Nucleoside transport and its significance for anticancer drug resistance

15
Nucleoside transport and its significance for anticancer drug resistance John R. Mackey,' Stephen A. Baldwin, 2 James D. Youngfl Carol E. Cass~ 'Department of Oncology, University of Alberta, and Cross Cancer Institute, Edmonton,Alberta, Canada;2Schoolof Biochemistry and Molecular Biology, University of Leeds, Leeds, UK; 3Department of Physiology, University of Alberta, Edmonton, Canada AbstractThis article discusses the role of nucleoside transport processes in the cytotoxicity of clinically important anticancer nucleosides.This article summarizes recent advances in the molecular biology of nucleoside transport proteins, review the current state of knowledge of the transportability of therapeutically useful anticancer nucleosides, and provide an overview of the role of nucleoside transport deficiency as a mechanism of resistance to nucleoside cytotoxicity are summarized. Several strategies for utilization of nucleoside transport processes to improve the therapeutic index of anticancer therapies, including the use of nucleoside-transport inhibitors to modulate toxicity of both nucleoside and non- nucleoside antimetabolite drugs are also presented. importance of nucleoside anticancer therapy has :teased recently as a result of the introduction of o_¢eral new nucleoside drugs into clinical use and the expansion of indications for anticancer nucleoside ther- apy into the arena of solid tumors. In parallel with these developments, there has been rapid progress in the under- standing of nucleoside transport (NT) processes, leading to the cloning and functional expression of cDNAs encoding the four major NT proteins of mammalian cells. In this review, recent developments are highlighted in the molecu- lar biology of NTs and the growing tmderstanding of the rela- tionship of NT to the problem of resistance to nucleoside drugs. Evidence is summarized linking membrane transport processes to the cytotoxic selectivity of nucleoside drugs and the concept of the 'nucleoside transport profile' as a determinant of anticancer nucleoside activity is introduced. Strategies for increasing the therapeutic index of nucleoside anticancer drugs by manipulation of NTs processes to achieve "pharmacologic resistance" of dose-limiting normal tissues are outlined, and the preclinical and clinical experi- ence with modulation of nucleoside salvage by inhibition of NT processes during antimetabolite therapy is reviewed. NUCLEOSIDE TRANSPORT PROCESSES OF HUMAN CELLS Seven functionally distinct NT processes have been described in human cells, of which four have been defined in molecular terms, through isolation and functional expression of cDNAs encoding the transporter proteins.l-6 The classifica- tion of NT activities is based on functional and pharmacolog- ical characteristics, including transport mechanisms (either equilibrative or concentrative), permeant selectivities and sensitivity to nanomolar concentrations of nucleoside and non-nucleoside inhibitors. 7-1~ The concentrative NTs are inwardly directed sodium/nucleoside symporters, which are capable of moving nucleosides against the concentration gradient through coupled movement of sodium down. its transmembran.e electrochemical gradient. The equilibrative (bidirectional) transporters accept both purine and pyrimi- dine nucleosides as permeants and are found in most, possi- bly all, cell types, whereas the concentrative transporters have relatively narrow selectivities for nucleoside permeants and are limited to specialized cell types. Equllibrative trans- porters have been subdivided on the basis of sensitivit 3, to nitrobenzylthioinosine (NBMPR: nitrobenzylmercaptopurine ribonucleoside; 6-[(4 nitrobenzyl)thio]-9-B-D-ribufuranosyl purine); one subtype, termed 'equilibrative sensitive' (es) is inhibited by nanomolar concentrations of NBMPR, as a result of high-affinity (Kd < 5 nM) binding of NBMPR at or near the permeant-binding site, whereas the other subtype, termed 'equllibrative insensitive' (et) is unaffected by low concentra- tions (< 1 btM) of NBMPR. Concentrative NTs comprise sev- eral functional subtypes that differ in their substrate selectivities and tissue distributions and, with some excep- tions, are unaffected by high concentrations (< 10 btM) of NBMPR. The chemical structure of NBMPR, together with those of the other classic inhibitors of equilibrative NT processes (dilazep, dipyridamole, draflazine), are shown in Figure 1. Understanding of relationships among NTs has been greatly advanced by the recent isolation and functional expression of cDNAs encoding rat (r) and human (h) nucleo- side transporter (NT) proteins.These NT proteins comprise two structurally unrelated protein families that are desig- nated ENT and CNT, depending on whether they mediate, respectively, equilibrative (E) or concentrative (C) NT processes. The two protein families differ substantially in their architectural design, based on predictions of topologi- cal orientation and organization of the transporter proteins in membranes (Fig. 2). The mammalian ENT proteins identi- fied thus far have approximately 450 amino acid residues 1.3,4,6,~3 and 11 predicted transmembrane domains, whereas the mammalian CNT proteins have approximately 650 amino acid residues and 14 predicted transmembrane domainsfl.5.1416 Members of the ENT family with functional characteristics of es-mediated transport processes are desig- nated ENT1 and those with functional characteristics of el- mediated transport processes are designated ENT2. Members of the CNT family with preference for pyrimidine nucleo- sides are designated CNT1 and those with preference for purine nucleosides are designated CNT2. In this nomencla- ture, the numbers represent the chronological order of dis- covery within a transporter subfamily, and a lower case letter precedes the transporter designation (e.g. hENT1, rENT1) to indicate the origin of the transporter protein (e.g. human, rat). The es transporter (hENT1) The first member of the ENT membrane protein superfamily, the human e quilibrative n_ucleoside transporter 0aENT1), was isolated from a human placental cDNA library by using Drug Resistance Updates (I 998) I, 310-324 © 1998 Harcourt Brace & Co. Lid

