Doxorubicin cardiomyopathy is associated with a decrease in calcium release channel of the sarcoplasmic reticulum in a chronic rabbit model

Doxorubicin cardiomyopathy is associated with

a decrease in calcium release channel of the

sarcoplasmic reticulum in a chronic rabbit

model.

D A Dodd, … , S Fleischer, R J Boucek Jr

J Clin Invest.

1993;

91(4)

:1697-1705.

https://doi.org/10.1172/JCI116379

.

Doxorubicin is a highly effective cancer chemotherapeutic agent that produces a

dose-dependent cardiomyopathy that limits its clinical usefulness. Clinical and animal studies of

morphological changes during the early stages of doxorubicin-induced cardiomyopathy

have suggested that the sarcoplasmic reticulum, the intracellular membrane system

responsible for myoplasmic calcium regulation in adult mammalian heart, may be the early

target of doxorubicin. To detect changes in the calcium pump protein or the calcium release

channel (ryanodine receptor) of the sarcoplasmic reticulum during chronic doxorubicin

treatment, rabbits were treated with intravenous doxorubicin (1 mg/kg) twice weekly for 12 to

18 doses. Pair-fed controls received intravenous normal saline. The severity of

cardiomyopathy was scored by light and electron microscopy of left ventricular papillary

muscles. Developed tension was measured in isolated atrial strips. In subcellular fractions

from heart, [3H]ryanodine binding was decreased in doxorubicin-treated rabbits (0.33

+/-0.03 pmol/mg) compared with control rabbits (0.66 +/- 0.02 pmol/mg; P < 0.0001). The

magnitude of the decrease in [3H]ryanodine binding correlated with both the severity of the

cardiomyopathy graded by pathology score (light and electron microscopy) and the

decrease in developed tension in isolated atrial strips. Bmax for [3H]ryanodine binding and

the amount of immunoreactive ryanodine receptor by Western blot analysis using

sequence-specific antibody were both decreased, consistent with a decrease in the amount

of calcium release channel of […]

Research Article

Doxorubicin

Cardiomyopathy Is Associated with

a

Decrease in

Calcium

Release

Channel

of the

Sarcoplasmic

Reticulum in

a

Chronic Rabbit Model

Debra A. Dodd, * James B. Atkinson,t Richard D. Olson,' Scott Buck, * BarryJ. Cusack,' SidneyFleischer,' and Robert J. Boucek, Jr.*I

*Division of Pediatric Cardiology, and Departments of

*Pathology,

§Molecular Biology, andlBiochemistry, Vanderbilt University MedicalCenter, Nashville, Tennessee37232;and'Department ofPharmacology, Veterans Administration, Boise, Idaho 83702

Abstract

Doxorubicin is a highly effective cancer chemotherapeutic agent thatproducesadose-dependent cardiomyopathy that lim-its lim-itsclinical usefulness. Clinicaland animal studies ofmorpho-logical changesduringthe early stages of doxorubicin-induced

cardiomyopathy have suggested that the sarcoplasmic reticu-lum,the intracellular membrane system responsible for myo-plasmic calcium regulation in adult mammalian heart, may be theearly targetofdoxorubicin. Todetectchanges in the

cal-ciumpumpproteinorthe calcium release channel(ryanodine receptor)of thesarcoplasmic reticulumduring chronic

doxoru-bicintreatment,rabbitsweretreatedwith intravenous

doxoru-bicin (1 mg/kg) twiceweekly for 12 to18doses.Pair-fed con-trols received intravenous normalsaline.Theseverity of

cardio-myopathy was scored by light and electronmicroscopy of left

ventricular papillary muscles. Developed tension was measured in isolated atrial strips. In subcellular fractions from heart,

I3Hjryanodine binding

was decreased in doxorubicin-treated rabbits(0.33±0.03 pmol/mg) compared with control rabbits

(0.66±0.02

pmol/mg,

P<0.0001).Themagnitudeof the de-creasein

I3Hjryanodine

bindingcorrelated with both the sever-ityofthecardiomyopathy gradedby pathology score(lightand electronmicroscopy)and the decrease in developed tension in isolated atrialstrips.

B.

for

[3Hlryanodine

binding and the amountofimmunoreactiveryanodinereceptorby Western blot

analysis using sequence-specific antibodywerebothdecreased,

consistent witha decrease in the amount of calcium release channelofsarcoplasmic reticulumindoxorubicin-treated

rab-bits. In contrast, there was nodecrease in theamount orthe

activity ofthe calcium pump

protein

ofthe

sarcoplasmic

reticu-lum indoxorubicin-treated rabbits. Doxorubicintreatmentdid notdecrease

[3H

_I

ryanodine bindingortheamountof immunore-active calcium release channel of

sarcoplasmic

reticulum in skeletal muscle. Since the

sarcoplasmic

reticulum regulates

musclecontraction bythecyclic uptakeand releaseofalarge

internalcalciumpool, altered functionofthecalciumrelease channel could leadtotheabnormalities of contractionand relax-ation observed in the doxorubicin

cardiomyopathy. (J.

Clin.

A portion of thispaperhas been presented inabstract form ( 1991. Biophys.J.59:lOla).

Address correspondenceto Dr.DebraA.Dodd,Vanderbilt

Univer-sityMedical Center, Division of Pediatric Cardiology, D-22 17MCN, Nashville,TN37232-2572.

Receivedfor publication21October1991and inrevisedform 19 October1992.

TheJournal ofClinicalInvestigation,Inc. Volume91,April 1993, 1697-1705

Invest. 1993. 91:1697-1705.) Key words: binding * cardiac-chemotherapy . radioligand-ryanodine receptor

Introduction

Doxorubicin is a highly effective cancer chemotherapeutic agent, butits clinical usefulness is limiteddue to the develop-mentofa dose-dependent cardiomyopathy (1-7). The total dose isusually limitedto450-500mg/m2body surface area,

sincetheincidence ofthecardiomyopathy is"low" below this dose. However, more than half ofthepatients could tolerate

highertotal doses(potentially neededto treattheir malignan-cies) without development of cardiomyopathy, whereas a

small percentofpatientswill develop the cardiomyopathyat eventheselow doses(8, 9). The abnormalities of

contraction

and relaxation associated with this cardiomyopathy are

unique. Forexample, some patients that develop

congestive

heartfailure duringdoxorubicinchemotherapyexhibitaslow

improvement in cardiac function, whichsuggests that the

doxo-rubicin-induced cardiac dysfunction may be reversible

(10-12). In contrast, a small percentage of patients with normal cardiac function after completing doxorubicin chemotherapy developmyocardial

dysfunction

years later(13, 14).

Preven-tion ofthiscardiomyopathywill

require

a moredetailed

charac-terization of its etiology.

