[CANCER RESEARCH 58. 4349-4357. October I. I998|

Functional Nucleoside Transporters Are Required for Gemcitabine Influx andManifestation of Toxicity in Cancer Cell Lines1

John R. Mackey2, Ra.jam S. Mani, Milada Seiner, Delores Mowles, James D. Young, Judith A. Belt,

Charles R. Crawford, and Carol E. CassDepartment of Oncology, University- of Alberta, unti Cross Cancer Institute, Edmonton, Alberili, T6G 1Z2 Canuda (J. R. M., R. S. M., M. S., D. M., C. E. C.¡;Department ofPhysiology, University of Alberiti, Edmonton, T6G 172 Canada ¡J.D. Y.I; and Department of Molecular Pharmacology. St. Jude Children's Research Hospital. Memphis.

Tennessee 38105 ¡].A. B.. C. K. CJ

ABSTRACT

Gemcitabine (2',2'-difluorodeoxycytidine) is a novel pyrimidine nucle-

oside drug with clinical efficacy in several common epithelial cancers. Wehave proposed that gemcitabine requires nucleoside transporter (NT)proteins to permeate the plasma membrane and to exhibit pharmacological activity. In humans, there are seven reported distinct NT activitiesvarying in substrate specificity, sodium dependence, and sensitivity toinhibition by nitrobenzylthioinosine (NBMPR) and dipyridamole. To determine which NTs are required for gemcitabine-dependent growth inhi

bition, cultures from a panel of 12 cell lines with defined plasma membrane NT activities were incubated with different concentrations ofgemcitabine. Cell proliferation was assessed by the sulforhodamine Bassay and cell enumeration to identify the concentrations of gemcitabinethat inhibited cell replication by 50% (IC50s). NT activity was a prerequisite for growth inhibition in vitro because: (a) the nucleoside transport-

deficient cells were highly resistant to gemcitabine; and (In treatment ofcells that exhibited only equilibrative NT activity with NBMPR or dipyridamole increased resistance to gemcitabine by 39- to 1800-fold. These

data suggested that the type of NT activities possessed by a cell may be animportant determinant of its sensitivity to gemcitabine and that NTdeficiency may confer significant gemcitabine resistance. We analyzed theuptake kinetics of [3H]gemcitabine by each of five human NT activities in

cell lines that exhibited a single NT activity in isolation; transient trans-

fection of the cDNAs encoding the human concentrative NT proteins(hCNTl and hCNT2) was used to study the rii and rif activities, respectively. The efficiency of gemcitabine uptake varied markedly among thecell lines with single NTs: es = dt > ei > db »> df. The transportability of [3H]gemcitabine was demonstrated by reconstitution of the

human es NT in proteoliposomes, confirming that gemcitabine permeationis a protein-mediated process.

INTRODUCTION

Gemcitabine (2',2'-difluorodeoxycytidine, dFdC,3 Gemzar) is a

novel pyrimidine nucleoside anticancer drug with demonstrated clinical activity in common epithelial cancers. Unlike other nucleosideanticancer drugs used in clinical practice [cytarabine ( l -ß-D-arabino-

Received 3/10/98: accepted 8/18/98.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 in accordance with18 U.S.C. Section 1734 solely to indicate this fact.

' Supported by a Canadian Cancer Society operating grant from the National CancerInstitute of Canada and a Pilot Project and New Investigator's Award from the Alberta

Cancer Board. C. E. C. is a Terry Fox Cancer Research Scientist of the National CancerInstitute of Canada, and J. D. Y. is a Heritage Medical Scientist of the Alberta HeritageFoundation for Medical Research. Isolation of the L12IO/DNC3 and L1210/DU-5 celllines was also supported by National Cancer Institute (USA) Research Grant CA55056.Cancer Center Support Grant CA21765. and the American Lebanese Syrian AssociatedCharities of St. Jude Children's Research Hospital (to J. A. B. and C. R. C.).

2 To whom requests for reprints should be addressed, at Cross Cancer Institute. 11560

University Avenue, Edmonton, Alberta, T6G 1Z2 Canada.1The abbreviations used are: dFdC. gemcitabine; CEM. CCRF-CEM; NT, nucleoside

transporter: ENT. equilibrative nucleoside transporter; CNT, concentrative nucleosidetransporter; NBMPR. nitrobenzylmercaptopurine ribonucleoside (nitrobenzylthioinosine);ei, equilibrative NBMPR-insensitive: es, equilibrative NBMPR-sensitive: cif, concentrative. insensitive to NBMPR and formycin B-selective; cit, concentrative, insensitive to

NBMPR and thymidine selective: cib, concentrative. insensitive to NBMPR and broadlyselective; OCTG. octylglucopyranoside; h. human; NBTGR. nitrobenzylmercaptoguanineribonucleoside (nitrobenzylthioguanosine).

furanoxylcytosine), cladribine (chlorodeoxyadenosine). and fludara-

bine (fluoroarabinosyladenine)], gemcitabine has demonstrated significant activity in Phase II and Phase III clinical trials in patients withadvanced breast (1,2), bladder (3), ovarian (4), non-small cell lung (5,6), pancreatic (7, 8), and head and neck cancers (9). The mecha-

nism(s) accounting for the unique and broad spectrum of gemcitabineactivity has not been fully determined. Although gemcitabine andcytarabine both require phosphorylation to produce the correspondingtoxic triphosphates, dFdCTP and ara-CTP, the cellular pharmacology

of gemcitabine differs from that of cytarabine, with relatively prolonged intracellular accumulation of gemcitabine triphosphate (10).dFdCTP can be incorporated into DNA, which is believed to be themain mechanism of cytotoxicity (11, 12). Gemcitabine is a substratefor deoxycytidine deaminase, producing the noncytotoxic compound2',2'-defluorodeoxyuridine (12).

For unknown reasons, some human tumors exhibit de novo gemcitabine resistance [as shown by low response rates in clinical trials ofpatients with advanced colon cancers (13)], whereas other initiallysensitive cancers may develop resistance during therapy. The in vivomechanisms of resistance to gemcitabine therapy are not defined,although several in vitro mechanisms of gemcitabine resistance havebeen identified. Gemcitabine resistance was induced by exposingA2780 human ovarian carcinoma cells to stepwise increases in gemcitabine concentration. The identified mechanism of resistance was adeficiency in deoxycytidine kinase, the enzyme necessary for theactivation of gemcitabine to gemcitabine monophosphate (14). Trans-

fection of CEM human leukemia cells with a vector that forcedexpression of cytidine deaminase conferred a 2.4-fold increase in

gemcitabine resistance (15). Gemcitabine efficacy was decreased infibrosarcoma cells induced by transfection to produce large quantitiesof heat shock protein 70 (16). Resistance to gemcitabine was notassociated with increased P-glycoprotein expression and was not seen

in cell lines resistant to alkylating agents and cisplatinum (17). Plasmamembrane transport has been studied in Chinese hamster ovary cells,where membrane transport did not appear to be a major factor contributing to the more potent cytotoxicity of gemcitabine when compared with cytarabine (10).

The biochemical targets for gemcitabine are intracellular, and theobligatory first step in cytotoxicity is permeation across the plasmamembrane. Because gemcitabine, like other anticancer nucleosides, ishydrophilic and would not be expected to readily permeate plasmamembranes by passive diffusion, its cellular uptake requires the presence of specialized plasma membrane NT proteins (18). Seven distinctNT activities have been demonstrated in human cells. Nucleosidetransport occurs by both sodium-independent, bidirectional equilibrative (ENT) and sodium-dependent, inwardly directed concentrative

(CNT) processes: es (broad permeant selectivity); ei (broad selectivity), cir, cib (broad permeant selectivity); cif (purine-nucleoside selectivity); csg (concentrative, NBMPR-insensitive, guanosine selectivity); and ci (concentrative, NBMPR-insensitive, unknown

selectivity). cDNAs encoding four human NTs have recently beencloned and functionally expressed: hENTl, which displays es activity

4349

on June 29, 2020. © 1998 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

(¡I-.MUTABINK C'YUITOXIflTY AND SI I I I-.OSID1 IRANSPOKT

(19); HENT2, which displays ei activity (20, 21); hCNTl, whichdisplays cit activity (22); and hSPNT, or in our terminology, hCNT2,which displays cit activity (23).4 The equilibrative NTs appear to be

widely distributed in different cell types and tissues, whereas theconcentrative NTs appear to be limited to specialized cells such asintestinal and renal epithelia. liver, and leukemic cells. It is likely thatdifferent human tissues and cell lines vary widely in NT activities, dueto variations in the number and type of NT proteins expressed on theplasma membrane (24).

