[CANCER RESEARCH 45, 4915-4920, October 1985]

Comparative Molecular Pharmacology in Leukemic L1210 Cells of theAnthracene Anticancer Drugs Mitoxantrone and Bisantrene1

George T. Bowden,2 Robin Roberts, David S. Alberts, Yei-Mei Peng, and Dava Garcia

Division of Radiation Oncology, Department of Radiology [G. T. B.], and Department of Medicine, Section of Hematology and Oncology [R. R., D. S. A., Y-M. P., D. G.J,College of Medicine, University of Arizona Health Sciences Center, Tucson, Arizona 85724

ABSTRACT

1,4-Dihydroxy-5,8-bis(2-[(2-hydroxyethyl)aminoethyl]aminoi-9,10-anthracenedione (mitoxantrone) and 9,10-anthracenedicar-boxaldehyde bis[(4,5-dihydro-1H-imidazol-2y)hydrazone] dihy-

drochloride (bisantrene) were evaluated for their abilities to causecytotoxicity and interact with cellular DNA using leukemic L1210cells. On a molar basis mitoxantrone has been found to be 7-

fold more toxic than bisantrene. Using a nucleoid sedimentationtechnique, bisantrene caused changes in DNA supercoiling whichwere characteristic of an intercalating drug, but mitoxantrone didnot induce these changes. Both drugs were found to interactwith cellular DNA with tight but noncovalent binding. Both drugsalso induced protein-associated double- and single-strand DNAbreaks, but with mitoxantrone only some of the DNA single-

strand breaks were protein associated, whereas with bisantreneall the DNA single-strand breaks were protein associated. The

cytotoxicity produced by bisantrene at a given frequency ofprotein-associated DNA strand breaks was low. However, with

mitoxantrone at an equivalent DNA strand break frequency, thecytotoxicity was high. Treatment of isolated L1210 nuclei witheither drug did not result in DNA single-strand breaks. It can be

concluded that bisantrene binds DNA in whole cells by an intercalâtive mechanism, whereas mitoxantrone binds DNA by anonintercalative, electrostatic interaction and induces non-protein-associated DNA strand breaks.

INTRODUCTION

Bisantrene3 and mitoxantrone (Chart 1) are two new anthra

cene drugs which have shown significant activity against a widevariety of animal tumor models (1-5). Both agents have provenactivity in patients with breast cancer (6-8); however, mitoxan

trone but not bisantrene appears effective in the treatment ofacute leukemia and lymphoma in Phase II clinical trials (9-12).

Although mitoxantrone and bisantrene were the first totally synthetic DNA binders to reach clinical trial and show promise ofuseful activity, the exact mechanism of action is unknown foreither of these drugs. Knowledge of their mechanism of actioncould aid in further development of the anthracene group ofcompounds.

The initial results of animal tumor studies as well as toxicolog-

ical and clinical investigations using bisantrene and mitoxantronecompared to doxorubicin suggested that these anthracene com-

1This research was supported by USPHS Grants CA-17343, CA-17094, andCA-23014.

2To whom requests for reprints should be addressed.'The abbreviations and trivial names used are: bisantrene, 9,10-anthracenedi-

carboxaldehyde bis[(4,5-dihydro-1H-imidazd-2-yl)hydrazone]dihydrochtoride; DSB,double-strand breaks; mitoxantrone, 1,4-dihydroxy-5,8-bis|2-[(2-riydroxyethyl)-aminoethyl]amino)-9-10-anthracenedione; SSB, single-strand breaks.

Received 8/6/84; revised 6/24/85; accepted 6/27/85.

pounds were similar to the anthracycline antibiotics. However,the anthracyclines and these anthracene drugs are quite differentin chemical characteristics, structure-activity relationships, and

dose responsiveness in terms of cytotoxicity (13). On thesebases, one would predict that their respective mechanisms ofaction are likely different.

Mitoxantrone has been shown to be distinctly different fromdoxorubicin or bisantrene in its cytotoxicity to P388 leukemia.Mitoxantrone was shown to be 20-fold more potent than doxorubicin and 100-fold more potent than bisantrene in a cytotoxicity

assay. These results suggest that there may be significantdifferences in the mechanism of cytotoxicity of mitoxantrone ascompared to bisantrene or doxorubicin.

We have observed similar differences between bisantrene andmitoxantrone in terms of cytotoxicity using mouse L1210 leukemia cells in culture. The planar structures of both drugssuggest that they may act as DNA intercalating agents andtherefore may induce protein-associated DNA strand scissions.

Using L1210 leukemia cells, we have explored potential differences between the two drugs in terms of their interactions withcellular DNA.

MATERIALS AND METHODS

Cell Cultures and Labeling Conditions. The L1210 cells used inthese experiments were supplied by R. A. G. Ewig, National CancerInstitute, Bethesda, MD. L1210 cells were grown in suspension cultureas described previously (14). Cultures were routinely tested for Myco-

plasma and found to be contamination free. Uniform labeling of the DNAwas obtained as described previously (14).