Endomyocardial biopsyhas been used to monitor for the

doxorubicin cardiomyopathy. Billingham ( 15) described the

morphological changes seen on biopsy specimens from

pa-tientsreceiving doxorubicin. The earliestchanges aredistended sarcoplasmic reticulum and early myofibrillarloss (15, 16). Laterchanges suggestdiffusecell damage withdegeneration of multiple cellular organelles. The morphological changes on

biopsy appearto precede overt heart failure( 15, 17). These

earlymorphologic abnormalities ofthesarcoplasmic reticulum

havebeen

described

inanimalmodels as well ( 18-23). Thesarcoplasmic reticulum regulates theintracellular

cal-ciumstoresonwhich adult mammalian cardiacmuscleis

de-pendent for contraction. The calciumpump

protein,

which is involvedinenergized calciumuptakeenablingmuscletorelax,

andthecalcium release channel, which mediates calcium re-lease to triggermuscle contraction, are the two key proteins

involved(24). Thus, a numberofstudieshavefocusedonthe invitro effects of anthracyclinesonfunction ofthe pumps and channels ofthis subcellular membrane system. Doxorubicin induces calciumrelease fromisolated

sarcoplasmic

reticulum vesicles and in skinned cardiac fibers (25-30). [

"C]-Doxorubicin bindstothe calcium release channel in fractions

enriched

interminalcistemae(31, 32). Doxorubicinalso in-creasesopen

probability

of calciumreleasechannelsin

doxorubicin, is a potent inhibitor of multiple intracellular pumps,including the calcium pump protein ofthe cardiac sar-coplasmic reticulum in isolated sarsar-coplasmic reticulum vesi-cles (36). However, higher anthracycline concentrations ap-pear to be required to mediate effects on the calcium-depen-dent ATPase than on the calcium release channel (28, 36). Thus,thereiscompellingevidencethatanthracyclinesalter the

functionof thesarcoplasmic reticuluminvitro,suggestingthat the contractiledysfunctionofanthracycline-induced cardiomy-opathymightbemediated by similar effects in vivo.

Recentstudieshave alsosuggestedthat acute in vitro and prolonged in vitro exposure have different effects. Single chan-nelstudieshavedemonstrated that although short exposures to doxorubicin open the channel,prolonged exposures

irrevers-iblyclose thechannel (35).Inisolatedratventricle itwas also demonstrated that acute in vitro exposure enhances [

3H]-ryanodine bindingand calcium release in a reversible manner

whereasprolonged ( 5-24 h)invitroexposureincreases doxoru-bicinsensitivityof the calcium release channel inan

irrevers-ible manner (28). However, the time course of these

"pro-longed"exposuresis still

relatively

short incomparisontothe time course of exposure to thedrugduringchemotherapy.The

relevance oftheseinvitro observationstothe

etiology

ofthe

chronic doxorubicin

cardiomyopathy

seen clinically remains tobe established.

Thepurposeof this

study

wastoassesstheeffectsof chronic

doxorubicin administration on

sarcoplasmic

reticulum

func-tionand to correlatechangesin

sarcoplasmic

reticulum

func-tion with funcfunc-tional and

microscopic

evidence of

cardiotox-icity.

Methods

Rabbits developadose-dependentcardiomyopathywith morphologic changes similartothosedescribed in humans (21, 37-41 ). Adult New ZealandWhite rabbits ( 1.9-3.7 kg)wererandomlyassignedto a doxo-rubicin treatment group (n=20)or apair-fedcontrol group (n=20). Doxorubicin ( 1 mg/kg)(AdriaLaboratories, Columbus, OH)was in-fused intoanearveinover30min twiceweekly foratotalof12 to 18 doses. The dosewaschosenasthat producingclinical effect as evi-dencedprimarilybyasignificantdrop-offinfood intake. Pair-fed

con-trol rabbitswereinfused with normal salineonthesameschedule. Bodyweight and food intakeweremonitored throughout the exper-iment. Pairfeedingwascarefullymaintainedto ensurethat therewas nosignificant contribution of malnutritiontoobserveddifferences

be-tweendoxorubicin-treated andcontrol-fed rabbits. Theweightof food consumedby the treated rabbitwasmeasured every 1-2dand deter-mined the amountof food given the control-fed rabbit atthe next

feeding.The treatedrabbitswerelessactive than their paired controls

but wereotherwise healthy withoutsigns of infection. Blood cell counts wereobtained on halfthe pairs and no significant differences in hemato-crit (treated: 33±1% vs. control:36±1%),white cell counts (treated:

8.5±0.2/mm2vs.control:7.6±0.7/mm2),orplatelet counts (treated: 544±106vs.control: 369±49) were observed. This is consistent with a lack of severe bone marrow suppression with the doxorubicin

treat-mentprotocol used.

Attheconclusionofthe treatment protocol ( 1-34 d after last dose), therabbitswereanesthetized with pentobarbital (50 mg/kg) and anti-coagulated with heparin (100 U/kg). The hearts were removed throughasternotomy andimmediately placed in Krebs bicarbonate buffer with thefollowingcomposition (mM): 5 dextrose, 142Na', 3.6 K+, 0.6 Mg+, 25 HCO-, and 2.5Ca2` at30'C. A left atrial strip was suspended fromajewelers chainon aforce transducer (Gould, Inc., Glen

Burnie,

MD)with the lower end anchoredagainst bipolar

stimu-luselectrodes, placed inabath with the Krebs bicarbonate buffer (as above)at30'Cbubbled with 95%02/5% C02,and stimulatedat0.5 Hz for 30 min toallow stabilization. The atriawere stretchedto a

resting force of 0.38 g. After stabilization, tracingswererecordedon a

stripchart at 100 mm/s paperspeed and werelaterdigitalized and analyzedby computer. The heart was then placed in normal saline solutiononice. A left ventricular papillary musclewasremoved and placed inglutaraldehyde solution for light microscopy and electron micrograph (EM)' examination. The atriawereremoved, frozen, and used forquantitation of doxorubicin and doxorubicinol levels. The remaining left ventricle and right ventriclewerehomogenized anda

sarcoplasmic reticulum-enriched microsomal fraction was isolated by using the method of Chamberlain and Fleischer (42) modified in this laboratoryfor rabbit heart. Modifications included initial homogeniza-tion withaPolytron (30sx2)and the inclusion of protease inhibitors (0.2 mM PMSF and 0.5 gg/ml leupeptin) in the homogenization buffer. The microsomal fractionswerestoredat-70'C(early in the study) orat -120'C (later in thestudy)until used for the specific assays. Protein concentrationsweredetermined using the bicinchon-inic acid method (Pierce Chemical Co., Rockford, IL).

With the last several rabbit pairs, a sarcoplasmic reticulum-enriched microsomal fractionwasalso isolated from rabbit back

mus-clebythe methods of Saitoetal.asdescribed in Chuetal. (43). Light and EM scoring. Left ventricular papillary muscles were scored inablinded fashion byacardiovascular pathologist (J.B. Atkin-son). The severity of doxorubicin-induced myopathic changes was graded by light microscopyandconfirmed by EM,accordingtothe method of Billingham ( 15). Overallscoresfor each rabbitwerederived byexamination of 5to 10 blocks.

Ryanodine binding.Recentidentification ofryanodineas aspecific high-affinityligand for the calcium release channel (24) provideda

methodto screenforchangesinregulationof the open stateof the calcium release channelaswellasforchangesinamountof calcium release channel inrelativelysmall andimpure preparations. Ryano-dinebindingis calciumdependent,felttoreflectthat channel opening is calciumdependent and ryanodine binds onlytothe open channel. In vitro,doxorubicin increasesryanodinebinding(and, byinference, cal-cium release channelopening) byincreasingcalciumsensitivityof the channel in subcellular fractions enriched insarcoplasmic reticulum,as

previouslydescribed(28).Weusedradioligand bindingtoquantitate and evaluate several parameters of the calcium release channel of po-tential functionalsignificanceinchronicallytreatedrabbits.

