[CANCER RESEARCH 49, 2790-2796, May 15, 1989)

P-Glycoprotein Expression in Multidrug-resistant Human Ovarian CarcinomaCell Lines1

Grace Bradley, Michael Naik, and Victor Ling2

The Ontario Cancer Institute, The Princess Margaret Hospital and the Department of Medical Biophysics fG. B., M. N., V. L.], and the Faculty of Dentistry [G. B.J,University of Toronto, Toronto, Ontario, Canada M4XIK9

ABSTRACT

Multiple selections with either vinblastine or vincristine in the humanovarian carcinoma cell line SKOV3 resulted in variants with increasingdegrees of multidrug resistance. SKOV3 derivatives that span a widerange in resistance (4- to 2000-fold) were obtained and analyzed for P-glycoprotein expression. In general, we observed a progressive increasein P-glycoprotein level (detected by Western blot) that paralleled theincrease in multidrug resistance. However, a more detailed analysis ofthe P-glycoprotein inRNA and gene level indicated that the amount of P-glycoprotein expressed may be under complex control. At low levels ofresistance, only an increase in P-glycoprotein niRN A and protein wasobserved. At intermediate to high levels of resistance P-glycoprotein geneamplification became evident. At the high level of resistance, an examplewas observed where only the amount of P-glycoprotein was increasedwithout a concomitant increase in mRNA or gene copy. The mechanismsthrough which the content of P-glycoprotein in the plasma membrane ismediated are not understood; it is possible that the resistant variantsidentified here represent perturbations at different levels of regulation.

INTRODUCTION

Tumor cell resistance to cytotoxic drugs is thought to be amajor cause of failure in chemotherapeutic treatment of malignant tumors. Current experience with combination chemotherapy is that in many types of cancer, initial response is followedby relapse and resistance to further chemotherapy (1). It ispostulated that selection and proliferation of subpopulations oftumor cells that are simultaneously resistant to diverse cytotoxic agents occur during combination chemotherapy and areresponsible for eventual treatment failure.

Drug resistance has been studied extensively in in vitro modelsystems, and a multidrug resistance phenotype has been frequently observed, where selection for resistance to one cytotoxicagent results in cross-resistance to a wide variety of structurallyand functionally unrelated compounds. This multidrug resistance phenotype is characterized by decreased accumulation ofdrugs and overexpression of a highly conserved, M, 170,000plasma membrane glycoprotein, termed P-glycoprotein. Cloning of the P-glycoprotein gene has allowed extensive investigations into the molecular biology of P-glycoprotein expression(for reviews, see Refs. 2-4). Several lines of evidence indicatethat overexpression of P-glycoprotein causes multidrug resistance (5-8). A widely accepted model of P-glycoprotein functionproposes that P-glycoprotein is a transmembrane, ATP-de-pendent drug efflux pump (2, 3, 9-12). Thus increased expression of this protein would result in a lowered net accumulationof the drugs involved.

There is evidence to indicate that P-glycoprotein-mediated

Received 8/4/88; revised 1/5/89; accepted 2/7/89.The costs of publication of this article were defrayed in part by the payment

of page charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

' This study was supported by the National Cancer Institute of Canada and byPublic Health Service Grant CA37130 from the National Institutes of Health.Part of the work was carried out while G. B. was a Fellow of the Medical ResearchCouncil of Canada.

2To whom requests for reprints should be addressed, at the Ontario Cancer

Institute, 500 Sherbourne St.. Toronto, Ontario, Canada M4X 1K9.

multidrug resistance occurs in human tumors. Overexpressionof P-glycoprotein has been detected in a variety of patient tumorsamples (13-16). In a small number of cases, increase in P-glycoprotein expression correlated with disease progression (13,15). Analysis of human multidrug-resistant cell lines will benecessary for a more detailed understanding of the mecha-nism(s) of multidrug resistance in human tumor cells (17-21).Since increased expression of P-glycoprotein has been observedpreviously in ascites cells from patients with ovarian carcinoma(13), we undertook in this study to develop multidrug-resistantderivatives of a human ovarian carcinoma cell line, SKOV3(22). We were interested in establishing a relevant model systemto examine the correlation between P-glycoprotein expressionand various aspects of the multidrug resistance phenotype. Thisreport describes the characteristics of two series of multidrug-resistant SKOV3 cell lines derived by selection with a widerange of concentrations (more than 100-fold) of vinblastine orvincristine. These cell lines vary not only in their degree of drugresistance but also in their cross-resistance profiles. P-glycoprotein overexpression was measured at both protein and mRNAlevels and amplification of P-glycoprotein genes was analyzed.Certain features observed in these cell lines appear novel, nothaving been reported previously in other multidrug-resistanthuman cell lines.

MATERIALS AND METHODS

Reagents. Vinblastine, vincristine, colchicine, and gramicidin D werepurchased from Sigma Chemical Co. Adriamycin was from AdriaLaboratories. Guanidine thiocyanate was from Fluka, cesium chloridewas from Terochem Laboratories, oligodeoxythymidylate cellulose(type 3) was from Collaborative Research, and RNA size standardswere from BRL. Nitrocellulose paper (BA-85; 0.45-í¿mpore size) wasfrom Schleicher & Schuell. Restriction enzymes were purchased fromBoehringer Mannheim. Deoxycytidine 5'-[32P]triphosphate (3000 Ci/

mmol) was from Amersham. All other chemicals were of reagent grade.Cell Lines and Tissue Culture Conditions. SKOV3 is a human ovarian

carcinoma cell line (22) obtained from the American Type CultureCollection by Dr. Ronald Buick, Ontario Cancer Institute. The SKVLB series of cell lines was obtained by selection of SKOV3 cells forgrowth in increasing concentrations of vinblastine, and similarly, theSK VCR cell lines were obtained by selection with vincristine. Drug-sensitive SKOV3 cells were seeded into Nunc T75 flasks (Gibco) at 5x IO5 cells per flask. After 24 h of incubation, the selecting drug

(vinblastine or vincristine) was added and the cells were incubated fora further 7 to 10 days. A range of drug concentrations was used andthe flask at the highest drug concentration which showed growth wasidentified and taken for further manipulation. The identified T75 flaskwas expanded into several T75 flasks and the cells were maintainedand subcultura) at the same drug concentration until their doublingtime was 48 h or less. This usually required several weeks. Aliquotswere frozen and stored at -80°Cand the selection process was repeated

on the remaining cells. Mutagens were not used during selection forincreasing drug resistance.