Drug Treatment. Unlabeled and 14C-radiolabeled bisantrene as wellas unlabeled and 3H- or 14C-radiolabeled mitoxantrone were obtained

from the Medical Research Division, American Cyanamide Co., PearlRiver, NY. Stock solutions of the drugs were made up, and the cellswere treated as described previously (14). Drug treatments were terminated by placing the cells at 4°Cand washing the cells twice with coldHanks' balanced salt solution. Cells were counted, and aliquots of the

appropriate cell number were used for the assays listed below.Cell Survival. Cells treated for 1 h with various concentrations of

bisantrene or mitoxantrone as described above were assayed for colonygrowth in soft agar by the method of MacPherson and Montagnier (15).Colonies were counted after 7 days of incubation at 37°C. Cloning

efficiency of control cells was 85 to 96%.Nucleoid Sedimentation on Neutral Sucrose Gradients. The sedi

mentation properties of nucleoids following drug treatment were assayedby a modified procedure (14) from Mattem and Painter (16).

Alkaline Elution Assay. The alkaline elution technique for the detection of protein-associated DNA single-strand breaks was carried out by

a modified procedure (14) described by Ross ef al. (17, 18). Data wereanalyzed to calculate DNA single-strand break frequencies using equa

tions developed by Ross ef al. (18). In this laboratory, it was found thatthe rate of elution of DNA from the filter was dependent on DNA single-strand size. When L1210 cells were X-irradiated on ice and then analyzed

CANCER RESEARCH VOL. 45 OCTOBER 1985

4915

on June 24, 2018. © 1985 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

MOLECULAR PHARMACOLOGY OF ANTHRACENES

MITOXANTRONE

OH _ NHCH2CH2NHCH2CH2OH

•2HCI

ÑHCH2CH2NHCH2CH2OH

BISANTRENE

2HCI

^H

. .. &HC=NN-

Chart 1. Structures of substituted anthraquinones.

by DNA alkaline elution (including proteinase K treatment), the fractionof DNA remaining on the filter at a fixed elution interval showed a first-order relationship to X-ray dose. Plots of the log of retention versos timeof elution were linear with X-irradiation. Alkaline sucrose sedimentationstudies have indicated that 1.064 breaks/109 nucleotides are produced

per rad of X-rays.

Neutral Elution Assay. A modification of the neutral elution assay fordetection of DNA double-strand breaks described by Bradley and Kohn

(19) was used.Isolation and Treatment of L1210 Nuclei. A modification of a pro

cedure described by Wozniak and Ross (20) and originally developed byFilipski and Kohn (21) was used to isolate L1210 nuclei. Briefly, wholecells were washed in cold Buffer A (1 mM KH2PO4-0.5 mM MgCI2-15 HIMNaCI-2 mM ethyleneglycol bis(j3-aminoethyl ether)-A/,W,N',A/'-tetraacetic

acid-4 mM disodium EDTA-0.3 mM spermine-1 mM spermidine) at pH

6.4. The cells were resuspended in 1 ml of the above buffer and werelysed with 9 ml of Buffer B (1 mM KH2PO4-0.5 mM MgClz-15 mM NaCI-1mM ethyleneglycol bis(/3-aminoethyl ether)-A/,A/,A/',W-tetraacetic acid)

plus 0.3% Triton X-100 (Eastman Kodak Co., Rochester, NY). Immedi

ately, 40 ml more of the original buffer were added, and the nuclei weresedimented by centrifugation at 900 x g for 15 min. The presence ofintact nuclei was confirmed by phase microscopy. The intact nuclei wereresuspended in 2 ml of the original buffer. The nuclei were then treatedwith drug for 1 h at 37°C, and the alkaline elution assay as described

above was performed to detect protein-associated DNA single-strand

breaks.CsCI Gradient Centrifugation of Cellular DNA. Whole L1210 cells

were treated with radiolabeled bisantrene or mitoxantrone for 1 h at37°C. The harvested cells were pelleted, and nuclei were isolated as

described above. Chromatin was isolated by sonicating the nuclei, andthe chromatin was centrifugea on a CsCI gradient (density = 1.7 g/ml).

The gradients were centrifuged at 33,000 rpm in a Beckman 50 Ti rotorfor 48 h. The gradients were fractionated from the bottom, the absor-

bance at 260 nm was read, and the radioactivity in each fraction wasdetermined by scintillation counting. Evidence for covalent binding of thedrug to DNA was sought by attempting extraction of the radioactivitywith normal butyl alcohol.

RESULTS

Cell Survival. Assessment of clonogenic survival of LI 210cells after a 1-h treatment with either bisantrene or mitoxantronewas made by studying anchorage-independent growth in soft

agar (Chart 2). On a molar basis, mitoxantrone was found to be7-fold more toxic than bisantrene (D0 bisantrene, 18 UM; D0

mitoxantrone, 2.5 ¿¿M).There was no evidence for a shoulder inthe survival curves for either drug.