Total ryanodine bindingto microsomal fractions was measured afterincubation with [3H] ryanodinefor 60 minat37°C (44). Nonspe-cificbindingwasmeasuredbyincubation with> 100-fold excess cold ryanodine. Specific high-affinity ryanodine binding was calculated as thedifference between totalandnonspecific binding. Vesicle-bound [3H]ryanodinewasseparatedfromfree[3H ]ryanodine by ultracentri-fugationinaTL-100 Tabletop System (Beckman Instruments, Palo Alto, CA)withaTLA-100rotor(44a) at 245,000 g for 10 min. The bindingbufferwas 1MKCI, 10mM Hepes, 25 ,uM CaCl (pCa 4.7) at pH 7.4,whichwasadjustedtovarious lower pCas by the addition of EGTA.Ryanodine bindingforeachheart (treated or control) was per-formedat 10 nM ryanodine concentration. The binding isotherms wereperformedonpooled treatedorpooledcontrol vesicle fractions.

Forthe invitro doxorubicin studies, doxorubicin was added to micro-somalfractions before the addition of ryanodine.

Westernblot with calcium releasechannel-specificantibody.A

se-quence-specific polyclonal antibodywasraised in rabbit that

cross-reactedwithboth theskeletal andcardiac calcium release channels. Cardiac(6 gg)and skeletal (4gg)microsomal fractionsorcardiac homogenate (24

MAg)

were electrophoresed underdenaturing condi-tionson gradient(5-15%) minigelsand weretransferredovernight

onto transfer membrane (Immobilon-P; Millipore Corp., Bedford,

MA).After preincubation in Blotto-based blocking solution for 60 min (45), the membranes wereincubated for 90 min with immune or preimmune serum in Blotto-based blocking solution followed by alka-line phosphatase-linked goat anti-rabbit antibody in Blotto-based blocking solution and were developed with p-nitro blue tetrazolium chloride and 5-bromo-4-chloro-3-indolyl phosphate-toluidine salt (Sigma Chemical Co., St. Louis, MO).

Quantitation ofthesarcoplasmic reticulum calcium pump and its activity.Microsomalfractions were run in duplicate or triplicate with treated and controlpairs on the same gradient polyacrylamide minigel (5-15%) and were stained with Coomassie blue. The relative amounts ofprotein in the 100kDband, corresponding to the calcium pump protein in treated versus control fractions, was estimated by densitome-tryusing the MicroScan Gel Analysis Program and Densitometer by Technology Resources Inc. (Nashville, TN).

Calcium-dependent ATPase and the ATP-dependent oxalate-stim-ulatedcalcium loading of the calcium pump protein were measured as previously described (36). ATP-stimulated calcium uptake was mea-sured by incubation inbuffer (20 mM Imidazole-HCl, 100 mMKCl,4 mMMgCl2,5 mMNaN3, 100 mMsucrose, 30MMEGTA,0.1 mM ouabain, 2.5 mM K2-oxalate, 66.2 gMCa2+or1.0mM EGTA, ±3.5 mM Na2ATP) for 2 min, followed by isolation of the vesicles by filtra-tion. The ATP-stimulated 45Ca loaded was then measured in a scintilla-tion counter.Simultaneously, calcium-stimulated ATPase activity was determined by thedifference in ATPase activity in presence ofCa2+ (66.2 MM)and in theabsenceofCa21(1.0 mM EGTA). Activity is expressed permilligram of total protein.

QuantitationofMg2+-dependentATPaseactivity and Na +K+ A TP-aseactivity. Mg2+-dependent ATPaseactivitywasmeasured inthe microsomal fraction by methods previously described (46). Magne-sium-dependent ATPaseactivitywasmeasuredby incubation in buffer (33 mMTris-acetate, pH 7.4, and 3.5mM Na2ATP) for4min at30°C in the presence and absenceofMg2"(1.7mM).

Ouabain-sensitive Na+K+ ATPase activitywasmeasured in the microsomalfraction by methods previously described (47). ATPase activitywasmeasuredby incubation in buffer with thefollowing

com-position (mM): 30imidazole-HCl, 120NaCl,20KC1, 3MgC12,0.5 EGTA, 5 NaN3, and 2 Na2ATPfor 10 minat37°C in the absence and presenceof ouabain(1 mM).

Measurement ofcardiac concentrationsofdoxorubicin and doxoru-bicinol. Tissue sampleswereblotted, weighed, and storedat-20°C untiltransportedondry ice foranalysisofdruglevels.Druglevelswere measuredinablinded fashion. The tissuesamples (50-100 mg)were

addedto 2g ammonium sulfate and 3ml0.9%NaCl,homogenized (1 min) and thenspiked with daunorubicin (5- 100

Mg

asrequired)as an

internal standard. 5 mlofisopropanol/chloroform (50:50, vol/vol)

wasadded tothehomogenization mixture,vortexedfor 3 min, and centrifugedat500 g for12min. Theorganiclayerwasextractedagain

usingchloroform/isopropanolcontaining3Msilvernitrate, vortexed,

andcentrifugedas above.Theorganiclayerwasseparated,dried under nitrogen,andresuspended in 500 Mi methanol before HPLC.

TheHPLC system (WatersAssociates,Milford,MA)uses a4,um phenyl Radial-Pak reversed-phasecolumn (WatersAssociates) with

gradient control of the mobile phases (initially 72:28, vol/vol 0.1% ammonium formate buffer (pH 4.0)/100% acetonitrile; and finally 100%acetonitrile) at a flow rate of 3 ml/min. Detection is by a fluores-cent detector(Kratos Analytical Instruments, Ramsey, NJ) with exci-tation at 470 nm and emission at 550 nm. Standard curves with varying doxorubicin or doxorubicinol concentrations were prepared using un-treatedrabbitventricular tissue.

Statisticalanalysis.Comparisons between treated and control data wereperformed by unpaired t test, with the exception that the pathol-ogy scores werecomparedusing rank-sumanalysis (NCSS Statistics Package, Kaysville, UT). Values are expressed as mean±standard error.Thedoxorubicin dose-response curves and the pCacurveswere generatedfrom the data using a computer program (Graphpad, San Diego, CA) to obtain the best-fit sigmoid curve.

Results

Characterization ofthe degree ofmyopathy. Food intake fell off significantly in the treated rabbits from the first week ( 155 g/d) tothe final week (70 g/d). By virtue of the design of the pair-feeding protocol, there was no significant difference in the weight of the control rabbits (3.31±0.08 kg) and the treated rabbits (3.06±0.10 kg) at the end of the treatment protocol. In contrast,the weight of the hearts (ventricles)wassignificantly higher in the treated rabbits (5.0±0.2 g) compared with the controlrabbits (4.4±0.1 g) (P<0.02), and the ratio of heart weight to body weight was also significantly higher for the treated rabbits (1.66±0.08) than the control rabbits (1.35±0.03) (P < 0.002). Heartweightanddegreeof myopa-thyarecorrelated (Table I).