All SKOV3 cell lines were routinely maintained in a-MEM3 (23)

3The abbreviations used are: a-MEM, o-minimal essential medium; cDNA,complementary DNA; SDS, sodium dodecyl sulfate; 1 x SSC, 0.15 M NaCl-0.015 M sodium citrate; PBS, 137 mM NaCl-2.7 mM KC1-8.1 raw Na2HPO4-1.3HIMKH2PO4-0.9 mM CaCh-0.3 mM MgCh.

2790

Research. on November 18, 2020. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from

P-GLYCOPROTEIN EXPRESSION IN MULTIDRUG-RESISTANT CELLS

supplemented with 15% fetal calf serum (Flow Laboratories). Drug-resistant cell lines were grown in medium containing the drug used forselection. Cells were kept at 37°Cin humidified atmosphere containing

5% CO2. These cell lines grew in monolayers and were subculturedweekly by standard trypsinization techniques (Bacto-Trypsin, Difco).

CEM, a human lymphoid cell line, and its vinblastine-resistantderivative, CEM VLB 0.1 (CEM/VLB,«,) were obtained from Dr. W.T. Beck, St. Jude Children's Research Hospital, Memphis, TN (18,

24). The CEM cell lines were grown in a-MEM supplemented with10% fetal calf serum as suspension cultures.

Drug Sensitivity Testing. Cell lines were tested for resistance to apanel of cytotoxic drugs using 24-well Linbro plates (Flow Laboratories), where one Linbro plate was used to test each cell line against eachdrug. Equal numbers of cells (approximately 1000 per well in 1 ml of«-MEMwith 15% fetal calf serum) were added to the wells and drugdissolved in 1 ml growth medium was then added in a serial 2-fold

dilution series through 23 wells. The remaining well contained 1000cells with no drug added as control. The plates were incubated at 37°C

for 14 to 21 days to adjust for different growth rates, until the controlwells for the resistant lines were equivalent to that of the parent line.The cells were then stained with méthylèneblue. Typically, wellscontaining low levels of drug showed a high density of colonies similarto that of the control well. At a fairly discrete drug concentration,inhibition of cell growth was noted and over a narrow range of drugconcentrations, there was a rapid fall in the number and size of thecolonies (Fig. 1). This allowed a quantitative determination of drugresistance by recording the drug concentration at which approximately90% inhibition of colony growth occurred. Duplicate experimentsshowed that this was a reproducible value for each cell line and eachdrug with up to a 2-fold error in determination of this value. Relative

resistance was calculated as the ratio of the inhibiting concentrationsfor the drug-resistant cell line and the parent line SKOV3.

DNA Preparation, Slot Blotting and Southern Blotting. High molecular weight genomic DNA was prepared by using standard techniques(25). For Southern blot analysis, 10 jig of DNA were digested with£coRI(5 units/Mg DNA) under conditions recommended by the manufacturer and were fractionated in a 0.6% agarose gel. Gel electropho-resis was carried out at 1.2 V/cm for 40 h at 4°C.The gel was stained

with ethidium bromide, denatured in 1.5 M NaCl, 0.5 N NaOH (twicefor 30 min) and neutralized in 1.5 M NaCl, l M Tris-HCI (twice for 30min). The DNA was transferred to nitrocellulose paper overnight in 20x SSC. The nitrocellulose paper was baked in vacuum for 2 h (25).DNA slot blots were prepared by using a Schleicher & Schuell slot blotapparatus according to manufacturer's instructions.

The probe used for detection of P-glycoprotein genomic sequenceswas pCHPl, a 660-base pair cDNA clone obtained from a Xgtllexpression library of a highly multidrug-resistant CHO cell line byscreening with a monoclonal antibody against P-glycoprotein (C219)(26, 27). The pCHPl sequence is within the most highly conservedregion of the P-glycoprotein cDNA. Sequence comparison betweenhamster and human cDNA clones showed extensive homology betweenpCHPl and the equivalent region in human P-glycoprotein clones (28).The pCHPl insert has been subcloned into PUC9 from which it wasexcised and was gel purified before use as a P-glycoprotein probe.

pCHPl was labeled with 32P by nick-translation (25) to a specificactivity of at least 2 x IO8cpm/Vg DNA. Southern blots and slot blotswere prehybridized for 6 to 18 h at 42°Cin 50% formamide, 5 x SSC,5 x Denhardt's solution (0.1% each of bovine serum albumin, polyvi-

nylpyrrolidone, Ficoll), 20 mM sodium phosphate, and 200 Mg/mlsalmon testes DNA. Hybridization was carried out for 20 h at 42°Cin50% formamide, 5 x SSC, 2 x Denhardt's solution, 20 mM sodiumphosphate, 200 Mg/ml salmon testes DNA, plus 2 x IO6 cpm/ml ofnick-translated pCHPl. After hybridization, blots were rinsed in 2 xSSC, 0.1% SDS at room temperature, washed in 2 x SSC, 0.1% SDS,0.1% sodium pyrophosphate at 50°C(twice for 30 min), and then in 1x SSC, 0.1% SDS, 0.1% sodium pyrophosphate at 50°C(twice for 30min). Blots were exposed to Kodak XAR film at —¿�70°Cwith intensifying screen. The degree of amplification of P-glycoprotein genes wasinitially estimated by autoradiographic exposure of Southern blots fordifferent lengths of time. This was then confirmed by densitornetry of