Transcript of Nucleoside transport and its significance for anticancer drug resistance

Nucleoside transport and its significance for anticancer drug resistance

John R. Mackey,' Stephen A. Baldwin, 2 James D. Youngfl Carol E. Cass ~

'Department of Oncology, University of Alberta, and Cross Cancer Institute, Edmonton,Alberta, Canada;2School of Biochemistry and Molecular Biology, University of Leeds, Leeds, UK; 3Department of Physiology, University of Alberta, Edmonton, Canada

AbstractThis article discusses the role of nucleoside transport processes in the cytotoxicity of clinically important anticancer nucleosides.This article summarizes recent advances in the molecular biology of nucleoside transport proteins, review the current state of knowledge of the transportability of therapeutically useful anticancer nucleosides, and provide an overview of the role of nucleoside transport deficiency as a mechanism of resistance to nucleoside cytotoxicity are summarized. Several strategies for utilization of nucleoside transport processes to improve the therapeutic index of anticancer therapies, including the use of nucleoside-transport inhibitors to modulate toxicity of both nucleoside and non- nucleoside antimetabolite drugs are also presented.

importance of nucleoside anticancer therapy has :teased recently as a result of the introduction of

o_¢eral new nucleoside drugs into clinical use and the expansion of indications for anticancer nucleoside ther- apy into the arena of solid tumors. In parallel with these developments, there has been rapid progress in the under- standing of nucleoside transport (NT) processes, leading to the cloning and functional expression of cDNAs encoding the four major NT proteins of mammalian cells. In this review, recent developments are highlighted in the molecu- lar biology of NTs and the growing tmderstanding of the rela- tionship of NT to the problem of resistance to nucleoside drugs. Evidence is summarized linking membrane transport processes to the cytotoxic selectivity of nucleoside drugs and the concept of the 'nucleoside transport profile' as a determinant of anticancer nucleoside activity is introduced. Strategies for increasing the therapeutic index of nucleoside anticancer drugs by manipulation of NTs processes to achieve "pharmacologic resistance" of dose-limiting normal tissues are outlined, and the preclinical and clinical experi- ence with modulation of nucleoside salvage by inhibition of NT processes during antimetabolite therapy is reviewed.