The primary microscopic changes observed in ventricles from treated rabbits was sarcotubular swelling without mito-chondrial or myofibrillar loss (Fig. 1).Comparingthe pathol-ogy scores for doxorubicincardiomyopathy,the left ventricular papillary muscles from the treated rabbits ( 1.31±0.14) demon-stratedearlychanges consistent with doxorubicin cardiomyopa-thy that were not observedin thepair-fed controls (0.15±0.03) (highly significant by rank-sum test). To allow comparison of

variousparameters withdegreeof myopathy, the rabbits were arbitrarily subdivided into mild,moderate,and severe myopa-thyaccording to theirmicroscopicscore(Table I).A score of <1 was considered mild, 1-2 wasmoderate,and>2was se-vere.

Heart(ventricles)and atrialweightin themildlyaffected

rabbitswere notsignificantlydifferent from thecontrols, but the moderate and severe groups demonstrated a progressive rise in theseweights (Table I). Severelyaffected heartswere visibly enlarged at the time of removal, whereas themildand moderate groups were similar in appearancetothecontrols.

Table I.Comparison

of

Severity

ofCardiomyopathy

Developed tension Developedtension [3H]Ryanodine

EMscore Heartweight Atriaweight (0.5Hz) (max) binding

g g g/mg g/mg pmol/mg

Control 0.15±0.03(15) 4.50±0.10(15) 0.58±0.03(15) 0.097±0.010

(13)

0.144±0.011

(13)

0.631±0.027(11) Mild 0.69±0.15(4) 4.44±0.26(4) 0.62±0.05(4) 0.058±0.006t(4) 0.125±0.023(4) 0.415±0.093(3) Moderate 1.26±0.09(8) 5.07±0.22*(8) 0.64±0.03(7) 0.043±0.013*(7) 0.135±0.023(7) 0.313±0.063t(4) Severe 2.23±0.12(3) 5.99±0.18§(3) 1.32±0.17*

(3)

0.013±0.006§ (3)

0.031±0.014§(3) 0.204±0.024§ (3)

Figure

1.

(a)

EMofacontrolrabbit

demonstrating

normalmyocyte

structure.

Bar,

1 gm.

(b)

EMofa treatedrabbit

_{demonstrating}

severe dilatation of the sarcotubular struc-tures

_(pathology

score =

_1.8).

This wasthecommon

_finding

in doxoru-bicin-treated

rabbits,

with

myofi-brillarloss

_being

rare._Bar, 1 ,um.

Functional changes

in

intact isolated muscles.

Left atrial stripswereofequalsize in the treated(8.2±0.9mg) and control

(8.8±1.2mg) rabbits.While stimulating at0.5 Hz, the devel-opedtension (g/mgatria)wasdecreasedin the treated versus the control rabbits

(Table

I).

Increasing severity

of the

myopa-thy by pathology score wasassociated with decreasing devel-opedtensionatthis

frequency.

Thus,there was good

correla-tion between the extent ofinhibition of contractile function

andseverity of microscopic lesions.

Characterization

ofthe

subcellular fractions.

Althoughthe

protein

yield

of the microsomal fraction was higher for the

treated than the control

_rabbits,

therewas nosignificant

differ-encein

45Ca2"

loading,

ATPase

activity,

orthecouplingratio between

loading

and ATPase

activity

in treatedversuscontrol rabbits

(Table

II).

There was no

significant

changein

45Ca2+

loading

or ATPase

_activity

with

increasing

pathology score.

Similarly,

therewas nodifference between the treated and

con-trol groups in theamountof calcium pumpproteinin cardiac

microsomal fractions determined by densitometryof Coomas-sie-stained

gels (Table II).

In thesesamemicrosomal

fractions,

doxorubicintreatmentdidnotalter activities of either the

Table I. Characterization ofMicrosomal Fractions from Treated and Control Rabbits

Control Treated

Protein yield (%) 1.09±0.06 (19) 1.36±0.06 (20) P=0.01

Pathology score 0.15±0.03 (19) 1.36±0.13 (20)

CaMg ATPase (nmol-mg-'M min') 192±10 (19) 179±9 (20) NS

Ca loading (nmol *mg-'. min-') 155±12 (19) 145±12 (20) NS

Coupling ratio 0.87±0.10 (19) 0.83±0.06 (20) NS

NaK ATPase (nmol *mg-' min')(SL) 171±17 (9) 146±26 (9) NS

Mg ATPase (nmol-mg-' min-')(mito) 563±43 (7) 665±39 (7) NS

100kDband (densitometry) Ratio treated/control 0.88±0.05 (13) NS

Parenthesesindicate n.

(Table II). All activities are expressed relative to total protein. Thus, despite the small difference in protein yield in micro-somalfractions fromtreated versus control rabbits, these frac-tions appear very similar with respect to the above assays.

Ryanodine binding scatchardanalysis. To screen for differ-ences ineither

_B.

or Kd,ryanodinebinding was measured in cardiacmicrosomal fractionsat a subsaturating concentration ofryanodine(10nM). Specificbinding at this single

concen-tration was markedly lower in the treated versus the control rabbits (Fig. 2). Complete binding isotherms using pooled

sam-plesfrom differenttreated or control rabbit hearts revealed that thedifferencenoted above was due to a significantly lower

_B.

without a significant change in Kd (Fig. 3). No difference in ryanodine binding between treated and control was seen after

shortcourses(two to three doses)ofdoxorubicin in the rabbit model (not shown).

Doxorubicin

sensitivity

ofryanodine

binding.Todetermine whether chronic doxorubicin exposure shifted the sensitivity to

in vitro doxorubicin exposure, microsomal fractions from

2

2

E

-.

0

z 0

w

ELI

z

D

0

z

A:

0 0.80 0.70 0.60 0.50 0.40

0.30 0.20

0.10 0.00

CONTROL TREATED

N=15 N=16

Figure2.Theeffect of chronic doxorubicin treatment on

[3H]-ryanodine binding. Bindingwasdeterminedat asinglepointonthe bindingisotherm for individualrabbitstodetectpotentialdifferences ineither

_Bm..

or_Kd.Microsomal fractions from each rabbit(N)were assayed intriplicatewith 10nM

[3H]ryanodine

in the presence (nonspecific binding)orabsence(totalbinding)ofan 100-foldexcess of unlabeledryanodine.Vesicle-bound[3H ] ryanodinewasseparated byultracentrifugation. Values shownaremean±SEfor thespecific

binding.Themeanpathologicalscorefor the treatedrabbitswas1.37 (moderate) in thisfigure.

treated and control hearts were exposed toincreasing doxoru-bicinconcentrationsatpCa 6.8. At this suboptimal pCa, doxo-rubicin will enhance binding by shifting the calcium sensitivity (28). At suboptimal pCa there was no difference in the doxoru-bicin concentrationrequired to reach half maximum for ryan-odine binding between control and treated rabbits (Fig. 4).

Cakium sensitivity of ryanodine binding. Todetermineif there was ashift in the calcium sensitivity ofthe calcium release channel indoxorubicin cardiomyopathy analogous to the shift inthe calcium sensitivity for ryanodine binding seen with in vitro doxorubicin exposure, pCa curves were determined for treated and control cardiac microsomal fractions (Fig. 5). There was a small leftward shift in the

_pCa"

in the treated versus control rabbits that was statistically significant. The

ap-parentdifference at very low pCa (i.e., 8) in the treated versus the control rabbits was not statistically significant.

