Sensitivity of K562 human chronic myelogenous leukemia blast cells transfected with a human multidrug resistance cDNA to cytotoxic drugs and differentiating agents

Sensitivity of K562 human chronic

myelogenous leukemia blast cells transfected

with a human multidrug resistance cDNA to

cytotoxic drugs and differentiating agents.

W N Hait, … , S Srimatkandada, J R Murren

J Clin Invest.

1993;

91(5)

:2207-2215.

https://doi.org/10.1172/JCI116447

.

The blast crisis of chronic myelogenous leukemia (CML) is refractory to most forms of

cancer chemotherapy, but may be amenable to drugs that differentiate rather than kill

leukemic cells. One mechanism implicated in resistance to cytodestructive drugs is

overexpression of P-glycoprotein, the MDR1 gene product. While several classes of drugs

sensitize multidrug-resistant (MDR) cells by interfering with the function of P-glycoprotein in

vitro, few sensitizers have been effective in vivo. We have developed a preclinical model of

MDR/CML uncomplicated by other mechanisms of drug resistance to evaluate the effects of

MDR1 overexpression on cytodestructive and differentiation therapy and the ability of

sensitizers to restore chemosensitivity in this disease. The CML-derived cell line K562 was

transfected with a human MDR1 cDNA from the pHaMDR1/A expression vector and

selected with vinblastine. Resistant K562 clones were 20-30-fold resistant to vinblastine,

were cross-resistant to doxorubicin and etoposide, and remained sensitive to cytosine

arabinoside, 6-thioguanine, hydroxyurea, and mechlorethamine. Resistance was

associated with decreased cellular accumulation of cytotoxic drug and was reversed by

cyclosporin A and trans-flupenthixol. The MDR phenotype did not adversely affect the ability

of K562 cells to produce fetal hemoglobin in response to hemin, and was associated with

increased responsiveness of cells to differentiate with cytosine arabinoside. Upon

differentiation, the resistant clones increased MDR1 mRNA and P-glycoprotein. These

studies suggest that the overexpression […]

Research Article

Find the latest version:

Sensitivity

of K562 Human Chronic

Myelogenous

Leukemia

Blast

Cells Transfected

with

a

Human

Multidrug

Resistance

cDNA to

Cytotoxic

Drugs

and

Differentiating Agents

William N. Hait, SipraChoudhury, SrinivasanSrimatkandada,and John R. Murren

Yale University School ofMedicineandComprehensiveCancerCenter, Departments of Medicine andPharmacology, NewHaven, Connecticut 06510

Abstract

Introduction

The blast

crisis

of chronic myelogenous leukemia (CML) is

refractory

to most

forms of

cancer

chemotherapy,

but may

be

amenable

to

drugs that differentiate

rather than kill leukemic cells.

One

mechanism

implicated

in

resistance

to

cytodestruc-tive

drugs

is

overexpression of P-glycoprotein, the

MDR]

gene

product.

While

several classes

of

drugs

sensitize

multidrug-re-sistant (MDR)

cells

by

interfering with

the

function of

P-glyco-protein in vitro,

few sensitizers

have been

effective

invivo. We have developed a

preclinical

model

of

MDR/ CML

uncompli-cated

by other mechanisms of

drug resistance

toevaluate the effects of MDR]

overexpression

on

cytodestructive

and

differ-entiation therapy

and the

ability of sensitizers

to restore

che-mosensitivity in

this disease.

The

CML-derived

cell lineK562

was

transfected with

a human MDR] cDNA from the

pHaMDR1

/A expression

vector and selected

with

vinblastine.

Resistant K562 clones

were

20-30-fold resistant

to

vinblastine,

were

cross-resistant

to

doxorubicin

and etoposide, and

re-mained sensitive

to

cytosine arabinoside, 6-thioguanine,

hy-droxyurea,

and

mechlorethamine.

Resistance

was

associated

with

decreased

cellular

accumulation

of cytotoxic

drug and was reversed

by

cyclosporin

A

and

trans-flupenthixol. The

MDR

phenotype

did

not adversely

affect the ability of

K562

cells to produce

fetal

hemoglobin

in response to

hemin,

and was

asso-ciated with increased

responsiveness

of cells

to

differentiate

with cytosine arabinoside. Upon differentiation,

the

resistant

clones

increased

MDR] mRNA

and

P-glycoprotein. These

studies

suggest that the

overexpression of

the MDR] gene in

CML

may not

adversely affect

the

ability

to

undergo

erythroid

differentiation and

that these

resistant

K562

cell

lines

are

good

models for

studying

drug

resistance

mediated

by

P-glycopro-tein

in

CML. (J. Clin.

Invest.

1993. 91:2207-2215.) Key

words:

multidrug

resistance *

differentiation

*cyclosporin

A.

trans-flupenthixol

This workwaspresentedin abstract format the 1991 meeting of the

American Association of CancerResearch, 21May, Houston, TX.

Address correspondence and reprint requests toDr. William N. Hait, M.D., Ph.D.,athisnewaddress: CenterforAdvanced

Biotechnol-ogy and Medicine, Cancer Institute ofNewJersey, 679 Hoes Lane, Piscataway, NJ 08854.

Receivedforpublication 27July1992and in revisedform 4

De-cember 1992.

The blast crisis of chronic myelogenous leukemia is resistant to treatmentwithstandard cancer chemotherapy ( 1 ). Cross-resis-tance

is

seen to

drugs from

a

variety of

pharmacological classes,

including

natural products such as

anthracyclines and vinca

alkaloids,

a

situation

reminiscent of the laboratory

phenome-non

of multidrug resistance.

Multidrug resistance

(MDR)'

is characterized by

cellular

resistance

toseveral classes

of chemotherapeutic

drugs and

is

attributable,

at

least in

part, to

the

overexpression of

P-glyco-protein (reviewed in reference 2). MDRJ is

a member

of a

family

ofgenesthat include MDRJ and MDR2 (also referred

to as MDR3) in humans, andappears tobe the only human gene

expressing

a

functional protein (3, 4). P-glycoprotein

has

structural and functional characteristics of

an

energy-depen-dent membrane

transporter,

and is believed

to

produce

drug

resis-tance

through

the

active

efflux of drugs from

tumor

cells (5).

Considerable attention has been given

to

identifying

drugs

that block

the

function of

P-glycoprotein (reviewed

in

refer-ence

6). Verapamil (7), trifluoperazine (7, 8),

trans-flupen-thixol

(9, 10), tamoxifen

(

l I

), and

cyclosporin

A

(

12,

13)

are

examples of compounds that sensitize multidrug-resistant cells

to

chemotherapeutic

drugs in vitro. However,

few drugs

have

activity in vivo (6). There

are

several

possible

explanations

for

this lack of in vivo

effectiveness, including

the

pharmacokinet-ics and

pharmacodynamics

ofthe

sensitizers,

or

the

presence

of

alternative mechanisms of drug resistance ( 14).

In

fact,

the

latter

has

been demonstrated

in several cell

lines selected for

resistance by continuous

exposure to

drugs

and may account

for

the

inability

of

most

modifiers

to

completely

restore

sen-sitivity

of the resistant cells

to

that

of

the

sensitive parental

lines (6).