.JO. rfV^. ..*u

^

Oii^i«

•¿�&^' 0 ®^'^?é;:éJÉ

BFig. 1. Drug-sensitivity assay in Linbro trays. A, drug-sensitive parent cell

line; li. drug-resistant derivative. Approximately 1000 cells in growth mediumwere added per well. Two-fold serial dilutions of drug were placed in the wells,beginning from Al and proceeding from left to right along each row. The lastwell, D6, contained cells in growth medium without drug, as control. The plateswere incubated at 37°Cuntil the control wells showed equivalent colony growth,and then were stained with méthylèneblue. In this example, a 260-fold resistancewas observed. This was calculated by comparing the drug concentrations requiredto reduce plating efficiency to 10% of the control well in each cell line (Well A2versus Well B4).

DNA slot blots with a Beckman DU-50 spectrophotometer. Slot blots

were stripped of the pCHPl probe and reprobed with a tubulin probeto confirm that equal amounts of DNA had been applied.

RNA Preparation and Northern Blotting. Cells grown in Nunc T175flasks were lysed with guanidine thiocyanate and RNA was separatedfrom other cellular constituents by equilibrium centrifugation in CsCl(29). Polyadenylated RNA was isolated by using oligodeoxythymidylatecellulose columns (25). For Northern blot analysis, 10 Mgof polyade-

nylated RNA from various cell lines were fractionated in a 1% agarosegel that contained 6.7% formaldehyde. Gel electrophoresis was carriedout in 10 HIMphosphate buffer with 6.7% formaldehyde at 2.5 V/cmfor 17 h at room temperature. RNA was transferred to nitrocellulosepaper overnight in 20 x SSC. The nitrocellulose paper was baked invacuum for 2 h (25). RNA slot blots were prepared by using a Schleicher& Schuell slot blot apparatus according to manufacturer's instructions.

P-glycoprotein mRNA was detected in Northern blots by using nick-

translated pCHPl under the same conditions as described for Southernblots (see above). The size of the P-glycoprotein mRNAs demonstrated

in the Northern blots was estimated by using RNA size standards. Therelative amounts of P-glycoprotein mRNA was initially estimated byautoradiographic exposure of Northern blots for different lengths oftime. This was then confirmed by densitometry of RNA slot blots witha Beckman DU-50 spectrophotometer. Slot blots were reprobed with a

2791

Research. on November 18, 2020. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from

P-GLYCOPROTEIN EXPRESSION IN MULTIDRUG-RESISTANT CELLS

tubulin probe to confirm that the required amounts of mRNA had beenapplied.

Plasma Membrane Preparation, SDS-Polyacrylamide Gel Electropho-resis, and Western Blotting. Trypsinized cells were washed and resus-pended in PBS and disrupted by using a Stanstead cell disruptor (ModelA09512WS, with a No. 716 disrupting valve; Stanstead Fluid PowerLtd.). A plasma membrane-enriched microsomal fraction was obtainedfrom the disrupted cells by differential centrifugation (14). Proteincontent of the plasma membrane preparation was determined by amodification of the technique of Lowry et al. (30).

For Western blot analysis, plasma membrane preparations werefractionated by SDS-polyacrylamide gel electrophoresis by using amodification of Fairbanks' technique, as described previously (14, 31).

Electrophoresis was carried out at constant power at 5 W/gel for 3.5h. Proteins were transferred onto nitrocellulose paper by electroblottingat 80 V for 4 h, essentially as described by Towbin et al. (32). The blotswere presoaked in 3% bovine serum albumin in PBS at 37"C overnightand stored at 4°Cin the same solution until use.

Western blots were probed for P-glycoprotein by using the monoclonal antibody C219 that has been shown to recognize a highlyconserved epitope in P-glycoprotein (27). Blots were incubated in 3%bovine serum albumin in PBS that contained 5 x IO5cpm/ml of 125I-labeled antibody for 20 h at 4°C,washed with 4 changes of PBS atroom temperature over 2 h, and exposed to Kodak XAR film at —¿�70°C

with intensifying screen. Relative amounts of P-glycoprotein were estimated by comparison with a dilution series of SK VLB 1 plasmamembrane preparation (see Fig. 2).

RESULTS

Drug Resistance of Cell Lines. Tables 1 and 2 show levels ofresistance to a panel of cytotoxic drugs of two series of SKO V3cell lines selected with increasing concentrations of two relatedVinca alkaloids, vinblastine and vincristine, respectively. Thereare a number of notable findings.

Multidrug resistance is observed in all cell lines, selected with

Table 1 Levels of drug resistance in SKOV3 cells selected with increasingconcentrations of vinblastine

Relative resistance"

CelllineSKVLB0.004CSK

VLB0.008SKVLB0.01''SKVLB0.03''SKVLB0.06aSK

VLB 0.18SKVLB0.4SKVLB

l"VLB*284644909809802,000VCR263162,6005,2005,20010,000ADR264887170260260CCH1225130260330510GRAM13146,30013,00025,000NA

" Relative resistance was calculated as the ratio of drug concentrations whichinhibited colony growth of each drug-resistant cell line and the drug-sensitiveparent (see "Materials and Methods"). The values for the selecting drug are

italicized.* VLB, vinblastine; VCR, vincristine; ADR, Adriamycin; CCH, colchicine;

GRAM, gramicidin D; NA, not assayed.' Numbers refer to selecting drug concentration in fig/ml.''Denotes cell lines that were analyzed for P-glycoprotein overexpression.