Changes in DNA Supercoiling. Neutral sucrose gradientswere used to determine if mitoxantrone in vivo induced changesin DNA supercoiling which were indicative of DNA intercalation.In a previous publication (14), it was shown that bisantreneinduced a biphasic change in the sedimentation coefficient of thenucleoids as a function of drug concentration. This biphasicchange is typical for the classical intercalating drug, ethidiumbromide, and indicated that bisantrene acts as an intercalatingdrug. When L1210 cells were treated in suspension culture for1 h at 37°Cwith mitoxantrone and the resulting nucleoids were

sedimented through sucrose gradients, no statistically significantchange in the relative sedimentation distance resulted (Chart 3).As a positive control, a similar experiment with bisantrene wascarried out, and a typical biphasic response in the relative sedimentation was observed (Chart 3). There, however, remainedthe possibility that mitoxantrone could be inducing changes inDNA supercoiling due to DNA intercalation which were obscuredby non-protein-associated DNA strand breaks (we will show

evidence for these DNA SSB in Table 1). To eliminate the non-protein-associated DNA SSB, we isolated nuclei from L1210

cells, incubated the nuclei with various concentrations of mitoxantrone at ice temperature, and then sedimented nucleoidsthrough sucrose gradients as described previously. Incubationof the nuclei at ice temperature prevents DNA SSB but wouldallow intercalation if it occurred. We found no significant changein the sedimentation of treated nucleoids compared to controlnucleoids (data not shown). A positive control experiment wascarried out in which isolated nuclei were treated with bisantrene

20 30 40 SO 60 70 80 9O

>iMOLAR CONCENTRATION

Chart 2. Effects of mitoxantrone and bisantrene on L1210 cell survival. Exponentially growing L1210 cells were treated for 1 h at 37°Cwith different concentra

tions of each drug. Clonogenic survival was assayed by anchorage-independentgrowth in soft agar as described in "Materials and Methods." Points, mean survival;

bars, SE. •,bisantrene; O, mitoxantrone.

CANCER RESEARCH VOL. 45 OCTOBER 1985

4916

on June 24, 2018. © 1985 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

MOLECULAR PHARMACOLOGY OF ANTHRACENES

2.0

IO 20 30

>jMOLAR CONCENTRATION

Charts. Effects of mitoxantrone and bisantrene on nudeoid sedimentation.L1210 cells were treated for 1 h with various concentrations of mitoxantrone orbisantrene, and nucleoids were sedimented in neutral sucrose gradients as described in "Materials and Methods." To obtain the relative distance sedimented,

the peak fraction number for each sample containing cells which were treated withmitoxantrone (O) or bisantrene (•)divided by the total number of fractions wasdivided by the same parameter calculated for the untreated control sample in thesame centrifugaron run. This quantity (i.e.. relative distance sedimented) is normalized from one run to another. Points, mean; oars, SE.

at ice temperature, and nucleoids were sedimented throughsucrose gradients. We found a typical biphasic response in therelative sedimentation of these nucleoids (data not shown).Therefore, we have observed both in vivo at 37°C and with

isolated nuclei at ice temperature that bisantrene inducedchanges in DMA supercoiling typical of classical intercalatingdrugs, while mitoxantrone did not.

CsCI Banding of Drug-bound Cellular DNA. To determine

the nature of any potential drug binding to cellular DNA, L1210cells were treated for 1 h with either 14C-radiolabeled mitoxan

trone or bisantrene. Nuclear DNA was isolated from treated cellsand banded on neutral CsCI gradients. Both absorbance at 260nm and 14C radioactivity banded at a buoyant density of 1.7 g/

ml (Chart 4) for each drug. The data for bisantrene were obtainedbut are not shown. To determine whether the drug binding wascovalent, the peak fractions were pooled and extracted withnormal butyl alcohol. Greater than 95% of the 14C radioactivity

and presumably the drug were extractable from the DNA usingnormal butyl alcohol (data not shown). These data indicate thatboth bisantrene and mitoxantrone bind tightly but noncovalentlyto cellular DNA.

DNA Strand Break Induction. The formation of DNA single-

strand breaks in L1210 cells was examined following exposureto various concentrations of both bisantrene and mitoxantronefor 1 h. DNA single-strand breaks were quantitated using the

alkaline elution assay with or without proteinase K treatment ofthe cell lysates on the filters. As has been shown previously, no

MITOXANTRONE:

FRACTION NUMBER

Chart 4. Banding of DNA from L1210 cells treated with ["CJmitoxantrone on

neutral CsCI gradients. Exponentially growing L1210 cells were treated for 1 h withradiolabeled drug. Nuclear DNA was isolated and banded on CsCI gradients asdescribed in "Materials and Methods." Absorbance (260 nm) and "C radioactivity

(cpm) are plotted against the fraction number.