Recently,

P-glycoprotein

has

been detected in

a

high

per-centage

of

patients

in

chronic myelogenous leukemia (CML)

blast

crisis ( 15, 16) and has been suggested

asa

possible

cause

of the

refractoriness

of this disease

to treatment.To

develop

a

model

for

identifying

effective drugs

against

CML blast cells

expressing P-glycoprotein and to study the effect of MDR1 gene

expression

on

cellular

differentiation,

we

have

transfected

K562

cells,

ahuman

CML

blast cell

line (17), with

ahuman

MDR1

cDNA

(

1

8).

These

studies

demonstratethe cross-resis-tance pattern

of

the MDR

clones,

identify

chemotherapeutic

drugs

thatare not

affected by

P-glycoprotein

inK562

cells,

and

highlight drugs

that can

sensitize these

cells to

chemotherapeu-1.Abbreviations usedinthispaper:CML,chronicmyelogenous leuke-mia;MDR, multidrug resistance; MTT, 3-(4,5-dimethylthiazole-2 yl)-2,5diphenyltetrazoliumbromide.

J.Clin. Invest.

© The AmericanSocietyfor ClinicalInvestigation,Inc.

0021-9738/93/05/2207/09

$2.00

tic agents. Furthermore, we found that the expression of the MDRI gene did not affect the ability of K562 cells to undergo erythroid differentiationinresponse to hemin orcytosine

ara-binoside, and that differentiationwasassociated withincreased expression of P-glycoprotein.

Methods

Cell culture.K562 human CML blast cellsweremaintained in

logarith-mic growth in RPMI 1640medium containingpenicillin(100 U/ml) andstreptomycin (100 ug/ml) supplemented with 10% FCS. They were checkedroutinely andfound to be free of contamination by my-coplasmaorfungi.Cellslineswerediscarded after 6moof continuous passage,andnewlineswerestartedfromfrozen stocks.

Liposome-mediated

transfection

ofK562 cells with

purified

plas-midDNA(pHaMDR

1/A).

PlasmidDNAwaspurifiedfrom DH Sa

harboring the pHaMDR I /A cDNA (map shown in reference 18) after

amplificationwithchloramphenicol (19). The DNA wasbanded in

cesiumchloride/ethidiumbromidegradients.TheplasmidDNAwas

prepared byphenol/chloroform extraction(19).

1 x I07K562 cellswerewashedtwice with serum-freemedium and

transfected with10

_gg

ofpurified plasmidDNAusing100

tLg

oflipofec-tin reagent thatwasprepared immediatelybeforeuse(20).The suspen-sion culturesweredilutedwith30mlofRPMI 1640medium

supple-mented with 10% FBS. Theviabilityofthe cellswasmonitoreddaily by

trypan blueexclusionand dead cellswereremovedusing lymphocyte separationmedium. Viable cellswerecloned in 0. 12% agarfor14din the presenceof10 nM vinblastine. Adozen cloneswerepickedand expanded inliquidculture. Three clones labeled

K562/Vbl,,

K562/ Vbl2,and K562/Vbl3 were randomly chosen forfurtherstudy.

DNA isolation and Southern blot

analysis.

Genomic DNA from cultured cellswas prepared by

phenol-chloroform

extraction (19). After digestionwithHindIII, 15

Mg

ofDNAfrom each cell linewere electrophoresedin0.8% agarosegelsand transferredtonitrocellulose

filters.Thefilterswereprehybridized overnightat42°Cin50% form-amide,5X Denhardt's reagent, SX SSC ( lx SSCcontains0.15 M NaCl and 0.015 Msodium citrate), 1%SDS,and100,ug/mldenatured sal-monsperm DNA.Hybridizationwasperformed for16 h inthe same solution with the addition of 10% dextran sulfate and 107 cpm of cDNAprobe, whichwasthe 4.1-kbpHaMDRl/A insertcontaining

the sequence for the

MDRJ

cDNA labeled by the random primer

method (21 ).After hybridization,filterswerewashedfor 30 min

se-quentiallywith 2X SSC and 0.5% SDSat roomtemperature, with 2x SSC and 0.5%SDSat60°C, with 0.5X SSC plus 0.1% SDSat60°C, thenwith 0.25x SSCplus0.05% SDSat60°C.The final wash

con-tained 0.25x SSCat roomtemperature. Blotswerevisualized by

expos-ingKodakX-Omat AR filmtothe filtersat-70°Cfor 12-24 husing

anintensifyingscreen.

Northern blotanalysis.Totalcellular RNAwasisolatedusing guan-idinethiocyanate-cesiumchloridegradients (22).RNAsamples

con-taining7%formaldehydeand50%formamideweredenaturedby

heat-ingat65°Cfor10 min ( 19).CytoplasmicRNAfromeach celllinewas

sizefractionated through1%agarose/7%formaldehyde gels and blot-tedontonitrocellulose sheets. Theprehybridization and hybridization conditionswerethesame asdescribed above.RNAsamples for dot blotswere treated ina similarmanner. Theprobe for the fetal Ay

globin genewas a0.6-kb cDNA labeled by the random primer method.

A

32P-labeled

yactin probewasusedto ensureequalloading ofRNA

samples ineachlane.

DetectionofP-glycoprotein. P-glycoprotein was detected by

West-ernblottingandby immunohistochemistry. Plasma membranes were preparedaspreviouslydescribed (23). Membrane proteins were dis-solved in sample buffer for 10 minat roomtemperature andseparated in 7.5% SDS-polyacrylamide gels. Transfer ofproteins to nitrocellulose

wascarried out by the methodof Towbin et al. (24). The blots were incubated in blocking solution consisting of5% nonfatdry milk in

TBST (10mMTris-HCl, pH 8.0, 150 mM NaCI, and 0.05% Tween 20)

at roomtemperaturefor1 hfollowed by incubation at room

tempera-ture overnight with the C219 monoclonal antibody (0.2

_Ag/ml

in TBST),which recognizes an epitope located on the internal plasma membrane. Theimmunoblots were visualized with goat anti-mouse IgGconjugatedtoalkaline-phosphatase.

1 x I07viable cells for flow cytometry, were washed with PBS by

centrifugationat900gfor10min,resuspended in 12 ml of 70% metha-nol for 30 minat-70'C, then washed in PBS. The cells were incubated in0.5 ml PBS and 5% normal rabbit serum with or without 3

_Ag/ml

C219 monoclonalantibody. Cells were mixed at40C for 1 h, then washed threetimes in ice-cold PBS containing 1% BSA, then reincu-bated in 0.5 ml PBS and 5% normal rabbit serum containing 5

_Al

of

fluorescein-conjugated goat anti-mouse antibody. Cells were incu-bated in the darkat40C for45min and then washed twice in PBS. Finally, cells were resuspended in 0.5 ml of PBS and analyzed using a

flowcytometer(FACS IV® Becton-Dickinson, San Jose, CA) using

anexcitation wavelength of 488 nm and an emission wavelength of 520 nm.

Cytocentrifuge preparations for immunohistochemistry were air dried and stored at -20'C in a desiccator until use. MRK-16 (Subclass

IgG2a)wasused as theprimaryantibody and the biotin-avidin conju-gatedimmunoperoxidasewasused as the method of detection (25). Cellswithaviabilityof>95%, determined by trypan blue exclusion, were suspendedincalcium and magnesium-free phosphate-buffered salinecontaining0.05% BSA.Theywereincubatedin a blocking solu-tion of dilute normal horseserum(1:200 vol/vol PBS) for 20 min at

4°C,thenfor 30 min with MRK- 16 or dilute normal mouse ascites in a concentrationof5-10

Mug/ml.