Table 2 Levels of drug resistance in SKOV3 cells selected with increasingconcentrations of vincristine

Relative resistance"

CelllineSKVCR0.004fSK

VCR0.006SKVCR0.01SKVCR0.015''SK

VCR0.04SKVCR0.1*SKVCR0.25"SK

VCR0.5SKVCR 2"VLB*222432260100020004100VCR222816645105101000ADR646163232260260510CCH2224686464260GRAM1112813041004100NA

1See Table 1 footnotes.

widely different concentrations of vinblastine or vincristine. Wewere interested to determine if collateral sensitivity to agentssuch as local anesthestics and steroids that was observed insome systems (33) was also present in multidrug-resistantSKOV3 cells. Sensitivity testing against 1-dehydrotestosteroneand acronycine revealed no evidence of collateral sensitivityagainst these compounds (data not shown).

There is a consistent rank-order correlation between resistance to the selecting drug and cross-resistance to other drugs.However, the detailed cross-resistance profile varies among celllines selected with the same drug. In addition, cross-resistanceto other drugs may greatly exceed resistance to the selectingdrug.

A small increment in the drug concentration used for selection may result in a large increase in drug resistance of the cellpopulation (for example, compare the drug resistance of SKVLB 0.03 and SK VLB 0.06). This suggests that the level ofresistance achieved is not governed simply by the concentrationsof drug used. Previous studies (17, 34) have indicated that thegenetic control of P-glycoprotein expression may be complex.This may, at least in part, account for the difficulty in predictinglevel of drug resistance from the drug concentration used forselection of a particular variant.

These multidrug-resistant cell lines are cross-resistant togramicidin D. Cross-resistance to gramicidin D has been previously described for multidrug-resistant CHO cells (33). Gramicidin D is a membrane ionophore and is thought to act at thecell surface, without requiring transport into cells for its cytotoxic action. The remarkable increase in gramicidin D resistance during the later steps of selection in our experiments couldperhaps be explained by a pleiotropic perturbation in membraneproperties as a result of the presence of large amounts of P-glycoprotein (3). As will be described in greater detail below,this increase corresponds to a gene amplification event.

P-glycoprotein Expression in SKOV3 Cell Lines. In Westernblots of plasma membrane preparations of the SKOV3 celllines, overexpressed P-glycoprotein could be detected as a M,170,000 band in all the multidrug-resistant SKOV3 derivativesthat were analyzed, with the exception of SK VCR 0.015 (Fig.2). P-glycoprotein could not be detected in the parental SKOV3cell line. The lower molecular weight bands seen in the morehighly drug-resistant cell lines were not consistently demonstrated in different experiments and probably represent P-glycoprotein degradation fragments (27). In general, there isgood correlation between the amount of P-glycoprotein anddrug resistance, although there are clear exceptions (for example, see values for amount of P-glycoprotein and vincristineresistance for SK VLB 0.06 and SK VCR 0.25; Table 3). Therelationship between amount of P-glycoprotein and drug resistance varies among the live drugs that were tested. Colchicineand Adriamycin resistance are least responsive to increase inP-glycoprotein while gramicidin D resistance is most responsive(Tables 1-3). The Western blot analyses were repeated withanother monoclonal antibody against P-glycoprotein that isavailable in our laboratory, with essentially the same results(data not shown).

The results of Western blot assay of SKOV3 cell lines suchas that shown in Fig. 2 may be compared with those obtainedpreviously from ovarian carcinoma samples from patients (13).The techniques and probes used in these studies were essentiallythe same. Although it was not possible to deduce the precisedegree of drug resistance of the ovarian carcinomas from theWestern blot, it was noted that the amount of P-glycoproteinthat was present in some samples of ovarian carcinoma was

2792

Research. on November 18, 2020. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from

P-GLYCOPROTEIN EXPRESSION IN MULTIDRUG-RESISTANT CELLS

*//*** * *

170 KO-

I

O I 3 6 IO 15 30 60

Fig. 2. Western blot analysis of SKOV3 and CEM cell lines and their multi-drug-resistant derivatives. Membrane protein (60 fig) was loaded in each lane andfractionated in a 5.6% polyacrylamide gel containing 2% SDS and 4.5 M urea.The Western blot was probed with '"I-labeled C219 antibody against P-glycopro-

tein. Inset, Western blot of serial dilutions of SK VLB I membrane preparations,probed with '"I-labeled C219. The numbers below each lane refer to the amount

(in fig) of SK VLB 1 membrane protein loaded per lane. Ordinate, molecularweight in thousands.

Table 3 Levels o/P-glycoprotein and its mRNA and DMA in multidrug-resistantSKOV3 cells

Cell line

Relative resistance"

VLB' VCR mRNA* DNA*

SKOV3SKVLB0.01SKVLB0.03SKVLB0.06SKVLB1SKVCR0.015SKVCR0.1SKVCR0.25SKVCR 214644902,00042601,0004,10013162,60010,0008645101,000ND<2210100ND5n50NDND*1111000.1356741111129NA187

" Relative resistance was calculated as the ratio of drug concentrations whichinhibited colony growth of each drug-resistant cell line and the drug-sensitiveparent (see "Materials and Methods").

* Relative amounts of P-glycoprotein and its mRNA and DNA were estimatedby probing with monoclonal antibodies and a cDNA probe for P-glycoprotein,respectively (see "Materials and Methods"). For protein and mRNA estimations,

the amount present in SK VLB 1 cells was arbitrarily set at 100, since this cellline exhibited the highest degree of P-glycoprotein expression in this study. Thiswas preferred to setting the values for the parent cell line SKOV3 at 1, since P-glycoprotein was not detected in standard Western or Northern blots for this cellline in the present study.

