzed oxidation of glucose. As shownin Fig. 3A, the addition of various amounts of glucose oxidaseto the cells revealed that the WT cells, compared to the ADRRcells, had a substantially decreased ability to resist a peroxide-induced oxidant stress. The ICso values for the sensitive anddrug-resistant cells were found to be 4 and 16 m units/ml ofglucose oxidase, respectively, for a 60-min exposure. The timecourse for this system (Fig. 3Â£f)also showed a significantdifference between the two cell lines in susceptibility to hydrogen peroxide.

Direct Addition of Hydrogen Peroxide. Hydrogen peroxideadded as a bolus to tumor cells was also very toxic, and as withthe xanthine oxidase and glucose oxidase systems, there was adivergence in the toxicity curves for the WT and ADRR cells(Fig. 4). The IC5o values for hydrogen peroxide (60-min exposure) were estimated to be about 30 and 85 pM, respectively,for the WT and ADRR breast tumor cells. In order to probe for

the participation of hydroxyl radical in hydrogen peroxidecytotoxicity, we added metal ion chelators to the medium priorto the addition of the peroxide, with the expectation thatchelators might block hydroxyl radical generation by removingmetal ions (iron and copper) that could catalyze the Fentonreaction (36). Neither diethylenetriaminepentaacetic acid at 10MMnor desferal as high as 1 mM had any sparing effect on thecytotoxicity of hydrogen peroxide to either cell line, althoughthere was some delayed toxicity resulting from the exposure todesferal (data not presented). Because these two highly charged

10




120

o 100

l

S "Â»O

40 60 80H202 (uM)

100 120

Fig. 4. Cytotoxicity caused by the direct addition of hydrogen peroxide tosensitive and drug-resistant MCF-7 breast tumor cells. In this experiment, theexposure to hydrogen peroxide was for 60 min and the cells were counted 3 dayslater. Data are expressed as mean Â±SD.

Sensitive cells

Eu

40 60 80

Hours

I IResistant cells

Hyperoxla

40 60 80Hours

Fig. 5. Growth of MCF-7 WT (A) and ADR* (B) tumor cells under a normal

atmosphere containing 21% oxygen 5% carbon dioxide (euoxia) and under anatmosphere containing 95% oxygen 5% carbon dioxide (hyperoxia) for variouslengths of time. Tumor cells were seeded at 50.000 cells/well in triplicate, andthe number of cells were counted at the indicated time points. Data are expressedas mean Â±SD.

chelators do not rapidly penetrate cell membranes, this lack ofprotection against hydrogen peroxide cytotoxicity suggests thatmetal ion-catalyzed conversion of hydrogen peroxide to hy-droxyl radical or other reactive metal-containing oxidant didnot occur outside of the cells.

Hyperoxia Toxicity. When cells are exposed to high concentrations of isobaric oxygen, the respiratory electron transferenzymes of the mitochondria and the flavin oxidoreductaseenzymes of the endoplasmic reticulum divert electrons to molecular oxygen. This process greatly enhances the leakage ofSuperoxide and hydrogen peroxide from these sources (20-23).Thus, hyperoxia exposure results in the intracellular productionof several reactive oxygen species. The cytotoxic effects ofexposure of MCF-7 WT and ADRR tumor cell lines to hyper

oxia for several days is shown in Fig. 5, A and B. After a periodof 24 h under 95% oxygen, there was only a marginal decreasein the number of WT and ADRR cells compared to euoxic

controls, indicating that this exposure time was not sufficientlytoxic to inhibit cell proliferation or to kill the cells and causethem to release from the monolayer. Longer exposure periodsof up to 96 h revealed that hyperoxia was selectively more toxicto the sensitive cells than to the drug-resistant subline. Itappeared that while WT cell proliferation was considerablyslowed during long-term hyperoxia (Fig. 5/4), the ADRR cells

were less susceptible to hyperoxia (Fig. 5B). These results areconsistent with the previous observations of differential susceptibility of these two cell lines to extracellular oxyradical exposure, as observed in the Superoxide and hydrogen peroxideexperiments described above.

Potentiation of Adriamycin Cytotoxicity by Hyperoxia. HecauseAdriamycin may be cytotoxic as the result of its abilityto redox cycle and generate reduced oxygen species, we reasoned that the combination of hyperoxia and Adriamycinshould result in additive or possibly synergistic toxicity. Indeed,when tumor cells were exposed to Adriamycin during hyperoxic(95% oxygen) conditions, the cytotoxicity of the drug relativeto euoxic (21% oxygen) controls was enhanced. As shown inFig. 6, hyperoxia potentiated the cytotoxicity of low concentrations of Adriamycin to the WT drug-sensitive cells, but incontrast, hyperoxia increased the cytotoxicity of Adriamycin tothe ADRR cells only at the higher drug concentrations. These

results appear to validate the hypothesis that oxyradicals areimportant to the mechanism of Adriamycin cytotoxicity to theMCF-7 WT cells. It is possible that oxygen radical generationmay be relatively less important in the cytotoxicity of the drugto ADRR cells than WT cells.