To determine whether gemcitabine resistance could be correlatedwith nucleoside transport capabilities, we examined the gemcitabinesensitivity of 12 murine and human cell lines with defined nucleosidetransport characteristics. To study the effect of pharmacological manipulation of nucleoside transport on gemcitabine-dependent inhibition of proliferation, the transport inhibitors NBMPR and dipyri-

damole were used to specifically inhibit, respectively, es NT activityor both es and ei NT activities. To further define the transportabilityof gemcitabine, we analyzed the uptake kinetics of [^HJgemcitabine

by five human nucleoside transporters acting in cell lines exhibiting asingle NT activity in isolation. To confirm the mediated character of| 'H]gemcitabine permeation in a noncellular system, the human es NT

was solubilized from membranes of cultured human BeWo chorio-

carcinoma cells and functionally reconstituted in proteoliposomes,which were used to study ['HJgemcitabine uptake.

MATERIALS AND METHODS

Materials. ['H]Gemcitabine was a kind gift of Eli Lilly. Inc. (Indianapolis.Indiana). |G-'H|NBMPR (22.5 Ci/mmol) and [5,6-'H]uridine (40 Ci/mmol)

were purchased from Moravek Biochemicals. Inc. (Brea. CA). Radiolabelednucleosides were purified by high-pressure liquid chromatography using wa-ter-methanol gradients on a C,8 reverse phase column; repurification ofI'Hlgemcitabine was essential because some preparations contained 'H degradation producís. ['''CICholesteryl oleate (0.1 mCi/ml) was obtained from

Amersham Canada. Ltd. (Oakville. Ontario, Canada). Asolectin (soybeanphospholipids) was purchased from Associated Concentrates (Woodside. NY)and stored under N:. Other phospholipids and cholesterol were purchased fromAvanti Polar Lipids, Inc. (Alabaster, AL). OCTG, adenosine, uridine,NBMPR. and dipyridamole were purchased from Sigma Chemical Co., andSephadex G-50 (fine and medium) was from Pharmacia, Canada. DilazepK/V,/V'-bis|3-(3,4.5-trimethoxy-ben/oyloxy)-propyl] homopiperazine] was a

gift from F. Hoffman La Roche and Co. (Basel. Switzerland). All otherchemicals were of analytical grade and commercially available.

Cell Culture. Murine cell lines were grown in RPMI and 10% horse serum,and human cell lines were cultured in RPMI 1640 and 10% fetal bovine serumas described elsewhere (25-27). Experiments were initiated with cells in

exponential growth phase. S49 mouse lymphoma cells possess only es nucleoside transport activity, which can be completely inhibited by either NBMPR(1 /¿M)or dipyridamol (10 /IM: Ref. 28). AE1 is a NT-deficient S49 mutant

(29). LI 210 murine leukemia cells have three NT activities: es, ei, andc//(30).B23.1 is double mutant derived from the parental L1210 line by mutagenexposure and selection with 7 /xM cytarabine. erythro-9-[3-(2-hydroxynony-l)]adenine. and NBMPR; it possesses only es transport activity (25). MA =

27.1 is a mutant L1210 line that possesses only ci/activity and was isolated bymutagen exposure and selection with tubercidin and cytarabine (31). LI210/DNC3 is a transport-deficient cell line isolated from L1210/B23.1 after mu-lagenesis and selection with 0.25 JU.Mtubercidin plus 1.0 ¡UM1-ß-D-arabino-

furanoxylcytosine and lacks all three of the nucleoside transport activitiesfound in wild-type LI210 cells. L12IO/DU-5 is a stable cell line expressingrecombinant rat CNT1 (32) in the transport-incompetent background of LI 210/

DNC3. rCNTl was subcloned into the mammalian expression vector pcDNA3and electroporated into LI210/DNC3 cells. Transport-competent cells were

4 M. W. L. Ril/el. S. Y. M. Yao. A. M. L. Ng, J. R. Mackey, C. E. Cass. and J. D.

Young. Molecular cloning, functional expression and chromosomal localization of acDNA encoding a human Na*/nucleoside cotransporter (hCNT2) selective for purine

nucleosides and uridine. suhmitled for purification.

selected using the de novo uridylate synthesis inhibitor Brequinar in thepresence of the salvageable nucleoside uridine. The LI210/DU-5 cell line was

cloned from this selection and had a single copy of rCNTl inserted into itsgenome (33). K562 is a human erythroid leukemia cell line that possesses esand ei transport activities (27). CEM is a human T-lymphoblast cell line with

es transport activity originally derived from a patient with acute lymphocyticleukemia (34). The NT-defective ARAC-8C cell line is a clonal derivative ofa hypoxanthine-guanine phosphoribosyltransferase-deficient CEM cell line

originally isolated by mutagen exposure and selection for resistance to 8 /UMcytarabine. For the present investigations, the absence of hypoxanthine-guanine phosphoribosyltransferase. a purine-metabolizing enzyme, was not pertinent to the metabolism or cytotoxicity of gemcitabine (35). CaCo-2 is a human

colon carcinoma cell line that possesses es, ei, and cib activities (24). HeLa isa human cervical carcinoma cell line that possesses es and ei activities (26).BeWo is a human choriocarcinoma cell line that possesses es and ei activities(36, 37).

Chemosensitivity Testing. The chemosensitivity of suspension cell lineswas assessed using the sulforhodamine B assay, essentially as describedpreviously (38. 39). Cells were seeded in multiples of eight wells in 96-wellcell culture plates, with —¿�20,000cells/well, then exposed to drug. Drug

exposures were performed in growth media with 10% horse serum for murinecell lines and 10% fetal bovine serum for human cell lines. Gemcitabine wasadded to final concentrations that ranged from I0~4-10~9 M. Cells were

enumerated with the sulforhodamine B assay at the time of drug addition andat 48 h after drug addition. All adherent cell lines were plated in triplicate 24 hbefore drug exposure, then similarly assessed for growth inhibition bytrypsinization and cell enumeration with an electronic cell counter after 72 h ofdrug exposure. Chemosensitivity was expressed as the concentration of gemcitabine that inhibited cell proliferation by 50% (IC50), as determined bynonlinear regression analysis using GraphPad Prism 2.01 (GraphPad Software,San Diego. CA). All experiments were performed independently on twooccasions, and mean IC5(>values are reported.

Generation of Expression Constructs. For expression in HeLa cells, thefull coding sequences of hCNTl (22) and hCNT24 were removed from theoriginal cloning vector pGEM3Z by Afrlll/Xhal digestion. Subsequently, the 5'

overhang was blunt-ended with the Klenow fragment and ligated into pCD-NAI/Amp at the EcoRV site. The 3' ends of hCNTl and hCNT2 were ligated

into the vector at the Xbal site. The structures of pCDNAI/Amp-hCNTl andpCDNAI/Amp-hCNT2 were confirmed by restriction mapping and sequence

analysis using an ABI Prism 310 Genetic Analyser (Norwalk, CT).Transient Expression of hCNTl and hCNT2 in HeLa Cells. The trans

port characteristics of recombinant hCNTl (cit activity) and hCNT2 (cifactivity) were studied using a method described previously (40) for recombinant rCNTl produced in COS-1 cells. HeLa cells were plated (5.0 X 10" per

100-mm plate) and grown to 50% confluency in RPMI with 10% calf serum.Transfection was performed by a modification of the DEAE-dextran method

(40) in which cells in DMEM and 10% NuSerum with 200 ¡IMchloroquine (5ml) were exposed to 100 ¿ilof DEAE-dextran DNA mixture, using 5 /j,g ofDNA per plate. Cells were incubated at 37°Cin 5% CO, for 2 h, then 2 ml of

DMSO 10% solution were added. Plates were incubated for 2 min at roomtemperature, washed three times with PBS. and placed in growth media for24 h. Cells were then trypsinized, pooled, and replated on 60-mm plates to

eliminate any differences in transfection efficiencies among the individualcultures. Cells were used for uptake assays 72 h after transfection. Transfectionefficiencies (determined by counting stained and unstained cells in culturesthat had been independently transfected with ß-galactosidase) ranged from15% to 30%. Cells were treated with 10 /J.Mdilazep to inhibit endogenous es-and e/-mediated NT activity prior to uptake assays.

[3H]Gemcitabine and l'11irridine Cellular Uptake Assays. Nucleoside

uptake assays were conducted at room temperature under zero-trans conditionsin either sodium-containing transport buffer (20 mM Tris/HCl. 3 mM K,HPO4.1 mM MgCl,-6H,O. 2 mM CaCU, 5 mM glucose, and 130 mM NaCl, pH 7.4;

300 ± 15 mOsm) or sodium-free transport buffer (20 mM Tris/HCL, 3 mMK2HPO4. 1 mM MgCl:-6H,O. 2 mM CaCU, 5 mM glucose, and 130 mMN-methyl-D-glucamine/HCl, pH 7.4; 300 ± 15 mOsm). Cells were washed

once with the appropriate transport buffer and then processed immediately, orin some experiments, incubated with NBMPR or dilazep at room temperaturefor 30 min before the uptake assay. For suspension cell lines (27), each samplewas processed individually, and permeant fluxes were terminated using the

4350

on June 29, 2020. © 1998 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

GEMCITABINE CYTOTOXICITY AND NUCLEOSIDE TRANSPORT

"inhibitor-oil" stop method. Dilazep dihydrochloride was used at 200 /UMasthe stopping reagent. Uptake at time "zero" was determined by the addition of

cold (4°C)dilazep to the assay mixtures, followed by the addition of cells and

immediate centrifugaron (16,000 X #). Cell-associated radioactivity wasdetermined by liquid scintillation counting using a xylene-detergent scintillant.

