[CANÅ’R Kl-SHAKOI 55. 4(I7.V4(I7S. September 15. IW5|

Transduction of NIH 3T3 Cells with a Retrovirus Carrying Both Human MDRland Glutathione 5-Transferase TTProduces Broad-Range Multidrug Resistance1

James H. Doroshow, Marianne Z. Metz, Linda Matsumoto, Katherine A. Winters, Masaharu Sakai,Musami Muramatsu, and Susan E. Kane2

Departments of Medical Oncology ami Therapeutics Research ¡J.H. l)., L M., K. A. W.I and Cell anil Tumor Biology ¡M.Z. M., S. K. K.¡,City of Hope National Medical Center,Dunne, Californio VIO10: Department of Biocheinislr\, Hokkaido Università' School of Medicine, N15. W7, Sapporo 060. Japan ¡M.S.1; and Department of HiochcniiMn;

Sidliima Medical School. .f<VMorolion^o. Morovama. Intma-^nn 350-04. Sailanta, Japan ¡M.M.J

ABSTRACT

In these experiments, we examined the ability of a retroviral vector,pHaMASV, to encode two potential chemoprotective genes on separatetranscription units. We previously described the pHaMSV vector, whichincludes the human MURI gene as a selectable marker and chemoprotective gene, plus an internal SV40 promoter for expressing a second hcter-

ologous gene along with MDRl [M. E. Met/, D. M. Best, and S. E. Kane.Virology, 20H: 634-643, 1995]. To test the ability of this vector to deliver

two therapeutic genes simultaneously, the cDNA for human glutathione.S'-transferase IT iGST.-r, the most abundant member of the glutathionc.S'-transferase family in human tumor cells) was inserted into pHaMASV,

and this plasmiti was transfected into ecotropic packaging cells. Theresulting pHaMASV.GSTir ecotropic retrovirus, which was produced at atiter of 2 x 10'*colony-forming units/ml, was used to transduce NIH 3T3

cells. After initial selection in 60 ng/ml colchicine, a population of transduced cells was exposed It)stepwise increasing colchicine concentrations toselect for amplified expression of MDRl. As MDRl expression increased,the expression of GSTw increased in concert, as demonstrated by Northern analysis. Western analysis, and measurements of glutathione .S'-trans

ferase activity. Transduced cells growing in 1280 ng/ml colchicine hadabout 3-fold higher total glutathione .S'-transferase activity than nontrans-

duccd cells and 2.5-fold higher activity than transduced cells growing in 60ng/ml colchicine. Northern hybridizations demonstrated a 3-5-fold increase in both the full-length retroviral message encoding MDRl and the

subgenomic mKNA encoding ( ÃŒSIrr after amplification of resistance from60 to I2KO ng/ml colchicine. The cytotoxic effects of several xenobioticswere evaluated in NIH 3T3 cells transfected with MDRl (3T3.MDR) ortransduced with the MDRl-GSTti retrovirus (3T3.GST640 or

3T3.GST1280) to evaluate the ability of our vector to produce a spectrumof drug resistances specific for the genes expressed. 3T3.MDR and3T3.GST1280 cells expressing equivalent levels of MDRl had identicallevels of resistance to doxorubicin or colchicine. These results suggest thatGST7T expression did not contribute to doxorubicin resistance in thismodel system. However, 3T3.GST640 cells were about 4-fold resistant toethacrynic acid and l-chloro-2,4-dinitrobenzene compared to cells ex

pressing MDRl alone, consistent with the ability of GST IT to conjugateboth of these cytotoxins. Increases in drug resistance paralleled increasesin gene-specific mRNA and rccombinant protein levels in all cases. Thus,

our studies suggest that the amplifiahle coexpression of MDRl plus asecond potentially therapeutic gene of interest is a feasible strategy for thedelivery of multiple drug resistance genes to normal cells for protectionagainst the toxic side effects of combination chemotherapy.

INTRODUCTION

Despite the availability of hematopoictic growth factors, hcmuto-

logical toxicities associated with combination chemotherapy remain asignificant therapeutic challenge to the application of dose-intensive

treatment regimens (1, 2), including those that utilize autologous bone

Received 5/27W; accepled 7/13/45.The costs (if publication of this article were defrayed in part hy the payment of page

charges. This article must therefore he hereby marked advertisement in accordance withIS U.S.C. Section 1734 solely to indicate Ihis fact.

1Supported by National Cancer Institute Cirants CA5')3()8 and CA33572.2 To whom requests for reprints should be addressed, at Department of Cell and Tumor

Biology. City of Hope National Medical Center. I5IXI I-ast Duarte Road. Duartc, CA

91010.

marrow or peripheral blood progenitor cell support (3). Consideringthe importance of dose in the therapeutic outcome of most chemo-

thcrapeutic programs (4, 5), clinically applicable strategies to diminish the toxicity of cancer chemotherapy for the bone marrow beyondthose possible by transplantation or treatment with bone marrowcytokines could substantially enhance the therapeutic index of currently available cytotoxic agents.

Several recent studies have focused on using the MDRl gene toprotect hematopoietic stem cells (6, 7). The human MOKI gene,which encodes the P-glycoprotein multidrug transporter, confers re

sistance to a wide variety of structurally dissimilar chemotherapeuticagents including taxol, etoposidc, and the anthracycline antibiotics(8). Whereas recent experiments have demonstrated low levels ofP-glycoprotein in some populations of early bone marrow progenitors

(9), this degree of expression does not persist in more highly committed cells and is insufficient to protect most hematopoietic precursors from drug-induced toxicity. Studies with A//)/?/-transgenic mice

(that constitutively express high levels of MDRl in bone marrowelements) and mice reconstituted with MDRl -transduced bone mar

row support the notion that MDRl expression can protect mice fromthe myelosuppressive side effects of drugs participating in the multi-

drug resistance phenotype (7, 10, 11). A similar approach has provedsuccessful in studies of the dihydrofolate reducíasegene, which hasbeen shown to confer resistance to methotrexate after retroviral trans-

duction into murine bone marrow (12, 13).Murine bone marrow reconstitution experiments suggest further that it

is possible to select for MDRl in vivo to enrich for a minor population ofstem cells transduced with MDRl within a background of nontransducedcells (10, 11). These findings are consistent with the activity oíMDRl invitro as a selectable and amplifiable marker in gene transfcction experiments (14, 15), and suggest that MDRl might also be useful as aselectable marker in the clinic for the expression of genes with potentialtherapeutic use that are not themselves selectable. Because there aremany classes of chemotherapeutic agents that are not substrates for theP-glycoprotein efflux pump (including the alkylating agents and the

antimctabolites), an important expansion of studies with MDRl in thebone marrow would be the combination of one or more drug resistancegenes with MDRl for the protection of hematopoietic precursors againstthe toxic effects of several families of anticancer drugs.