1.00

-BMAX

= _{.79 pmol/mg}

la ~~~~KD

= _{2.1 nM}

0 8MAX = .37 pmol/mg

w 0.40

0

0

z

a: #

*~~~~~~~~~

Control (n=41

t

~~~~~~~*

Treated (n=41

0.00

-RYANODINE UNBOUND (nM)

Figure 3. The effect of chronic doxorubicintreatmenton

[3H]-ryanodine binding. Microsomal fractionswerepooledtoobtain treatedsamples (n=4)and controlsamples (n=4)of sufficient volume to run abindingcurve oneachsample.Nonspecificand total ryanodinebindingateight differentryanodineconcentrations,in du-plicateortriplicate,weredetermined for each ofthepooledsamples.

Themean±SE of thespecificbindingforthetreated(n=₄₎and control (n=4) samplesis shown.Bindingisothermsweregenerated

usingthe computer programGraphpad.Thesecurveswerethen con-vertedtoScatchardplotstodetermineBm,, orKd.Nonspecific bind-ingwasidentical for the control and treatedrabbits. Themean

100

-75

-50

-25

0

* Control (N=2) * Dox (N=2)

EC1/2 = 18 _piM

* /

,

E E x E 0 (9 z 0 ;-l co :U pCa 6.8

10 nM 3H-Ryanodine

-8.0 -7.0 -6.0 -5.0 -4.0 -3.0

Log [Doxorubicin](M)

Figure4. The effect ofchronicdoxorubicintreatmentoninvitro doxorubicinsensitivityofryanodine binding.To determine if chronic

treatmentwithdoxorubicin alters thepropertiesof thecalcium release

channel, [3H] ryanodine bindingintwomicrosomalsamplesfrom

treated andcontrolheartswerecomparedatsuboptimalcalcium

concentration(pCa 6.8)in thepresenceof five differentdoxorubicin concentrations.Eachassaywasperformedintriplicate.Thecurves for controlandtreated rabbitsrepresentthemeanspecific binding (percentofmaximum)for thetwosamples, plottedas afunctionof thelogarithmofthe doxorubicin concentration. Maximumspecific bindingwas0.75pmol/mgforcontrol and 0.30pmol/mgfor treated

hearts.Doxorubicin increased specific bindingin thecontrol hearts

asexpectedfrompriorinvitrostudies(not shown).Doxorubicin

in-creasedspecific bindingin thedoxorubicin-treated hearts,withan

EC,/2

of18jM,whichwassimilartotheresponseinthecontrol

hearts.Themeanpathologicalscorefor treated rabbitswas 1.6

(moderate)forthisfigure.

75

-50

-25

-0

pCa5o 6.42

* Treated (N=3)

* Control (N=5)

10 nM 3H-Ryanodine

8 7 6

pCa

5 4

Figure5.Theeffect of chronic doxorubicintreatmentonthe

Ca2"

sensitivityofryanodine binding.Todetermine if chronictreatment with doxorubicin alters the propertiesof the calcium releasechannel, pCacurves werecomparedfor treated and control hearts. Cardiac

microsomal fractions from control (C) andtreated(n)rabbitswere

incubated with[3H]ryanodineinthepresenceofvariableEGTA

concentrations.Specific [3H]ryanodinebound(percentofmaximum

binding)is shownas afunction ofpCa.Maximumspecific binding

was0.71±0.01 pmol/mg for these control rabbits and 0.28±0.05

pmol/mg for these treated rabbits (P<0.01). The controlcurveisa

compositeofassaysperformedonfive differentmicrosomalsamples. Thetreatedcurveisacomposite ofassaysperformedonthree

differ-entmicrosomalsamples.SE barsareshownforpointswheren >2.

ThepCa50for the control rabbitswas6.29±0.04 andthat forthe

treated rabbitswas6.42±0.05(P=0.05, Mann-Whitney).Themean morphologicalscorefortreated rabbitswas1.6 (moderate) for this

figure.

Correlation of pathology score withfunctional data and

bindingdata. Tocorrelate theseverityof thecardiomyopathy

as determined by the pathology score with other

measure-ments, the treated rabbits were divided into mildly affected

(pathology score < 1), moderately affected (pathologyscore

1-2), and severely affected (pathologyscore >2). As demon-stratedinTableIandFig. 6, developedtensionat0.5Hz de-creasedasthepathologyscore increased. Ryanodine binding also decreasedasthepathologyscore rose.The linesfitting the

data for developed tension and binding are nearly identical (Fig. 6).

Estimateofcalcium release channel by anti-calcium release

channel antibody. Western blot analysis ofcardiac

homoge-natesenabledustodemonstratethat thedecrease in calcium release channel in treated versus control rabbit microsomes wasnotanaberration referabletosubcellularfractionation of normaland treated rabbits (Fig.

7)..

Westernblot analysis ofcardiac microsomal fractions dem-onstratedqualitatively loweramountsof immunoreactive cal-ciumreleasechannel in the treated rabbitscompared with the controlsjudged by intensity ofcolorindicator(not shown).

Comparison of effects on skeletal muscle. To determine

whether the effect of doxorubicintreatment onsarcoplasmic

reticulumwasspecifictocardiac muscle, skeletalmuscle micro-somalfractionswereisolated from four pairs of doxorubicin-treated andcontrolrabbits. Despiteamarked decrease in the ryanodine binding in these cardiac microsomal fractions, there

was nodifferenceobserved in ryanodine binding in the skeletal microsomal fractionscompared with their controls (Fig. 8).

Using the anti-calcium release channel antibody that

cross-reactswith bothcardiac and skeletal calcium release channel, theamount ofcalcium releasechannel in the skeletal micro-somal fractions fromtreated andcontrolrabbitswas

qualita-tively similar, despiteasignificant decrease in calcium release

0.75 -0a E 0.60 E -a Z 0.45 -z m w z 0.30

-a 0 z > _0.15 0 0.00 CONTROL MODERATE SEVERE

0.0 0.5 1.0 1.5 2.0 2.5

- 0.10

0 0

- 0.08 c - 0.06 o

en

N

- 0.04

-(a 3

3 - 0.02 S

- 0.00

PATHOLOGY SCORE

Figure6.Comparison of ryanodine binding (solid symbols)(pmol/

mg) and developed tension (DT)(opensymbols)(g/mgat0.5 Hz)

as afunction ofcardiomyopathyseverity.Valuesrepresentmean±SE.

Figure7.Representative Western blot of ho-+ C R C mogenates

using

calcium release

channel-10 11 12

~~~~~~~~~~specific

antibody. Each sample is

run

in

du-1

2

3 4

5

8

7

8

9

10

11

12

plicate.Samplesfrom rabbits with different degrees ofcardiomyopathyby EM score were comparedwiththeir paired controls. LanesIand 2=severe (EM score=2.4;ryanodinebinding=0.19pmol/mg); lanes 5 and 6=moderate (EM score= 1.3;ryanodine binding=0.28 pmol/mg); lanes 9, 10=mild (EM score=0.9;ryanodine binding=0.59 pmol/mg). Lanes 3, 4, 7,8, 11, and12represent the controls.

channel in thecardiac preparations fromthosetreated rabbits

(notshown).