Table 1DNA SSB frequencies for mitoxantrone-treated cells with and without proteinase K

treatment

L1210 cells were treated for 1 h with the indicated dose of mitoxantrone. Celllysates were or were not treated with proteinase K before elution.

[Mitoxantrone](MM)10

102020No.

ofDNASSB/10"ProteinaseK nudeotides0.1

15 ±0.056*

+ 0.300 ±0.0560.376 ±0.022

+ 1.233 ±0.50a Mean ±SE of the SSB frequency from at least three experiments.

Table 2Comparison of cytotoxicity and frequency of DNA single-strand breaks between

mitoxantrone and bisantrene at equal molar concentrationsA comparison was made between the cytotoxicity and frequency of drug-induced

DNA single-strand breaks for mitoxantrone and bisantrene at equal molar concentrations.

DrugBisantrene

MitoxantroneMM10 10Survival

(%)49.9

2.4No.

ofDNASSB/10«

nudeotides0.331

0.311

detectable strand breaks were found with bisantrene withoutproteinase K treatment. However, DNA single-strand breaks

were detected without proteinase K when the cells were treatedwith mitoxantrone (Table 1). Approximately 66% of the strandbreaks were protein associated with mitoxantrone, whereas themajority of the strand breaks induced by bisantrene were proteinassociated (14). Mitoxantrone was also shown to induce DNAsingle-strand breaks in a dose-dependent manner (data notshown). A comparison of toxicity and frequency of DNA single-

strand breaks between bisantrene and mitoxantrone at equalmolar doses (Table 2) indicates that treatment with the drugsresulted in the same level of DNA single-strand breaks but

completely different levels of reduction in cell survival. Mitoxantrone was 20-fold more toxic than bisantrene at the same dose

level and at the same level of DNA strand breaks. The relationshipbetween DNA single- and double-strand breaks was also deter

mined for both drugs utilizing a newly developed neutral elutionassay to measure the frequencies of the double-strand breaks.

CANCER RESEARCH VOL. 45 OCTOBER 1985

4917

on June 24, 2018. © 1985 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

MOLECULAR PHARMACOLOGY OF ANTHRACENES

Due to the nature of the neutral elution assay, it was not possibleto determine the absolute DNA double-strand break frequency.

Instead, this frequency was determined in ionizing radiation radequivalents. With this information, it was possible to determinethe absolute ratio of DNA single-strand breaks to DNA double-

strand breaks by comparing the rad equivalents for the twotypes of DNA strand breaks and knowing the estimated ratio ofDNA single-strand to double-strand breaks produced by ionizing

radiation as determined by a gradient technique. The reportedvalues for the ratio of X-ray-induced single- to double-strand

breaks using alkaline and neutral sucrose gradients have beenin the range of 10 to 40. DNA single-strand break frequenciesdetermined using alkaline elution assay include DNA single-strand breaks arising from DNA double-strand breaks as well astrue single-strand breaks. The relationship between measured

and true break frequencies is given in Table 3 for bisantrene andmitoxantrone. For bisantrene these ratios ranged from 0.3 to7.2, and for mitoxantrone the values ranged from 1.0 to 10.

Treatment of Isolated L1210 Nuclei with Bisantrene andMitoxantrone. To investigate further the potential difference inthe mechanisms involved in the production of DNA single-strandbreaks induced by bisantrene and mitoxantrone, the productionof single-strand breaks in treated, isolated nuclei was studied.Isolated nuclei were treated at 37°Cfor 1 h, and the nuclei were

lysed and treated on the filters with proteinase K. The results ofthese experiments (Chart 5) indicated that both bisantrene andmitoxantrone require the presence of cytoplasm to produce DNAstrand breaks. Published evidence indicates that intercalatingagents at high concentrations can actually inhibit strand breakage by topoisomerase (22), and thus it was possible that failureto observe strand breaks in the isolated nuclei was related tothe nature of the dose-response curve and not a requirement for

cytoplasm. Therefore, we treated isolated nuclei with lowerconcentrations (0.1 and 1 /¿M)of bisantrene and mitoxantrone.Again we observed no significant DNA SSB even at these towconcentrations (data not shown). A control experiment wascarried out to show that the nuclei isolation procedure did notinterfere with the detection of DNA strand breaks. Intact cellswere treated with bisantrene or mitoxantrone, and nuclei were

Table 3Relationship between DNA single-strand breaks and double-strand breaks

The frequencies of DNA SSB and DSB induced by bisantrene and mitoxantronewere determined for two different concentrations of each drug as described in'Materials and Methods."