One such class of drug resistance genes is the glutathione transferasefamily (16). Members of this family of cnzmcs (including the TT.«,andju, isoenzymes) arc found ubiquitously in human tumor cells (17) andhave been shown to contribute to the resistance of these cells to a varietyof xenobiotics as well as commonly used chemotherapeutic agents (18,19). Thus, in the experiments presented here, we chose to evaluate therole of GST-ny1 the most abundant glutathione transferase in human

tumors, in concert with MDRl in producing xenobiotic resistance.Whether for chemoprotection with multiple drug resistance genes

or for the coexpression of selectable and nonselectable therapeutic genes.

1The abbreviations used are: GST-rr. glutathione .S'-transferase TT:MUR. multidrugresistance; nt. nucleotide; LTR. long terminal repeat; C'DNH. l-chloro-2.4-dinilrohen«:nc;

RA. ethacrynic acid.

4073

on July 31, 2021. © 1995 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

TWO-GENE TRANSDUCTION WITH MIÃŒRI

it is necessary to package MDR1 with other gene(s) of interest into asingle delivery vehicle suitable for gene therapy. In early experimentswith MDRI, this was accomplished by creating a retrovirus encoding afusion gene between MDRI and the ADA gene (20). Cells transduced inculture with the MDR-ADA virus express a chimeric gene product that is

localized to the plasma membrane, confers multidrug resistance, and hasADA activity. This retrovirus has the potential for selection in vivo (byvirtue of MDRI expression) as well as the therapeutic potential to correctthe metabolic defect caused by ADA deficiency. However, the use of afusion gene including MDRI prevents independent regulation of the twogenes and might dictate the (»localizationof the fused gene products tothe plasma membrane, potentially limiting the function of the non-MDRIgene product. We recently reported a modified HaMSV-based retroviral

vector, pHaMASV, which encodes MDRI plus a second, heterologousgene regulated by an internal SV40 promoter (21). Using the humanMDRI and GST-rr genes inserted into pHaMASV, we now demonstrate

the simultaneous transfer, expression, and activity of two drug resistancegenes in NIH 3T3 cells. These studies suggest that the coexpression ofMDRI with a second therapeutic gene of interest is a feasible strategy fora variety of approaches to the gene therapy of cancer in the clinic.

MATERIALS AND METHODS

Cell Lines. NIH 3T3 cells were maintained in DMEM containing 10%bovine serum (Irvine Scientific, Irvine, CA), 5 mM glutamine, 50 units/ml ofpenicillin, and 50 /xg/ml of streptomycin at 37°Cin a humidified atmosphere

of 5% CO2 in air. The 3T3.MDR cell line, which has been described previously(15), was developed by transfection of NIH 3T3 cells with the MDRI gene;these cells were maintained in (issue culture medium containing either 640 or1280 ng/ml of colchicine (Sigma Chemical Co., St. Louis, MO).

Construction of pHaMASV.GSTn. Vector pHaMASV.GSTir was derived from pHaMASV, the construction of which is described elsewhere (21).Cloning of the human GSTir cDNA has been described previously (22). Toobtain an appropriate fragment of GSTîrfor cloning into pHaMASV, PCR wasperformed on the GSTir template using manufacturer's recommended reaction

conditions for Taq\ polymerase (Perkin Elmer Cetus), 200 /J.Mof each dNTP.and 30 cycles at 94°C(1 min), 42°C(I min), and 72°C(30 sec) for each cycle.

The upstream PCR primer spanned nt -6 to +6, and the downstream primerwas complementary It) nt 673—684just upstream of the GSTir polyadenylation

signal (numbering according lo Kano et al. (22)]. Sail restriction sites wereincluded at the 5' ends of each of these primers. The GSTTT PCR product was

digested with Sail, purified, and ligated into the Sail site of pHaMASV. Theresulting plasmid, designated pHaMASV.GSTir, is shown in Fig. 1.

Retrovirus Production and Transduction of NIH 3T3 Cells. PlasmidpHaMASV.GSTir was transfcctcd into GP + E86 ecotropic packaging cellsthat were obtained from Dr. A. Bank (23). Cells were selected in the presenceof 20 ng/ml colchicine: drug-resistant cells were pooled and grown to 80%

confluence. Fresh medium lacking colchicine was added to the cells, and viruswas collected after 20-24 h. Virus titer, assayed on NIH 3T3 cells, wasdetermined to be 2 X 10'' colony-forming units/ml. For transduction of NIH3T3 cells, 5 x It)5 cells were plated in 35-mm dishes on day 0. On day 1,

1 ml of virus was added to the cells in the presence of 8 fig/ml polybrene.On day 3, 60 ng/ml colchicine was added to select for transduced cells.Colchicine-resistant cells were pooled, and drug resistance was amplified by

growing cells in increasing concentrations of colchicine as described previously (14, 15). Amplification was continued until cells were growing in640 or 1280 ng/ml of colchicine.