Slideswerewashedwith PBS, incubated with biotinylated rabbit anti-mouse IgG for 30min,then with biotin-avidin reagentfollowedbyincubationin peroxidase solution, and fi-nally counterstained withhematoxylin.

Inhibitionofcell growth. The effect ofdrugs on the growth of K562

cellswasdeterminedbythe3-(4,5-dimethylthiazole-2yl)-2,5

diphenyl-tetrazolium bromide (MTT) assay(26). Briefly, 3 x 104cells/well wereplatedin 100

Ml

of growth medium in 96-wellmicrotiterplates andallowedtoequilibrate for2 h. Drugs weredilutedinmediumand added in 100Ml volumes.After72h, 50 MlMTTsolution (2mg/ml

PBS)wasadded,and cellswereincubatedanadditional4h at37°C. Themediumwasaspiratedand150

Mil

of100%dimethylsulfoxidewas added andplateswerethoroughly mixedfor10 min. Theoptical

den-sityof each wellwasdeterminedbyabsorbance spectrophotometryat

550nMusinganautomaticplatereader (DynatechLaboratories,Inc.,

Chantilly, VA).Assayswereperformedin-quadruplicate.

ICm

values represent the concentration ofdrugorcombinationsof drugs that pro-duced 50%inhibitionof

_Abs5m

ascomparedtovehicle treated controls. Theeffectof drugs aloneonthe MTTreactionwastaken into account in each assay.

Theeffect of chemosensitizersondrugresistancewasstudiedby

exposing

cellsto aconcentrationof chemosensitizer that alone pro-duced <10% inhibition of growth. Dose response curves were

corrected for theinhibitionof cellgrowthcausedbychemosensitizers alone. The "fold reversal" of MDRforeachdrug plus chemosensitizer

wascalculatedbydividing the IC50 obtained for chemotherapeutic drug aloneby the

_IC50

obtained forchemotherapeuticdrug plus chemosensi-tizer.

Cellularaccumulation of doxorubicin. 2.5 x 106cells in a total volumeof 1.5mlwereincubatedat37°Cwith 0-20MMdoxorubicin for 3 h in the presence and absence ofcyclosporin A(0.8

AM)

or trans-flupenthixol(5

MM).

Cellswerewashedthree times in PBSby

centrifugationat900 g for 10 min, resuspended in 1.5 ml of 0.3N

HCl

in50%ethanol, sonicated with 10pulsesat200-Wswithacell

sonica-tor(Tekmar/ Hereaus,Cincinnati, OH),and thencentrifugedat1,000

g for30 min. Thesupernatant fractionwasremoved andassayedfor doxorubicincontentwith aspectrofluorometer (model 512; Perkin-ElmerCorp., Norwalk, CT)usinganexcitationwavelengthof475nm

andanemissionwavelengthof575nm(8).

Effect

ofdrugs

oncellular

differentiation.

The

expression

offetalY

globin

mRNAandconcentration of

hemoglobin

wasusedasmarkers of

erythroid

differentiation. The

ability

ofheminand

cytosine

arabino-side toinduce

hemoglobin production

in K562 cellswas

assayed

as previously described (27,

28).

Cellswere grown in RPMI 1640

me-diumat370Cfor

varying periods

of time withorwithoutheminor

cytosine arabinoside,

then collected

by

centrifugation

at900 g for 10 min, washedoncewith ice-cold

PBS, resuspended

at aconcentrationof 1 X I05cells per

Al,

and

lysed

in 5 mM sodium

phosphate

buffer (pH

7.4)

containing

1% NP-40. The

lysates

were

centrifuged

at

12,000

g for 30 minat40Cand

hemoglobin

measured

by

the

tetramethylbenzidine

method

(28).

Statistical

analysis.

Statistical

analysis

ofeach doseresponsecurve was

performed

bythe methodof Finney (29).The

IC,0

values±stan-dard errorsfortheinhibitionof cellular

proliferation

by

drugs

aloneor

incombinationwere determined

by

linear

regression analysis

ofthe

logit-transformed

data. The

significance

ofeach"fold reversal" was

thendetermined

by

Student'sttestandwas

expressed

intermsof P values.

Drugs and chemicals.

Cyclosporin

Aand

trans-flupenthixol

were

supplied by

SandozLaboratories

(Sandoz

Pharmaceuticals,

East Han-over,

NJ)

and Dr.John

Hyttel

ofH. Lundbeck

(Copenhagen,

Den-mark),

respectively.

C2 19 and MRK- 16 monoclonalantibodies

recog-nizing P-glycoprotein,

weregenerous

gifts

from Centocor

Diagnostics,

Inc.

(Malvern,

PA)and Dr. T.

Tsuruo,

JapaneseFoundationofCancer Research

(Tokyo),

respectively.

The

following compounds

were pur-chased fromcommercialsources

including: vinblastine,

Eli

Lilly

Co.

(Indianapolis,

IN); doxorubicin,

Ben Venue Laboratories

(Bedford,

OH);

etoposide,

Bristol Laboratories

(Evansville,

IN);

Lambda

HindIII,

alkaline

phosphatase-conjugated

goat anti-mouseIgG, Pro-mega

(Madison, WI);

arandom

primer

DNA

labeling kit,

Boehringer

Mannheim

(Mannheim, Germany); [a-32P]-dCTP (3,000

Ci/

mmol),

Amersham

Corporation

(Arlington Heights, IL);

afetal A-y

globin

gene 0.6-kb cDNAwas

kindly

supplied byDr.Ajay Bhargava

(Yale

University,

NewHaven,

CT);

fluorescein-conjugated

goat

anti-mouse

antibody

from Becton

Dickinson;

MTT,

dimethylsulfoxide,

mechlorethamine,

hydroxyurea,

hemin,

hemoglobin,

3,3',5,5'-tetra-methyl

benzidine,

and

cytosine

arabinoside, SigmaChemicalCo., (St.

Louis, MO);

2-amino-6-purinethiol,

Aldrich Chemical Co.

(Milwau-kee, WI);

lipofectin

TM reagent, Bethesda Research

Laboratories,

(Gaithersburg,

MD) and;yactin probe,AmericanTypeCulture Col-lection

(Rockville,

MD).Allother chemicalswereof reagentgradeand

purchased

fromcommercialsources.

Results

Transfection of

K562 cells

by

pHaMDR 1/A

DNA.

The

trans-fected clones

were

20-30-fold

resistant

to

vinblastine,

23-35-fold

resistant

to

doxorubicin,

and three-

to

fivefold

resistant

to

etoposide

(Table I).

The

transfected

clones

retained their

sensi-tivity

to

mecholrethamine, hydroxyurea,

and

6-thioguanine.

The

nontransfected

resistant

clone,

K562

/

Vbl2,

showed a

simi-lar

pattern

of

resistance

to

that

of the transfected cells.

Transfection

of the

pHaMDRl

/A

DNA

resulted in the

in-corporation

of

the MDRI

cDNA

into

genomic

DNA

in

two

of

the three

clones

(Fig. 1).