0 10 20 30 40 50

Adriamycin ( nM )

0 5 10 15 20 25

Adriamycin ( M )

Fig. 6. Potentiation of the cytotoxicity of Adriamycin to WT and ADR" tumor

cells by concomitant exposure to 95% oxygen. Tumor cells were plated and grownunder euoxic and hyperoxic conditions as described in the previous figure, butAdriamycin was added at the indicated concentrations just before the tumor cellswere exposed to 95% oxygen (hyperoxia) for a period of 26 h. The number ofsurviving tumor cells were determined 3 days after the hyperoxic exposure. Dataare expressed as mean Â±SE.

Proliteration assay Clonogenic assay

A\8l\,nslti%1/eRe^.*\^sisla^II B|^\Re^1

Â¡startensit^va

â€¢1000 100 200 300 400 500 600

Radiation dose (rads)0 100 200 300 400 500 600

radiation dose ( rads )

Fig. 7. Inhibition of proliferation of WT and ADR" breast tumor cells by

gamma irradiation. Tumor cells suspended in complete medium were irradiatedat 3 and 5 Gy, then inoculated into 6-well plates in triplicate for both the growthinhibition assay (.I) and the clonogenic assay (B). Tumor cell colonies were stainedwith mÃ©thylÃ¨neblue after a 12-day growth period and colonies with more than10 cells were counted. Data are expressed as mean Â±SD.

7 Irradiation. There is considerable evidence that ionizingradiation causes cell damage in part by an o\\ radical-mediatedmechanism (37). When tumor cells were exposed to 137Csas asource of 7 radiation, survival of both WT and ADR" cells at

3 and 5 Gy was almost identical, as determined by both theinhibition of proliferation and by the clonogenic assays (Fig.7). These results showing no resistance to irradiation by theADRR cells contrast with the previously observed decreasedcytotoxicity of Superoxide and hydrogen peroxide to this sub-line. It may be that oxygen radicals generated within the nucleusof tumor cells are unable to be scavenged by cytosolic enzymaticoxyradical defenses that are known to be elevated in the resistant cell line (8, 16). An alternative explanation is that the

11




radiolysis of water produces hydroxyl radical directly, and thispathway is not linked to the sequential single electron reductionof molecular oxygen process that occurs in enzymatic systems.

Biochemical Parameters of Sensitive and Drug-ResistantMCF-7 Tumor Cells. In a previous publication it was reportedthat several biochemical parameters relevant to Adriamycincytotoxicity were elevated in the ADRR tumor subline comparedto the drug-sensitive WT parental line (8,16). Because the IC5ovalue for Adriamycin toward the WT tumor cells appeared tohave decreased slightly from the previously reported value, andbecause of the possibility that extended culture of both sensitiveand resistant cell lines might have resulted in altered biochemistry, we examined a number of enzymatic and biochemicalvariables that might influence either the activation of Adriamycin or the detoxification of reactive forms of oxygen. Inaddition, several of the enzymatic activities were previouslymeasured in sonicates of tumor cells, and we decided to moreclosely examine select enzymes in purified subcellular fractionsto determine if there were subtle differences between the sensitive and drug-resistant tumor cells. These biochemical andenzymatic measurements are summarized in Table 1. The activities of two enzymes known to activate Adriamycin,NADPH-cytochrome P-450 reducÃase and NADH-dehydro-genase, were roughly equivalent in microsomes and mitochondria isolated from the WT and ADRR tumor cells. The slightlylower activity of cytochrome P-450 reducÃasein the resislantcells (65% of the sensilive cell value) would noi be expecled loresult in dramatically diminished Adriamycin activation to itsfree radical intermediale. NADH-dehydrogenase aclivily wasnearly two-fold higher in purified milochondria from ADRR

lumor cells, ruling oui any diminished ability to actÃvaleAdriamycin by Ihis milochondrial enzyme in Ihe drug-resistanlsubline. The activily of Superoxide dismulase was found lo bealmosl twofold higher in Ihe ADRR cells, suggesling Ihe differ-

Table 1 Comparison of several enzymatic and biochemical parameters in MCF-7sensitive and Adriamycin-resistant breast tumor cells

Tumor cells were sonicated in buffer or prior to the isolation of microsomesor mitochondria by differential centrifugaron the cells were first lyscd in glassdistilled water. The NAD(P)H-oxidoreductase enzyme activities were assayed inthe microsomal and mitochondria! fractions, and all other parameters exceptprotein were assayed in the cytosolic fractions. The measurements of the enzymeactivities were as described in the text.

Parameter"Microsomal

NADPH-cytochrome P-450reducÃaseMitochondrial

NADH-dehydrogenaseGlutathione

peroxidasewithhydrogenperoxidewitheumenehydroperoxideGlutathione

S-aryltransferaseSuperoxidedismutaseCatataseReduced

glutathione(nmol/10'cells)(nmoi/mg

protein)ProteinOig/10' cells)Sensitive

tumors66Â±9*53

Â±20.74

Â±0.221.71Â±0.610.4

Â±2.03.0Â±0.714.9Â±3.816

Â±545Â±11366Â±47Resistant

tumors49

Â±10e96Â±3C8.9

Â±1.1'11.6Â±1.3C144

Â±45C5.4Â±1.0e10.2

Â±0.86.1

Â±2.3C19Â±6C321

Â±27R/S0.741.8126.8141.80.70.380.420.88" Enzyme activities are expressed as nmol of product formed/mg of protein/

min, except Superoxide dismutase activity is expressed units/mg protein whereone unit is defined as the amount of enzyme necessary to inhibit by half thesuperoxide-dependent reduction of acetylcytochrome c, catalase is expressed asiimul of hydrogen peroxide decomposed/run protein/min and glutathione peroxidase activity is expressed as the rate of NADPH oxidation (nmol/mg protein/min) in the glutathione reducÃasecoupled assay using either hydrogen peroxide(0.073 HIM)or eumene hydroperoxide (1.2 HIM)as ciÂ»substrates.