' VBL, vinblastine; VCR, vincristine; ND, not detected; NA, not assayed.d For this cell line, P-glycoprotein mRNA was not detected above the back

ground of nonspecific hybridization in slot blots, but could be clearly detected inNorthern blots (see Fig. 3).

similar to that in SKOV3 cell lines with greater than 50-foldresistance to several clinically useful chemotherapeutic agents.

Expression of P-glycoprotein mRNA. In Northern blot analysis, each of the multidrug-resistant SKOV3 derivatives showedoverexpression of P-glycoprotein mRNA of 4.4-5 kilobases insize (Fig. 3). P-glycoprotein mRNA was not detectable for thedrug-sensitive parental line SKOV3. Additional mRNA speciesof larger size were evident in the Northern blots. In particular,

an mRNA of 5.5-6 kilobases was clearly seen in most of themultidrug-resistant cell lines, being approximately 2-fold lessintense than the major (4.4-5 kilobases) mRNA. The presenceof P-glycoprotein mRNA of different sizes within one cell linehas been reported previously (17, 35, 36), but the presentfinding of two distinct P-glycoprotein mRNAs that are relatively similar in size and abundance has not been described inprevious studies. Multiple transcription initiation sites for P-glycoprotein mRNA have been described in multidrug-resistanthuman KB/HeLa cell lines (37). This finding is unlikely toaccount for the two distinct P-glycoprotein mRNAs in themultidrug-resistant SKOV3 cell lines, because the largest difference in position among the different initiation sites is lessthan 500 bases (37). Moreover, Northern blot analysis of theKB/HeLa cell lines did not reveal two distinct P-glycoproteinmRNA bands differing in size by about 1 kilobase (17). Thesize of the major mRNA species (4.4-5 kilobases) in our studyis consistent with the estimated molecular weight of 140,000for the polypeptide portion of P-glycoprotein (38). The largermRNA species of 5.5-6 kilobases may be a precursor of the4.4-5 kilobase mRNA, but the significance of its presence isnot clear.

The relative amounts of P-glycoprotein mRNA in differentSKOV3 derivatives were estimated (Fig. 5B; Table 3). Theamount of P-glycoprotein mRNA generally correlated with theamount of P-glycoprotein as demonstrated by Western blot, butwe observed an exception. The cell line SK VCR 2 which wasderived from SK VCR 0.25 in two steps showed virtually noincrease in P-glycoprotein mRNA content in repeated analysesbut it clearly has increased P-glycoprotein and increased resistance to all five tested drugs. This suggests that P-glycoproteincontent and multidrug resistance may be regulated at the protein level, without alteration of mRNA content.

P-glycoprotein Gene Amplification. Southern blot analysis of

o- o- o- ^ o- .N

4I I

4- 4 41 4 4"I I

to*

kb

9.49 -7.46 -

4.40-

2.37-

1.35-

Fig. 3. Northern blot analysis of SKOV3 and its multidrug-resistant derivatives. Polyadenylated RNA (10 fig) was loaded in each lane and fractionated in a1% agarose-formaldehydegel. The Northern blot was probed with nick-translatedpCHPl for P-glycoprotein sequences. This is a composite of films exposed fordifferent lengths of time to show the pattern of bands clearly for each cell line.Lanes 1-3 and 6-7, 8 days of exposure; lanes 4-5 and 8-9, 22 hours of exposure.Size markers were the BRL RNA ladder, kb, kilobases.

2793

Research. on November 18, 2020. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from

P-GLYCOPROTEIN EXPRESSION IN MULTIDRUG-RESISTANT CELLS

\ 0 Ox < O o0' Q- V 0- Ov O" v

f f i

lililí////v /

l l lkb

23.1-

9.4-

6.7-

4.4-

2.3-2.0-

Fig. 4. Southern blot analysis of SKOV3 and CEM cell lines and theirmultidrug-resistant derivatives. EcoRI-digested genomic DNA (10 fig) was loadedin each lane and fractionated in a 0.6% agarose gel. The Southern blot was probedwith nick-translated pCHPl for P-glycoprotein sequences. Size markers were////nil 11fragments of XDNA. Aré,kilobase pairs.

genomic DNA with pCHPl as a probe for P-glycoproteinsequences showed a complex pattern of eight EcoRl fragmentsfor the SKOV3 cells (Fig. 4). The same complement of eightbands hybridizing to pCHPl was previously described for thehuman lymphoid cell line CEM (26, 39). The multiplicity ofEcoRl fragments that hybridize to pCHPl in Southern blots isstrongly suggestive of a multigene family that encodes P-glycoprotein, since pCHPl is a relatively short cDNA sequencewith no internal EcoRl site (26). Recent analysis of P-glycoprotein clones from human liver cDNA libraries has establishedthe existence of two distinct but homologous P-glycoproteingenes (34). Comparison of the nucleotide sequence of the hamster-derived pCHPl probe with the equivalent regions of thehuman cDNA clones showed extensive sequence homologybetween pCHPl and both cloned human P-glycoprotein genes(data not shown). Thus the pCHPl-homologous sequencesshown in Fig. 4 must represent at least two human P-glycoprotein genes.

Initial steps of selection with vincristine and vinblastine didnot result in amplification of P-glycoprotein genes, althoughthere was progressive increase in P-glycoprotein mRNA andprotein (compare Fig. 4 with Figs. 2 and 3). Thus, in multidrug-resistant SKOV3 cell lines, selected with vinblastine or vincristine, P-glycoprotein overexpression precedes gene amplification. A similar finding has been previously reported for colchi-cine-selected human KB/HeLa cell lines in which mRNA levelsand amplification of mdrl sequences (P-glycoprotein sequences) were studied (17).