Northern Analysis. For Northern analysis, total RNA was extracted fromcell lines with RNAzol B (Cinna/Biotecx Laboratories International) accordingto manufacturer's instructions. Ten /ig of each RNA were loaded onto dupli

cate lanes of a 1% agarose-formaldehyde gel; separated RNAs were transferred

to a Hybond N+ nylon membrane (Amcrsham). Filters were hybridized witheither an A/DA/-specific probe spanning nt 1919-2188 (where +1 is at theinitiating ATG) or the full-length GSTir-specific Sail fragment from pHaMASV. GST7T. Hybridizations were performed in Rapid-hyb buffer (Amer-sham) at 65°Covernight using 2 X IO5 cpm/ml of random primer-labeled

probe. Filters were washed according to a standard procedure with the final

wash at 0.1 X SSC plus 0.1% SDS (1 x SSC is 150 mM NaCl-15 mM sodiumcitrate, pH 7.0) at 65°Cfor 20 min. Dried filters were exposed to XAR-5 filmat -8()°Cwith an intensifying screen. For quantification, filters were exposed

to a Molecular Dynamics Phosphorlmager screen for 72 h and analyzed withthe Image Ouant software. Band intensities were normalized to the intensities

of the endogenous GSTir mRNAs in each cell line to control for different

amounts of RNA loaded in the gel lanes.Flow Cytometry. Flow cytomctry was performed on live cells as described

elsewhere (21). Primary antibody was mAb 4E3 (Signet Laboratories, Inc.;Ref. 24), which detects an external epitope of the MDRI gene product.Secondary antibody was fluorcscein-conjugated goat antimouse IgG (H + L)F(ab')2 (Jackson ImmunoResearch).

Western Analysis. Cells were lysed in 50 mM Tris (pH 8.0)-5 mM EDTA-150 mM NaCl-0.5% NP40-1 mM phenylmethylsulfonyl fluoride by sonication

on ice. Protein concentrations were determined using the BioRad DC proteinassay. One hundred /xg of protein from each lysate were run on 12% SDS-

PAGE then electroblotted onto Immobilon P polyvinylidene difluoride membrane (Millipore). The filter was blocked overnight with 10% Carnation driedmilk in PBS containing 0.1% Tween 20 and then incubated for 1 h with aGSTw-specific polyclonal antibody (BioGenex) diluted 1:100 in PBS, 0.1%

Tween 20, and 1% BSA. Hybridized antibody was visualized using proteinA-peroxidase (Boehringer Mannheim) and the enhanced chemiluminescence

Western detection system from Amersham. As a loading control, filters werestripped and subsequently reacted with mAb directed against actin (data notshown). Hybridization signals were quantified using a GS-250 Molecular

Immager (BioRad) and Phosphor Analyst/PC software.Enzyme Activities in Transduced Cells. Enzyme activities were deter

mined from cytosolic preparations of NIH 3T3 cells, 3T3.MDR cells, and apopulation of NIH 3T3 cells transduced with retroviral supernatants, selected for

colchicine resistance, and grown to 1280 ng/ml colchicine resistance. Glutalhioneand glutathione reducíasewere purchased from Sigma; bovine liver catalasc(devoid of Superoxide dismutase activity) was obtained from Boehringer Mann

heim (Indianapolis, IN). To determine enzymatic activities, cells in logarithmicphase growth were harvested with trypsin-EDTA typically 4 days after plating.Approximately 6 x 10'' cells were harvested for each experiment. The cells were

washed twice in PBS (pH 7.4). resuspended in 1 ml of 0.25 M sucrose buffercontaining 5 mM HEPES (pH 7.4), and then disrupted at 4°Cwith 20-25 bursts of

1-2 s each with a Branson ultrasonicator. The cytosolic fraction from the cells wasseparated by centrifugation of the sonicates for 20 min at 1(X),000X g at 4°C.The

resulting supernatants were assayed for enzymatic activity. GST, glutathioneperoxidase (with hydrogen peroxide as substrate), glutathione reducíase,glucose-

6-phosphate dehydrogenase, and catalase activities were determined by established

methods as described previously (25). Glutathione levels were determined as totalnonprotein thiols using —1x IO7 cells as reported previously (26, 27). Protein

concentrations of the cytosols were determined by the method of Lowry et al. (28).

Cytotoxicity Assays. Clonogenic cytotoxicity assays were performed using logarithmically growing cells that had been taken out of colchicine for atleast one passage; the drug of interest was added to tissue culture flasks for 1h. Colchicine, CDNB, and EA were obtained from Sigma. Doxorubicin waspurchased from Adria Laboratories (Columbus, OH), cisplatinum was fromBristol-Myers Squibb Co. (Princeton, NJ), and chlorambucil was obtained

from Burroughs Wellcome Co. (Research Triangle Park, NC). Clinical formulations were used for all cancer chemotherapeutic agents. Cells were harvestedwith trypsin-EDTA, washed with PBS, and 2500 cells were plated in

60 X 15-mm dishes, in triplicate. Colonies of >50 cells were counted afterincubation for 8-10 days at 37°Cin a humidified atmosphere of 5% C02 in air.

The percent clonogenic survival has been expressed as the number of coloniesproduced by drug-treated cells divided by the number of colonies produced by

untreated control cells x 100. Assays were performed on nontransduced NIH

3T3 cells, on 3T3.MDR cells carrying only the MDRI gene, and on3T3.GST640 or 3T3.GST1280 cells carrying transduced MDRI plus GSTirgenes. 3T3.GST1280 cells expressed about 45% more functional MDRImRNA than 3T3.GST640 cells, and the two cell lines had similar GSTirmRNA and protein levels. The doubling times and plating efficiencies (15-

25%) of these cell lines did not differ significantly.Statistical Analyses. Cytotoxicity assays were analyzed using Student's

two-tailed r test for independent means; significance required P < 0.05 (29).

4074

on July 31, 2021. © 1995 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

TWO-OENE TRASSI)! l HON Willi MURI

RESULTS

Retroviral Transduction of NIH 3T3 Cells. A map of the retro-

viral vector used in this study is shown in Fig. 1. Briefly, the vectorcontains LTRs and a viral packaging signal derived from Harveymurine sarcoma virus (30); the human MDR1 cDNA is under thecontrol of the upstream LTR, and the SV40 early promoter andenhancer drive the expression of the human GSTTr cDNA. The plas-

mid is designated pHaMASV.GSTîr.The pHaMASV.GSTîrplasmidwas used to produce ecotropic retrovirus as described in "Materialsand Methods," virus was used to transduce NIH 3T3 cells at a

multiplicity of infection of 1, and transduced cells were selected with60 ng/ml of colchicine to select for cells expressing the MORI gene.A pooled population of drug resistant cells was grown in stepwiseincreasing concentrations of colchicine, up to 640 or 1280 ng/ml, toselect for amplified MORI expression (14, 15).