* Data are expressed as the mean Â±SD for three to 12 determinations.'' Denotes a statistically significant difference when compared to WT values at

P < 0.05.

eniial susceplibilily of ihese cells lo Superoxide formed in ihexanlhine oxidase system might be attributed lo Ihe increasedabilily lo scavenge Superoxide by Ihese lumor cells. The moslinleresling enzymalic differences belween Ihe WT and IheADRR tumor cells were the marked increases in the activiliesof glutathione S-aryl Iransferase (14-fold) and glutathione peroxidase (7 lo 12-fold). Il has been previously reported lhal IheseIwo enzymes were greally elevaled in Ihe ADRR cells (8), bul

Ihe absolute values were slightly different. Because Ihe increasedperoxidase aclivily of Ihe ADRR lumor cells was nearly equiv-

alenl wilh eilher hydrogen peroxide or eumene hydroperoxideas cosubslrale, il appears Ihis activily is predominanlly associated wilh Ihe selenium-dependenl glulalhione-peroxidase. Theincreased aclivily of glutathione peroxidase would theoreticallyconfer upon ihe resislanl cells an increased capability lo delox-ify Adriamycin free radical-enhanced or glucose oxidase-gen-

eraled hydrogen peroxide. We found Ihe calalase aclivity lo besimilar in bolh cell lypes, as previously reported (16). Of noie,is Ihe measurement lhal Ihe resislanl cells had nearly threefoldlower levels of reduced glutalhione, Ihe cofaclor for Ihe Iransferase and peroxidase enzymes. However, Ihe concenlralion ofglutathione in ihe resislanl cells (6.1 Â±2.3 nmol/million cells,or aboul 3 nm) would siili be sufficient ly high lo provide amplereduced glutalhione lo bolh glutathione Iransferase and glutathione peroxidase. No imporlanl differences were found in Iheprolein conlenl of Ihe WT and ADRR cells.

DISCUSSION

The developmenl of acquired multi-drug resistance by malig-nanl cells is undoubtedly a major barrier lo Ihe effeclive chemo-Iherapy of neoplastic diseases, and although there have beenfragmentary advances in our underslanding of Ihe biochemistryand genetics of Ihe resislance phenomenon, Ihe known mechanisms of drug resislance have noi provided clinically usefulstrategies for Ihe reversal of Ihis problem. While numerous invitro models of drug resislance by lumor cells in culture havebeen developed, usually by ihe seleclive pressure of anlicancerdrug exposures, no consislenl universal mechanism for resistance has emerged, and the relationships of the pharmacological,biochemical, and genelic characteristics of these lumor cellmodels lo clinical mult Â¡drugresistance are only recently becoming known.

In the presenl sludy, we have hypolhesized lhal Ihe emergence of resislance lo Adriamycin by MCF-7 breasl lumor cellsmighl be related lo Ihe melabolism of Ihe drug lo an activatedsemiquinone free radical intermediate and Ihe subsequenl production of reaclive oxygen species lhal results from ihe redoxcycling of Ihe drug semiquinone (10, 11). Superoxide andhydroxyl free radicals and hydrogen peroxide are generated byIhis process, and Ihese species avidly reacl with and damage abroad range of cellular conslituents including membrane phos-pholipids, enzymes, and DNA (38). Oxygen free radicals areknown lo cause single strand breaks in DNA (39) and lo induceoxidalive modifications in DNA bases (40). These molecularevenls have been postulated lo accounl for ihe mulagenicity ofoxyradicals lo bacteria (41) and lo mammalian cells (42). Fur-Ihermore, il has been clearly shown lhal Adriamycin can cleaveDNA by an oxygen free radical-mediated process (11, 43, 44).We Ihus reasoned thai if oxyradicals played a role in ihedevelopment of resistance lo Adriamycin, Ihen lumor cellsselected for resislance lo Adriamycin should also demonstrateresislance lo Ihe direcl cyloloxic effects of oxyradicals, therebyproviding a tesi of Ihe hypolhesis.