A 4.1-kb MDR] cDNA

fragment

prepared

from

the

pHaMDR1/A

vector

by

digestion with

SacII

and

XhoI

strongly

hybridized

to

4.3-kb

and 2.2-kb DNA

fragments

in

_K562/Vbl,

and

Vbl3 clones.

Therewas no detect-able

hybridization

tothese

fragments

in the

parental line

orin the

Vbl2

clone. The

hybridization

pattern revealed no

struc-tural

differences

in the

endogenous

MDRI gene in

drug-sensi-tive and

-resistant

cells.

Resistant

K562 clones also

overex-pressed

MDR] mRNA whereas no

significant

hybridization

TableI.Effectof Chemotherapeutic

Druigs

onSensitive and Resistant K562 Clones

Cell line

Drug K562/S K56S/Vbl, K562/Vbl2 K562/Vbl3 IC50*

Vinblastine (nM) 2.7±1.8 55±9 80±15 70±5.3 Doxorubicin(,uM) 0.4±0.1 9.2±4.5 9.9±3.6 14±6.5 Mechlorethamine

(AM)

2.2±0.5 2.2±0.3 2.0±0.2 6.7±1.9 Etoposide (tiM) 5.6±0.6 16±2.5 15±2.5 30±0.5 Hydroxyurea(mM) 3.0±1.1 4.0+0.6 6.4±0.9 4.6±1.1 6-Thioguanine (MM) 1.2±0.2 0.5±0.2 0.9±0.3 0.8±0.5

*_{Valuesrepresent the}

_mean±SEM

_{fromtwo tothreeexperiments.}

was seen with mRNAs from the parental cells (data not

shown).

Resistant clones overexpressed P-glycoproteinasmeasured

by

severaltechniques.

Fig.

2 Ademonstrates the results

ofWest-ernblots

using

the C219

antibody,

which

recognizes

an inter-nalP-glycoprotein epitope. All three clones expressed the

pro-tein, although

the nontransfected

clone, Vbl2,

showed less

reac-tivity. Results of flowcytometryusing the C2 19 antibody and cell permeabilization are shown in Fig. 2 B. Sensitive cells showednoadditional fluorescenceoverbackground in the pres-enceof antibody, whereas the resistant clone, Vbl3,wasclearly identifiable. Cells incubated with C2 19 alone,orwith fluores-cein-conjugatedgoatanti-mouse antibody alone producedno

change comparedtobackground fluorescence. Results of

im-cm~~~c

iCD. CD; C-D

CD~.

LO L Lo

_Y)

I)

YY Y I

I:.

....

;

^ |

*

S-4.3

Kb

-2.2 Kb

Figure1. Identification of the MDR] cDNA in sensitive and resistant K562 cells by Southern blotanalysis. 15Mgof

genomicDNAderived from each cell linewere

digestedtocompletion

withHindIII, electro-phoresed in 0.8%

aga-rosegels,transferredto nitrocellulose,and

hy-bridized with the 4. 1-kb

32P-labeledMDR]

cDNAprobeas de-scribedin Methods. The

A

[IC

-0,

U-11I

MW(kD) 2

200- 116-

97-Cl) > > > (0 (0 (0 (0

LO 10 L 1O

Y Y Y Ye

66

-45

-B

(1)K562/S

(2) K562/S+C219 (3) K562/S+C219+Ig FITC

(4) K562/S+Ig FITC

FluorescenceIntensity

(1) K562Nbl3 (2) K562Nbl3+C219 (3) K562Nbl3+C219+Ig FITC

(1) (4) K562NbI3+IgFITC

(4)

100 lo 102

FluorescenceIntensity

Figure2.Expression of P-glycoproteininsensitive andresistantK562 cells.(A)Immunoblots: Membraneproteins (50-100jg perlane)

wereresolvedby SDS-PAGE,transferredtonitrocellulose, and

probedwith theC2 19 monoclonalantibodyasdescribedinMethods.

Sizemarkerswere

myosin, 200,000;

Escherichia coli

f3-galactosidase,

116,250;Rabbit musclephosphorylase b, 97,400; BSA, 66,200;hen

eggwhiteovalbumin, 45,000;bovine carbonicanhydrase, 31,000; soybean trypsin inhibitor, 21,500;and heneggwhitelysozyme,

14,400. MCF-7/ADrR cellswereusedaspositive controls. (B) FACSOanalysis. 1 x 107 cellswerefixedinmethanol, incubated in PBS containing 3 jg/ml of C219 antibody then counterstained with 5

Al

ofafluorescein-conjugatedgoatanti-mouseantibodyas

de-scribedinMethods.FACS® analysiswascarriedoutusing excitation

andemissionwavelengthsof 488nmand520nm,respectively. Top

panel, K562/S; bottom panel, K562/Vbl3. (C) (Continuedonnext

page) Immunohistochemistry. Cytocentrifuge preparationsof sensi-tiveand resistantcellswereairdried thenincubatedinPBS contain-ing5-10,ig/mlofMRK-16monoclonalantibody and visualized by

theavidin-biotin reactionasdescribed in Methods. Toppanel,

K562/Vbl3; bottom panel, K562/S.

munohistochemistry, using

the MRK- 16

antibody

that

recog-nizes

an

external P-glycoprotein epitope

areshown in Fig. 2 C. All

clones demonstrated immunoreactivity.

In

contrast,

the

sensitive parental line, K562/S,

showed

no

evidence

of

expres-sion of

P-glycoprotein

by Western blots

or

immunohistochem-istry.

Cross-resistance patterns and

effects

of chemosensitizers.

Trans-flupenthixol

and

cyclosporin

A

sensitized

the

resistant

cells

to

doxorubicin

and

vinblastine

(Table II). Cyclosporin

A was more potent and more

effective

than

trans-flupenthixol.

For

example, 0.8

,M

cyclosporin

A

sensitized

the

resistant

cells to

doxorubicin

by 27-47-fold and

to

vinblastine

by

12-26-fold.

In contrast,

5

_qM

trans-flupenthixol

sensitized the resistant

cells

to

doxorubicin

by

6-12-fold

and

to

vinblastine

by

8-10-fold.

The

chemosensitizers

had

no

significant effect

on

the

pa-rental line.

Drug accumulation

in

the

transfected clones. Sensitive

K562 cells accumulated twice

as

much doxorubicin

as

resistant

cells

(Fig.

3).

Trans-flupenthixol

and cyclosporin

A

increased

the cellular accumulation

of

doxorubicin

to

concentrations

ap-proximately equal

to

that

of the parental line (Fig. 3).

Chemo-sensitizers

had

no

effect

on

the

accumulation ofdoxorubicin

in

the sensitive

lines.

Influence of the

MDR

phenotype on erythroid

differentia-tion.

Sensitive

and

resistant K562 cells

were

equally

sensitive

to

the effects of hemin

on

erythroid differentiation (Fig.

4).

Fig.

4 A

demonstrates that fetal

Sy

globin

mRNA

increased

in both

K562/S and

K562/Vbl3

cells

after

exposure to

hemin.

Hemo-globin

content

increased with

increasing concentrations of

he-min in both cell lines until the

concentration of hemin

ex-ceeded 25

,M

(Fig.