To measure the extent of amplified MORI expression in transducedcells, we used flow cytometry to detect P-glycoprotein on the cell surfaceof those cells. Results arc shown in Fig. 2. P-glycoprotein expression

increased with increasing selective pressure up to 640 ng/ml of colchicine, with an additional slight increase at 1280 ng/ml. Furthermore, thelevel of P-glycoprotein on 3T3.GST1280 cells was comparable to that on

control 3T3.MDR cells that had been transfected previously with just theMDRI gene and grown in 1280 ng/ml colchicine (15).

Enzyme Activities and Gene Expression in Transduced Cells.To determine whether the GST ITgene was transduced along with MDRI,we measured total GST activity in control and transduced NIH 3T3 cellsas well as in the 3T3.MDR line. The results are shown in Fig. 3 and Table1. The population of A/D/?/-GST-n--transduced cells had about 2.5-fold

higher total GST activity (assayed with CDNB) than either nontrans-

duced NIH 3T3 cells or cells transfected with MDRI and independentlyselected for resistance to 1280 ng/ml of colchicine (P < 0.02). Furthermore, the level of GST increased progressively with selection in higherconcentrations of colchicine (Fig. 3). Whereas the total GST activityincreased in concert with the degree of colchicine resistance, the activitiesof other glutathione-dependent enzymes and of glutathione itself were

unchanged in response to selection in colchicine, even at the highest levelof colchicine resistance (Table 1). The level of glucose-6-phosphate

dehydrogenase, however, was lower in cells transfected with MDRI thanin control NIH 3T3 cells or in transduced cells. The explanation for thisresult is not clear.

To demonstrate that the elevated GST activity was derived from thetransduced GSTir gene, we performed Northern analysis on totalRNA extracted from these cell lines. As illustrated in Fig. 4, with anMDR/-specific probe (sec "Materials and Methods" and Fig. 1), we

detected the expected full-length mRNA of about 7000 nt, transcribed

from the upstream LTR and terminating in the downstream LTR (asdiagrammed in Fig. 1). This message was present in all cell linestransduced with pHaMASV.GST-n- but not in nontransduced NIH 3T3

cells. With a GSTîr-specificprobe and pHaMASV.GSTîr-transducedcells, we detected both the full-length message, which includes GSTTr

sequences, and a subgenomic mRNA of about 1500 nt that was

GST 1280

10'

B900

800

700

, 600O^ 500"H>O 400

300

200

100

NIH3T3

GST 1280

\10° IO1

Fluorescence Intensity

Fig. 2. Flow cytometry. Flow cytometry was performed as described in "Materials andMethods." Primary antibody was 4E3 mAb (24) that detects an external epitope of

P-glycoprotcin, the MDRI gene product. A, cell surface expression of P-glycoprotein in

nontransduced NIH 3T3 cells and in pHaMASV.GSTTMransduced cells growing at 60,640. or 1280 ng/ml colchicine, as indicated. B. expression levels in cells transfected withthe MDRI gene (3T3.MDR) and transduced 3T3.GST cells, both growing at 1280 ng/mlcolchicine.

Table 1 ¿'iicvmcactivities in noniransditceLl NIH 3T3 cells ami in 3T3.MDR ami

3T3.GST1280 cells selected for resistance to 1280 n%/tnl calchicineEnzyme activities were determined from cytosolic preparations as outlined in "Mate

rials and Methods"; the data represent the results of at least 3 separate experiments (except

glucose 6-phosphate dehydrogenase measurements in NIH 3T3 cells, which were deter

mined once) and are expressed as mean ±SE.

EnzymeactivityGlutathionei'-lransferase(nmol/min/mg)Catalase

(/xmol/min/mg)Glucose-6-phosphatedehydrogenase(nmol/min/mg)Glutathione

peroxidase(nmol/min/mg)Glutalhioncreducíase(nmol/min/mg)Glutathione(nmol/106 cells)NIH

3T345

±72.58±0.1253."3.8

±1.69.5±0.57.4+ 0.63T3.MDR51

±92.34±0.1930.7±1.31.8

+0.212.9+3.38.6

±0.63T3.GST1280115

+11°2.51

±0.2354.8±2.9*1.8

+0.212.2+2.310.9

±1.7" P < 0.02 versus 3T3.MDR and NIH 3T3.* P < 0.02 versus 3T3.MDR.

Ho-llRMDR/A-fsv]| GST-ir|_.Ho-LTR3SOnt

Fig. 1. Linear representation of pHaMASV.GSTir. Vector pHaMASV.GSTir wasderived from pHaMASV. Expected transcripts are shown as dashed lines below the map.MDRI is encoded by the full-length message expressed from the upstream HaMSV LTR(Ha-LTR). MDR/A, MDRI thai lacks its own polyadenylation signal. GSTir is translated

from the suhgenomic mRNA derived from the internal SV40 promoter (5V). Black bars,probes used in Northern hybridizations.

transcribed from the internal SV40 promoter and terminated in thedownstream LTR. Endogenous GST-Tr mRNA was detected as anapproximately 800-nt band in nontransduced and transduced cell

lines. Phosphorlmager quantification of the Northern hybridizationindicated that there was a 5-fold increase in the full-length messageand a nearly 3-fold increase in the subgenomic message after ampli

fication from 60 to 1280 ng/ml colchicine resistance. There was nearly4-fold more of the transduced, subgenomic GSTTT mRNA than of

endogenous GST77 mRNA in the cell line growing in 1280 ng/mlcolchicine.