S/DDrug

(MM)Bisantrene

1020

Mitoxantrone5

10[SSB]8

[DSB]0.23

0.150.24

0.30krskrd0.3

-0.50.41.0krskrd"4°7.24.07.610.0

* [SSB], measured SSB frequency in SSB rad equivalents; [DSB], measured

DSB frequency in DSB rad equivalents; S, frequency of true SSB (not includingthose arising from DSB); D, true DSB frequency; krs and krd, efficiencies of SSBand DSB production per roentgen of X-ray; S/D calculated according to thefollowing equation

S = krs [SSB] _D ~ krd [DSB]

taken from Zwellmg et al. (32).

234 i 6 7 8 9 IO II

NO. of 4 min FRACTIONS

Charts. Effects of bisantrene on alkaline elution kinetics from drug-treatedisolated nuclei or drug-treated whole cells followed by nuclei isolation. IsolatedL1210 nuclei were treated for 1 h with bisantrene (O. O MM;O, 10 MM;A, 20 MM;and D, 40 MM) before carrying out the alkaline elution assay with proteinase Ktreatment as described in "Materials and Methods." In a control experiment, whole

cells were treated for 1 h with bisantrene (0, 0 MM;•,10 MM;A, 20 MM;and •40MM), and then nuclei were isolated. The nuclei were then run with the alkalineelution assay using proteinase K.

isolated from these cells. DNA single-strand breaks were de

tected in these nuclei (Chart 5). Therefore, the nuclei isolationprocedure did not interfere with the detection of DNA strandbreaks.

DISCUSSION

Bisantrene and mitoxantrone are two new anthracene derivatives which are undergoing clinical trial for treatment of humancancers. Both drugs have proven activity in patients with meta-static breast cancer (6-8), but only mitoxantrone has also shown

significant activity against acute lymphocytic and nonlymphocyticleukemia, malignant lymphomas, and hepatomas (9-12). Al

though the mechanism(s) of action of these drugs is not understood, a number of investigators have clearly established thatthe drugs inhibit RNA and DNA synthesis and intercalate DNA inin vitro reactions (23-25). The purpose of our investigations was

to study in vivo the nature of the interactions of these two drugswith cellular DNA.

Bisantrene does behave as a classical intercalating drug invivo (14), since we have shown that bisantrene causes changesin DNA supercoiling typical of DNA intercalating agents andprotein-associated DNA breaks which are detectable by neither

alkaline elution nor nucleoid sedimentation in the absence ofproteinase K. These protein-associated DNA breaks could therefore correspond to topoisomerase II-DNA complexes trapped by

the intercalator (26, 27). In contrast, we have shown that mitoxantrone, which on a molar basis is more toxic than bisantrene,induces DNA SSB which are non-protein associated. We havealso shown that mitoxantrone does not act as a classical inter-

CANCER RESEARCH VOL.45 OCTOBER 1985

4918

on June 24, 2018. © 1985 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

MOLECULAR PHARMACOLOGY OF ANTHRACENES

calating drug in changing DNA supercoiling, but instead maycause in vivo DNA compaction due to electrostatic interactionwith the anionic exterior of the DNA helix. Therefore, we haveshown that there are major differences in the way bisantreneand mitoxantrone interact with cellular DNA.

The question arises as to why DNA strand breakage does notaffect nucleoid sedimentation when cells are treated at 37°C

with bisantrene or mitoxantrone. Since DNA strand breaks areprotein associated with bisantrene, it is possible that covalentlinkage of integral proteins to DNA in the nucleoid would preventDNA strand breaks from affecting supercoiling. Since the majorityof the DNA strand breaks induced by mitoxantrone are proteinassociated, the non-protein-associated DNA strand breaks which

could affect supercoiling may not be at a level detectable by thenucleoid sedimentation technique.

Mitoxantrone is structurally related to a seríesof substitutedanthraquinones and has features known to be essential for DNAintercalation (planar, electron-rich chromophore). The presence

of side chains bearing basic amino groups which could bindelectrostatically to the phosphate groups of DNA may enhanceintercalation. Indeed, mitoxantrone and related drugs have beenshown to intercalate purified DNA in the test tube (23,28). Mostof these drugs show strong DNA binding as evidenced by ATmvalues equal to or greater than those produced by the anthra-

cycline antibiotics. However, the ability of these drugs to increasethe melting temperature of DNA and to inhibit nucleic acidsynthesis in vitro does not always correlate with in vivo antitumoractivity (23). This raises the possibility of a mechanism of actionother than or in addition to DNA intercalation.

Data presented in this paper suggest that mitoxantrone whenadded to intact cells does not act as a classical intercalatingdrug as measured by the nucleoid sedimentation technique.However, we have reported in this paper the rather strongbinding of mitoxantrone to cellular DNA using CsCI gradientsedimentation of DNA extracted from intact cells treated withmitoxantrone. This tight binding could be associated with theelectrostatic interaction of mitoxantrone with the anionic exteriorof the DNA. Scatchard plots of mitoxantrone binding to DNA byequilibrium dialysis analysis indicated two binding sites. Thestrong binding site may be associated with intercalation into DNAbase pairs, and the weaker binding site may possibly be associated with secondary binding of the side chains by electrostaticinteraction with the anionic exterior of the DNA helix. Therefore,there remains the possibility that even though mitoxantronecontains the requisite planar chromophore, the presence of twoextended alkyl residues in positions 1 and 4 may preclude thesmooth incorporation of all parts of the molecule. As pointed outby Durr ef al. (23), this is supported by the observed greaterintercalative binding of the less hindered ethidium molecule.These observations have led some investigators to speculatethat DNA binding includes both partial intercalation and externalbinding.