For adherent cell lines. ~ 106 cells were plated on 60-mm plates 48 h before

transport studies. Uptake intervals were started by adding 1.5 ml of transportbuffer containing 'H-labeled permeant and inhibitors. Uptake intervals were

ended by rapidly aspirating the permeant solution and immersing the culturedishes in 1.5 I of ice-cold transport buffer containing dilazep. The dishes weredrained, and the cells were solubilized in 1 ml of 5% (v/v) Triton X-100 and

combined with 5.0 ml of Ecoute scintillant for radioactivity measurements.Uptake at time zero was determined by incubating cells for 10 min at 4°Cwith

transport buffer that contained 10 JU.Mdilazep and. immediately thereafter, for£3 s with ice-cold transport buffer that contained 3H-labeled permeant and

either 10 /UMNBMPR or 10 /J.Mdilazep. For kinetic studies with transfectedHeLa cells, ['H]uridine uptake was measured both in sodium-containing and

sodium-free transport buffer in cells transfected with the vector alone or that

contained the appropriate cDNA constructs. Initial rates of uptake were derivedby linear regression analysis of time courses of uptake of radioactive nucleo-

sides. Kinetic parameters were estimated from nonlinear regression analysisusing GraphPad Prism 2.01. and the resulting Km and Vmail values wereaveraged.

Isolation and Soin bili/al ion of Be Wo Membranes. BeWo cells werechosen as a source of es NT due to their unusually high NT activity and highnumbers of high-affinity NBMPR-binding sites (26). Cells (5 X IO6 per flask)

were grown in 15 Corning disposable tissue culture flasks and. althoughactively proliferating, the harvest yielded 8 x 10S-I X IO9 cells. Plasma

membranes were prepared as described previously (41 ). In brief, washed cellswere swollen by suspension in 1 mM ZnCU, fragmented using a Polytron

homogenizer (Brinkman Instruments. Westbury, NY), and resuspended in9.25% sucrose with removal of nuclei by centrifugation (2 min. 900 X g).Plasma membranes were then separated on a 15-45% (w/w) sucrose gradientand stored at -80°C in 15% (v/v) DMSO. The fractions collected at 25%

sucrose were used for reconstitution experiments because they were enrichedin plasma membranes when assayed for 5'-nucleotidase (EC 3.1.3.5) activity(Sigma Diagnostics 5'-nucleotidase assay kit).

Membranes were thawed and washed three times in ice-cold reconstitution

buffer (pH 7.4) composed of 100 mM KC1. 10 mM Tris. 0.1 niM MgCU and 0.1mM CaCU. Membranes were then combined with 1% OCTG and 0.15%asolectin in reconstitution buffer and incubated on ice for l h with frequentmixing. The insoluble material was removed by centrifugation (60 min,100.000 X g. 4°C).and the supernatants were retained on ice until use. The

membrane protein concentrations were monitored using the modified Bradfordassay (42) and adjusted to —¿�200¿igof protein/ml.

Proteoliposome Reconstitution Procedure and 3H-Labeled Nucleoside

Uptake Assays. The reconstitution technique is described in detail elsewhere(43, 44). Solubilized BeWo membranes (2 ml) were supplemented with asonicated preparation of lipids (0.3 ml) consisting of phosphatidylcholine(bovine brain), cholesterol, phosphatidylethanolamine (egg), and phosphati-

dylserine (bovine brain) in molar ratios of 33:33:26:8, respectively, plus a trace(10s dpm/ml) of [l4C]cholesteryl oleate. The detergent was removed by

filtration on Sephadex G-50. and the resulting proteoliposomes were pooled,frozen in ethanol and dry ice. and stored at -80°C for up to 1 month.

Proteoliposomes were thawed at room temperature, centrifuged (40.000 X g,20 min), resuspended in ~ 1 ml of reconstitution buffer, sonicated for 15 s in

a cylindrical tank sonicator (Laboratory Supplies Company Inc.. Hicksville.NY), and held on ice until use.

The uptake assays were initiated by the rapid addition of 100 ¿ilofproteoliposome suspension (~5 fig protein, with or without inhibitors) to 25 /¿I

Table 1 Gemcittlhine i'heniosensiíivity in cell lines wilh tiffined NT íl(7/w'//í'.v

Cultures were exposed to graded concentrations of gemcitabine (10 -10 M)for 48 or 72 h. as described in "Materials and Methods." Cells were enumerated, and chemosensitivity

was expressed as the concentration of gemcitabine that inhibited cell replication by 50% (IC5(,). All studies were performed Iwice in independent experimenls. and each IC50 is reported.

Nucleoside transport activities

CelllineS49AEIL12IOB23.1DU-5MA27.1DNC-3K562K562

+dipyridamoleCEMCEM

+NBMPRCEM

-1-dipyridamoleCEM-ARAC8CHeLaHeLa

+NBMPRHeLa

+dipyridamoleCaCo-2CaCo-2

+NBMPRCaCo-2

+ dipyridamoleOrigin

eseiMurine

lymphoma+S49

NT-deficientmutantMurine

leukemia 4-+L1210

mutant+L1210rCNTl

stabletransfectantL1210

mutantLI210

NT-deficientmutantErylhroid

leukemia ++Erythroid

leukemiaLymphoblastic

leukemia+Lymphoblastic

leukemiaLymphoblastic

leukemiaNT-deficient

CEMmutantCervical

cancer ++Cervical

cancer+Cervical

cancerColon

cancer ++Colon

cancer+Colon

cancercil

cif cib IC50(fiM)0.0130.0091.61.0+

0.00870.00550.1140.070+

0.0150.009+

6.04.3150.0570.049>IOO>IOO0.0320.0305.84.48.77.9>100>1000.0360.0200.0390.0231.40.8+

0.00370.0025+

0.00390.0029+

0.00410.0035Relative

gemcitabineresistance1.0118l.o131.773018001.0>18001.0164267>32001.01.1391.01.11.3

4351

on June 29, 2020. © 1998 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

GEMCITABINE CYTOTOXICITY AND NUCLEOSIDE TRANSPORT

of reconstitution buffer with 20 /IM "H-labeled permeant. After specific incu

bation times, 100 ß\of the reaction mixtures were layered on ice-cold Seph-adex minicolumns ( 1-ml syringes fitted with a polyethylene filter), equilibratedin reconstitution buffer containing 10 /¿Mdila/.ep and NBMPR. and centri-

fuged (45 s. 7(X) x #). The effluents were collected in preweighed tubes fromwhich portions were removed for the determination of protein. 3H. and I4C

contents. Estimates of/ero-time uptake values were obtained by measuring theuptake of 'H at =2 s in the presence of ice-cold solutions containing 10 itiM

adenosine. 10 /IM dipyridamole, and 10 /IM NBTGR. Protein was measuredusing BSA as the standard with appropriate corrections for detergent interference (42. 45).

RESULTS

Growth Inhibition Studies. We have compared the ability ofgemcitabine to inhibit proliferation of murine and human cancer celllines exhibiting a spectrum of NT activities. In the experimentssummari/.ed in Table 1, cells were exposed to graded concentrations

10.0-,

Fig. 3. Initial rates of uptake of ['Hlgemcitabine by nucleoside-transport competent(CEM) and incompetent (CEM-AraC-8Cl cells. Cells were exposed to 10 fiM ['H|gcm-

citahine for the lime periods indicated, permeant fluxes were terminated using theinhibitor-oil slop method, and cell-associated radioactivity was determined (see "Materialsand Methods"). Each point represents the mean of experiments performed in triplicate;

bars, SE (not shown where smaller than data points). A single representative experimentof four is shown. •¿�.CEM cells: O, AraC-C8 cells.

* -WC^—D0)C0«1,>

/\•1— v

*Jo1£10075-

50-25-0-1IJ

1X

!1b-Q-o-o7

-15 -13 -11 -9 -7 -5

Log gemcitabineconcentration (M)

Fig. I. A comparison of the anliptulifcralive action of gemcitabine against nucleosidetransport-incompetent (CEM-AraC-8C) and nucleoside transport-competent (CEM) cells.Cells were exposed lo a range of gemcitabine concentrations for 48 h and then quantitatedby the sulf'orhodaniinc B assay. Each point represents the mean of eight replicates; bars.

SE (not shown where smaller than the size of data points). A single representativeexperiment is shown. O. CEM cells; •¿�AraC-C8 cells.

=0)u

I§o.