4

B).

At

50

,M,

hemin

was more

effective

in

inhibiting the growth

of

parental cells

(data

not

shown)

while being

more

effective in inducing

hemoglobin

production

in

resistant

cells

(Fig.

4

B).

We also

studied the

effects

of

cytosine arabinoside

on

ery-throid differentiation

in K562 cells.

Fig.

5

demonstrates that

the

sensitive

K562/S

cells and

the

resistant

K562/Vbl3

cells

were

equally

sensitive

to

growth inhibition by cytosine

arabino-side

(Fig.

5

A),

and that

the resistant

cells

weremore

sensitive

to

the

effects

of

this drug

on

erythroid

differentiation

(Fig.

5

B).

To

study the

effect

of

differentiation

on

the

expression of

the MDRI

gene,we

measured the

expression

ofMDRI

mRNA

and

P-glycoprotein

after

exposureto

hemin

or

cytosine

arabino-side.

These

compounds

had

no

effect

on

the

expression

of

MDRJ

mRNAor

P-glycoprotein

in the

sensitive

cell line.

In contrast,

hemin

increased the

expression

of

MDR]

mRNA

by

130-290%

(Fig.

6

A) and

cytosine arabinoside

increased the

expression

of

MDRJ

mRNA

by

140%

(Fig.

7

A).

Similarly,

treatment

with

hemin

(Fig.

6

B)

or

cytosine arabinoside

(Fig.

7

B) selectively increased the

expression

of

P-glycoprotein

in the

resistant cells.

Discussion

The

transfection of K562 cells with

ahuman

MDRI

cDNA

produced

clones

of cells

that

display

the MDR

phenotype.

During

the

selection

process, we

also isolated

a

clone that

was

multidrug resistant,

but

unlike

K562/Vbl1

and

K562/Vbl3,

clone

K562/Vbl2

hadnot

incorporated

the

transfected

MDRJ

cDNA

(Fig. 1).

Allclones express

P-glycoprotein

when

mea-2210 W.N.Hait,S.Choudhury,S.Srimatkandada,andJ.R.Murren

750

D 500

E

z

)

_250n

a)

E

z

Figure

2(Continued)

sured

by

immunoblotting (Fig.

2

A),

by

flow

cytometry

(Fig. 2

All three clones display the

MDR phenotype (Table I) and B), or by

immunohistochemistry

(Fig.

2

C).

These results

maintain

the

phenotype

for<2 mo when grown in drug-free

point

out

that

even a

brief

exposureto

vinblastine

can

give rise

medium.

The pattern

of cross-resistance

in the

transfected

and to

resistant

clones

of

cells

expressing

the endogenous

MDRJ

the

nontransfected

clones are

similar.

They are most

resistant

gene. The

evaluation

of P-glycoprotein by

immunohistochem-

to

vinblastine

and

doxorubicin

and are less

resistant

to

etopo-istry

in these

lines

is

simple

and

sensitive,

since the two

antibod-

side.

ies against P-glycoprotein

are

of

mouse

origin,

and

therefore,

These

studies demonstrate

that K562 human

CML

blasts the

_{secondary antibodies directed against}

the mouse

immuno-

expressing P-glycoprotein

are as

sensitive

to

cytosine

arabino-globulin

do not produce

high background staining of

the hu-

side

(Fig.

5

A),

mechlorethamine, 6-thioguanine,

and

hydroxy-mancells. urea as the parental

line (Table

I).

Since

these

drugs

fail

to

t4,

.. I

..

.i ,.l.. ;" VIVL

:: x4.:.

..%`;.:,. r-..

4

_46---.q

7.

Table II. Effect of

Trans-flupenthixol

and Cyclosporin A on the Sensitivity ofParental andResistantK562 Clones to Doxorubicin and Vinblastine

Cell line Drug Foldreversal

Trans-flupenthixol CyclosporinA

K562/S Doxorubicin 0.6*

1.3*

K562/Vbl1

Doxorubicin 10§ 4711

K562/Vbl2 Doxorubicin 6§ 33§

K562/Vbl3 Doxorubicin 12$ 27§

K562/S Vinblastine 0.4* 1.0*

K562/Vbl1

Vinblastine 81" 26§

K562/Vbl2 Vinblastine

10o

15t

K562/Vbl3 Vinblastine 9* 12§

Cells were exposed to doxorubicin (0.1-33

jIM)

orvinblastine

(0.1-200 nM) in the absence or presence of trans-flupenthixol (5

MM)

or

cyclosporinA(0.8 WM).Fold reversal represents the IC50 of the

cyto-toxicdrug alonedividedby the IC50 of the cytotoxic drug in the presenceofthechemosensitizer.Each value represents the mean of atleastthreeexperiments. *Notsignificant compared to cytotoxic drug alone. tP<0.05. §P<0.01. 1"P <0.001.

produce sustained remissions in patients with

CML

in

blast

crisis,

these data strongly

suggest

that additional mechanisms

of resistance

arepresent

in the

clinical setting, and that

over-coming resistance mediated by P-glycoprotein

may

only

be the

first

step

toward improving

treatment. An

important feature of

the

resistant K562 clones, whether transfected

or

selected, is

that

resistance

mediated by

P-glycoprotein

is

not

complicated

by increased cellular

content

of glutathione

or altered

topo-isomerase

II

(Hait,

W. N.,

and

Y.-c.Cheng,

unpublished

obser-vations),

two

other mechanisms

known to produce cross-resis-tance to

chemotherapeutic drugs (30,

31

).

Since selection of

the clones

required

growth in low concentrations of vinblastine

for short

periods

of

time,

we

have

not

totally

excluded other

changes such

as

altered microtubular

function

(32).

In

addi-tion,

because of the low

IC50

for doxorubicin

in

the

sensitive

clone

(0.4

,M), it

was not

possible

to compare intracellular

accumulation

of doxorubicin between sensitive and resistant

clones

at

equitoxic

drug

concentrations.

Finally,

since

PCR

analysis

was not

done,

one cannot

definitively

rule

out

the

possibility

that

differences

between

sensitive and resistant

clones

were

caused

purely by

differences

in

the

expression

of

P-glycoprotein.

However, the

pattern

of

cross-resistance,

the

defect

in cellular

accumulation of

drug (Fig. 3),

and the

restora-tion of

sensitivity

by the chemosensitizers

(Table II) strongly

suggest that

expression of

P-glycoprotein

is

the

major

cause

of

the MDR

phenotype

in

these

cells.

Trans-flupenthixol

and

cyclosporin

A

increase sensitivity

of resistant K562 cells

to

vinblastine, doxorubicin,

and

etopo-side (Table II).

Cyclosporin

A

fully sensitized

both the

trans-fected

and

nontransfected K562 cell

lines,

whereas

trans-flu-penthixol

was somewhat less

effective (Table II).

At

higher

concentrations, trans-flupenthixol displayed intrinsic

cytotox-icity, making it impossible

to

distinguish

this effect from its

effect

on

P-glycoprotein mediated drug resistance.

The pattern

of reversal

seen in

transfected

clones

did

not

consistently differ from that

observed in theselected clone. For

._

o0

-0o

= E

=)

E

A Sensitive

0 10 20 30

Dox

_(gM)

3

2

.,

x0 _

0

- Q

0|

C T

B

Resistant

0 10 20 30

Dox

(gM)

Figure 3.