It has also been shown that mitoxantrone is capable of inducing condensation and compaction of isolated chromatin. Thishas been attributed to extensive supercoiling of the DNA (29).Factors in addition to intercalation may play a role, since ethidiumbromide is inactive in this system. These include ionic interactionsof inter- and intrastrand cross-links in addition to DNA intercala

tion. Lown ef al. (30), using flow microfluorimetry measurementson relaxed PM2 DNA reacted with mitoxantrone, have shown

the formation of lace-like networks of DNA linked together in a

compacted fashion. This phenomenon has been attributed tointer-DNA linking by the charged side arms of mitoxantrone.

The production of protein-associated DNA single- and double-

strand breaks has been associated with the action of DNAintercalating drugs. We have previously shown (14) that bisantrene induces both protein-associated DNA single-strand breaksas well as DNA-protein cross-links in a ratio of one to one.Bisantrene-induced DNA strand breaks appear to be protein

associated. In contrast, we have found that mitoxantrone induces both protein-associated and non-protein-associated DNA

strand breaks. Of interest is the fact that mitoxantrone containsa hydroquinone structure which could potentially generate freeradicals during NADPH and 02 reaction with reducíaseenzymes,and DNA strand breaks can be generated by this mechanism(3D-

Besides the non-protein-associated DNA strand breaks, mitoxantrone induces protein-associated strand breaks which

could have resulted from the partial insertion of the mitoxantronemolecule into the DNA. There is also in vitro cell-free experimental

evidence that drugs with either weak or no intercalating activityhave the ability to induce the formation of a topoisomerase II-

DNA complex (26). This same complex could be formed in a celltreated with these drugs.

Experimental results reported by Zwelling ef a/. (32) and Rossand Bradley (33) have indicated that there are wide differencesbetween ratios of SSB to DSB produced by different intercalatingagents in the order acridinyl anisidide > Adriamycin > ellipticine.It has been suggested that ellipticine produces DNA DSB almostexclusively (17). Our results obtained with both bisantrene andmitoxantrone indicate ratios of SSB to DSB very similar to thosefound for Adriamycin and certainly less than those found foracridinyl anisidide. It has been speculated that, if the DNA strandbreaks were induced enzymatically, the intercalators may differwidely in their relative stimulation of topoisomerases producingsingle-strand versus double-strand breaks.

Besides identifying the types of DNA damage induced bybisantrene and mitoxantrone in treated whole cells, we havedetermined the action of the drugs on isolated nuclei. We foundthat neither drug induced DNA single-strand breaks (protein ornon-protein associated) in isolated nuclei. Thus, it would appear

that the cytoplasm of intact cells is necessary for drug action.However, we cannot exclude the possibility that the failure tosee DNA strand breakage in isolated nuclei could be a functionof the buffer system used for these experiments. Recently Woz-niak and Ross (20) showed that the podophyllotoxin VP-16 [4'-

demethylepipodophyllotoxin-9-(4,6-O-ethylidene-/S-D-glucopy-ranoside) (etoposide)] does induce both DNA SSBs and double-strand breaks in isolated nuclei, indicating that for VP-16 cyto-

plasmic components are not required for drug action.As other investigators have reported, we have not found a

positive con-elation between the frequency of protein-associated

DNA strand breaks and the level of cytotoxicity induced bybisantrene and mitoxantrone. Pommier ef a/. (27) have reportedin studies of the modification of intercalator-induced protein-

associated DNA strand breaks by thiourea and dimethyl sulfoxidethat alterations in intercalator-induced DNA scission were not

accompanied by corresponding alterations in cytotoxicity. Ourdata taken together with Pommier's data would suggest a clear

dissociation of intercalator-induced strand break production and

CANCER RESEARCH VOL. 45 OCTOBER 1985

4919

on June 24, 2018. © 1985 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

MOLECULAR PHARMACOLOGY OF ANTHRACENES

lethality.In conclusion, data have been presented which are consistent

with the idea that the two anthracene drugs, mitoxantrone andbisantrene, differ from each other in the way they interact withDMA in the intact cell. Though not directly addressed in thispaper, this difference could indicate that mitoxantrone and bisantrene have different mechanisms of cytotoxic action. Thedifferences in mechanisms of action could explain differences inthe spectrum of activity of the drugs in preclinical and clinicaltrials.