Doxorubicin anddoxorubicinol concentrations. The atrial

tissueconcentration of doxorubicinwas 0.42±0.1 0 ng/mg wet

tissue weightandof doxorubicinol was 0.07±0.01 ng/mg wet tissueweightin treatedrabbits.Thiscorresponds to a doxoru-bicin concentration of 1 uM

assuming

that 1

kg

of tissue

contains - _{1literoffluid.}

Discussion

Our studies demonstrate that thecardiomyopathy associated with chronic doxorubicin exposure is accompanied by a de-crease in the amountofcalcium releasechannelofthe

sarco-plasmic reticulumdetectedby a decreasein radioligand

bind-ing usbind-ing thespecific high affinity ligand ryanodine (Fig. 2)

and byareduction inimmunoreactive calcium release channel (Fig. 7). Further, thedecrease in amountof calciumrelease

channel correlates with the severity of the

cardiomyopathy

scoredbythemicroscopic criteria of Billingham ( 15)andwith

the decrease indeveloped tensioninatrialstrips fromthe

doxo-rubicin-treatedrabbits (Fig. 6). Thisdecrease in thecalcium release channelwasinitiallydetected and has been best

charac-terized in the

sarcoplasmic

reticulum-enriched microsomal

fractions. Thesefractions fromtreatedand controlrabbitsare

otherwiseequivalent withrespect to amountof calciumpump

protein ofthesarcoplasmic reticulum, including calcium-de-pendent ATPase

activity

and

ATP-dependent

oxalate-stimu-latedcalciumloading

activity

(TableII).Thedecrease in

cal-0.80

-E

-

0.60-0

E

0. 0m

C ~5 0.40 -0

._ 0

c 0.20

-0.00

-r 2.40

-2.00 =

0

- _{1.60 ,}

w

- 1.20 a :i

-0.80 3

0

3 - 0.40 S

-0.00 CARDIAC SKELETAL

Figure 8.Ryanodinebinding in cardiacvs.skeletalmicrosomal frac-tions from control and treated rabbits.[3H ] ryanodine bindingin

mi-crosomalpreparations fromtreated(n=_3)andcontrolrabbits(n = _{3)wasdeterminedin 10 nMryanodine forbothcardiacand} skele-tal muscle. Values represent the mean±SE of threesamplesin

tripli-cate.The meanpathologicscorefor treated rabbitswas1.83 (moder-ate) in thisfigure.

cium release channel can be detected in both the microsomal fraction and the homogenate oftreated rabbits by Western blot analysis. Therefore, the decrease in calcium release channel detected in the microsomal fractions of treated rabbits is not referable to selective loss of the terminal cisternae (the calcium releasechannel-rich region ofthe sarcoplasmic reticulum) dur-ing subcellularfractionation. Inaddition, the calcium release channels present in the treated rabbits do not appear

signifi-cantlyaltered comparedwiththe channelsof the control rab-bits with regard to: Kd for ryanodine binding (Fig. 3) or sensitiv-ity of ryanodine binding to doxorubicin added in vitro (Fig. 4). There was a small difference in

_pCam

for ryanodine binding, withthe calcium release channel from treated rabbits requiring less calcium for binding (Fig. 5). Finally, the decrease in cal-cium release channel was specific to cardiac muscle, leaving skeletal muscleessentially unaffected (Fig. 8).

Thisstudyis unique in several aspects. The first is the use of pair-fed controls. Since variousnutritionaldeficiencies,such as

selenium deficiency,are known to causetheirownmyopathic changes(48), we felt it wascrucialtohave controls that were better matched in nutritional status. Second, this study is a

multilevelstudycorrelating changes inmorphologyscored by

lightand electronmicroscopy,changesin intact muscle func-tion, and changes in subcellular functions resulting from chronic doxorubicin exposure.Finally, it isoneofthefew

stud-iestoevaluatefunction of variouspumps and channels in the chronic setting and the only one to evaluate the calcium release channel.

The results support the hypothesis that the sarcoplasmic reticulum is effectedearly inthe doxorubicin cardiomyopathy

assuggested by the sarcotubulardilationnoted on histopathol-ogy(Fig. 1) ( 15, 16). It appears to be a specific effect on the calcium release channel of the sarcoplasmic reticulum and does not appear toaffect the calciumpumpprotein ofthe

sar-coplasmicreticulum. The decrease inryanodine bindingwould beconsistent with severalexplanations:adecreasedamountof

calcium release channel; a decrease in high

affinity

binding

secondary to an alteredbindingsite or due to modification of thechannel at a regulatory site distant from the ryanodine-binding site. The decrease in receptor was also detected by

Westernblotanalysis usingaspecific

antibody

to the calcium

release channel. Togetherthese two

findings,

decreased ryano-dinebindinganddecreasedimmunoreactive channel,suggest an actual decrease in theamountof channel rather than the presenceofanaltered channelnotdetectedby eithermethod.

Thisdecrease in amountofcalcium release channel

result-ing from chronic doxorubicintreatmentseems,atfirst

glance,

to be very different from the increased calcium

sensitivity

of the channelresulting frominvitrodoxorubicinexposure

(28).

chronic stimulation due to increased calcium sensitivity. We

did detect a small shift in calcium sensitivityin the chronic model. Theexpectedshift in

pCa50

would be small at the mea-sured doxorubicin tissue levels (- 1 1M). The shifts in

cal-cium sensitivitymay be more marked at peak drug levels after infusion.

Although thechangesseen inthe calcium release channel of

thesarcoplasmicreticulum are not proven to be causal for the doxorubicin cardiomyopathy, the correlation with degree of myopathy andtheagreementwithprior morphologicaland in vitrostudiessuggesting the calciumrelease channelmightbea target make thisplausible.Mechanisms for doxorubicin

toxic-itysuggestedby earlierstudies include: free radicalgeneration

and lipid peroxidation (49-52), reactive sulfhydryl groups (53, 54), binding to channel regulatory sites (25, 28, 31,32), orinhibitedmRNA/protein synthesis (55-58).The results to date have beenconflicting, possibly complicated bythe

poten-tial oftwo differentmechanisms,one foracutetoxicityand one for the late(chronic) toxicity; orbythepotentialof different

mechanisms during therapeutic dosing versus much higher dosing;orby differentmechanisms for doxorubicintoxicityin

variousinvitro conditions. Furthermore, somemechanisms,

such asinhibition ofmRNAsynthesis, mightbeexpectedtobe lessspecific forthecalciumrelease channelthandirect interac-tionswith the calciumrelease channel.Therefore,

understand-ingthemechanismbehind the apparent decrease incalcium

release channel may alsohelp elucidatethedegree ofspecificity ofthesarcoplasmic reticulumasanearly site ofinjury.

Acknowledgments

Wethank Dr. Andrew Marks andDr.PaulTempstforprovidingthe peptide used in the making of the polyclonal antibodytothecalcium release channel.

Supported by NIH F32 HL08006 (D. Dodd), CA 48930(R.J. Boucek, Jr.), HL32711 (S.Fleischer), and Department of Veteran's Affairs (R. D. Olson).

References

1.Bristow,M.R.,M. E.Billingham, J.W.Mason, and J.R.Daniels. 1978. Clinicalspectrumofanthracycline antibiotic cardiotoxicity. Cancer Treat. Rep. 62:873-879.