NUCLEOSIDE TRANSPORT PROCESSES OF HUMAN CELLS

Seven functionally distinct NT processes have been described in human cells, of which four have been defined in molecular terms, through isolation and functional expression of cDNAs encoding the transporter proteins.l-6 The classifica- tion of NT activities is based on functional and pharmacolog- ical characteristics, including transport mechanisms (either

equilibrative or concentrative), permeant selectivities and sensitivity to nanomolar concentrations of nucleoside and non-nucleoside inhibitors. 7-1~ The concentrative NTs are inwardly directed sodium/nucleoside symporters, which are capable of moving nucleosides against the concentration gradient through coupled movement of sodium down. its transmembran.e electrochemical gradient. The equilibrative (bidirectional) transporters accept both purine and pyrimi- dine nucleosides as permeants and are found in most, possi- bly all, cell types, whereas the concentrative transporters have relatively narrow selectivities for nucleoside permeants and are limited to specialized cell types. Equllibrative trans- porters have been subdivided on the basis of sensitivit 3, to nitrobenzylthioinosine (NBMPR: nitrobenzylmercaptopurine ribonucleoside; 6-[(4 nitrobenzyl)thio]-9-B-D-ribufuranosyl purine); one subtype, termed 'equilibrative sensitive' (es) is inhibited by nanomolar concentrations of NBMPR, as a result of high-affinity (K d < 5 nM) binding of NBMPR at or near the permeant-binding site, whereas the other subtype, termed 'equllibrative insensitive' (et) is unaffected by low concentra- tions (< 1 btM) of NBMPR. Concentrative NTs comprise sev- eral functional subtypes that differ in their substrate selectivities and tissue distributions and, with some excep- tions, are unaffected by high concentrations (< 10 btM) of NBMPR. The chemical structure of NBMPR, together with those of the other classic inhibitors of equilibrative NT processes (dilazep, dipyridamole, draflazine), are shown in Figure 1.

Understanding of relationships among NTs has been greatly advanced by the recent isolation and functional expression of cDNAs encoding rat (r) and human (h) nucleo- side transporter (NT) proteins.These NT proteins comprise two structurally unrelated protein families that are desig- nated ENT and CNT, depending on whether they mediate, respectively, equilibrative (E) or concentrative (C) NT processes. The two protein families differ substantially in their architectural design, based on predictions of topologi- cal orientation and organization of the transporter proteins in membranes (Fig. 2). The mammalian ENT proteins identi- fied thus far have approximately 450 amino acid residues 1.3,4,6,~3 and 11 predicted transmembrane domains, whereas the mammalian CNT proteins have approximately 650 amino acid residues and 14 predicted transmembrane domainsfl.5.1416 Members of the ENT family with functional characteristics of es-mediated transport processes are desig- nated ENT1 and those with functional characteristics of el-

mediated transport processes are designated ENT2. Members of the CNT family with preference for pyrimidine nucleo- sides are designated CNT1 and those with preference for purine nucleosides are designated CNT2. In this nomencla- ture, the numbers represent the chronological order of dis- covery within a transporter subfamily, and a lower case letter precedes the transporter designation (e.g. hENT1, rENT1) to indicate the origin of the transporter protein (e.g. human, rat).

The e s t r a n s p o r t e r (hENT1) The first member of the ENT membrane protein superfamily, the human e quilibrative n_ucleoside transporter 0aENT1), was isolated from a human placental cDNA library by using

Drug Resistance Updates (I 998) I, 310-324 © 1998 Harcourt Brace & Co. Lid

Inhibitors of equilibrative nucleoside transport

o2 -iH2

HO~,~ H 0 H NBMPR

• ~N. N C CH2CIuI20H

N~ y "7" ~.~o.~o. HOCHxCH2% l [ l [

HOCH2Cn: ) Nt '~N 4 J ' ~ N

Dipytidamo/e

CH3Ox ~ OCH~

c H 3 ° ~ c ° ° ( c n 9 3 ~N,~_~N -'(CH2)~ OOc - @ O C H 3

CHfl>" ~.OCH3 Dil~p

H2 O I ~

C

Drntlazine

~ 2

"Z CONH(CH2)sCONH(CH2)2S~ H2 I

Pyrimidine anticancer nucleoside analogs NH2

O H O

SAENTA-tluorescein

NH2 Purine anticancer nucleoside analogs

HO O

H H

Cytarabine Cladribine

NH2

o:+ O

H F

Gemcitabine oH v

NH2

0 H HO

Fludarabine

Chemical structures of transport inhibitors and nucleoside analogs.