REFERENCES

1. Citarella. R. B., Wallace, R. E., Murdock, K. C., Angter, R. B., and Darr, R. E.Antitumor activity of CI 216,942: 9,10-anthracenedicarboxaldehyde bis[(4,5-dihydrc-1 H-imidazol-2-yl)hydrazone] dihydrochtoride. In: Abstract of the 20thInterscience Conference on Antimicrobial Agents and Chemotherapy. Be-

thesda, MO: American Society for Microbiology, 1980.2. Corbe«, T. H., Roberts, B. J., Trader, M. W„Laster, W. R., Jr., GriswokJ, D.

P., Jr., and Schabe!, F. M., Jr. Response of transplantable tumors of mice toanthracenedione derivatives alone or in combination with clinically usefulagents. Cancer Treat. Rep., 66: 11876-12000,1982.

3. Fujimoto. S., and Ogawa, M. Antitumor activity of mitoxantrone against murineexperimental tumors: comparative analysis against various antitumor antibiotics. Cancer Chemother. Pharmacol., 8:157-162,1982.

4. Johnson, R. K., Zee-Cheng, R. K. Y., Lee, W. W., Acton, E. M., Henry, D. W.,and Cheng, C. C. Experimental antitumor activity of aminoanthraquinones.Cancer Treat. Rep., 63: 425-439,1979.

5. Schädel,F. M., Corbe«, T. H., GriswokJ, D. P., Laster, W. R., and Trader, M.W. Therapeutic activity of mitoxantrone and ametantrone against murinetumors. Cancer Treat. Rev., 10 (Suppl. BV 12-21,1983.

6. Osbom, C. K., Von Hoff, D. D., Cowan, J. D., and Sandbach, J. Bisantrene,an active drug in patients with advanced breast cancer. Cancer Treat. Rep.,68:357-360,1984.

7. Yap, H-Y., Blumenschein, G. R., Schell, F. C., Buzdar. A., Valdivieso, M., andBodey, G. P. Dihydroxyanthracenedione: a promising new drug in the treatment of metastatic breast cancer. Ann. Int. Med., 95: 694-697,1981.

8. Yap, H-Y., Yap, B-S., Blumenschein, G. R., Barnes, B. C., Schell, F. C., andBodey, G. P. Bisantrene, an active new drug in the treatment of metastaticbreast cancer. Cancer Res., 43:1402-1404,1983.

9. Coltman, C. A., Jr., McDaniel, T. M., Balcerzak, S. P., Morrison, F. S., andVon Hoff, D. D. Mitoxantrone hydrochloride (NSC-310739) in lymphoma.Invest. New Drugs, 1: 65-70,1983.

10. Estey, E. H., Keating, M. J., McCredie, K. B., Bodey, G. P.. and Freireich, E.J. Phase II trial of dihydroxyanthracenedione in acute leukemia. Proc. Am.Assoc. Cancer Res.. 23: 113, 1982.

11. Gerad, H., Van Echo, D. A., Donehower, M., Pocelmko. R., and Wiemik, P. H.Phase l-ll trial of anthracenedicarboxaldehyde in patients with refractory adultacute leukemia. Cancer Treat. Rep., 68: 535-537,1984.

12. Van Echo, D. A., Shulman, P. N., Ferrari, A., Budman, D., Markus, S. D., andWiemik, P. H. A Phase II trial of mitoxantrone (DHAD, NSC 301739) in adultacute leukemia (AL). Proc. Am. Soc. Clin. Oncol., 7:132,1982.

13. Johnson, R. K., Broome, M. G., Howard, W. S., Evans, S. F., and Pritchard,D. F. Experimental therapeutic and biochemical studies of anthracenedionederivatives. In: M. Rozencweig. D. D. VanHoff, and M. J. Stagne! (eds.), NewAnticancer Drugs: Mitoxantrone and Bisantrene. New York: Raven Press,1983.

14. Bowden, G. T., Garcia, D., Peng, Y. M., and Alberts, D. S. Molecular pharmacology of the anthracycline drug 9,10-anthracenedicarboxaldehyde bis[(4,5-

dihydro-1H-imidazol-2-yl)hydrazone]dihydrochloride (CI 216, 942). CancerRes., 42: 2660-2665,1982.

15. MacPherson, I., and Montagnier, L. Agar suspension culture for the selectiveassay of cells transformed by polyoma virus. Virology, 23: 291-294,1964.

16. Mattem, M. R., and Painter, R. B. Dependence of mammalian DNA replicationon DNA supercoiling. I. Effects of ethidium bromide on DNA synthesis inpermeable Chinese hamster ovary cells. Biochim. Biophys. Acta, 563: 293-305,1979.

17. Ross, W. E., Glaubiger, D. E., and Kohn, K. W. Protein-associated DNA breaksin cells treated with Adriamycin or ellipticine. Biochim. Biophys. Acta, 519:23-30.1978.

18. Ross, W. E., Glaubiger, D. E., and Kohn, K. W. Qualitative and quantitativeaspects of intercalator-induced DNA strand breaks. Biochim. Biophys. Acta,562: 41-50,1979.