Effect

ofchemosensitizers onaccumulation of doxorubicin insensitive and resistant K562 cells. 2.5X

1O6

cells were incubated for 3 h with 0-20

MM

doxorubicinintheabsence orpresence of0.8

MMcyclosporinA or 5

AM

trans-flupenthixol.

Cell-associated doxoru-bicin(pmol/

106

cells)wasdetermined

spectrofluorometrically

as

de-scribedinMethods. Each

point

isthe meanof

duplicate

determina-tions.- -,

Doxorubicin;

-. -, doxorubicin+

cyclosporin

A; -o-, doxorubicin+

flupenxithol.

example,

Vbl2

(selected)

was

less affected

by trans-flupenthixol

plus

doxorubicin

than

Vbl3

(transfected),

yet

Vbl2

was more

sensitive

to

cyclosporin plus

vinblastine

than

Vbl3.

Further-more,

Vbl2

was at

least

as

sensitive

to

trans-flupenthixol

plus

vinblastine

as

both transfectants and

as

sensitive

to

cyclosporin

plus vinblastine

as

Vbl3.

Trans-flupenthixol

and

cyclosporin

A

produce identical

in-crements in

drug

accumulation

in the

Vbl3

cells

(Fig. 3),

yet

cyclosporin

A

supersensitizes

these cells

to

chemotherapeutic

drugs

(Table

II).

Similar

resultswere

obtained

with all clones

studied.

Trans-flupenthixol

and

cyclosporin

Aare

believed

to

work

through interference with

the

function of

P-glycoprotein,

since both

chemosensitizers

can

displace

[3H

]

azidopine

from

P-glycoprotein

(

10, 33).

The current

results

support

previous

A K562/S K562/Vbl3

Hemin(25,tm) _± +

Hemin(25,im)

2

pg

of RNA 1- * * *. , 0.5- .

0.25-

0.125-Hemin (25KLm)

Hemin(gM)

Figure4. Effect of hemin on the differentiation of K562 cells. (A) Effect on fetalyglobinRNAin sensitive and resistant K562 cells:

Afterexposureof sensitive and resistant K562 cells to 25MM hemin for4d,RNA waspreparedasdescribed in Methods. 0. 125-4Mgof RNA was blotted ontonitrocellulose and probed using a fetal

Ay

globin0.6-kb cDNA labeledby the random primer method. (B) Ef-fectonhemoglobin content: 1.5 xI05cells/ml wereincubatedfor

4d in the presence of 6.25-50MMheminorvehicle. Hemoglobin

content wasdeterminedusing tetramethylbenzidine as described in Methods. Eachpoint represents the mean±SEM of three experiments eachrunintriplicate.-_o-_{K562/S; -i -,}K562/Vbl3.

C/)

C(j (0

Ii

:5

(.0 LO)

Ne

I

c'j

(0o I n

MCF-7

(AdrR)

K562/Vbl3

K562/VbI3

(AdrR)~

~

C) C) 03 CD

MW_(kD) C O LIO C

200- 116-

97-

66-

45-observations by

us and others that

cyclosporin

A appears to have

additional

mechanisms of action.

For

example,

cyclo-sporin

A

increased the

sensitivity of "nonresistant" cells

to

A

0 1 10 100 1,000 Cytosinearabinoslie(nM)

E-ao

r

CX

D zz

00

Figure 5. Effect of cyto-sine arabinosideonthe growth and differentia-tionof K562 cells. 1.5 X I05cells/ml were

in-A _cubatedfor_{4d in the}

presence of0.36-3,600

nM orvehicle(media). After4d, cell number wasdetermined by elec-B troniccounting (A) and

hemoglobin content with tetramethylbenzi-dine (B)asdescribed in Methods. Each point represents the

mean±SEM of three

ex-perimentseachrunin

triplicate.

Figure6.Effect ofheminon MDR]geneexpressioninsensitive and resistant K562 cells. (A)MDR] mRNA:After exposure of sensitive

andresistant K562 cellsto25MMhemin for4d,RNA wasprepared

asdescribed in Methods. 20Mgof RNAwaselectrophoresedin 1% agarose/7%formaldehydegels then transferredtonitrocellulose. Blots wereprobed usinga 4.1-kbMDRIcDNAinsertprepared from the pHaMDRI/A retroviralvector.(B)P-glycoprotein. K562/Vbl3cells wereexposed to 25

MM

heminfor4d.Membraneswerepreparedas

described in Methods. Proteinsweresolublized andelectrophoresed

on7.5%SDSgels, transferredtonitrocellulose andprobedwith the C219antibody. MCF-7/ADrRcellswereusedaspositivecontrols.

anthracyclines (34, 35)

and altered membrane

potentials

(36) and

synthesis of nucleolar proteins (37).

Expression

of P-glycoprotein in K562

cells does not inter-fere with

erythroid

differentiation. This wastruein both the

transfected

_(K562/Vbli

and

Vbl3)

andnontransfected clones

(K562/Vbl2).

In

fact,

the

resistant cells produce

more

hemo-globin

in response to

cytosine arabinoside

than the

parental

cells

(Fig.

5

B).

This suggests that MDRI gene

expression

per sedoes not

interfere with early

steps in

hematopoietic

cellular

differentiation. This

data may be relevantto

ongoing

attempts

to create

multidrug

resistant

bone marrow

by transfection

of the MDRI gene

into early

hematopoietic

precursors, since

suc-A

f:E

C\j

[y

+ - + - +

0

a

-30

8=.

20a

8_co

20-5 10 0 E

Id

B

0)

(C

C)

1odA

0

so

E

I

60

40

20

10 100 1,000 1000

Cytosinearabinoside(nlU)

A

U)

LO

r-(0

LO y

r- -.]

Cytosine Arabinoside - + - + (1 Onrn)

B

CytosineArabinoside

(10 nm) MCF-7

(AdrR)

K562/Vbl3

EJCY) 03)

0) 0 0 0

C0 0 LO) 0

K562/Vbl3

0)

IC)LO

MW(kD)

200

- 116-97

-66

-45

-Figure 7.Effect of cytosine arabinosideon MDR]geneexpressionin

sensitiveandresistantK562 cells.(A)MDR] mRNA.Afterexposure

of sensitiveandresistantK562 cells to 10 nMcytosinearabinosidefor

4d,20,ugRNAwaspreparedand blotswere

probed

asdescribedin Methods and inFig.6. (B)P-glycoprotein. K562/Vbl3cells were

ex-posedto 10 nMcytosine arabinoside for4d.Membraneswere

pre-pared, blotted,andprobedasdescribedin

Fig.

6.

leagues, who

found

that multidrug resistant neuroblastoma

cells induced

to

differentiate with retinoic acid

increased

ex-pression of P-glycoprotein

(41), and with those of Fojo's group,

who

found increased P-glycoprotein in

coloncancer

cells induced

to

differentiate with sodium

butyrate (42). These

data

also show that induction of differentiation

in certain cells

that

do

not

intrinsically

overexpress MDRI

(i.e.