© 1998 Harcourt Brace & Co. Ltd Drug Resistance Updates (I 998) I, 310-324

CNT Family

N H 2 ~ C O O H

ENT Family

~ COOH

Fig. 2 The concentrative nucleoside transporter and equilibrative nucleoside transporter protein families.

amino acid sequence information obtained from N-terminal sequencing of the purified human erythrocyte es trans- porter. 1 The hENT1 cDNA encodes a 50-kDa protein, com- prised of 456 amino acids wi th 11 predicted t ransmembrane domains and three potential glycosylation sites of which one site lies in the predicted extracellular loop be tween trans- membrane domains 1 and 2. The native human erythrocyte es transporter is known to be glycosylated near one of its ends. .7 hENT1 possesses a number of potential phosphoryla- tion consensus sites, although it has not been determined if hENT1 is phosphorylated in vivo.The hENT1 gene has been localized to chromosome 6p21.1-p21.2) 8 The rat homo- logue (rENT1), which has 457 amino acids, also has 11 pre- dicted transmembrane domains but contains three potential glycosylation sites in the loop be tween transmembrane domains 1 and 2. ~3 hENT1 and rENT1 are highly conserved, wi th 78% identity and 88% similarity at the amino acid level.

Recombinant hENT1 has been produced in Xenopus lae- vis oocytes m and shown to mediate transport of uridine (Km

value, 0.24 raM) (Table 1). Uridine transport in hENTl-con- taining oocytes is inhibited by physiological purine and pyrimidine nucleosides (but not by uracil) and by cladribine, cTtarabine, fludarabine and gemcitabine (for drug structures, see Fig. 1). Uridine transport is also inhibited by NBMPR (ICs0 value, 3.6 nM) and by the vasoactive drugs, dipyridamole, dilazep and draflazine (for structures, see Fig. 1). Chimeric studies in which various t ransmembrane domains of hENT1 and rENT1 have been interchanged have established that dipyridamole and dilazep (and thus, probably, also NBMPR) bind to a region of hENT1 be tween transmembrane domains 3 and 6. ~9 The tissue distribution of hENT1 is broad, suggest- ing that the transporter may be ubiquitous; hENT1 tran- scripts have been observed in several human cancer cell lines probed with the bENT1 cDNA 2° and expressed sequence tags wi th identity to hENT1 from many different normal and neoplastic human tissues have been registered in the genomic data bases.

The e i transporter (hENT2) A second human equilibrative nucleoside transporter (hENT2) with ei-type activity has been identified by isolation

and functional expression in Xenopus oocytes of a hENT1- related cDNA from human placenta. ~ hENT2, a 50-kDa pro- tein, consists of 456 amino acids and is 49% identical and 69% similar to hENT1, hENT2 also has 11 predicted trans- membrane domains.Two of its three potential glycosylation sites are on the predicted extracellular loop be tween trans- membrane domains 1 and 2. An identical transporter was identified in cultured HeLa ceils by functional expression cloning of a HeLa cDNA in a nucleoside-transport deficient leukemia cell line. 3 A related rat homolog (rENT2), which was isolated from a rat jejunal cDNA library, 13 has 456 amino acids, 11 predicted transmembrane domains and two poten- tial glycosylation sites in the extracellular loop be tween transmembrane domains 1 and 2. bENT2 and rENT2 are 93% similar and 88% identical at the amino acid level.