_0)£

Log gemcitabine concentration (M)

Fig. 2. The effects of inhibitors of equilibrative nucleoside transport on (he antipro-liferative action of gemcitabine against CaCo-2 human colon cancer cells. Cells wereplated in triplicate for 24 h and then exposed to graded concentrations of gemcitabine for72 h. lrypsini/ed. and enumerated. Each point represents the mean; burs. SE (not shownwhere smaller than data points). •¿�.gemcitahine alone; Q. gemcitabine after preincuhationwith I(K) nM NBMPR; O. gemcitabine after preincubation with 10 ¿IMdipyridamole. Asingle representative experiment is shown.

of gemcitabine: (a) for 48 h (suspension cultures) and quantitatedusing the sultbrhodamine B assay (38); or (b) for 72 h (adherentcultures) and quantitated by trypsinization and electronic enumeration. Chemosensitivity was expressed as the concentration of gemcitabine that inhibited cell proliferation by 50% (IC50). and results aresummarized in Table 1. We first compared chemosensitivity in pairsof cell lines that were either transport competent (parental line) ortransport incompetent (mutant line) in experiments similar to the oneshown in Fig. 1 for CEM, a human T-lymphoblast cell line with extransport activity (34). and ARAC-8C. a NT-incompetent clonal de

rivative that was isolated by mutagen exposure and selection forresistance to 8 /¿Mcytarabine (35). The ARAC-8C cell line was highly

gemcitabine resistant with an IC50 exceeding 100 /AMcompared withan IC5(Iof 0.031 /AMfor the CEM line (Fig. 1). The other transport-

incompetent cell lines produced genetically by mutation (AE1,DNC-3) were also significantly less sensitive to gemcitabine than the

matched parental cell lines (S49, LI 210). The IC5,,s of gemcitabine inthe transport-incompetent cell lines, when related to the IC5()s inparental cell lines, increased by 118-fold to greater than 3200-fold.The high levels of resistance seen in transport-incompetent cells

indicated that uptake of gemcitabine in the absence of functional NTsin the plasma membrane was inefficient and suggested that diffusionaluptake of gemcitabine through the plasma membrane was capable ofproducing growth-inhibitory concentrations inside cells only in the

presence of large transmembrane concentration gradients.To determine the contributions of the two bidirectional equilibra

tive NT activities to the antiproliferative action of gemcitabine, wenext examined the growth-inhibitory effects of gemcitabine in three

human cancer cell lines (K562, CEM, and HeLa) in the presence andabsence of pharmacological inhibitors of es and ei (Table I ). At theconcentrations used in these studies, NBMPR (100 nM) selectivelyinhibits ev-mediated transport, whereas dipyridamole (10 /¿M)inhibitsboth ex- and ii-mediated transport. Although NBMPR exposuresconferred 164-fold resistance to gemcitabine in CEM cells, which

exhibit es NT activity in isolation, NBMPR was not capable ofprotecting HeLa cells, which exhibit both es and ei NT activities (26).In contrast, dipyridamole exposures conferred 39-fold resistance inHeLa cells, indicating that «'-mediated uptake alone was sufficient to

achieve intracellular concentrations of gemcitabine that inhibited proliferation. Dipyridamole increased gemcitabine resistance by 39- to> 1800-fold in cell lines with only equilibrative NT activities (K562,

CEM, and HeLa), demonstrating that pharmacological inhibition ofNT activities can significantly modulate the antiproliferative action ofgemcitabine.

4352

on June 29, 2020. © 1998 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

GEMCITABINE CYTOTOXICITY AND NUCLEOS1DE TRANSPORT

Table 2 Kinetic parameters of ¡"Hfgemcitabine transport in cell lines with defined NT adivines

Cells (Iransfected and/or inhibitor treated as indicated) were exposed to graded concentrations of [ HJgemcitabine. and initial uptake rates were determined. Assays were conductedat room temperature using short gemcitabine exposures terminated by inhibitor/oil stop methods (CEM cells) or inhibilor/cold stop methods (all other cell types). Kinetic parameterswere estimated from nonlinear regression analysis. Each value represents the mean ±SE of values from independent experiments conducted in triplicale.

TransportereseicitcifdbCell line andtreatmentCEMHeLa

+ 100 HMNBMPRRecombinanthCNTl in HeLacellsRecombinanthCNT2 in HeLa cells

CaCo2 colon cancer cells + dilazepKm

(MM)329

±91832±20418.3±7.2No

measurable transportRate insufficient for accurate kineticsVniax(pmol/s/IO6

cells)17.0

±4.34.2±1.30.94±0.22NA"

Rate insufficient for accurate kinetics«JVmM20.1

±5.9215±2022.8±9.7NA

NANo.

of independentexperiments43323

' NA, not assessable.

In addition to the equilibrative NTs described above, mammaliancells possess sodium-coupled inwardly directed concentrative NT

activities (18). To determine the contributions of the concentrativeNTs to gemcitabine activity, we examined gemcitabine-induced

growth inhibition in cell lines that exhibited a single concentrative NTactivity (cit, cif, or cib; Table 1).

The cit NT exhibits sodium dependence, selectivity for pyrimidine-

nucleosides, and insensitivity to inhibition by NBMPR, and cit NT

20,

-8 10-1o

Q.

1000 2000 3000

(MM)

B

"=3o

o

1000 2000

(MM)

3000

1.0,

«41-—o>_ ooE

0.5.

0.00 25 ¿0 /5

(uM)

100 125

Fig. 4. Kinetic analysis of initial uptake of [ 'HJgemcitabine by human es, ei, and cit NTactivities. Cells were exposed to graded concentrations of ['HJgemcitabine. and the initialrates of update were determined (see "Materials and Methods"). Each point represents the

mean of the rate derived from by linear analysis of experiments performed in triplicate;burs, SE (not shown where smaller that data points). A single representative experimentis shown for each transporter. Nonlinear regression analysis was used to analyze uptakekinetics. A, kinetics of «-mediated ['HJgemcitabine uptake in CEM cells. Initial rates ofuptake were determined as described in "Materials and Methods." Analysis of initial[3H]gemcitabine uptake kinetics revealed a Km of 410 yM and a Vmaxof 16.1 pmol/s/106 cells. B, kinetics of f/-medialed ['HJgemcitabine uptake in HeLa cells. Initial rates ofuptake were determined as described in "Materials and Methods." and endogenous esactivity was inhibited by 100 nM NBMPR. Analysis of initial ['HJgemcitabine uptakekinetics reveals a Afmof 737 JIM and a Vm„of 1.99 pmol/s/10 6 cells. C, kinetics of(•«-mediated['HJgemcitabine uptake in HeLa cells transiently transfected with hCNTl.HeLa cells were transiently transfected as described in "Materials and Methods." After

72 h cells were preincubated with 10 JIM dilaz.ep to inhibit endogenous equilibrative NTactivity. Analysis of initial [•'HJgemcitabineuptake kinetics reveals a Km of 31.8 JAManda ym«of 0.82 pmol/s/106 cells.

activity has been observed in intestinal enterocytes, brush-border

membrane vesicles of several species, and rabbit choroid plexus (18).Because cit activity has not yet been demonstrated in cultured celllines, we used L1210/DU-5, a stable transfectant that produces re-combinant rat CNT1 (32) in the transport-incompetent background ofL1210/DNC3. L1210/DU-5 cells exhibit sodium-dependent transport

of thymidine that is inhibited by the pyrimidine nucleosides uridineand cytidine but not by the purine nucleosides inosine or guanosine(33). Gemcitabine inhibited proliferation of DU-5 cells (IC5() 0.092/J.M) in contrast to the parental, transport-incompetent DNC-3 cells

(IC5() 13 /LIM),suggesting that the cit transporter mediated gemcitabineinflux.

The cif NT exhibits sodium dependence, selectivity for purine-

nucleosides, and uridine, and insensitivity to inhibition by NBMPR.and «/NT activity has been observed in mouse splenocytes, macrophages, rat macrophages, and hepatocytes and in several cultured celltypes, including mouse L1210 leukemia cells (18). The L1210/MA27cell line, which exhibits only cif NT activity, was 730-fold less

sensitive to gemcitabine (IC5() 5.2 /J.M)than the parental L1210 cellline (IC50 0.0071 /J.M),suggesting that gemcitabine is a poor substratefor the a/NT.

The cib NT has the ability to transport a wide range of both purineand pyrimidine nucleosides and has been described in the HL-60

cultured human promyelocytic leukemia cell line after induction ofdifferentiation (46) and in the CaCo-2 colon carcinoma cell line (24).