19. Bradley, M. 0., and Kohn, K. W. X-ray induced DNA double-strand break

production and repair in mammalian cells as measured by neutral filter elution.Nucteic Acid Res., 7: 793-804, 1979.

20. Wozniak, A., and Ross, W. E. DNA damage as a basis for 4'-demethylepipo-dophyllotoxin 9-(4.6-O-ethylidene-.>'-D-glucopysanosideXetoposide) cytotoxic-

ity. Cancer Res., 43:120-124,1983.21. Filipski, J., and Kohn, K. W. Ellipticme-mduced protein-associated DNA breaks

in isolated L1210 nuclei. Biochim. Biophys. Acta, 698: 280-286, 1982.22. Pommier, Y., Schwartz, R., Minford, J., Kohn, K. W., and Zwelling, L A.

Effects Of 2-methyl-9~hydroxyellipticinium (HOEME') on DNA double-strand

break (DSB) formation by a type II DNA topoisomerase activity in isolatednuclei from L1210 cells. Proc. Am. Assoc. Cancer Res., 25: 296, 1984.

23. Durr, F. E., Wallace, R. E., and Citarella. R. V. Molecular and biochemicalpharmacology of mitoxantrone. Cancer Treat. Rev., 10 (Suppl. B): 3-11,1983.

24. Kapuscinski, J., Darzynkiewicz, Z., Tráganos, F., and Melamed. M. R. Interactions of a new agent, 1.4-dihydroxy-5.8-bis|(2-(2-nydroxyethyl)amino)-ethylamino]-9,10-anthracenedione, with nucleic acids. Biochem. Pharmacol.,30:234-240,1981.

25. Plumbridge, T. W., Knight, V., Patel. K. L., and Braun, J. R. Mechanism ofDNA-binding of some aminoalkylamino-derivatives of anthraquinone and na-phthacenequinone. J. Pharmacol., 32: 78-80,1980.

26. Nelson, E. M., Tewey, K. M., and Uu, L. F. Mechanism of antitumor drugaction: poisoning of mammalian DNA topoisomerase II in DNA by 4'-(9-acndinylamino)methanesulton-m-anisidide Proc. Na«.Acad. Sci. USA, 87:1361-1365,1984.

27. Pommier, Y., Zwelling, L. A., Mattem, M. R., Erickson, L. C., Kerrigan, D.,Schwartz. R„and Kohn, K. W. Effects of dimethyl sulfoxide and thiourea uponintercalator-induced DNA single-strand breaks ¡nmouse leukemia (L1210)cells. Cancer Res., 43: 5718-5724, 1983.

28. Foge, W. G., Vajrogupta, 0. P. A., and Sengupta, S. K. DNA-binding specificityand RNA polymerase inhibitory activity of bis(ammo alkyl)anthraquinones andbis(methylthio)vinylquinone iodides. J. Pharm. Sci. 77:253-257,1982.

29. Waldes, H., and Center, M. S. Adriamycin-induced compaction of isolatedchromatin. Biochem. Pharmacol., 37:1057-1061,1982.

30. Lown, J. W., Hanstock, C. C., Bradley, R. D., and Scraba. D. G. Interactionsof the antitumor agents, mitoxantrone and bisantrene, with deoxynbonucleicacids studied by electron microscopy. Mol. Pharmacol., 25:178-184,1984.

31. Berlin, V., and Haseltine, W. A. Reduction of Adriamycin to a semiquinone-freeradical by NADPH cytochrorne P-450 reducíaseproduces DNA cleavage in areaction mediated by molecular oxygen. J. Btol. Chem., 256: 4747-4756,

1981.32. Zwelling, L. A., Michaels, S., Erickson, L. C., Ungerleider. R. S., Nichols, M.,

and Kohn, K. W. Protein-associated deoxynbonucleic acid strand breaks inLI210 cells treated with the deoxynbonucleic acid intercalating agents 4'-(9-

acrindinylamino)methanesulfon—m-anisidide and Adriamycin. Biochemistry,20.6553-6563,1981.

33. Ross, W. E., and Bradley, M. O. DNA double strand breaks in mammalian cellsafter exposure to intercalating agents. Biochim. Biophys. Acta, 654:129-134,

1981.

CANCER RESEARCH VOL. 45 OCTOBER 1985

4920

on June 24, 2018. © 1985 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1985;45:4915-4920. Cancer Res   George T. Bowden, Robin Roberts, David S. Alberts, et al.   Bisantreneof the Anthracene Anticancer Drugs Mitoxantrone and Comparative Molecular Pharmacology in Leukemic L1210 Cells

  Updated version

  http://cancerres.aacrjournals.org/content/45/10/4915

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

  .pubs@aacr.orgDepartment 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/45/10/4915To request permission to re-use all or part of this article, use this link

on June 24, 2018. © 1985 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from