4075

on July 31, 2021. © 1995 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

TWO-GENE TRANSDUCTION WITH Mimi

130120

S 110co 100

- -H 90

fso„ 70-| 60-| 50o 40| 30

S 20

100

«P<0.02vs3T3or3T3NDR**P<0.02vs3T3

illNIH3T3 GST-60 GST-1280

3T3HDR1280CST-640

Fig. 3. GST-specific activities in cytusnlic preparations of nonlransduccd NIH 3T3cells. NIH 3T3 cells transfected with MDRI and growing in 1280 ng/nil colchicine(3T3MDRI2m¡. and NIH 3T3 cells transduced with the pHaMASV.GST-rr ecolropic

retrovirus and growing in increasing (M), MO. and 12X0 ng/ml) concentrations of colchicine (GST-M), GST-MO, and GST-I2W).

MDR GST*

— MDRl

Fig. 4. Northern analysis. Northern hybridizations were performed on total RNA asdescribed in "Materials and Methods." Cell lines are nontransduced NIH 3T3 cells (3T3)

or cells transduced with pHaMASV.GSTTT virus and grown in 60, 640. or 1280 ng/mlcolchicine. as labeled. Membranes were hybridi/ed with MOKI or GSTir-specific probes,as indicated. Probes are described in "Materials and Methods" and shown in Fig. I.

Arntws. full-length relroviral message (MDRI) and subgenomic message expressed fromthe internal SV40 promoter (GST7T) as illustrated in Fig. 1. *. position of endogenous

mouse GSTTr mRNA expressed by all four cell lines analyzed.

To confirm that transduced cells expressed elevated GST-rr protein,

we performed Western analysis. Results arc shown in Fig. 5. Consistent with the Northern analysis, transduced cells growing at 60ng/ml colchicine contained marginally increased GST-Tr relative tonontransduced NIH 3T3 cells, but there was 3.6-fold enhanced ex

pression in cells grown in 1280 ng/ml colchicine; there was littledifference of GSTTT expression between 3T3.GST1280 and3T3.GST640 cells. Changes in GST-Trprotein paralleled changes inP-glycoprotein expression (compare Figs. 2 and 5).

Relative Resistance of Transduced Cells to Xenobiotics. Thegoal of these experiments was to determine whether transduction andcoexpression of two drug resistance genes produced cell lines thatwere resistant to a broad range of cytotoxic agents. To test thisconcept, we compared the toxicity of a variety of xenobiotics in ourWD/f/-GST-n--transduced cell lines (3T3.GST640 or 3T3.GST1280)

to that in nontransduced NIH 3T3 cells or in 3T3.MDR cells expressing MURI alone. When adapted to growth in 1280 ng/ml colchicine,3T3.MDR cells and the MD/?/-GSTTT-transduced line expressed comparable amounts of P-glycoprotein on their cell surface (Fig. 2) and

had equivalent plating efficiencies and growth rates (data not shown).As illustrated in Fig. 6 A and B, both these cell lines were ~ 100-fold

resistant to doxorubicin and about 10-fold resistant to colchicine by

comparison of IC5(, versus nontransduced NIH 3T3 cells, consistentwith the expression of the MDRI gene product in both lines. Therewas no apparent difference in the toxicity of colchicine or doxorubicin

between the transfected and transduced cells (Fig. 6 A and B). Furthermore, there was no effect of enhanced GSTir levels on thecytotoxicity of doxorubicin or cisplatinum (Fig. 6Q. These results areof interest in light of previous studies suggesting that GSTTToverex-

pression contributes to doxorubicin or cisplatinum resistance in tumorcell lines selected in vitro (31-34).

On the other hand, the A/ß/?/-GSTTT-transducedline (compared toNIH 3T3 and 3T3.MDR cells) was ~4-5-fold resistant to EA or CDNB

(Fig. 7 A and B). These findings are compatible with the relative levels ofGST activity and GSTTTprotein in these two cell lines and with the levelof transduced GSTTTmRNA compared to the endogenous message in thetransduced cells. As shown in Fig. 1C, enhanced expression of GSTTr hadno effect on the toxicity of chlorambucil.

DISCUSSION

Gene therapy with MDRI has been proposed as a means of protectingpatients against myelosuppression produced by dose-intensive chemo

therapy (10, 11). The feasibility of this concept has been demonstrated intransgcnic mice (8) and in mice reconstituted with AfD/?/-transduced

bone marrow cells (10, 11). Other drug resistance genes arc beingexplored for this same purpose. In particular, a modified version of theDHFR gene product has been shown to protect transplanted mice fromlethal doses of methotrexate (12, 13). These approaches aim to exploit thepotential of drug resistance genes to provide genetic selection for transduced hematopoietic cells in vivo, allowing them to survive under conditions which normally destroy their nontransduced counterparts. Anextension of this application would be to deliver two or more drugresistance genes simultaneously so as to protect transduced cells from awide variety of toxic agents. In addition, it should be possible to use thein vivo selectability of MDRI to deliver a second therapeutic, but non-

selectable, gene to appropriate target cells and sustain the expression ofthe therapeutic gene by maintaining selective pressure on transduced cellsvia the MDRI gene.

MDR] Retroviruses. Previous work with MDRI has used thepHaMDR/A vector for retrovirus production (10, 11, 35). A similarvector has proven capable in DNA-mediated transfcctions of cultured

cell lines of mediating expression and amplification of MDRI underthe selective pressure of high cytotoxic drug concentrations (14, 15).MDRI retrovirus encoded by pHaMDR/A has also been used inmouse bone marrow transduction and reconstitution experiments,during which A/D/i/-transduced marrow is protected from the myelo-

suppressive effects of taxol (10, 11). In addition, the vector was usedto create a fusion gene encoding a chimeric, bifunctional MDRI-ADA

gene product (20). Finally, we have recently described a modified,reduced-size version of pHaMDR/A that is capable of encoding

MDRI plus a second gene under the independent control of aninternal SV40 promoter (21). In the current work, we created

toro

31 -

25 -

18 -

Fig. 5. Western analysis. One hundred fig of protein from whole cell lysates of NIH3T3 cells (3T3). transfected 3T3.MDR cells growing at 1280 ng/ml colchicine(3TS.MDR), and transduced cells growning at 60. 640. or 1280 ng/ml colchicine (GSTfiO,GSTMO, and GSTI2XÜ,respectively) were run on 129ÕSDS-PAGE. Western hybridiza

tion was performed with polyclonal antibody specific for the GSTTr iso/yme and visualized as described in "Materials and Methods."