2.Lefrak,E. A., J.Pitha, S. Rosenheim, andJ. A.Gottlief. 1973.A

clinico-pathologic analysisof adriamycincardiotoxicity.Cancer(Phila).32:302. 3.Lipshultz,S. E., S. D. Colan, R. D.Gelber,A.R.Perez-Atayde, S. E. Sallan, and S.P.Sanders. 1991. Latecardiac effects of doxorubicin therapy foracute

lymphoblasticleukemia in childhood.NewEngl.J.Med.324:808-845. 4.Schwartz,R.G.,W. B.McKenzie,J. Alexander, P.Sager,A. D'Souza, A.

Manatunga,P. E.Schwartz, H.J. Berger, J.Setaro,L.Surkin,et al. 1987.

Conges-tive heart failure and left ventriculardysfunctioncomplicating doxorubicin

ther-apy. Am. J.Med. 82:1109-1118.

5.Singal,P.K., C.W.Deally,and L. E.Weinberg.1987. Subcellulareffectsof

adriamycin intheheart:aconcise review.J.Mol. Cell. Cardiol. 19:817-828. 6. VonHoff,D.D., M. W. Layard, P. Basa, H. L. Davis, A. L. VonHoff, M.

Rozencweig,F.M.Muggia.1979. Risk factors for doxorubicin-induced

conges-tive heartfailure.Ann. Intern.Med.91:710-717.

7.Young, R. C., R. F. Ozols, and C. E. Myers. 1981. The anthracycline

antineoplasticdrugs. N.Engl.J. Med. 305:139-153.

8.Jakacki, R.,J.Silber,R.Larsen, G. Barber, J. Goldwein, and A. Meadows.

1991. Cardiac dysfunction following low-risk cardiotoxic treatment for

child-hood malignancy.Pediatr.Res. 29:143A.

9. Watts, R. G. 1991. Severeandfatalanthracycline cardiotoxicity at

cumula-tive doses below 400mg/m2:evidence forenhanced toxicity with multiagent

chemotherapy.Am. J.Hematol.36:217-218.

10.Goorin,A.M., K. M. Borow, A. Godman, R. G. Williams,I.C. Hender-son,S.E.Sallan,H.Cohen, andN.Jaffe. 1981. Congestiveheartfailuredue to

adriamycin cardiotoxicity: its natural history in children.Cancer. 47:2810-2816.

11.Lewis,A.B.,V. L.Crouse,W.Evans,M.Takahashi,and S. E.Siegel.

1981.Recoveryof left ventricularfunctionfollowing discontinuation of anthra-cycline chemotherapy in children. Pediatrics. 68:67-72.

12. Saini,J., M. W. Rich, and A. P. Lyss. 1987. Reversibility ofsevereleft ventriculardysfunctiondue to doxorubicin cardiotoxicity.Ann.Intern.Med.

106:8 14-8 16.

13. Freter, C.E.,T. C.Lee,M. E.Billingham,L.Chak,and M. R.Bristow. 1986. Doxorubicincardiactoxicity manifestingsevenyearsafter treatment. Am.

J. Med. 80:483-485.

14. Goorin, A. M., A. R. Chauvenet, A. R. Perez-Atayde, J. Cruz, R. McKone, and S.E.Lipshultz. 1990. Initialcongestiveheartfailure,six to ten

yearsafter doxorubicinchemotherapy for childhoodcancer.J. Pediatr. 1

6:144-147.

15.Billingham,M. E. 1991. Role ofendomyocardial biopsyindiagnosis and

treatmentof heart disease.InCardiovascularPathology (M.D.Silver, editor). 2nd ed. ChurchillLivingstone,NewYork,1465-1486.

16.Unverferth, D. V.,R. D.Magorien,B. P.Unverferth,R. L.Talley,S. P.

Balcerzak,and N.Baba. 1981. Humanmyocardial morphologicandfunctional

changesin the first 24 hours after doxorubicin administration.Cancer Treat.Rep.

65:1093-1097.

17.Mortensen, S. A.,H.S.Olsen,andU.Baandrup.1986. Chronic anthracy-cline cardiotoxicity: haemodynamicandhistopathologicmanifestations

suggest-ingarestrictiveendomyocardialdisease. Br. Heart. J. 55:274-282.

18.Bellini, O., andE.Solcia. 1985. Early and latesarcoplasmicreticulum

changes in doxorubicin cardiomyopathy. Virchows Arch. B (Cell Pathol.). 49:137-152.

19.Fischer,V.W.,G. M.Wang,and N. H. Hobart. 1986.Mitigationof an

anthracycline-inducedcardiomyopathy bypretreatment withrazoxane: a quanti-tativemorphologicalassessment. Virchows Arch. B(CellPathol.). 51:353-361.

20.Jaenke,R. S. 1976.Delayed andprogressive myocardiallesions after

adriamycin administrationin the rabbit. Cancer Res. 36:2958-2966.

21.Olson,H.M.,D. M.Young,D.J.Prieur,A.F.LeRoy, andR.L. Reagan. 1974.Electrolyte andmorphologicalterations ofmyocardiumin adriamycin-treatedrabbits.Am.J. Pathol. 77:439-454.

22.Solcia, E., L. Ballerini, 0. Bellini,U.Magrini, C. Bertazzoli, G. Tosana, L. Sala,F.Balconi,and F.Rallo. 1981.Cardiomyopathyof doxorubicin in experi-mentalanimals. Factorsaffectingtheseverity, distribution and evolution of myo-cardial lesions. Tumori. 67:461-472.

23. Van Vleet, J.F.,V.J.Ferrans, andW. E.Weirich. 1980. Cardiac disease induced by chronicadriamycin administration in dogs and an evaluation of vitaminEandselenium as cardioprotectants. Am. J. Pathol. 99:13-42.

24.Fleischer, S., and M. Inui. 1989. Biochemistry and biophysics of excita-tion-contractioncoupling.Annu.Rev.Biophys. Biophys.Chem. 18:333-364.

25. Abramson,J.J.,E.Buck, G. Salama, J. E. Casida, and I. N. Pessah. 1988. Mechanism ofanthraquinone-induced calcium release from skeletal muscle

sar-coplasmicreticulum. J.Biol. Chem.263:18750-18758.

26.Kim,D.H.,A. B.Landry, Y. S. Lee, and A. M. Katz. 1989. Doxorubicin-induced calcium release from cardiac sarcoplasmic reticulum vesicles. J. Mol. Cell. Cardiol. 21:433-436.

27. Palade, P. 1987. Drug-inducedCa2" release from isolated sarcoplasmic reticulum. J.Biol. Chem.262:6142-6148.

28. Pessah, I. N.,E. L.Durie, M. J. Schiedt, andI. Zimanyi. 1990.

Anthra-quinone-sensitizedCa2"release channelfrom rat cardiac sarcoplasmic reticulum:

possible receptor-mediatedmechanism of doxorubicin cardiomyopathy. Mol. Pharmacol. 37:503-514.

29.Salviati, G., andP.Volpe. 1988.Ca2"release from sarcoplasmic reticulum ofskinned fast-andslow-twitch musclefibers. Am. J. Physiol. 254:C459-C465. 30. Zorzato, F., G. Salviati,T.Facchinetti, and P. Volpe. 1985. Doxorubicin induces calcium release from terminal cisternae ofskeletal muscle. J. Biol. Chem. 260:7349-7355.