12




The involvement of oxyradicals and oxidative stress in themechanism of tumor cell killing by Adriamycin is controversial,but convincing, nascent evidence supporting this possibility isbeginning to become available. For example, first Sato et al.(12) and later Doroshow et al. (13) have shown that Ehrlichascites tumor cells and MCF-7 human breast tumor cells (14)generate increased amounts of reduced oxygen metabolites asthe result of Adriamycin activation. Moreover, exogenous su-peroxide dismutase, catalase, and hydroxyl radical scavengerswere shown to partially protect MCF-7 tumor cells from Adriamycin cytotoxicity (14), thus supporting the involvement ofoxyradicals in the mechanism of cell kill. There is also someevidence that Adriamycin may be selectively toxic to well-oxygenated tumor cells (45), and another report indicated thatdifferential cytotoxicity of daunomycin in tumor cells was related to their ability to detoxify hydrogen peroxide by a gluta-thione-dependent mechanism (46). Most recently, Sinha et al.(15) demonstrated by electron spin resonance spin-trappingtechniques, that Adriamycin increased hydroxyl free radicalproduction in drug-sensitive MCF-7 breast tumor cells, but incontrast, hydroxyl radical generation in 200-fold Adriamycin-resistant MCF-7 tumor cells was nearly absent. This techniquespecifically identifies hydroxyl free radical. In a separate report,Sinha et al. (16) demonstrated that a 5-fold differential hydroxylradical formation in WT and ADRR tumor cells exposed to

Adriamycin was not the result of diminished activation ofAdriamycin by flavin monooxygenases, but could be attributedto enhanced activities of cytosolic reactive oxygen detoxifyingenzymes in the resistant tumors. These investigators also reported partial protection against Adriamycin toxicity in bothsensitive and resistant tumors by preincubating the cells withSuperoxide dismutase and catalase (16). Thus, it can be concluded, at least in MCF-7 breast tumor cells, that Adriamycincytotoxicity can be linked to the metabolic overproduction ofSuperoxide radical and hydrogen peroxide with conversion ofthese species to hydroxyl free radical, and possibly alkoxyradicals or metal-bound oxidants such as ferryl or perferryl ironspecies. We believe it is most likely hydroxyl radical thatdamages critical biomolecules because its reactivity is limitedonly by its diffusion rate.

The results presented in the present paper extend previousobservations of oxyradical participation in tumor killing byAdriamycin into the area of oxyradical involvement in thedevelopment of resistance to this agent. Clearly, ADRR tumor

cells selected for resistance to Adriamycin are significantly lesssusceptible than the WT Adriamycin-sensitive line to the toxiceffects of exogenous Superoxide and hydrogen peroxide generated by the xanthine oxidase/xanthine system. This differentialsusceptibility may be attributed to a twofold elevation in super-oxide dismutase activity in the resistant cells, however, morelikely, the difference in toxicity results from an increased detoxification of hydrogen peroxide, produced from the dismu-tation of Superoxide free radical, by the 7- to 12-fold elevatedglutathione peroxidase activity of the Adriamycin-resistant line.This interpretation is supported by the additional observationthat Adriamycin-resistant cells are similarly resistant to extracellular hydrogen peroxide, as demonstrated by the experimentswith the glucose oxidase system or by the direct addition ofhydrogen peroxide to the tumor cells.

The effectiveness of glutathione peroxidase in detoxifyinghydrogen peroxide by reducing it to water lies in its ability toscavenge extremely low concentrations of hydrogen peroxide,and further, glutathione peroxidase is capable of reducing secondary toxic aliphatic and aromatic peroxides to relatively inert

alcohols (47). The peroxidase activity of the ADRR tumors isassociated with the selenium-containing glutathione-peroxidase

based on its ability to reduce both hydrogen peroxide andeumene hydroperoxide. Non-selenium glutathione peroxidase

(glutathione transferase) cannot use hydrogen peroxide as asubstrate. This observation with WT and ADRR tumor cells

has also been reported by Kramer et al. (48). Recently, otherinvestigators in our laboratory have used a cDNA probe specificfor the human selenium-dependent peroxidase to show thatADRR tumor cells contain increased mRNA for this enzyme,

revealing that the increased peroxidase activity of these cells isassociated with the selenium-containing peroxidase.3

The comparative effectiveness of Superoxide dismutase, catalase, and glutathione peroxidase in protecting cells from thetoxic effects of oxyradicals was recently tested by quantitativemicroinjection of these enzymes into fibroblasts prior to exposing the fibroblasts to hyperoxia, and it was concluded thatglutathione peroxidase offered much more protection fromhyperoxia than the other two enzymes (49). In a variety of othertest systems, the importance of glutathione peroxidase as acellular defense against oxyradicals has been emphasized (50),thus we conclude that this enzyme is most likely responsiblefor the resistance toward oxyradicals by ADRR tumor cells.

While the xanthine oxidase and the glucose oxidase systemswere used to expose attached tumor cells to extracellular reactive forms of oxygen, the hyperoxia experiments compared thesusceptibility of the WT and the ADRR tumors to elevated

intracellular generation of oxyradicals. In these experiments,the difference in oxyradical toxicity between the two cell typeswas not apparent until several days of exposure to the elevatedoxygen concentrations. The divergence of the proliferationcurves of the WT line under euoxic and hyperoxic conditionsmay be the consequence of the inability of these cells to detoxifya constant increased rate of production of Superoxide andhydrogen peroxide by electron transport enzymes. It is likelythat the ADRR tumors were better able to withstand the elevated

intracellular production of oxyradicals because of their higherlevels of Superoxide dismutase and glutathione peroxidase. Atthis time we do not know what biochemical targets are involvedin hyperoxia toxicity to MCF-7 tumor cells or whether the

targets for oxyradicals generated inside the cells are the sameas those for oxyradicals formed outside the cells. The mostconspicuous target for extracellular oxyradicals would be thecell membrane; intracellular oxyradicals could damage multiplecritical biomolecules and could even attack and damage the cellmembrane from inside.