4076

on July 31, 2021. © 1995 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

TWO-GENE TRANSDUCTION WITH MIÃŒRI

Fig. A. Clonogcnic survival in NIH 3T3 (3T3.dashed lines), 3T3.MDR (solid lines), and3T3.GSTI280 (doited lines) cells after a l-h exposure to doxoruhicin ¡Ai,colchicinc (#J, or cisplati-num (C) was determined in triplicate.

00 05 1.0 li 2.0!5 K

Doxorubicin

O 10 20 30 40 50 60 70 80 90 100

Colchicine (u.a/mlì

»3T3•3T3MDR•3T3GST1280

pHaMASV.GSTTT to demonstrate efficient retrovirus-mediated

transduction and expression of two potentially therapeutic genes inNIH 3T3 cells.

MDRl and GSTw Coselection in Transduced Cells. NIH 3T3cells transduced with the pHaMASV.GSTTr retrovirus were efficientlyselected with colchicine, and drug resistance was readily amplified byexposing cells to increasing concentrations of this drug. As drugresistance increased, expression of both MDRl and the transducedGSTîrincreased, as demonstrated by Northern analysis, protein (flowcytometry and Western) analysis, and enzyme activity measurements.The ability to obtain high GST-Tr expression by selecting for high

colchicine drug resistance mimics the previously reported role ofMDRl as a selectable and amplifiable marker (14, 15). These resultssuggest that if MDRl is selectable and amplifiable in viro, it will alsoserve as a means for achieving elevated, long-term expression of

therapeutic sequences in gene therapy.The levels of total GST activity and GSTir expression in our

transduced cells are comparable to or exceed those reported instudies using GSTir transfections to investigate mechanisms ofdrug resistance (18, 36), suggesting that expression from the pHa-

MASV.GSTir vector is good. Nevertheless, the levels of GSTirobtained are relatively modest compared with heterol ogous geneexpression achieved in previous work with the MD/f/-based trans-

fcction, selection, and amplification system (14, 15). It is possible,though speculative, that very high levels of GST activity do notcarry a survival advantage and that high expressers do not survivethe colchicine amplification process. It is also possible that amplification of transduced copies of MDRl is not as efficient asamplification of transfected, tandemly integrated copies of MDRl.This issue has not yet been explored in the literature, but it couldhave an impact on how gene therapy with in vivo selectablemarkers will proceed in the future.

Transduction of Multiple Drug Resistance Genes. With the advent of gene therapy, it has become theoretically possible to deliverdrug resistance genes to normal cells for protection against the toxicside effects of chemotherapy. The MDRl gene can confer resistance

to a wide variety of chemotherapeutic agents including taxol andetoposide, which have dose-limiting bone marrow toxicities. How

ever, there are a number of common drugs, in particular the alkylatingagents and antimetabolites, that are not substrates for transport byP-glycoprotein. Additional gene products will be needed if marrow

cells are to be made resistant to a broad range of potential chemotherapeutic agents. The advantage of this approach to gene therapy isthat it is not necessary for the therapeutic drug resistance genes to beroutinely responsible for multidrug resistance in tumors. It is sufficient that these genes are able to act as dominant genetic factorscapable of conferring resistance to the chemotherapeutic drugs ofinterest in normal host tissues.

We have used the human MDRl and GST-rr genes as a model for

this approach and have obtained transduced cells which are simultaneously resistant to both MDRl- and GSTTT-related drugs. Forthe AiD/?/-related drugs doxorubicin and colchicine, our transduced cells were ~ 100-fold and about 10-fold resistant, respec

tively, relative to nontransduced NIH 3T3 cells. Although somelaboratories have found low levels of resistance to doxorubicinafter transfection of GSTTr into suitable host cells (19, 37), ourexperiments are consistent with those of Fairchild et al. (36), whofailed to demonstrate alterations in doxorubicin cytotoxicity forMCF-7 human breast cancer cells overexpressing GSTTr. Although

GSTTr levels are frequently elevated in tumor tissue, the absence ofconvincing data demonstrating that doxorubicin is a substrate forGSTir-mediated conjugation reactions probably explains the re

sults presented in Fig. 6.4.On the other hand, the toxicity of EA and CDNB for NIH 3T3

cells was modulated by GST-rr expression (Fig. 7 A and B). Be

cause EA and CDNB are efficiently conjugated by GSTTr (38, 39),and because tumor cells selected for resistance to EA demonstrateincreased GSTTr mRNA at a level consistent with the degree ofresistance (40), it is not surprising that our 3T3.GST640 cells were4-5-fold resistant to EA and CDNB. Because GST-n- is quite

inefficient at utilizing chlorambucil as a substrate (19), the similar

Fig. 7. Clonogenic survival in NIH 3T3 f.?7.?.dashed lines), 3T3.MDR (solid lines), and3T3.GST640 (¡hncil lines) cells after a l-h exposure to EA (A). CDNB <B>. or chlorambucil <C>was determined in triplicate.

P< 0.05•3T3MOR•3T3GST640»3T3

100 200

Ethocrynic Acid

80

60

100 f.

0.0

i P<005i 3T3MDRi 3T3GST640•3T3

25 5.0 75

CDNB (MM)

•3T3MDR•3T3GST640«3T3

10 20 30

Chlorambucil (

4077

on July 31, 2021. © 1995 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

TWO-GENK TRANSDUCTION WITH MDkl

cytotoxicity profile for this agent in N1H 3T3, 3T3.MDR and3T3.GST640 cells was also expected.