Recombinant hENT2 has been functionally characterized in Xenopus oocytes 4 and in a NT-defective line of cultured human leukemia cells. 3 The recombinant protein 's insensitiv- ity to inhibition by NBMPR and relatively broad permeant selectivity led to the functional classification of bENT2 as an ei-type transporter. In oocytes, bENT2 exhibits a K m value for uridine transport of 0.20 mM (Table 1). Studies undertaken to identify potential permeants and/or inhibitors, in which 1 mM test compound was assessed for its ability to block inward transport of 10 btM 3H-uridine,3 demonstrated com- plete inhibition of transport wi th adenosine, inosine, or thymidine, partial inhibition wi th guanosine, cytidine or hypoxanthine, and no inhibition wi th adenine or uracil. Recombinant hENT2 is inhibited by dipyridamole, dilazep and draflazine at 1 btM? ,4 hENT2-encoded mRNA transcripts have been observed in a variety of tissues, including brain, heart, lung, thymus, prostate, pancreas and skeletal muscle. 3 The level of expression of bENT2 mRNA is greatest in skele- tal muscle, raising the possibility of differential expression of related isoforms?

Table I Kinetic parameters of nucleoside influx by recombinant nucleoside transporters

NT Substrate K m~p Vm~ × (pmol/ Vrnax:Kmapp Reference

(~M) oocyte.min -~)

hENT I Uridine 240 3.6 0.015 I

rENTI Uridine 150 18 0.120 13

hENT2 Uridine 200 6.4 0.032 4

rENT2 Uridine 300 14 0.047 13

hCNT I Uridine 45 26 0.58 2

rCNT I Uridine 37 21 0.57 16

Adenosine 26 0.07 0.003 14

hCNT2 Uridine 40 0.74 0.02 6

Adenosine 8 0.49 0.06 6

hSPNTI Uridine 80 0.53 0.0007 5

Inosine 5 0.19 0.04 5

rCNT2 Uridine 31 0.86 0.03 6

Adenosine 14 0.90 0.06 6

SPNT Adenosine 6 0.46 0.08 15

Drug Resistance Updates (I 998) I, 310-324 © 1998 Harcourt Brace & Co. Ltd

T h e cit t r a n s p o r t e r (hCNT1) The NT process designated as 'cit' was first described in freshly isolated mouse intestinal epithelial cells as a process that is -concentrative, _insensitive to inhibition by NBMPR and capable of transporting _thymidine. 2' NT processes wi th cit- type activity exhibit a preference for pyrimidine nucleosides and, because they are inhibited by low concentrat ions of adenosine, were originally believed to also mediate transport of adenosine. TM The proteins responsible for the cit process were identified w h e n cDNA encoding a pyrimidine nucleo- side-selective, sodium-dependent (i.e. concentrative) trans- por ter was isolated from rat jejunum by functional expression cloning in Xenopus oocytes. 1622 The human homologue of rCNT1 was subsequently identified by hybridization/RT-PCR (reverse transcriptase polymerase chain reaction) cloning and functional expression of two almost identical cDNAs from human kidney.2The two human cDNAs encode the same protein (of 649 and 650 amino acids) te rmed hCNT1 that exhibits cit-type NT activity w h e n produced in Xenopus oocytes, hCNT1 has 83% identity wi th rCNT1 in amino ~tcid sequence, encoding a 71-kDa protein that is predicted to possess 14 t ransmembrane domains.The hCNT1 gene maps to chromosome 15q25-26. Recombinant hCNT1 produced in Xenopus oocytes exhibits cit-type activ- ity, has a K m for uridine uptake of 42 btM (Table 1), mediates uptake of zidovudine (AZT) and zalcitabine (ddC), and is inhibited by adenosine, thymidine, cytidine and uridine, but not by guanosine or inosine. Direct measurements of the kinetics of adenosine uptake by recombinant hCNT1 in Xenopus oocytes have shown that adenosine binds wi th high affinity to the transporter, presmnably at the permeant- binding site, yet is t ransported at very low rates (Table 1). Since physiologic adenosine concentrat ions rarely exceed 5-10 ~tM, it is possible that adenosine regulates hCNT1 activ- ity in vivo, acting to inhibit transport of pyrimidine nucleo- sides (e.g. uridine).