In the experiment of Fig. 2, we examined gemcitabine inhibition ofgrowth in CaCo-2 cells that were treated with dipyridamole to block

endogenous es and ei activities. There were no differences in theresponses to gemcitabine between untreated and inhibitor-treated

cells. It has been shown previously that dipyridamole enhanced thetoxicity of 9-ß-D-arabinofuranosyladenine against the L1210/C2 cellline (47), which has es, ei, and «'/'transport activities (30). Although

the initial rate of drug uptake was reduced by dipyridamole, there wasan intracellular accumulation of drug above the steady-state intracel-

lular concentration because of the elimination of drug efflux via thebidirectional equilibrative transporters. That dipyridamole treatmentof CaCo-2 cells did not increase gemcitabine sensitivity suggested

that release of gemcitabine back into the growth medium via thebidirectional equilibrative NTs was not a major component of netgemcitabine retention.

[3H]Gemcitabine and [3H]Uridine Cellular Uptake Assays. To

maximize the potential clinical relevance of this work, gemcitabinetransport by each of the human NTs was examined individually, eitherin transport-competent cell lines or in human cell lines transientlytransfected with cloned human NT cDNAs. The transport-incompetent CEM-AraC-8C cell line was studied first to establish whether

gemcitabine was capable of entry into cells in the absence of afunctional nucleoside transporter (i.e., by diffusion). No significant[-'H]gemcitabine uptake occurred (Fig. 3), because uptake rates did

not significantly vary from zero. The parental cell line CEM, fromwhich CEM-AraC8C was derived, possesses only ex activity, with

4353

on June 29, 2020. © 1998 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

OEMCITABINE CYTOTOXICITY AND NUCl.HOSIPE TRANSI'OKI

8"o

10.0,

7.5-

5.0-

10 12

time (s)

B

£

o«DO

!Q.

10.0,

7.5.

5.0.

2.5.

0.0.0 4 è

time (s)

10 12

Fig. 5. Initial rales of uptake of | 'HJgemcitabinc and |'H)uridine in HeLa cellstransiently Iransfeclcd with hCNTl (i-ii NT) or HCNT2 («yNT). Cells were transiently

transfected with vector alone (controls) or that contained either hCNTl M) or hCNT2 (ß)cDNAs as described in "Materials and Methods." After 72 h. cells were preincubated with

IO /IM dila/cp to inhibit endogenous equilibrarne NT activity. Cells were then exposed toeither K) JIM [ 'H|genicitabine or [lH)uridine for the indicated time periods, and cell-associated radioactivity was determined (see "Materials and Methods"). Each point

represents the mean of experiments performed in triplicate; Aurs. SE (not shown wheresmaller than data points). A single representative experiment is shown for each recombinant transporter. A. |'H ¡undineuptake (•)or [ 'Hjgenicilabinc uptake (Q) in transfectedcell with hCNTl cDNA; |'H]uridinc uptake (O) or ['H]gemcitabine uptake <•>in cellstransfected with vector alone. R, | lH)uridine uptake (•)or [ 'Hjgemcitabine uptake (O) incells transfected with hCNT2 cDNA. Values for uptake of ['Hjuridine and ['Hlgemcit-

abine by cells transfected with vector alone were not significantly different from zero andare thus not shown.

-3.3 X IO5 NBMPR-binding sites/cell (48). Uptake of ['HJgemcit-

abine by CEM cells was rapid, with kinetic analyses of initial uptakerates demonstrating a Km of 329 ±91 /¿Mand Vmaxof 17.0 ±4.3pmol/s/10(' cells (mean ±SE of four independent experiments per

formed in triplicate: Table 2: Fig. 4/1).Human cells that exhibit ei NT activity in isolation have not been

described. However, HeLa cells possess es and ei without concentra-

tive NT activity. By selectively inhibiting HeLa es activity withNBMPR I(X)nM, the NBMPR-insensitive endogenous ei activity wasstudied in isolation. Uptake of [3H]gemcitabine was linear over 10 s,

and kinetic analysis revealed a Km of 832 ±204 /¿Mand Vmaxof4.3 ±1.3 pmol/s/106 cells (mean ±SE of three independent exper

iments performed in triplicate; Table 2; Fig. 4fi).Although human cell lines that exhibit at or d/NT activities have

not been described, the recent isolation of cDNAs encoding humanNTs with either cil (22) or cif activities4 has provided an approach to

kinetic analysis of these NTs in intact cells. We studied gemcitabinetransport by recombinant cit or cif after transient transfection of HeLacells, respectively, with hCNTl or hCNT2 cDNA. Transport studieswere performed in the presence of 10 /AMdilazep to inhibit endogenous equilibralive NT activities normally present in HeLa cells. Theresults of Fig. 5/4 demonstrated: (a) the lack of mediated uptake of'H-labeled nucleoside uptake in HeLa cells transfected with vectoralone: and (b) uptake of both ['Hjgemcitabine and [3H]uridine in

HeLa cells transfected with the hCNTl vector. Gemcitabine was agood permeant for recombinant hCNTl, with a Km of 18.3 ±7.2 /J.Mand Vmaxof 0.94 ±0.22 pmol/s/106 cells (Table 2; Fig. 4C). Con

versely, transfection of the hCNT2 cDNA into HeLa cells had noeffect on uptake of [3H]gemcitabine, although the expected [3H]uri-

dine uptake was observed (Fig. 5fi). This result was consistent withthe observed gemcitabine resistance in L1210/MA27 cells, whichexhibit only the purine-nucleoside selective cif NT.

Analysis of de-dependent transport of gemcitabine was difficult. Ahuman cDNA encoding a sodium-dependent NT with broad permeantselectivity has not yet been reported, and the CaCo-2 cell line exhibits

relatively low levels of endogenous cib activity in addition to ex andei activities (24). Uptake of [3H]gemcitabine was examined (not

shown) in CaCo-2 cells in the presence of 10 ¡ÕMdilazep to block es

and ei activities. Gemcitabine entry was mediated by the cib transporter, but uptake rates were too low for kinetic analysis. The demonstration of «'¿-mediatedgemcitabine uptake explains our earlier

observation (Fig. 2) of high-level gemcitabine activity in CaCo-2

despite the presence of inhibitors of equilibrative NT.Transport efficiency can be expressed as A"m/Vmax,where low

values indicate efficient transporter-mediated permeation. The values

of Table 2 indicate that the es (Arm/Vmax,20.1) and dt (KJVm„„22.8)NT activities were the most efficient mediators of gemcitabine influx,followed by the ei (Km/Vmax, 215) NT activity. The «/NT did notmediate permeation of gemcitabine across plasma membranes.

Because uridine is accepted as a permeant by all of the reportedhuman NTs, it is considered to be a "universal permeant" and was

chosen as a reference nucleoside with which to compare gemcitabineuptake rates. To allow an assessment of efficiency of gemcitabineuptake among the human NTs at the therapeutically and physiologically relevant concentration of 10 /J.M,uptake rates of both gemcitabine and uridine were determined for each of the human NTs (Table3). Relative uptake rates (gemcitabine/uridine) were 0.12 for es, 0.26for ei, 0.10 for cit. and 0.58 for cib (there was no gemcitabine uptakefor cif). Thus, although gemcitabine was transported by four of thefive NTs, the physiological substrate uridine was transported by allfive with much greater efficiency. In the HeLa cell line, known topossess both es and ei activities (26), the relative contributions toinitial ['HJgemcitabine uptake at a concentration of 10 /J.Mwere 72.4

and 27.6%, respectively (Table 4), indicating that the major component of gemcitabine influx was mediated by the es NT.

[•'HJGemcitabineand l'il|l ridine Uptake by Proteoliposomes

Prepared from Detergent-solubilized Membrane Fractions Enriched in NBMPR-binding Activity. The representative experimentshown in Fig. 6 examined the initial rates of uptake of 20 H.M[3H]gemcitabine by proteoliposomes prepared from BeWo cells by

Table 3 Comparison of initial rules of uptake of [' Hìgeineìtahìneanil /" Hlitritline bv

human mtcleositlt' transporters

Cells (transfected or inhibitor treated as indicated) were exposed to either 10 /AM["Hjgemcitabine or [ H|uridine. Initial uptake rates (transport) were determined at room

temperature using short gemcitabine exposures terminated by inhibitor/oil stop methods(CEM cells) or inhibitor/cold stop methods (other cell types). Each transport valuerepresents the mean ±SE of values from two to four independent experiments, eachconducted in triplicate.

NTtypees

eicitcif

cibCell

typeCEM

HeLa + NBMPRRecombinant hCNTl

in HeLaRecomhinant hCNT2

in HeLaCaCo-2 + dilazepGemcilabine

transport(pmol/s/IO6cells)0.109

±0.0490.034 ±0.0080.337 ±0.153No

measurable uptake

0.064 ±0.01Uridine

transport(pmol/s/106cells)0.894

±0.1450.131 ±0.022

3.25 ±1.480.76

±0.09

0.11 ±0.01Ratio"0.12

0.260.100

0.58' Gemcitabine:uridine transport rates.