Other drug resistance or drug-metabolizing genes are potential

candidates for use in gene therapy. These include the peroxide andalkylating agent detoxifying genes glutathione peroxidase and GST a;untimetabolite target genes such as thymidylate synthase and DHFR;and a variety of DNA repair genes. In concert with MDRI, theselective coexpression of several drug resistance-enhancing genes in

the hone marrow or other normal tissues has the potential to decreaseor eliminate much of the host toxicity associated with combinationchemotherapy. The experiments outlined herein provide a first steptoward the attainment of that goal.

ACKNOWLEDGMENTS

The authors lhank A. Bank and M. Gottesman for providing cell lines andplasmids. L. Brown and J. Bolen for assistance with the flow cytometry, andJ. Momand for advice on the Western analysis.

REFERENCES

1. Neidhart, J., Mangalik, A.. Köhler,W.. Slidley. C.. Saiki. J.. Duncan, P., Souza. L..and Downing. M. Grunulocytc colony-stimulating factor stimulates recovery ofgranulocytes in palicnls receiving dose-intensive chemotherapy without hone marrowtransplantation. J. C'lin. Oncol., 7: 1685-1692. 1989.

2. Hryniuk. W. The importance of dose intensity in the outcome of chemotherapy. In:S. Hcllman. V. T. De Vita. Jr., and S. A. Rosenberg (eds.). Important Advances inOncology, pp. 121-141. Philadelphia: J. B. Lippincott Co.. 1988.

3. Sonilo. G.. Doroshow, J. H., Forman, S. J., Leong, L. A., Margolin, K. A.. Morgan.R. J.. Jr.. Raschko. J. W., Akman. S. A.. Ahn. C, Schmid!. G. M.. Brcnt, J.. Rincón.A-, and Nagasawa. S. High-dosc doxoruhicin. eloposide, and cyclophosphamide withautologous stem cell rcinfusion in patients with responsive metastatic or high-riskprimary breast cancer. Cancer (Phila.). 73: 1678-1685, 1994.

4. Frei. E. 111.and Candios. G. P. Dose: a critical factor in cancer chemotherapy. Am.J. Med., 69: 585-594, 1980.

5. Levin. L., and Hryniuk. W. M. Dose intensity analysis of chemotherapy regimens inovarian carcinoma. J. Clin. Oncol.. 5: 756-767. I'M?.

6. McLachlin, J. R.. Eglitis. M. A.. Ueda, K.. Kantoff. P. W., Pastan. I. H.. Anderson.W. F.. and Gottesman, M. M. Expression of a human complementary DNA for themullidrug resistance gene in murine hemalopoietic precursor cells with the use ofrelroviral gene transfer. J. Nati. Cancer Inst.. ,S2: 1260-1263, IWO.

7. MickiNch. (1. II.. l.ichl. T., Merlino, G. T., Cìottcsman.M. M., and Pastan, I.Chemotherapy and chemoscnsitization of transgcnic mice which express the humanmultidrug resistance gene in hone marrow: efficacy, potency, and toxicity. CancerRes., 51: 5417-5424. 1W1.

8. Kane. S. E.. Pastan. I., and Gottesman, M. M. Genetic basis of multidrug resistanceof tumor cells. J. Bioencrg. Biomembr.. 22: 593-618, 1990.

9. C'haudhary, P. M.. and Roninson. I. B. Expression and activity of P-giycoprotein. a

multidrug efflux pump, in human hematopoietic stem cells. Cell, 66: 85-94. 1991.

10. Sorrentino. B. P.. Brandt. S. J.. Bodine. D.. Gottesman, M.. Pastan, I.. Cline. A., andNienhuis. A. W. Selection of drug-resistant hone marrow cells in vim after retroviraltransfer of human MDRI. Science (Washington DC), 257.- 99-103, 1992.

11. Podda. S.. Ward. M., Himelstcin, A., Richardson. C.. DelaFlor-Weiss, E., Smith, L..Gottesman. M.. Pastan. I., and Bank. A. Transfer and expression of the humanmultiple drug resistance gene into live mice. Proc. Nail. Acad. Sci. USA. XV:9676-9680. IW2.

12. Williams, D. A.. Hsieh. K.. Desilva. A., and Mulligan. R. A. Protection of bonemarrow transplant recipients from lethal doses of mctholrcxatc by the generation ofmethotrexate resistant bone marrow. J. Exp. Med.. 166: 210-218. 1987.

13. Zhao, S-C., Li, M-X., Bancrjee. D., Schweitzer. B. I., Mineishi, S.. Gilboa. E., andBerlino. J. R. Long-term protection of recipient mice from lethal doses of melhotrex-ale by marrow infected with a double-copy vector retrovirus containing mutantdihvdrololatc reducíase.Cancer Gene Ther.. /: 27-33. 1994.

14. Kane. S. E.. and Gottesman, M. M. Use of the multidrug resistance gene in mammalian expression vectors. Methods Enzymol.. 2/7: 34-47, 1993.

15. Kane. S. E., Reinhard. D. H., Fordis, C. M.. Pastan, L, and Gottesman, M. M. Anew vector using the human multidrug resistance gene as a selectable markerenables overexpression of foreign genes in eukaryotic cells. Gene. 84: 439-446.

1989.

16. Mannervik. B.. and Danielson. U. H. Glutalhione transferases: structure and catalyticactivity. Crit. Rev. Biochem.. 23: 283-337, 1988.

17. Castro, V. M., Soderstrom. M.. Carlbcrg. I.. Widersten. M.. Platz, A., and Mannervik,B. Differences among human tumor cell lines in the expression of glutathionetransferases and other glutathione-linked enzymes. Carcinogenesis (Lond.), //:1569-1576, 1990.

18. Nakagawa. K.. Saijo. N.. Tsuchida. S.. Sakai, M.. Tsunokawa, Y., Yokota, J.,Muramalsu. M.. Sato, K., Terada, M.. and Tew. K. D. Glutathione .S'-transferase ITas

a determinant of drug resistance in transfeclant cell lines. J. Biol. Chem., 265:4296-4301, 1990.

19. Ciaccio. P. J.. Tew, K. D.. and LaCreta. F. P. Enzymatic conjugation of chloramhuci!with glutathione by human glutathione 5-transferases and inhibition by clhacrynicacid. Biochem. Pharmacol.. 42: 1504-1507, 1991.