The production of Superoxide and hydrogen peroxide by bothcell types under the hyperoxic conditions would be expected tobe similar based upon the nearly equivalent activities ofNADPH-cytochrome P-450 reducÃase and NADH-dehydro-genase. However, it is possible that secondary biochemicalchanges occurred in the MCF-7 tumor cells during several daysgrowth under hyperoxic conditions. Among the first biochemical alterations in response to oxidative stress are the rapid lossof reduced glutathione and the inactivation of thiol-containingenzymes as a result of sulfhydryl group oxidation (21,51). Afterseveral days of hyperoxia, the levels of antioxidant defensesmight have increased selectively in the resistant tumor cells.These possibilities are currently under examination in both theWT and ADRR cells.

Oxyradical production by Adriamycin redox cycling is dependent upon the availability of adequate oxygen concentra-

3A. Townsend and K. Cowan, personal communication.

13




lions. Mimnaugh et al. (52) have shown that Adriamycin redoxcycling leading to greatly enhanced lipid peroxidation in amicrosomal system was very oxygen-dependent, with more than5-fold greater lipid peroxidation at 100% than at 21% oxygen.There is also evidence in intact tissue for an oxygen-dependenceof Adriamycin-caused toxicity, especially with respect to car-diotoxicity (53). In comparison, the pulmonary (54) and tumorcytotoxicity (55) of bleomycin and the pulmonary toxicity ofparaquat (56), agents believed to kill cells by generating oxyrad-icals, are enhanced when animals or tumor cells in culture areconcomitantly exposed to elevated (>21%) oxygen levels. Thus,there is an experimental rationale for the expectation thathyperoxia would potentiate the cytotoxicity of Adriamycin.This expected result was observed with both the WT and theADRR tumor cells, and the potentiation of Adriamycin cytotox

icity was more apparent with the sensitive line. The ICso valuefor Adriamycin with the WT cells under hyperoxic conditions(about 3 nM) was approximately one third of the value undereuoxic conditions, and considerably below the pharmacologicalplasma levels of Adriamycin in humans. Although there was ameasurable increase in the toxicity of Adriamycin to the resistant cells at 10 pM drug, hyperoxia was not especially valuablein reversing the resistance to Adriamycin by these cells. Again,the most obvious explanation for these differences is attributedto the increased ability of the ADRR cells to detoxify Superoxide

and hydrogen peroxide by the actions of Superoxide dismutaseand selenium-dependent glutathione peroxidase.

There were clearly differences in the susceptibility of thesensitive and drug-resistant tumor cells to Superoxide, hydrogenperoxide, hyperoxia, and especially Adriamycin, but both tumorlines were equally sensitive to gamma irradiation. Similar results were reported recently by Mitchell et al. (57), in that 30-fold daunomycin-resistant Chinese hamster ovary cells wereinherently as radiosensitive as the parental line, even thoughthe subline had approximately twofold higher levels of reducedglutathione, glutathione peroxidase, and catalase. An explanation for this apparent discrepancy may be that the damagingeffects of ionizing irradiation were directed preferentially toward DNA. Any increased ability of the ADRR cells or Chinese

hamster ovary cells to resist oxyradicals may have been bypassed because of the cytosolic localization of glutathione andglutathione peroxidase. Although Superoxide, hydrogen peroxide, and hydroxyl radical are all produced by irradiation, theradiolysis of water can produce hydroxyl radical without super-oxide and hydrogen peroxide intermediates (37). Neither su-peroxide dismutase nor glutathione peroxidase are capable ofenzymatically scavenging hydroxyl radical. Alternatively, the"oxygen-fixation" hypothesis (58) suggests that hydroxyl radi

cals formed by ionizing radiation abstract protons form criticaltargets such as DNA bases to form secondary radicals. Thesetarget radicals then react with molecular oxygen to form peroxyradicals and other secondary oxidative products (58, 59). Inthis scenario, enzymatic defenses against Superoxide or hydrogen peroxide would have no effect. This may explain why theADRR tumor cells were as sensitive as the WT cells to irradiation, even though the ADRR cells were better protected against

Superoxide and hydrogen peroxide.Results presented in this paper show the ADRR tumor cells

to be resistant to oxyradicals, however, the degree of resistancetoward Superoxide or hydrogen peroxide differed markedlyfrom the degree of resistance toward Adriamycin. We offer fourpossible interpretations of these observations. The first is thatAdriamycin cytotoxicity is only partially mediated by the generation of oxyradicals, and other mechanisms of toxicity not

only occur, but they may be additive and relatively more important than oxyradicals. The second is that Adriamycin-gen-erated oxyradicals may be important in the cytotoxicity of thedrug to WT tumor cells, especially because 10 to 20 nM concentrations of Adriamycin effectively kill most of the cells, andbecause WT cells generate much more hydroxyl radical thanADRR cells, at least at equivalent concentrations of Adriamycin

in the range of 50 to 100 ^M (16). By producing considerablymore hydroxyl radical as the result of the redox cycling ofAdriamycin, the WT tumors are killed by less than 10 nM drug;the ADRR tumor cells are not killed at these drug levels. Much