4354

on June 29, 2020. © 1998 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

GEMCITABINE CYTOTOXICITY AND NUCLEOSIDE TRANSTORT

sucrose-density centrifugation and examined in the presence and

absence of a mixture of inhibitors (adenosine, NBTGR, and dilazep)oíes- and e/'-mediated transport. We have shown previously (41) that

proteoliposomes so prepared exhibit only es NT activity, despite thepresence of both es and ei activities in BeWo crude membranes. Thevalue from four separate experiments for the uninhibited initial rate ofgemcitabine uptake was 20 ±2 pmol/mg protein/s, and preincubationof proteoliposome with the mixture of transport inhibitors decreasedthis rate to 13 ±2 pmol/mg protein/s, or 65% of total uptake. Themediated component of [3H]gemcitabine uptake, which was obtained

by subtracting the inhibited value from the total value, was 7 ±2pmol/mg protein/s. In a separate study conducted similarly, the initialrate of uptake of 20 JU.M[3H]uridine (six separate experiments) was45.0 ±5.0 pmol/mg proteins/s (41). The ratio (0.15) of [3H]gemcit-abine to [3H]uridine uptake rates by the reconstituted BeWo es NT

was very similar to that (0.12) for the native es NT of CEM cells.These results support the use of functional reconstitution of individualhuman NT proteins as an appropriate means of studying human NTactivities in isolation; this is the first report of gemcitabine uptakemediated by a reconstituted NT.

DISCUSSION

Although resistance to gemcitabine because of decreased druginflux has not been reported previously, there is evidence that nucle-oside-transport deficiency may be a mechanism of resistance to anti-

cancer nucleosides both in vitro and in vivo. Decreased es NT activitycorrelated with in vivo cytarabine resistance in human acute myelog-

enous leukemia (49). Resistance to cytarabine therapy has also beenshown to be associated with decreased es activity in freshly isolatedhuman leukemia cells (50, 51). Estimates of the numbers of estransporters have been inferred from quantitative analysis of cellularNBMPR binding. Decreases in NBMPR-binding sites have beenobserved during treatment of leukemia; in a patient with T-cell acute

lymphoblastic leukemia, the leukemic blasts examined at relapseexhibited a 75% reduction in NBMPR binding when compared withblasts examined before initiation of cytarabine treatment (52). Murineerythroleukemia cells that were selected in vitro for resistance tocytotoxic purine nucleosides exhibited decreased levels of the estransporter, as indicated by a reduction in NBMPR-binding sites (53).Similarly, CEM T lymphoblasts exhibited a 75% decease in 2',3'-

dideoxyctidine sensitivity when transmembrane nucleoside transportactivity was decreased by 80% (35).

The mechanisms of gemcitabine permeation have not been defined.The present study demonstrated that transport-incompetent cells, pro

duced either by mutation or by pharmacological inhibition of NTactivity, were resistant to gemcitabine inhibition, and in some celllines, the lack of mediated uptake conferred greater than 1000-fold

resistance to the antiproliferative action of gemcitabine. These observations suggested that passive diffusion of gemcitabine through

Table 4 /" H'¡Cemcitabine initial uptake rates b\ HeLtt fells in the presence and

absence of inhibitors of nucleoside transponHeLa cells were exposed to 10 ^.M [3H]gemcitabine in the presence or absence of 100

nm NBMPR or 10 JIM dilazep. Initial uptake rales were determined at room temperatureusing short gemcitabine exposures terminated by inhibitor/cold stop methods. The finalvalues represent the mean ±SE from four independen! experiments, each conducted intriplicate. Uptake rates (pmol/s/106 cells) plus SE are recorded.

Uptake(no inhibitors)0.504

0.2140.3310.387

0.359 ±0.060Uptake

(NBMPR)0.110

0.0780.0450.161

0.099 ±0.024Uptake

(dilazep)000

00

= a>

||jS »üiE ooíEgì

1.00,

0.75.

0.50.

025-

0.00.O 2"5 5"0 7*5 100 125 150

time (s)Fig. 6. Time course of ['HJgemcitahine uptake by proteoliposomes prepared from

detergenl-solubilized membrane fractions from BeWo cells. Proteoliposomes were prepared from OCTG-solubilizcd membrane fractions that were isolated by density-gradientcentrifugation in 25% sucrose as described in "Materials and Methods." The proteoliposomes were incubated with 20 (ÃŒM['H|gemcitabine for the indicated times in the presence

(O. non-mediated uptake) and absence (•.total uptake) of 10 nut adenosine. 10 (ÌM

NBTGR. and 10 fiM dipyridamolc. Values for mediated uptake (•)were calculated fromthe difference between total and nonmedialed uptake. Assays were conducted in duplicateand plots from one of four separate experiments are shown.

plasma membranes was a relatively minor component of gemcitabineuptake. By studying gemcitabine-dependent inhibition of proliferation

in a panel of murine and human cell lines with defined NT activities,we were able to infer that gemcitabine uptake was mediated by the es,ei, cit, and cib NTs but not by the cif NT. The es, ei, and cib NTsexhibit broad permeant selectivity, accepting both purine and pyrim-

idine nucleosides, whereas the cit and cif NTs accept uridine andexhibit selectivity, respectively, for pyrimidine and purine nucleosides(18).

To confirm the transportability of gemcitabine suggested by thegrowth-inhibition studies, we performed detailed kinetic studies of[3H]gemcitabine uptake in human cells. Initial rates of uptake were

measured in cells in which NT activity was absent or in which only asingle NT activity was present (either naturally or by transient trans-

fection with the relevant cDNA). The results of the transport studieswere entirely concordant with those of the growth-inhibition studies.

Diffusion was not a measurable component of gemcitabine cellularinflux, and gemcitabine uptake was mediated by several NT activities.However, the NT activities exhibited by the various cell lines variedmarkedly in their apparent efficiencies of gemcitabine permeation: es= cit > ei > cib »> cif. Because the absolute numbers of

functional transporters (which probably differ among the various celltypes) could not be determined, it was not possible to directly comparethe intrinsic transportability of gemcitabine by the different NTs.

Uridine is accepted as a permeant by all of the known human NTsand was therefore used as a standard with which to compare NT-

mediated gemcitabine uptake. Initial uptake rates for a concentration(10 /XM)that is therapeutically relevant for gemcitabine and physiologically relevant for uridine were determined for the five NT activities (see Table 3). The ratios of gemcitabine:uridine uptake ratevaried among the NTs, and gemcitabine rates were consistently lessthan uridine rates for the es, ei, cit, and cib NTs. The cif NT, whichmediates transport of uridine and purine nucleosides, includingcladribine (54), did not mediate gemcitabine influx.

Because the equilibrative NTs are bidirectional, net gemcitabineuptake would be expected to represent the combined contributions ofNT-mediated influx and efflux. Therefore, the antiproliferative action

of gemcitabine might be potentiated by inhibitors of equilibrativetransport, if a concentrative NT was also present. This possibility wasexplored in the CaCo-2 cell line, which possesses es, ei, and cib;however, dipyridamole exposure did not potentiate gemcitabine cy-

totoxicity, suggesting that drug efflux by the equilibrative NTs wasnot a major component of net gemcitabine retention.

4355

on June 29, 2020. © 1998 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

GEMCITABINE CYTOTOXICITY AND NUC'LEOSIDE TRANSPORT

The pharmacokinetic properties of gemcitabine suggest that efficient uptake of gemcitabine may be required for anticancer efficacy invivo. Gemcitabine is typically administered to humans as an i.V. bolusinfusion lasting 30 min, in doses ranging from 800 to 1200 mg/m2.

The most common dosage schedule is weekly administration on thefirst, eighth, and fifteenth day of a 28-day cycle. Peak serum gemcit

abine concentrations are dose dependent and generally do not exceed50 /IM (55). However, after bolus injections, enzyme-mediated deami-

nation causes a rapid decrease in plasma concentrations, and within2 h, the plasma concentration of the parent drug falls below quantifiable levels (56). These observations suggest that cells in vivo may beexposed to significant gemcitabine concentrations for only short periods and raise the possibility that inefficient cellular uptake may be amechanism underlying the observed resistance to gemcitabine in somesolid tumors.

Although gemcitabine was shown to be a substrate for four NTs, themajor mediators of gemcitabine uptake in human tissues are probablythe equilibrative NTs, because human cil activity has, thus far, onlybeen demonstrated in kidney, liver, and intestinal epithelium (22, 57),and human cib activity has only been demonstrated in human myeloidleukemic cell lines, freshly isolated myeloblasts (46), and the coloncancer cell line used in this report, CaCo-2 (24). By implication,

deficiency of equilibrative NT in tumor cells may be a mechanism ofgemcitabine resistance in vivo. Given that both deoxycytidine kinasedeficiency and NT deficiency confer high-level resistance to gemcit

abine cytotoxicity in vitro, either mechanism could be the basis of theacquired or de novo resistance to gemcitabine observed in clinicaltrials.