K562/S cells)

does

not

induce the

expression of this

gene product. Okabe-Kado et

al. had

previously

shownthat K562 cells selected for

resistance by chronic

exposure to

vincristine

could

differentiate

in

response to

hemin, but did

not

quantitate

the changes in

hemoglobin production

orthe

effects

on

P-glycoprotein

(43).

Furthermore, unlike previous studies

the current results are

unlikely

to

be

confounded

by cellular

alterations

that occur whencell

lines

are selected for long periods oftime in cytotoxic

drugs.

An

alternative explanation of

these

findings

is that the

drugs that induce differentiation

are

substrates for

P-glycopro-tein,

and

therefore, further upregulate the drug-efflux pump

as a

protective

mechanism, rather than

as aconsequence

ofdiffer-entiation.

In

fact, the resistant cells

were less

sensitive

to the

growth inhibitory effects of hemin than the resistant

ones. How-ever,

this

explanation

seems

unlikely, since

sensitivity

to

cyto-sine

arabinoside

is

not

affected

by

P-glycoprotein

(Fig.

5 A), yet

this

drug also

increased

MDR]

expression (Fig. 7).

Studies with transfected

K562 cells

may

provide direction

for

selecting

drugs

for

the

treatment

of

the blast

crisis of

CML.

For

example,

it

would

be

important

to

determine whether

treatment

of

patients

with

cyclosporin

A

plus

an

anthracycline

is

superior

to

that

of the

anthracycline alone,

or

if

a

combina-tion of non-cross-resistant

drugs, such

as

cytosine arabinoside

and

6-thioguanine given

with

an

anthracycline,

or

etoposide

plus

a

chemosensitizer, is superior

to

currently

available

treat-ments. In

addition,

these results

suggest

that

the

MDR

pheno-type

does

not

produce

cross-resistance

to

certain

differentiating

agents,

and

offers

a

potential alternative

to

cytodestructive

ther-apy

in

drug-resistant

CML.

Therefore,

the ultimate

test

of the

current

observations

will

require carefully designed clinical

trials

that

rigorously

evaluate the

presence

of

P-glycoprotein

in

CML blasts and the

effect

of chemosensitizers used with

drugs

affected by this form of

drug resistance.

Acknowledgments

cessful

marrow

reconstitution

would

depend

on

normal

prolif-eration and

differentiation.

Because

of the

ability

to

induce

erythroid

differentiation in

the

resistant

K562

clones,

we were

able

to

study

the

effect of

differentiation

onthe

expression

of

the

MDRJ

gene.

Differen-tiation along

the

erythropoietic

pathway increased

the

produc-tion

of MDRI

mRNA

(Figs.

6

A

and

7

A)

and the

synthesis

of

P-glycoprotein

in

resistant cells, but had

no

effect

on

the

ex-pression of this

gene

in the sensitive line

(Figs.

6

B

and

7

B).

The

increase

in MDR] was most

pronounced

in

clone

Vbl1.

The

induction of

MDRJ

is

complex

and may be

related

to

cellular

stress

induced

by

heat

shock, heavy metals,

orDNA

damage

as

described by Chin

etal.

(38,

39),

or

through

the

induction

of protein kinase

C

(40).

It will be

interesting

to

determine

whether these responses

differ

in the

_Vbl,

clone

compared

to

the

others.

These

studies

are

consistent with

those

of Biedler and

col-Wewish to thank Dr. Michael M. Gottesman and Dr. Ira Pastan for

kindly supplyingthe

pHaMDRl

/A retroviral expressionvectorand for manyhelpful discussionsand Dr.EdBenz,Dr.RobertHeimer,and Dr.RogerStrairfor

reviewing

themanuscript.

This work wassupported byPublic Health Service grants CA 43888 andCA 08341.Dr.William Hait isaBurroughs-WellcomeScholarin

ClinicalPharmacology.

References

1.Fialkow,P.J., and J. W.Singer.1989.Chronicleukemias. In Cancer Princi-ples and Practice of Oncology.V.T.DeVita,S.Hellman, and S.A.Rosenberg,

editors. J.B.Lippincott Co.,Philadelphia.p. 1841.

2.Gerlach,J.H.,N.Kartner,D.R.Bell, andV.Ling. 1986.Multidrug

resis-tance. CancerSurv.5:2546.

3.Gros,P.,Y. B.Neriah,J. M.Croop,andD. E.Housman. 1986.Isolation

andexpressionofacomplementaryDNAthat confersmultidrugresistance.

Na-ture(Lond.).323:728-731.

4.Ueda,K.,C.Cardarelli,M. M.Gottesman,andI.Pastan. 1987.Expression

ofafull-lengthcDNAfor the human "MDR1" gene confers resistanceto colchi-cine, doxorubicin and vinblastine. Proc.Nati.Acad.Sci. USA. 84:3004-3008.

5.Fojo, A. T.,S.-I. Akiyama,M.M.Gottesman, andI.Pastan.1985. Reduced drugaccumulation in multiple drug-resistant human KB carcinoma cell lines. Cancer Res.45:3002-3007.

6. Ford, J. M., and W. N. Hait. 1990. Pharmacology of drugs that alter multi-drug resistance in cancer.PharmacolRev.42:155-199.

7.Tsuruo, T., H. lida, S. Tsukagoshi, and Y. Sakurai. 1982. Increased

accu-mulation of vincristineandadriamycin in drug resistant P388 tumor cells follow-ing incubationwith calcium antagonists and calmodulin inhibitors. Cancer Res. 42:4730-4733.

8.Ganapathi, R., and D.Grabowski. 1983. Enhancementofsensitivity to

adriamycininresistant P388 leukemiaby the calmodulin inhibitor trifluopera-zine. Cancer Res. 43:3696-3699.

9.Ford, J. M., W. C.Prozialeck,and W. N. Hait. 1989. Structural features

determining activityofphenothiazines and related drugs for inhibition of cell growth and reversalof multidrug resistance. Mol.Pharmacol. 35:105-115.

10. Ford, J. M., E. P. Bruggemann, I. Pastan, M. M. Gottesman, and W. N. Hait. 1990. Cellular and biochemical characterization of thioxanthenes for rever-salofmultidrug resistancein human andmurinecelllines. CancerRes.

50:1748-1756.

1 1.Ramu,A., D. Glaubiger, and Z. Fuks. 1984. Reversalofacquired resis-tance todoxorubicin inP388 murine leukemia cellsbytamoxifenand other

triparanolanalogues. CancerRes.44:4392-4395.

12. Slater, L. M., P. Sweet, M. Stupecky, M. W. Wetzel, and S. Gupta. 1986.

CyclosporinAcorrectsdaunorubicin resistance in Ehrlich ascites carcinoma.Br. J.Cancer. 54:235-238.

13.Hait, W. N., J. M. Stein, A. Koletsky, M. W. Harding, and R. E. Handschu-macher. 1989.Activityof cyclosporin A and a non-immunosuppressive

cyclo-sporinagainst multidrug resistant leukemiccelllines.Cancer Commun. 1:35-43. 14.Murren, J. A., and W. N.Hait. 1992. Why haven't we curedmultidrug resistantcancer?Oncol.Res. 1:1-5.

15. Goldstein, L.J., H.Galski, A. T. Fojo, M. Willingham, S.-L. Lai, A.

Gazdar,R.Pirker,A.Green, W.Crist,G. M. Brodeur, et al. 1989.Expression ofa multidrug resistance gene in human cancer. J.Natl.Cancer Inst.81:116-124.