higher concentrations of Adriamycin are necessary for equivalent cytotoxicity to the ADRR cells, on the order of 5 to 10 ÃŸM

drug in continuous exposure, and it is possible that secondaryor tertiary mechanisms become important as the extracellularconcentration of Adriamycin is increased to these levels. Third,it is possible that Adriamycin redox cycles and generates oxyradicals at very specific sites in the tumor cells, perhaps withinthe mitochondria or even within the nucleus (60). Adriamycinjuxtaposed to DNA would be expected to cause more criticaldamage than at other, less vital, locations. A fourth explanationis that oxyradical generation, or other pharmacological eventsin the WT tumors exposed to Adriamycin, initiated biochemicaland genetic alterations that include increases in Superoxidedismutase and glutathione peroxidase activities only as part ofa catalogue of changes that collectively lead to markedly increased resistance to the drug. The cooperative functioning ofall of the changes in the Adriamycin-selected resistant tumors,that is, the P-glycoprotein-mediated decreased net accumulation of Adriamycin, the increased ability to detoxify oxyradicalsand secondary oxyradical products, and possibly other biochemical factors yet to be described such as differences in repairmechanisms or altered levels of targets, allow the ADRR tumor

cells to survive in and even to slowly proliferate in 10 nMAdriamycin. The key to understanding the phenomenon of drugresistance may lie in elucidating those other factors.

In conclusion, breast tumor cells selected for acquired resistance to Adriamycin in culture have concurrently developed atolerance to Superoxide and hydrogen peroxide, most likelybecause of elevated activities of enzymatic defenses againstoxyradicals. If the increased levels of enzymes that conferprotection against oxyradicals in the drug-resistant tumor cellscan be specifically inhibited, then at least a partial reversal ofAdriamycin resistance might result, and the therapeutic benefitof Adriamycin might be improved.

REFERENCES

1. Curt, G. A., Clendeninn, N. J., and Chabner, B. A. Drug resistance in cancer.Cancer Treat. Rep., 68: 87-99, 1984.

2. Beck, W. T. The cell biology of multiple drug resistance. Biochem. Pharma-col., 36: 2879-2887, 1987.

3. Myers, C., Cowan, K., Sinha, B., and Chabner, B. The phenomena ofpleiotropic drug resistance. In: V. T. DeVita, S. Hellman, and S. A. Rosenberg (oils.). Important Advances in Oncology 1987, pp. 27-38. Philadelphia:J. B. Lippincott Co., Inc., 1987.

4. Juliano, R. L., and Ling, V. A surface glycoprotein modulating drug permeability in Chinese hamster ovary cell mutants. Biochem. Biophys. Acta, Â¥55:152-162, 1976.

5. Somfai-Relle, S., Suzukake, K., Vistica, B. P., and Vistica, D. T. Reductionin cellular glutathione by buthionine sulfoximine and sensitization of murinetumor cells resistant to L-phenylalanine mustard. Biochem. Pharmacol., 33:485-490, 1984.

6. Hamilton, T. C., Winkler, M. A., Louie, K. G., Batist, G., Behrens, B.,Tsuruo, T., Grotzinger, K. R., Mckoy, W. M., Young, R. C., and Ozols, R.F. Augmentation of Adriamycin, melphalan and cisplatin cytotoxicity indrug-resistant and -sensitive human ovarian cancer cell lines by buthioninesulfoximine-mediated glutathione depletion. Biochem. Pharmacol., 34:2583-2589, 1985.

7. Mungikar, A., Chitnis, M., and Gothoskar, B. Mixed function oxidase

14




enzymes in Adriamycin-sensitive and -resistant sublines of P388 leukemia.Chem.-Biol. Interact., 35: 119-124, 1981.

8. Batist, G., Tulpule, A., Sinha, B. K., Katki, A. G., Myers, C. E., and Cowan,K. H. Overexpression of a novel anionic glutathione transferase in multidrug-resistant human breast cancer cells. J. Biol. Chem., 261:15544-15549,1986.

9. Cowan, K. H., Batist, G., Tulpule, A., Sinha, B. K., and Myers, C. E. Similarbiochemical changes associated with multidrug resistance in human breastcancer cells and carcinogen-induced resistance to xenobiotics in rats. Proc.Nati. Acad. Sci. USA, 83:9328-9332, 1986.

10. Abdella, B. R. J., and Fisher, J. A chemical perspective on the anthracyclineantitumor antibiotics. Environ. Health Perspect., 64: 3-18, 1985.

11. Lown, J. W. Molecular mechanisms of action of anticancer agents involvingfree radical intermediates. Adv. Free Radical Biol. Med., 1: 225-264, 1985.

12. Sato, S., Iwaizumi, M., Handa, K., and Tamura, Y. Electron spin resonancestudy on the mode of generation of free radicals of daunomycin, adriamycinand carboquone in NAD(P)H-microsome system. Gann, 68:603-608, 1977.

13. Doroshow, J. Role of hydrogen peroxide and hydroxyl radical formation inthe killing of Ehrlich tumor cells by anticancer quiÃ±ones.Proc. Nati. Acad.Sci., Â«5:4514-4518, 1986.