31. Zorzato, F., A. Chu, and P. Volpe. 1989. Antibodies to junctional

sarco-plasmic reticulum proteins: probes for theCa2"-releasechannel. Biochem. J.

261:863-870.

32. Zorzato, F., A. Margreth, and P. Volpe. 1985. Direct photoaffinity label-ing ofjunctional sarcoplasmic reticulum with [4C]Doxorubicin. J. Biol. Chem. 261:13252-13257.

33.Holmberg, S. R. M., P. A. Poole-Wilson, and A. J. Williams. 1989. Inter-actions of doxorubicin and mitoxantrone with the calcium release channel from cardiac sarcoplasmic reticulum-a possible basis for cardiotoxicity. Circulation. 80:II-141.

34. Nagasaki, K., and S. Fleischer. 1989. Modulation of the calcium release channel of sarcoplasmic reticulum by adriamycin and other drugs. Cell Calcium. 10:63-70.

35.Ondrias,K., L. Borgatta,D. H.Kim, and B. E. Ehrlich. 1990. Biphasic

effectsof doxorubicin on the calcium release channel from sarcoplasmic reticu-lum ofcardiac muscle. Circ. Res. 67:1167-1174.

36.Boucek,R.J.,R.D. Olson, D. E. Brenner, E. M. Ogunbunmi, M. Inui,

andS.Fleischer. 1987. The major metabolite ofdoxorubicin is a potent inhibitor of membrane-associated ion pumps. J. Biol. Chem. 262:15851-15856.

cardiac failurewithactivation of vasoconstrictor mechanisms. Cardiovasc.Res. 19:378-382.

38.Bachur, N. R., R. C.Hildebrand,and R.S. Jaenke. 1974.Adriamycinand daunorubicindispositionintherabbit.J.Pharmacol.Exp.Ther. 191:331-339.

39. McGrath, B. P. B. F. Jover, L. Trigg, and L. F. Arnolda. 1987. Adriamy-cin-induced cardiomyopathic heart failure in the rabbit. In Cardiomyopathy Up-date 1:Pathogenesis ofMyocarditisandCardiomyopathy. C. Kawai and W. Abelmann,editors.University Tokyo Press, Tokyo.

40. Shenasa, H., J.-L.Rouleau, M. Vermeulen, C. Bier, H. Eid, P. Paradis, C. Juneau, J. deChamplain. 1987. Chronic doxorubicin induced cardiotoxicity in rabbits: increased catecholamine responsiveness, shorter contraction and action

potentialduration. J.Am.Coll. Cardiol.9:147A.

41.Vescovo,G., S. E.Harding,R.Wanless, L. DallaLibera,and P. A. Poole-Wilson. 1989. Does a reduction in myosin content account for the depressed

contractility inadriamycin-induced cardiomyopathy? J. Mol. Cell. Cardiol.

21(Suppl. III):S.74.

42.Chamberlain,B., and S.Fleischer.1988.Isolationofcanine cardiac sarco-plasmic reticulum.Methods Enzymol. 157:91-99.

43.Chu, A., M. C.Dixon,A.Saito,S.Seiler,andS.Fleischer.1988.Isolation of sarcoplasmicreticulumfractions referabletolongitudinaltubules and junc-tionalterminal cisternae from rabbit skeletal muscle.Methods Enzymol. 157:36-46.

44.McGrew,S. G., C. Wolleben, K. P. Siegl, M. Inui, and S. Fleischer. 1989. Positivecooperativity of ryanodine bindingtothecalciumreleasechannelof sarcoplasmicreticulum from heart and skeletal muscle.Biochemistry. 28:1686-1691.

44a.Chadwick, C. C.,A.Saito,andS.Fleischer.1990.Isolationand character-ization of the inositoltriphosphatereceptorfromsmooth muscle. Proc.Nail.

Acad.Sci. USA. 87:2132-2136.

45.Johnson, D. A., J. W.Gautsch,J. R.Sportsman,and J. H. Elder. 1984. Improvedtechniqueutilizing nonfatdry milk foranalysis ofproteinsandnucleic

acidstransferredtonitrocellulose. Genet.Anal. Tech.Appl.1:3-8.

46.Fleischer, S.,and B. Fleischer. 1967. Removal andbindingofpolarlipids

inmitochondria and other membrane systems.MethodsEnzymol. 10:406-433.

47.Seiler, S.,and S. Fleischer. 1982. Isolation ofplasmamembrane vesicles

from rabbit skeletal muscle and their use in ion transport studies. J. Biol. Chem. 257:13862-13871.

48.Xia, Y., K. E. Hill, and R. F. Burk. 1989. Biochemical studies of a sele-nium-deficient population in China: measurement of selenium, glutathione per-oxidase and other oxidantdefense indices in blood. J. Nutr. 119:1318-1326.

49.Harris, R. N., and J. H. Doroshow. 1985. Effect ofdoxorubicin-enhanced hydrogenperoxide and hydroxyl radical formation on calcium sequestration by

cardiacsarcoplasmicreticulum.Biochem.Biophys.Res.Commun. 130:739-745. 50.JulicherR.H. M., L. Sterrenberg, A. Bast, R.0.W. M. Riksen, J. M. Koomen, and J. Noordhoek.1985. The roleoflipid peroxidation in acute doxoru-bicin-inducedcardiotoxicityasstudied in ratisolateheart.J.Pharm. Pharmacol.

38:277-282.

51. MyersC. E., W. P. McGuire, R. H. Liss, I. Ifrim, K. Grotzinger, and R. C. Young. 1977.Adriamycin: the role of lipid peroxidation in cardiac toxicity and tumorresponse.Science(Wash.DC). 197:165-167.

52. Thornalley, P. J., and N. J. F. Dodd. 1985. Free radical production from normal and adriamycin-treated rat cardiacsarcosomes. Biochem. Pharmacol. 34:669-74.

53. Abramson, J. J., and G.Salama. 1989. Critical Sulhydryls Regulate Cal-cium ReleasefromSarcoplasmic Reticulum. J. Bioenerg.Biomembr. 21:283-294.

54.Trimm J. L., G. Salama, and J. J. Abramson. 1986. Sulfhydryl oxidation inducesrapid calcium release from sarcoplasmic reticulum vesicles. J. Biol.

Chem. 261:16092-16098.

55. Ito H., S. C.Miller, M. E. Billingham, H. Akimoto, S. V. Torti, R. Wade,

R.Gahlmann, G. Lyons,L.Kedes, andF. M.Torti. 1990. Doxorubicin

selec-tively inhibitsmuscle geneexpressionincardiacmuscle cells invivo and invitro.

Proc.Natl. Acad.Sci. USA. 87:4275-4279.

56.Lewis,W., and B. Gonzalez. 1990. ActinisoformmRNAalterations in-ducedby doxorubicin in cultured heart cells.Lab.Invest.62:69-76.

57.Papoian, T., and W. Lewis. 1990.Adriamycincardiotoxicity in vivo-se-lective alterations inratcardiac mRNAs.Am.JPathol. 136:1201-1207.

58.Zahringer,J.,R.Kandolf,and W.Raum. 1981.Decreaseofmyocardial