In summary, we have demonstrated that functional NTs are required for manifestation of gemcitabine toxicity in vitro. Conversely,a deficiency in nucleoside-transport activity was capable of conferringhigh-level resistance to the toxicity of gemcitabine. We have per

formed detailed kinetic studies of gemcitabine uptake in human celllines exhibiting a range of NT activities and have demonstratedmarked variability in the capacity of the different NTs to mediategemcitabine influx. Finally, we have demonstrated mediated uptake ofgemcitabine by the human o NT reconstituted in proteoliposomes.Although the potential for clinical application of these observations ishigh, efforts to rationally modulate the antitumor and normal tissuetoxicities of gemcitabine must await a clearer understanding of thetumor and tissue distribution of the NT proteins. The recent cloning ofcDNAs encoding the four major human NTs ( 19-23)4 has led to new

molecular and immunological probes to study NT distribution which,in turn, may lead to transport-based strategies capable of improving

the therapeutic indices of anticancer nucleosides.

ACKNOWLEDGMENTS

We acknowledge Dr. Kathryn Graham and Thack Lang for the provision of

expression constructs used in this work.

REFERENCES

10.

15.

16.

18.

19.

20.

1. Blackslcin. M.. Vogel, C. L., and Amhiner. R. Phase 11study of gemcitabine inpatients with mctastalic breast eancer. Proc. Am. Soc. Clin. Oncol.. /5: 135. 1996. 28.

2. Cantilenaci. J.. Possinger. K.. Phillip. P.. Beykirch, M.. Kerr. H.. Walling, J.. andHarris. A. L. Advanced breast cancer: a phase II trial with gemcitabine. J. Clin.Oncol., 13: 2731-2736. 1995. 29.

3. Pollera. C. F.. Cerihelli. A.. Crecco. M.. and Calabresi. F. Weekly gemcitabine inadvanced bladder cancer: a preliminary report from a phase I study. Ann. Oncol.. 5:182-184, 1994.

4. Lund. B.. Hansen. O. P.. Thcilade. K.. Hansen. M., and Neijt. J. P. PhaseII study of 30.gemcitabine (2'.2'-difluor(xleoxycytidine) in previously treated ovarian cancer pa

tients. J. Nail. Cancer Inst.. M: I530-1533, 1994.5. Anderson. H . Lund. B.. Bach. F.. Thalcher. N.. Walling. J.. and Hansen. H. H. 31.

Single-agent activity of weekly gemcitabine in advanced non-small-cell lung cancer:a phase II study. J. Clin. Oncol.. 12: 1821-1826. 1994.

4356

Ahratt. R. P.. Bezwoda. W. R.. Falkson. G.. Goedhals. L.. Hacking. D.. and Rugg.T. A. Efficacy and safety profile of gemcitabine in non-small-cell lung cancer: a phaseH study. J. Clin. Oncol., 12: 1535-1540. 1994.Casper. E. S.. Green. M. R.. Kelsen. D. P.. Heelan. R. T.. Brown. T. D.. Flombaum.C. D.. Trochanowski. B.. and Tarassoff. P. G. Phase II trial of gemcitabine (2,2'-

difluorodeoxycytidine) in patients with adenocarcinoma of the pancreas.Invest. NewDrugs. 12: 29-34, 1994.

Moore. M. Activity of gemcitabine in patients with advanced pancreatic carcinoma.Cancer (Phila.). 78: 633-638. 1996.Catimel. G.. Vermorken. J. B.. Clavel. M., de Mulder. P.. Judson. I.. Sessa, C.,Piccarl. M., Bruntsch. U.. Verweij. J.. Wanders, J.. Franklin. H.. and Kay. S. B. Aphase II study of gemcitabine (LY 188011) in patients with advanced squamouscellcarcinoma of the head and neck. Ann. Oncol.. 5: 543-547. 1994.Hcinemann. V.. Hertel. L. W.. Grindey. G. B.. and Plunkett. W. Comparison of thecellular pharmacokinelics and toxicity of 2',2'-difluorodeoxycytidine and l-ß-i>

arabinofuranosylcytosine. Cancer Res.. 48: 4024-4031. 1988.Ruiz van Haperen. V. W.. Veerman. G., Vermorken. J. B.. and Peters, G. J. 2',2'-

Difluoro-deoxycytidine (gemcitabine) incorporation into RNA and DNA of tumourcell lines. Biochem. Pharmacol.. 46: 762-766. 1993.

Plunkett. W.. Huang. P.. Xu. Y. Z.. Heinemann. V.. Grunewald. R.. and Gandhi. V.Gemcitabine: metabolism, mechanisms of action, and self-potentiation. Semin. Oncol., 22: 3-10. 1995.

Fink. U.. Molle. B., and Daschner. H. Phase II study of gemcitabine in metastatiecolorectal cancer. Proc. Am. Soc. Clin. Oncol.. Il: A507. 1992.Ruiz van Haperen, V. W., Veerman. G.. Eriksson, S„Boven. E., Stegmann, A. P.,Hermsen. M.. Vermorken. J. B.. Pinedo. H. M., and Peters. G. J. Development andmolecular characterization of a 2'.2'-difluorodcoxycytidine-rcsistant variant of the

human ovarian carcinoma cell line A2780. Cancer Res.. 54: 4138-4143. 1994.Neff, T.. and Blau. C. A. Forced expression of cytidine deaminase confers resistanceto l-/3-r>-arabinofuranosylcytosine and gemcitabine. Exp. Hemalol.. 24: 1340-1346.1996.Sliutz. G.. Karlseder. J.. Tempfer. C.. Orel. L.. Holzer. G., and Simon. M. M. Drugresistance against gemcitabine and topotecan mediated by constitutive hsp70 over-expression in vitro: implication of qucrcetin as sensitiser in chemotherapy. Br. J.Cancer, 74: 172-177. 1996.

Waud. W. R., Gilbert. K. S.. Grindey. G. B., and Worzalla. J. F. Lack of in vimcrossresistancewith gemcitabine against drug-resistant murine P388 leukemias. Cancer Chemother. Pharmacol.. 38: 178-180. 1996.

Cass. C. E. Nucleoside transport. In: N. H. Georgopapadakou (ed.). Drug Transportin Antimicrobial Therapy and Anticancer Therapy, pp. 403-451. New York: MarcelDekker. 1995.Griffiths. M.. Beaumont. N., Yao, S. Y., Sundaram. M.. Boumah. C. E.. Davies. A..Kwong. F. Y.. Coe, I., Cass, C. E., Young, J. D.. and Baldwin. S. A. Cloning of ahuman nucleosidc transporter implicated in the cellular uptake of adenosine andchemolherapeutic drugs. Nat. Med.. 3: 89-93. 1997.Crawford. C. R.. Palei. D. H.. Naeve. C.. and Belt. J. A. Cloning of the humanequilibrative. nitrobenzylmercaptopurine riboside (NBMPR I-insensitive nucleosidetransporter ei by functional expression in a transport-deficient cell line. J. Biol.Chem.. 273: 5288-5293. 1998.Griffiths. M.. Yao, S. Y. M., Abidi. F.. Phillips, S. E. V.. Cass. C. E.. Young. J. D..and Baldwin. S. A. Molecular cloning and characterization of a nilroben/ylthioi-nosine-insensitive (ei) equilibrative nucleoside transporter from human placenta.Biochem. J.. 328: 739-743. 1997.Ritzel. M. W. L.. Yao. S. Y. M.. Huang. M. Y.. Elliott. J. F., Cass.C. E., and Young.J. D. Molecular cloning and functional expression of cDNAs encoding a humanNa*-nucleoside cotransporter (hCNTl). Am. J. Physiol.. 272: C707-C7I4. 1997.