14. Doroshow, J. Prevention of Doxorubicin-induced killing of MCF-7 humanbreast cancer cells by oxygen radical scavengers and iron chelating agents.Biochem. Biophys. Res. Commun., 135: 330-335, 1986.

15. Sinha, B. K., Katki, A. G., Batist, G., Cowan, K. H., and Myers, C. E.Adriamycin-stimulated hydroxyl radical formation in human breast tumorcells. Biochem. Pharmacol., 36: 793-796, 1987.

16. Sinha, B. K., Katki, A. G., Batist, G., Cowan, K. H., and Myers, C. E.Differential formation of hydroxyl radical by Adriamycin in sensitive andresistant MCF-7 human breast tumor cells: implications for the mechanismof action. Biochemistry, 26: 3776-3781,1987.

17. Fairchild, C. R., Ivy, S. P., Sao-Shan, C. S., Whang-Peng, J., Rosen, N.,Israel, M. A., Melera, P. W., Cowan, K. H., and Goldsmith, M. E. Isolationof amplified and overexpressed DNA sequences from Adriamycin-resistanthuman breast cancer cells. Cancer Res., 47: 5141-5148, 1987.

18. Lyon, R. C., Cohen, J. S., Faustino, P. J., Megnin, F., and Myers, C. E.Glucose metabolism in drug-sensitive and drug-resistant human breast cancercells monitored by magnetic resonance spectroscopy. Cancer Res., 48: 870-877, 1988.

19. Yeh, G. C., Occhipimi, W. J., Cowan, K. H., Chabner, B. A., and Myers, C.E. Adriamycin resistance in human tumor cells associated with markedalterations in the regulation of the hexose monophosphate shunt and itsresponse to oxidant stress. Cancer Res., 47: 5994-5999, 1987.

20. Turrens, J. F., Freeman, B. A., and Crapo, J. D. Hyperoxia increaseshydrogen peroxide release by lung mitochondria and microsomes. Arch.Biochem. Biophys., 217:411-421, 1982.

21. Freeman, B. A., and Crapo, J. D. Biology of disease: free radicals and tissueinjuiry. Lab. Invest., 47:412-426, 1982.

22. Jamieson, D., Chance, B., Cadenas, E., and Boveris, A. The relation of freeradical production to hyperoxia. Am. Rev. Physiol., 48: 703-719, 1986.

23. Fridovich, I. Biological effects of the Superoxide radical. Arch. Biochem.Biophys., 247:1-11, 1986.

24. Gram, T. E., Litters!, C. I... and Mimnaugh, E. G. Enzymatic conjugationof foreign chemical compounds by rabbit lung and liver. Drug Metab. Dispos..2:254-258, 1974.

25. Mimnaugh, E. G., Trush, M. A., Bhatnagar, M., and Gram, T. E. Enhancement of reactive oxygen-dependent mitochondria! membrane lipid peroxi-dation by the anticancer drug Adriamycin. Biochem. Pharmacol., 34: 847-856, 1985.

26. Williams, C. H., Jr., and Kamin, H. Microsomal triphosphopyridine nucleo-tide-cytochrome c reducÃaseof liver. J. Biol. Chem., 237: 587-595, 1962.

27. Paglia, D. E., and Valentine, W. N. Studies on the quantitative and qualitativecharacterization of erythrocyte glutathione peroxidase. J. Lab. Clin. Med.,70:158-169, 1967.

28. Habig, W. H., Pabst, M. J., and Jakoby, W. B. Glutathione S-transferases.The first enzymatic step in mercapturic acid formation. J. Biol. Chem., 249:7130-7139, 1974.

29. Ellman, G. L. Tissue sulfhydryl groups. Arch. Biochem. Biophys., 82: 70-77, 1959.

30. McCord, J. M., and Fridovich, I. Superoxide dismutase an enzymatic function for erythrocuprein (hemocuprein). J. Biol. Chem., 244: 6049-6055,1969.

31. Azzi, S., Montecucco, C., and Richter, C. The use of acetylated ferricyto-chrome c for the detection of Superoxide radicals produced in biologicalmembranes. Biochem. Biophys. Res. Commun., 65: 597-603, 1975.

32. Lowry, O. H., Rosebrough, N. J., Fair, A. L., and Randall, R. J. Protein

measurement with the Folin phenol reagent. J. Biol. Chem., 193: 265-275,1951.

33. Snedecor, G. W. Statistical Methods, p. 45. Ames, IA: Iowa State UniversityPress, 1956.

34. Fridovich, I. The biology of oxygen radicals. Science (Wash. DC), 201: 875-880, 1978.

35. DiGuiseppi, J., and Fridovich, I. The toxicology of molecular oxygen. CRCCrit. Rev. Toxicol., 12: 315-342, 1984.

36. Halliwell, B., and Gutteridge, J. M. C. Oxygen toxicity, oxygen radicals,transition metals and disease. Biochem. J., 214: 1-14, 1984.

37. Greenstock, C. L. Oxy-radicals and the radiobiological oxygen effect. IsraelJ. Chem., 24: 1-10, 1984.

38. Slater, T. F. Free-radical mechanisms in tissue injury. Biochem. J., 222: 1-15, 1984.

39. Brawn, K., and Fridovich, I. DNA strand scisson by enzymatically generatedoxygen radicals. Arch. Biochem. Biophys., 206:414-419, 1981.