16. Tsuruo, T., Y. Sugimoto, H. Hamada, I. Roninson, M.Okumura, K. Adachi, Y.Morishima,and R.Ohno. 1987. Detection of multidrug resistance markers,P-glycoprotein andmdrl mRNA, in human leukemia cells. Jpn. J. Cancer Res.78:1415-1419.

17.Lozzio, C. B.,and B. B. Lozzio. 1975. Human chronic myelogenous

leukemiacell line with positive Philadelphia chromosome.Blood.45:321-334. 18. Pastan,I.,M. M.Gottesman, K. Ueda, E. Lovelace, A. V. Rutherford, and M.C.Willingham. 1988. Aretrovirus carryingan MDR] cDNAconfers multi-drugresistanceandpolarized expression of P-glycoproteinin MDCK cells. Proc.

Natl.Acad.Sci. USA. 85:4486-4490.

19.Maniatis,T., E. F. Fritsch, and J. Sambrook. 1989. Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. 20. Chang, A. C. Y., and D. G. Brenner. 1988. Cationic liposome-mediated

transfection:a newmethodfortheintroductionof DNA into mammalian cells. Focus(Idaho). 10:66-69.

21.Feinberg, A.P., and B. Vogelstein. 1983. a technique for radiolabeling DNArestrictionendonucleasefragments to high specific activity. Anal. Biochem. 132:6- 13.

22.Chirgwin,J. M., A. E. Przybyla, R. J. MacDonald, and W. J. Rutter. 1979.

Isolationofbiologicallyactive ribonucleicacid from sources enriched in ribonu-clease.Biochemistry.18:5294-5299.

23.Gerlach, J.H., D. R. Bell,C. Karakousis, H. K. Slocum, N. Kartner, Y.M.,Rustum,V.Ling, and R. M. Baker. 1987. P-glycoprotein in human sar-coma:evidence formultidrug resistance. J. Clin. Oncol.5:1452-1460.

24.Towbin,H., T. Staehelin, and J. Gordon. 1979. Electrophoretic transfer of

proteins frompolyacrylamide gels to nitrocellulose sheets: procedure and some

applications.Proc.Natl.Acad.Sci. USA. 76:4350-4354.

25.Warnke,R., and R. Levy. 1980.Detection of T and B cell antigens with

hybridoma monoclonal antibodies: a biotin-avidin horseradish peroxidase method. J.Histol. Chem.Cytochem. 28:771-777.

26.Mosmann, T. 1983. Rapid colorimetricassayfor cellular growth and survival:application to proliferation and cytotoxicity assays. J. Immunol. Meth-ods. 65:55-63.

27.Rutherford, T. R., J. B. Clegg and D. J. Weathrall. 1979. K562 human leukaemic cells synthesize embryonic haemoglobin in response to haemin. Na-ture(Lond.).280:164-165.

28.Conscience, J. R., R. A. Miller, J. Henry, and F. H. Ruddle. 1977. Acetyl-cholinesterase, carbonic anhydrase, and catalaseactivityinFriend

erythroleuke-mic cells,non-erythroid mouse cell lines and their somatic hybrids. Exp.CellRes. 105:401-414.

29.Finney, D. J. 1978. Statistical Methods in Biological Assay. Charles

Grif-fin and Co., London.

30.Dusre, L., E. G. Mimnaugh, C. E. Myers, and B. K.Sinha.1989.

Potentia-tion of doxorubicincytotoxicitybybuthionine sulfoximinein

multidrug-resis-tanthuman breast tumorcells. CancerRes.49:511-515.

31.Ganapathi,R., D.Grabowski,J. M.Ford, C.Heiss,D.Kerrigan,and Y.

Pommier.1989.Progressiveresistancetodoxorubicininmurine leukemia L 1210 cellswith multidrug resistancephenotype: reductions indrug-induced topoiso-merase11-mediatedDNAcleavage. Cancer Commun. 1:217-224.

32. Pain, J., F. M. Sirotnak, J. R. Barrueco, C-H Yang, and J. L.Biedler. 1988.

Altered molecularproperties of tubulinin amultidrug-resistant variant of

Chi-nese hamster cells selectedforresistance to vinca alkaloids. J. Cell. Physiol.

136:341-347.

33. Foxwell, B. M. J., A. Mackie, V. Ling, and B.Ryffel. 1989.Identification

ofthemultidrug resistance-relatedP-glycoproteinas acyclosporin binding

pro-tein. Mol. Pharmacol.36:543-546.

34.Chambers,S. K.,W.N.Hait,B.M.Kacinski,S. R.Keyes,and R. E. Handschumacher. 1989. Enhancementofanthracycline growthinhibition in par-entandmultidrug-resistant Chinesehamster ovary cellsbycyclosporinAand its analogues. Cancer Res.49:6275-6279.

35. Meador, J., P. Sweet, M. Stupecky, M. Wetzel, S. Murray, S. Gupta, and L. M.Slater. 1987. Enhancement bycyclosporinAof daunorubicinefficacy in Ehrlich's ascitescarcinomaandmurinehepatoma 129. Cancer Res. 47:6216-6219.

36.Vayuvegula, B., L.M.Slater,J.Meador,andS.Gupta.1988.Correction ofalteredplasmamembranepotentials.Apossiblemechanism ofcyclosporinA

andverapamilreversalofpleiotropicdrug resistance inneoplasia.Cancer Che-mother.Pharmacol. 22:163-168.

37.Sweet, P., P. K. Chan, and L. M. Slater. 1989.CyclosporinAand

vera-pamilenhancementofdaunorubicin-producednucleolarproteinB23

transloca-tion indaunorubicin-resistant and-sensitivehuman and murinetumorcells.

CancerRes.49:677-680.

38.Chin, K-V,T.Shiego,G.Darlington, 1. Pastan, and M. M. Gottesman. 1990. Heat shock and arsenite increase expression of the multidrug resistance

(MDRJ)geneinhumanrenal cellcarcinoma.J.Biol. Chem.265:221-226. 39.Chin,K-V, S. S. Chauhan,I.Pastan, and M. M. Gottesman. 1990. Regula-tionof mdrRNAlevelsin response to cytotoxic drugs in rodent cells.CellGrowth

&Differ. 1:361-365.

40.Chaudhary,P.M., andI.B.Roninson. 1992. Activation ofMDR-1

(P-gly-coprotein)geneexpressionin human cells by protein kinase C agonists.Oncol.

Res.4:281-290.

41.Biedler, J. L., D. Casals, T.-D. Chang, M. B. Meyers, B. A. Spengler, and R. A.Ross.1991.Multidrugresistant human neuroblastoma cells are more dif-ferentiated than controls andretinoic acid further induces lineage-specific differ-entiation. In Advances inNeuroblastomaResearch.Wiley-Liss,Inc., NewYork.

181-191.

42.Mickley,L. A., S. E.Bates, N. D.Richert,S.Currier,S. Tanaka, F. Foss, N.Rosen, and A.T.Fojo. 1989. Modulation oftheexpression ofamultidrug

resistancegene (mdr- 1 /P-glycoprotein) by differentiating agents.J. Biol.Chem. 264:18031-18040.