Further steps of selection with either vincristine or vinblastineresulted in the emergence of cell populations that have amplified P-glycoprotein genes (Fig. 4). Seven of the eight pCHPl-homologous EcoRl fragments seen in the parental line SKOV3are amplified. The pattern of DNA amplification is similar in

SKOV3

SK VLB 0.01

SK VLB 0.03

SK VLB 0.06

SK VLB I

SK VCR O.I

SK VCR 0.25

SK VCR 2

- 5

- 5

- 5

-0.6

- 0.3

- 5

- 1.7

- l .7

P-Glycoprotein Tubulin /igDNA

SKOV3

SK VCR 0.015

SK VCR O.I

SK VCR 0.25

SK VCR 2

SK VLB 0.01

SK VLB 0.03

SK VLB 0.06

SK VLB I

- 10

- 10

- 2

- l

- l

- 10

- 6.7

- 1.3

- l

P-Glycoprotein fubulin /¿gRNA Q

Fig. 5. Representative DNA (A) and RNA (B) slot blots used for determinationof relative P-glycoprotein DNA and mRNA content. The slot blots were initiallyprobed with nick-translated pCHPl for P-glycoprotein sequences. The pCHPlprobe was then removed and the slot blots were reprobed with a nick-translatedtubulin cDNA probe. Different amounts of genomic DNA (A) or polyadenylatedRNA (//) were loaded for different cell lines in order to produce autoradiographicsignals that were more comparable in intensity, to facilitate densitometric analysis.The amounts of P-glycoprotein DNA or mRNA as determined from densitometrywere normalized against the tubulin contents.

multidrug-resistant SKOV3 derivatives that were selected withdifferent drugs (vinblastine or vincristine) and have differentcross-resistance profiles. The pattern of DNA amplification inCEM VLB 0.1, a multidrug-resistant human lymphoid cell lineselected with vinblastine, is also shown in Fig. 4. CEM VLB0.1 showed a distinctly different pattern of DNA amplification,where only four of the eight pCHPI-homologous fragments areamplified (Fig. 4, Bands 1, 3, 6, and 7). The differential amplification of pCHPl-homologous EcoRl fragments is consistentwith the assignment of Bands 1, 3, 6, and 7 to one P-glycoprotein gene and Bands 2,5, and Ato a second P-glycoprotein gene(Fig. 4). Thus, our findings confirm the recent report of differential amplification of two P-glycoprotein genes in humanmultidrug-resistant cell lines (40). The significance of Hand 4is not clear as it hybridizes only weakly to pCHPl and is notamplified in either the SKOV3 or the CEM cell lines.

The degree of P-glycoprotein gene amplification in the multidrug-resistant SK.OV3 cell lines was estimated (Table 3).Autoradiographic exposure of Southern blots for differentlengths of time indicated that DNA bands corresponding to thetwo P-glycoprotein genes are coordinately amplified in thehighly drug-resistant SKOV3 derivatives selected with eithervinblastine or vincristine. Slot blot analysis was then used toconfirm estimates of the degree of P-glycoprotein gene ampli-

2794

Research. on November 18, 2020. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from

P-GLYCOPROTEIN EXPRESSION IN MULTIDRUG-RESISTANT CELLS

fication (Fig. 5/1). The cell line SK VLB 0.06 was derived fromSK VLB 0.03 by doubling the selecting concentration of vin-blastine and has an 11-fold amplification of both P-glycoproteingenes. Similarly, the cell line SK VCR 0.25 was derived fromSK VCR 0.1 by a 2.5-fold increase in vincristine concentrationand has an 8-fold amplification of P-glycoprotein genes. Thus,increase in P-glycoprotein gene copy number may occur in large

steps during selection for drug resistance.Additional steps of selection with vinblastine resulted in the

cell line SK VLB 1 which shows further P-glycoprotein geneamplification. Additional steps of selection with vincristineresulted in the cell line SK VCR 2 which shows essentially thesame degree of P-glycoprotein gene amplification and P-glycoprotein mRNA expression as SK VCR 0.25, while a distinctincrease in amount of P-glycoprotein is evident from the West

ern blot assay (Table 3).

DISCUSSION

Multidrug resistance has been studied extensively in rodentand human cell lines. Overexpression of P-glycoprotein hasbeen demonstrated at either the protein level or the mRNAlevel in the majority of multidrug-resistant cell lines. In somecases, P-glycoprotein gene amplification has been shown toaccompany overexpression. However, the progression of P-glycoprotein overexpression during selection for increasing degrees of drug resistance has only been carefully examined in afew systems. Multidrug-resistant CHO cell lines obtained bystepwise clonal selection show an increase in P-glycoproteingene copy number, mRNA content, and protein content inevery step of selection (26). This is in contrast to serial multi-drug-resistant derivatives of the human KB/HeLa cell line,where overexpression of P-glycoprotein mRNA (mdrl sequences) precedes gene amplification (17). The present studycorrelates P-glycoprotein gene amplification, P-glycoproteinmRNA overexpression, and protein overexpression during selection for increasing multidrug resistance in the human ovariancarcinoma cell line SKOV3. The results indicate that the presence of increased amounts of P-glycoprotein in the plasmamembrane may be the result of three types of alteration in thecontrol of P-glycoprotein expression: (a) overexpression ofmRNA and protein without gene amplification; (¿>)gene amplification and overexpression of mRNA and protein; and (c)overexpression of protein without a change in mRNA or genelevel (Table 3).

Our finding that, at low levels of multidrug resistance, P-glycoprotein mRNA overexpression occurs without gene amplification agrees with previous reports of other human multi-drug-resistant cell lines (17, 35). In addition, during selectionfor high levels of vincristine resistance, we also observed anincrease in resistance to multiple drugs and increase in P-glycoprotein content in a cell line (SK VCR 2) without acommensurate change in P-glycoprotein mRNA and gene level(Table 3). This indicates that the amount of P-glycoproteinmay be regulated at the translational or posttranslational level.This type of control of P-glycoprotein content has not beenpreviously described and its occurrence and mechanism meritfurther study. If it proves to be of common occurrence, measurement of the actual P-glycoprotein content, such as withWestern blot or immunocytochemical staining, should be preferred over mRNA measurements as an accurate indicator ofdegree of multidrug resistance.