Wang. J.. Su, S-F.. Dresser. M. J.. Schnaner. M. E.. Washington. C. B., andGiacomini. K. M. Na ' -dependent purine nucleoside transporter from human kidney:

cloning and functional characterization. Am. J. Physiol.. 27.?; F1058-F1065. 1997.Belt. J. A.. Marina. N. M.. Phelps. D. A., and Crawford. C. R. Nucleoside transportin normal and neoplastic cells. Adv. Enzyme Regul.. 33: 235-252. 1993.Vijayalakshmi. D.. Dagnino. L.. Belt, J. A.. Gali. W. P., Cass. C. E.. and Palerson,A. R. L1210/B23.1 cells expressequilibrative. inhibitor-sensitive nucleoside transportactivity and lack two parental nucleoside transport activities. J. Biol. Chem.. 267:16951-16956. 1992.Boumah. C. E.. Hogue. D. L.. and Cass, C. E. Expression of high levels of nitroben-zyllhioinosine-scnsitive nucleoside transport in cultured human choriocarcinoma(BeWo) cells. Biochem. J.. 288: 987-996. 1992.Boleti. H.. Coe. I. R.. Baldwin. S. A.. Young. J. D.. and Cass, C. E. Molecularidentification of the equilibrative NBMPR-sensilive (es) nucteoside transporter anddemonstration of an equilihrative NBMPR- insensilive (ei) transport activity inhuman erythroleukemia (K562) cells. Neuropharmacology. 36: 1167-1179. 1997.Belt. J. A., and Noel. L. D. Nucleoside transport in Walker 256 rat carcinosarcomaand S49 mouse lymphoma cells. Differences in sensitivity to nitrobenzylthioinosineand thiol reagents. Biochem. J.. 2.J2: 681-688. 1985.Cass. C. E.. Kolassa. N.. Uehara. Y.. Dahlig-Harlcy. E.. Harley. E. R.. and Paterson.A. R. Absence of binding sites for the transport inhibitor nitrobenzylthioinosine onnueleoside transport-deficient mouse lymphoma cells. Biochim. Biophys. Acta, 649:769-777. 1981.Crawford, C. R., Ng, C. Y.. Noel, L. D., and Belt. J. A. Nucleoside transport in LI210murine leukemia cells. Evidence for three transporters. J. Biol. Chem.. 265: 9732-9736. 1990.Crawford, C. R.. Ng. C. Y.. and Belt, J. A. Isolation and characterization of an L1210cell line retaining the sodium-dependent carrier cif as its sole nueleoside transportactivity. J. Biol. Chem.. 265: 13730-13734. 1990.

on June 29, 2020. © 1998 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

OEMCITABINE CYTOTOXICITY AND NUCLEOSIDE TRANSPORT

32. Huang. Q. Q., Yao. S. Y.. Ritzel, M. W.. Paterson. A. R., Cass, C. E., and Young, J. D.Cloning and functional expression of a complementary DNA encoding a mammaliannucleoside transport protein. J. Biol. Chem.. 269: 17757-17760, 1994.

33. Crawford. C. R., Cass. C. E.. Young, J. D.. and Bell, J. A. Stable expression of arecombinant sodium-dependent, pyrimidine-selective nucleoside transporter (rCNTI )in a transport-deficient mouse leukemia cell line. Biochem. Cell Biol.. in press, 1998.

34. Crawford, C. R., Ng, C. Y., Ullman. B., and Belt, J. A. Identification and reconstitution of the nucleoside transporter of CEM human leukemia cells. Biochim. Biophys.Acta, 1024: 289-297, 1990.

35. Ullman. B. Dideoxycytidine metabolism in wild type and mutant CEM cells deficientin nucleoside transport or deoxycytidine kinase. Adv. Exp. Med. Biol., 253B: 415-

420, 1989.36. Pattillo, R. A., and Gey, G. O. The establishmenl of a cell line of human hormone-

synthesizing trophoblastic cells in vitro. Cancer Res., 28: 1231-1236, 1968.37. Burres, N. S., and Cass, C. E. Density-dependent inhibition of expression of syncy-

tiotrophoblastic markers by cultured human choriocarcinoma (BeWo) cells. J. CellPhysiol., 128: 375-382, 1986.

38. Skehan, P.. Storeng, R.. Scudiero, D., Monks, A., McMahon, J., Vistica. D., Warren.J. T., Bokesch. H., Kenney. S., and Boyd. M. R. New colorimetrie cylotoxicity assayfor anticancer-drug screening. J. Nati. Cancer Inst., 82: 1107-1112, 1990.

39. Rubenstein. L. V.. Shoemaker, R. H.. Paull, K. D., Simon, R. M., Tosini, S.,Skehan. P., Scudiero, D. A.. Monks. A., and Boyd, M. R. Comparison of in vitroanticancer-drug-screening data generated with a tetrazolium assay versus a protein assay against a diverse panel of human tumor cell lines. J. Nati. Cancer Inst.,82: 1113-1118, 1990.

40. Fang, X., Parkinson, F. E., Mowles, D. A., Young, J. D., and Cass, C. E. Functionalcharacterization of a recombinant sodium-dependent nucleoside transporter with

selectivity for pyrimidine nucleosides (cNTIrat) by transient expression in culturedmammalian cells. Biochem. J., 317: 457-465, 1996.

41. Mani. R. S.. Hammond. J. R.. Young. J. D.. and Cass, C. E. Demonstration ofhigh-affinity binding of nitrobenzylthioinosine and nitrobenzylthioinosine-sensitivem him.- transport in nuclear membranes of cultured human choriocarcinoma (BeWo)

cells by functional reconstitution in proteoliposomes. J. Biol. Chem., in press. 1998.42. Markwell. M. A. K.. Hass, S. M., Tolbert, N. E.. and Bieber. L. L. Protein determi

nation in membrane and lipoprotein samples: manual and automated procedures.Methods Enzymol.. 72: 296-303. 1981.

43. Hammond, J. R. Functional reconstitution of pharmacologically distinct subtypes ofnucleoside transporters in liposomal membranes. J. Pharmacol. Exp. Ther., 277:906-917, 1994.

44. Mimms. L. T.. Zampighi. G., Nozaki. Y.. Tanford. C., and Reynolds, J. A. Phospho-lipid vesicle formation and transmembrane protein incorporation using octylglu-coside. Biochemistry, 20: 833-840, 1981.

45. Tse, C. M., and Young, J. D. Passive and carrier-mediated permeation of different

nucleosides through the reconstituted nucleoside transporter. Biochim. Biophys. Acta,985: 343-346. 1989.

46. Belt, J. A.. Harper. E. H.. Byl. J. A., and Noel. L. D. Sodium-dependent nucleoside

transport inhuman myeloid leukemic cell lines and freshly isolated myeloblasls. Proc.Am. Assoc. Cancer Res.. 33: 20, 1992.

47. Dagnino, L., and Paterson, A. R. Sodium-dependent and equilihrative nucleosidetransport systems m I L'I u mouse leukemia cells: effect of inhibitors of equilibrative

systems on the content and retention of nucleosides. Cancer Res.. 50: 6549-6553.

1990.48. Cass, C. E.. King, K. M., Montano, J. T., and Janowska-Wieczorek. A. A comparison

of the abilities of nitrobenzylthioinosine, dilazep, and dipyridamole to protect humanhematopoietic cells from 7-deazaadenosine (tubercidin). Cancer Res.. 52: 5879-

5886. 1992.49. White, J. C.. Rathmell. J. P.. and Capizzi. R. L. Membrane transport influences the

rate of accumulation of l-j3-D-arabinofuranosylcytosine in human leukemia cells.J. Clin. Invest., 79: 380-387. 1987.

50. Wiley, J. S.. Jones, S. P., Sawyer, W. H., and Paterson, A. R. I-/3-D-Arabinofurano-sylcytosine influx and nucleoside transport sites in acute leukemia. J. Clin. Investig..69: 479-489. 1982.

51. Gali, W. P., Paterson. A. R., Larra»,L. M.. Turner. A. R., and Belch. A. R. Sensitivityof acute leukemia cells to cytarabine is a correlate of cellular es nucleoside transportersite content measured by flow cylometry with SAENTA-fluorescein. Blood. 90:346-353, 1997.

52. Wiley, J. S.. Woodruff, R. K., Jamieson, G. P.. Firkin. F. C.. and Sawyer, W. H.I-/3-D-Arabinofuranosylcytosine in the treatment of T-cell acute lymphoblastic leukemia. Aust. NZ J. Med., 17: 379-386. 1987.

53. Hoffman. J. Murine erythroleukemia cells resistant to periodate-oxidized adenosinehave lowered levels of nucleoside transporter. Adv. Exp. Med. Biol., 309A: 443-446.

1991.54. King, K. M., and Cass, C. E. Membrane transport of 2-chloro-2'-deoxyadenosine and

2-chloro-2'-arabinofluoro-2'-deoxyadenosine is required for cytotoxicity. Proc. Am.

Assoc. Cancer Res.. 35: A3436. 1994.55. Cheson, B. D. Miscellaneous chemotherapeutic agents. In: V. T. Devita (ed.) Cancer:

principles and practice of oncology. Ed. 5, Vol. 1, pp. 490-498. Philadelphia:Lippincott-Raven, 1997.

56. Allerheiligen, S., Johnson, R.. and Hatcher, B. Gemcitabine pharmacokinetics areinfluenced by gender, body surface area, and duration of infusion. Proc. Am. Soc.Clin. Oncol., 13: A136. 1994.

57. Patii, S. D., and Unadkat, J. D. Sodium dependent nucleoside transport in the humanintestinal brush border membrane. Am. J. Physiol. Castro. Liver Physiol. 35: Gl 314-

G1320. 1997.

4357

on June 29, 2020. © 1998 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1998;58:4349-4357. Cancer Res   John R. Mackey, Rajam S. Mani, Milada Selner, et al.   LinesGemcitabine Influx and Manifestation of Toxicity in Cancer Cell Functional Nucleoside Transporters Are Required for

  Updated version

  http://cancerres.aacrjournals.org/content/58/19/4349

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/58/19/4349To request permission to re-use all or part of this article, use this link

on June 29, 2020. © 1998 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from