40. Myers, L. S., Jr. Free radical damage of nucleic acids and their componentsby ionizing radiation. Fed. Proc., 1882-1894. 1973.

41. Moody, C. S., and Hassan, H. M. Mutagenicity of oxygen free radicals. Proc.Nati. Acad. Sci. USA, 79: 2855-2859, 1982.

42. Cunningham, M. L., and Lokesh, B. R. Superoxide aniÃ³n generated bypotassium Superoxide is cytotoxic and mutagenic to Chinese hamster ovarycells. MutÃ¢t.Res., 121: 299-304, 1983.

43. Berlin, V., and Haseltine, W. A. Reduction of Adriamycin to a semiquinonefree radical by NADPH-cytochrome P-450 reducÃaseproduces DNA cleavagein a reaction mediated by molecular oxygen. J. Biol. Chem., 256:4747-4756,1981.

44. Eliot, H., Gianni, L., and Myers, C. E. Oxidative destruction of DNA by theAdriamycin-iron complex. Biochemistry, 23:928-936, 1984.

45. Tannock, I. Response of aerobic and hypoxic cells in a solid tumor toAdriamycin and cyclophosphamide and interaction of the drugs with radiation. Cancer Res., 42:4921-4926, 1982.

46. Bozzi, A., Mavelli, I., Mondovi, B., Strom, R., and Rotilio, G. Differentialcytotoxicity of daunomycin in tumour cells is related to glutathione-depend-ent hydrogen peroxide metabolism. Biochem. J., 194: 369-372, 1981.

47. Chance, B., Sies, H., and Boveris, A. Hydroperoxide metabolism in mammalian organs. Physiol. Rev.. 59: 527-606. 1979.

48. Kramer, R. A., Zakher, J., and Kim, G. Role of glutathione redox cycle inacquired and de novo multidrug resistance. Science (Wash. DC), 241: 694-697, 1988.

49. Raes, M., Michiels, C., and Remacle, J. Comparative study of the enzymaticdefense systems against oxygen-derived free radicals: the key role of glutathione peroxidase. Free Radical. Biol. Med., 3: 3-7, 1987.

50. Flohe, L. Glutathione peroxidase brought into focus. Pryor, W. A. (ed.), FreeRadicals in Biology, Vol. 5, pp. 223-354, New York: Academic Press, Inc.,1982.

51. Frank, L., and Massaro, D. Oxygen toxicity., Am. J. Med., 69: 117-126,1980.

52. Mimnaugh, E. G., Gram, T. E., and Trush, M. A. Stimulation of mouseheart and liver microsomal lipid peroxidation by anthracycline anticancerdrugs: characterization and effects of reactive oxygen scavengers. J. Pharmacol. Exp. Ther., 226:806-816, 1983.

53. Gianni, L., Cordon, B. J., and Myers, C. E. The biochemical basis ofanthracycline toxicity and antitumor action. Rev. Biochem. Toxicol., 5: 1-82, 1983.

54. Tryka, A. F., Godleski, J. J., and Brain, J. D. Differences in effects ofimmediate and delayed hyperoxia exposure on bleomycin-induced pulmonaryinjury. Cancer Treat. Rep., 68: 759-764, 1984.

55. Yamauchi, T., Raffin, T. A., Yang, P., and Sikic, B. I. Differential protectiveeffects of varying degrees of hypoxia on the cytotoxicities of etoposide andbleomycin. Cancer Chemother. Pharmacol., 19: 282-286, 1987.

56. Fisher, H. K., Clements, J. A., and Wright, R. R. Enhancement of oxygentoxicity by the herbicide paraquat. Am. Rev. Response Dispos., 107: 246-252, 1973.

57. Mitchell, J. B., Gamson, J., Russo, A., Friedman, N., DeGraff, W., Carmi-chael, J., and Glatstein, E. Chinese hamster pleiotropic multidrug-resistantcells are not radioresistant. Nati. Cancer Inst. Monogr., 6: 187-191, 1988.

58. Alper, T., and Howard-Flanders, P. Role of oxygen in modifying the nidiosensitivity of E. coli B. Nature (Lond.), 178: 978-979, 1956.

59. Schulte-Frohlinde, D., and VonSonntag, C. Radiolysis of DNA and modelsystems in the presence of oxygen, Â¡n;H. Sies (ed.), Oxidative Stress, pp.11-40. Orlando, FL: Academic Press, Inc., 1985.

60. Bachur, N. R., Gee, M. V., and Friedman, R. D. Nuclear catalyzed antibioticfree radical formation. Cancer Res., 42: 1078-1081, 1982.

15



1989;49:8-15. Cancer Res Edward G. Mimnaugh, Lata Dusre, John Atwell, et al. Tumor CellsAdriamycin-sensitive and -resistant MCF-7 Human Breast Differential Oxygen Radical Susceptibility of

Updated version

http://cancerres.aacrjournals.org/content/49/1/8

Access the most recent version of this article at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/49/1/8To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]



differential oxygen radical susceptibility of adriamycin ... · superoxide dismutase (3050 units/mg...

Documents