The onset of gene amplification in our cell lines is abrupt andis accompanied by a large increment in multidrug resistance

which is disproportionate to the increase in selecting drugconcentration (Tables 1-3). The rise in cross-resistance to themembrane ionophore, gramicidin D, is particularly remarkable.It is possible that an analogous phenomenon may occur inclinical chemotherapy, where exposure of neoplastic cells to acritical drug concentration may result in the selective outgrowthof a subpopulation with markedly increased resistance to multiple drugs. Such multidrug-resistant cell populations may bedifficult to eradicate with conventional chemotherapy, but maybe susceptible to novel therapeutic agents, because of their highcontent of P-glycoprotein and altered membrane properties.

The multidrug-resistant human ovarian cell lines describedhere illustrate the complexity of multidrug resistance both atthe phenotypic and at the macromolecular level. These cell linesspan a wide spectrum of multidrug resistance and they shouldbe useful in future studies of factors that determine the diversityof the cross-resistance profile and the mechanisms that regulatethe amount of P-glycoprotein expressed in the plasma membrane. Better understanding of P-glycoprotein-mediated multi-drug resistance in human tumor cells should help in the development of a rational approach toward overcoming resistanceto chemotherapy.

ACKNOWLEDGMENTS

We thank Deanna Evernden-Porelle for her technical assistance.

REFERENCES

1. Devita, V. T. Jr., I Id Ini.in. S., and Rosenberg, S. A. Cancer. Principles andPractice of Oncology, Vol. 1, pp. 262-267. Philadelphia: J. B. LippincottCo., 1985.

2. Pastan, I., and Gottesman, M. Multiple-drug resistance in human cancer. N.Engl. J. Med., 316: 1388-1393, 1987.

3. Bradley, <... Juranka, P. F., and Ling, V. Mechanism of multidrug resistance.Biochim. Biophys. Acta, 948: 87-128, 1988.

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

5. Ling. V.. Juranka, P. F.. Endicott. J. A., Deuchars, K. L., and Gerlach, J. H.Multidrug resistance and P-glycoprotein expression. In: P. V. Wooley andK. D. Tew (eds.). Mechanisms of Drug Resistance in Neoplastic Cells, Vol.9, pp. 197-209. New York: Academic Press, 1988.

6. Gros, P., Neriah, Y. B., Croop, J. M., and Housman, D. E. Isolation andexpression of a complementary DNA that confers multidrug resistance.Nature (Lond.), 323: 728-731, 1986.

7. Croop, J. M., Guild, B. C., Gros, P., Mulligan, R., and Housman, D. E.Genetics of multidrug resistance: relationship of a cloned gene to the complete multidrug resistance phenotype. Cancer Res., 47: 5982-5988, 1987.

8. Ueda, K., Cardarelli, C., Gottesman, M. M., and Pastan, I. Expression of afull-length cDNA for the human MUKI gene confers resistance to colchicine,doxorubicin and vinblastine. Proc. Nati. Acad. Sci. USA, 84: 3004-3008,1987.

9. Cornwell, M. M., Pastan, I., and Gottesman, M. M. Certain calcium channelblockers bind specifically to multidrug-resistanl human KB carcinoma membrane vesicles and inhibit drug binding to P-glycoprotein. J. Biol. Chem.,262:2166-2170, 1987.

10. Safa, A. R., Glover, C. J., Meyers, M. B., Biedler, J. L., and Felsted, R. L.Vinblastine photoaffinity labeling of a high molecular weight surface membrane glycoprotein specific for multidrug-resistant cells. J. Biol. Chem., 261:6137-6140, 1986.

11. Cornwell, M. M., Tsuruo, T., Gottesman, M. M., and Pastan, I. ATP-bindingproperties of P-glycoprotein from multidrug-resistant KB cells. FASEB J.,A-51-54, 1987.

12. I lanuda. H., and Tsuruo, T. Purification of the 170- to 180-kilodaltonmembrane glycoprotein associated with multidrug resistance. J. Biol. Chem.,263: 1454-1458, 1988.

13. Bell, D. R., Gerlach, J. H., Kartner, N., Buick, R. N., and Ling, V. Detectionof P-glycoprotein in ovarian cancer: a molecular marker associated withmultidrug resistance. J. Clin. Oncol., 3: 311-315, 1985.

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

15. Ma, D. D., Davey, R. A., Harman. D. H.. Isbister, J. P., Scurr, R. D.,Mackertich, S. M.. Dowden. G., and Bell, D. R. Detection of a multidrugresistant phenotype in acute non-lymphoblastic leukemia. Lancet, /. 135-137, 1987.

16. Fojo, A. T., Ueda, K., Slamon, D. J., Poplack, D. G., Gottesman, M. M.,

2795

Research. on November 18, 2020. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from

P-GLYCOPROTEIN EXPRESSION IN MULTIDRUG-RESISTANT CELLS

and Pastan, I. Expression of a multidrug-resistance gene in human tumorsand tissues. Proc. Nati. Acad. Sci. USA, 84: 265-269, 1987.

17. Shen, D. W., Fojo, A., Chin, J. E., Roninson, I. B., Richert, N., Pastan, 1.,and Gottesman, M. M. Human multidrug-resistant cell lines: increased mdrlexpression can precede gene amplification. Science (Wash. DC), 232: 643-645, 1986.

18. Beck, W. T., Mueller, T. J., and Tanzer, L. R. Altered surface membraneglycoproteins in Vinca alkaloid-resistant human leukemic lymphoblasts. Cancer Res., 39: 2070-2076, 1979.