20. Germann. U. A., Chin. K. V.. Pastan. I., and Gottesman, M. M. Retroviral transfer ofa chimeric multidrug resistance-adenosine deaminase gene. FASEB J.. 4: 1501-1507,1990.

21. Met/. M. E., Best, D. M., and Kane, S. E. Harvey murine sarcoma virus/MDRlretroviral vectors: efficient virus production and foreign gene transduction usingMDRI as a selectable marker. Virology. 208: 634-643. 1995.

22. Kano. T.. Sakai. M.. and Muramatsu. M. Structure and expression of a humanclass TIglutathione .S'-transferase messenger RNA. Cancer Res.. 47: 5626-5630,

1987.23. Markowilz. D.. Goff. S.. and Bank, A. A safe packaging line for gene transfer:

separating viral genes on two different plasmids. J. Virol.. 62: 1120-1124. 198S.

24. Arceci. R. J.. Stieglitz. K., Bras. J.. Schinkel. A., Baas, F.. and Croop. J. Monoclonalantibody to an external epitope of the human nutrìP-glycoprotein. Cancer Res.. 53:310-317. 1993.

25. Zwelling, L. A.. Slovak. M. L.. Doroshow. J. H.. Hinds, M.. Chan. D., Parker, E.,Mayes. J., Sic. K. L.. Meltzer, P. S., and Trent. J. M. HT1080/DR4: a P-glycoprolein-negative human fibrosarcoma cell line exhibiting resistance to topoisomerasc II-reactive drugs despite the presence of a drug-sensitive topoisomerase II. J. Nati.Cancer Inst.. X2: 1553-1561, 1990.

26. Sedlak. J.. and Lindsay. R. H. Estimation of total, protein-hound, and nonproteinsulllivdryl groups in tissue with Ellman's reagent. Anal. Biochem.. 25: 192-205.

1968.27. Doroshow, J. H.. Locker. G. Y., Ifrim. I., and Myers. C. E. Prevention of doxoruhicin

cardiac toxicity in the mouse by /V-acelylcysteinc. J. Clin. Invest., 6cS: 1053-1064.

1981.28. Lowry, O. H.. Rosebrough, N. J., Farr, A. L.. and Randall. R. Protein measurement

with the folin phenol reagent. J. Biol. Chem., IV3: 265-275, 1951.

29. Armitage, P. Statistical methods in medical research. Oxford, United Kingdom:Blackwell Scientific Publications. 1971.

30. Soeda, E., and Yasuda. S. Harvey murine sarcoma virus genome. In: R. Weiss, N.Teich. H. Varmus, and J. Coffin (eds.), RNA tumor viruses: molecular biology oftumor viruses, pp. 928-939. Cold Spring Harbor. NY: Cold Spring Harbor Labora

tory. 1985.31. Cole, S. P. C.. Downes, H. F.. Mirski, S. E. L., and Clements, D. J. Alterations in

glutathione and glutathione-related enzymes in a multidrug-resistant small celt lungcancer cell line. Mol. Pharmacol.. .Õ7:192-197. 1990.

32. Whelan, R. D. 11.. Ilosking, L. K.. Townscnd. A. J.. C'owan, K. II.. and Hill, B. T.Differential increases in glutathione .S'-transferase activities in a range of multidrug-

rcsistant human tumor cell lines. Cancer Commun.. /: 359-365, 1989.33. Meijer, C.. Mulder. N. H.. and de Vries, E. G. E. The role of detoxifying systems in

resistance of tumor cells to cisplatin and Adriamycin. Cancer Treat. Rev., 17:389-407, 1990.

34. Scheeler, R. L.. Woo. A.. Duong. M.. and Batist. Ci. In vim and in vilm mechanismsof drug resistance in a rat mammary carcinoma model. Cancer Res.. 51: 1434-1442.

1W1.35. Pastan, I., Gotlesman. M. M., Ueda, K., Lovelace, E., Rutherford. A. V., and

Willingham, M. A retrovirus carrying an MDRI cDNA confers multidrug resistanceand polarized expression of P-glycoprolcin in MDC'K cells. Proc. Nail. Acad. Sci.

USA, 85: 4486-4490. 1988.36. Fairchild. C. R.. Moscow. J. A.. O'Brien. E. E., and Cowan, K. H. Mullidrug

resistance in cells transfected with human genes encoding a variant P-glycoproteinand glutathione S-transferase-ir Mol. Pharmacol., 37: 801-809, 1990.

37. Puchalski. R. B., and Fahl, W. E. Expression of recombinant glulalhione .S'-transferase

•n.Ya, or Yb, confers resislance lo alkylating agents. Proc. Nati. Acad. Sci. USA, 87:2443-2447, 1990.

38. Mannervik. B. The isoenzymes of glulalhione Iransferasc. Adv. Enzymol.. 57:357-417, 1985.

39. Ploemen, J. H. T. M., Van Schanke, A.. Van Ommen. B.. and Van Bladeren. P. J.Reversible conjugation of ethacrynic acid with glulalhione and human glutathioneS-lransferase Pl-1. Cancer Res., 54: 915-919. 1994.

40. Kuzmich, S.. Vandcrveer. L. A.. Walsh. E. S., LaCrcla. F. P., and Tew. K. D.Increased levels of glulalhione .S'-lransferase TTtranscript as a mechanism of resistance

to elhacrvnic acid. Biochem. J.. 281: 219-224. 1992.

407N

on July 31, 2021. © 1995 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1995;55:4073-4078. Cancer Res   James H. Doroshow, Marianne Z. Metz, Linda Matsumoto, et al.   Broad-Range Multidrug Resistance

Producesπ-Transferase S and Glutathione MDR1Human Transduction of NIH 3T3 Cells with a Retrovirus Carrying Both

  Updated version

  http://cancerres.aacrjournals.org/content/55/18/4073

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/55/18/4073To request permission to re-use all or part of this article, use this link

on July 31, 2021. © 1995 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from