19. Batist, G., Tulpule, A., Sinha, B. K., Katki, A. G., Myers, C. E., and Cowan,K. H. Overexpression of a novel anionic glutathione transferase in multidrug-resistant human breast cancer cells. J. Biol. Chem., 26]: 15544-15549,1986.

20. Fairchild, C. R., Ivy, S. P., Goldsmith, M. E., Kao-Shan, C. S., Whang-Peng,J., Rosen, N., Israel, M. A., Melera, P. W., and Cowan, K. H. Isolation ofamplified and over-expressed DNA sequences from Adriamycin-resistanthuman breast cancer cells. Cancer Res., 47: 5141-5148, 1987.

21. Dalton. W. S., Durie, B. G. M., Alberts, D. S., Gerlach, J. H., and Cress, A.E. Characterization of a new drug-resistant human myeloma cell line thatexpresses P-glycoprotein. Cancer Res., 46: 5125-5130, 1986.

22. Fogh, J., and Trempe, G. New human tumor cell lines. In: J. Fogh (ed.).Human Tumor Cells In Vitro, pp. 115-159. New York: Plenum PublishingCorp., 1975.

23. Stanners, C. P., Elicieri, G., and Green, H. Two types of ribosome in mouse-hamster hybrid cells. Nature New Biol., 230: 52-54,1971.

24. Beck, W. T. Cellular pharmacology of t inni alkaloid resistance and itscircumvention. Adv. Enzyme Regul., 22:207-227, 1984.

25. Maniatis, T., Fritsch, E. F., and Sambrook, J. Molecular Cloning: A Laboratory Manual. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory,1982.

26. Riordan, J. R., Deuchars, K., Kartner, N., Alón,N., Trent, J., and Ling, V.Amplification of P-glycoprotein genes in multidrug-resistant mammalian celllines. Nature (Lond.), 316: 817-819, 1985.

27. Kartner, N., Evernden-Porelle, D., Bradley, G., and Ling, V. Detection of P-glycoprotein in multidrug-resistant cell lines by monoclonal antibodies. Nature (Lond.), 316: 820-823, 1985.

28. Endicott, J. A., Juranka, P. F., Sarangi, F., Gerlach, J. H., Deuchars, K. L.,and Ling, V. Simultaneous expression of two P-glycoprotein genes in drug-

sensitive Chinese hamster ovary cells. Mol. Cell. Biol., 7: 4075-4081, 1987.29. Chirgwin, J. M., Przybyla, A. E., MacDonald, R. J., and Rutter, W. J.

Isolation of biologically active ribonucleic acid from sources enriched inribonuclease. Biochemistry, 18: 5294-5299, 1979.

30. Peterson, G. L. A simplification of the protein assay method of Lowry et al.which is more generally applicable. Anal. Biochem., 83: 346-356, 1977.

31. Fairbanks, G., Steck, T. L.. and Wallach, D. F. H. Electrophoretic analysisof the major polypeptides of the human erythrocyte membrane. Biochemistry,10: 2606-2617, 1971.

32. Towbin, H., Staehelin, T., and Gordon, J. Electrophoretic transfer of proteinsfrom polyacrylamide gels to nitrocellulose sheets: procedures and someapplications. Proc. Nati. Acad. Sci. USA, 76:4350-4354, 1979.

33. Bech-Hansen, N. T., Till, J. E., and Ling, V. Pleiotropic phenotype ofcolchicine-resistant CHO cells: cross-resistance and collateral sensitivity. J.Cell. Physiol., 88: 23-31, 1976.

34. Van der Bliek, A. M., Baas, F., Ten Houle de Lange, T., Kooiman. P. M.,Van der Velde-Koerts, T., and Borst, P. The human mdrìgene encodes anovel P-glycoprotein homologue and gives rise to alternatively splicedmRNAs in liver. EMBO J., 6: 3325-3331,1987.

35. Fuqua, S. A. W., Moretti-Rojas, I. M., Schneider, S. L., and McGuire, W.L. P-glycoprotein expression in human breast cancer cells. Cancer Res., 47:2103-2106, 1987.

36. Scotto, K. W., Biedler, J. L., and Melera, P. W. Amplification and expressionof genes associated with multidrug resistance in mammalian cells. Science(Wash. DC), 232: 751-755, 1986.

37. Ueda, K., Clark, D. P., Chen, C., Roninson, I. B., Gottesman, M. M., andPastan, I. The human multidrug resistance (marl) gene. cDNA cloning andtranscription initiation. J. Biol. Chem., 262: 505-508, 1987.

38. Ling, V., Kartner, N., Sudo, T., Siminovitch, L., and Riordan, J. R. Themultidrug resistance phenotype in Chinese hamster ovary cells. Cancer Treat.Rep., 67:869-874, 1983.

39. Bell, D. R., Trent, J. M., Willard, H. F., Riordan, J. R., and Ling, V.Chromosomal location of human P-glycoprotein gene sequences. CancerGenet. Cytogenet., 25: 141-148, 1987.

40. Van der Bliek, A. M., Baas, F., Van der Velde-Koerts, T., Biedler, J. L.,Meyers, M. B., Ozols, R. F., Hamilton, T. C., Joenje, H., and Borst, P.Genes amplified and overexpressed in human multidrug-resistant cell lines.Cancer Res., 48: 5927-5932, 1988.

2796

Research. on November 18, 2020. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from

1989;49:2790-2796. Cancer Res   Grace Bradley, Michael Naik and Victor Ling  Ovarian Carcinoma Cell LinesP-Glycoprotein Expression in Multidrug-resistant Human

  Updated version

  http://cancerres.aacrjournals.org/content/49/10/2790

Access the most recent version of this article at:

   

   

   

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

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

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

  Permissions

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

.http://cancerres.aacrjournals.org/content/49/10/2790To request permission to re-use all or part of this article, use this link

Research. on November 18, 2020. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from