T h e c / f t r a n s p o r t e r (hCNT2) The NT process designated as ' c / f was first described in freshly isolated mouse intestinal epithelial cells 21 as a process that is sodium-dependent, insensitive to inhibition by NBMPR and capable of transporting formycin B (the C-nucle- oside analog of inosine). NT processes wi th c/fltype activity were subsequently shown to be concentrat ive and capable of transporting a variety of purine nucleosides and uridine. TM

cDNAs encoding c ~ t y p e transporters were first isolated from rat liver by functional expression cloning in Xenopus oocytes 's and subsequently by RT-PCR from rat jejunum. ~4 The open reading frames of the liver and intestinal cDNAs predicted an identical 72-kDa protein, te rmed SPNT is and rCNT2.'4 The transporter protein, hereafter te rmed rCNT2, is predicted to have 659 amino acids wi th 14 t ransmembrane domains and is 64% identical to rCNT1, wi th the greatest divergence in the N- and C-terminal portions, cDNAs encod- ing a human homolog (hSPNT1 or hCNT2) were isolated from kidney s and intestine 6 by hybridization cloning/RT-PCR amplification. The human homolog, hereafter te rmed hCNT2, has 658 amino acids and, like the CNT1 proteins, is predicted to possess 14 t ransmembrane domains. Northern analyses suggest expression in a variety of human tissues,

including heart, liver, skeletal muscle, kidney, intestine, pan- creas, placenta, brain and lung.The hCNT2 (or hSPNT1) gene is located on chromosome 15. 5,6

Recombinant hCNT2 in Xenopus oocytes exhibits K m val- ues of 5, 8 and 40 J, tM, respectively, for inward transport of inosine, adenosine and uridine (Table 1). Other hCNT2 per- meants include 2 '-deoxyadenosine (adenosine > 2'- deoxyadenosine), guanosine and didanosine (ddD, but not thymidine, cytidine, uracil, zidovudine or zalcitabine. A c / f type NT process has been described in the acute promyelo- cytic NB4 leukemia cell line; 23 it mediates adenosine and uridine uptake with K m values, respectively, of 10 ~tM and 30 gM and thymidine is not accepted as a permeant.

O t h e r n u c l e o s i d e t r a n s p o r t ac t iv i t i es i n h u m a n ce l l s A N T process te rmed 'cib' that is concentrative, insensitive to NBMPR, and possessing _road permeant selectivity for both purine and pyrimidine nucleosides has been described in freshly isolated human leukemic blasts 24 and human colon cancer CaCo-2 cells. 25 The technical difficulties in function- ally distinguishing be tween the various concentrat ive trans- por ter types w h e n multiple types are present in single tissues or cells, together wi th the low levels of cib-type activ- ity, have limited the characterization of cib transport processes in human cells.The cib transporter protein has not been identified.

A sodium-dependent (i.e. concentrative) NBMPR-_sensitive NT process wi th selectivity for guanosine and thus terned 'csg' has recently been repor ted in NB4 acute promyelocytic leukemia cells. 26 Although 2'-deoxyguanosine inhibits the uptake of radiolabelled guanosine by approximately 50%, it is unknown whe the r csg mediates transport of both guano- sine and 2'-deoxyguanosine, or whe the r 2 '-deoxyguanosine is a compet i t ive inhibitor of guanosine transport. Further study is required to define the relevance of the csg transport process to cancer chemotherapy, as this transporter has only been described in a single cell line, and its selectivity for anti- cancer nucleosides is not known.The protein responsible for mediating the csg process has not been identified.

A second sodium-dependent (i.e. concentrat ive) NBMPR- _sensitive NT process, te rmed 'cs', has been described in freshly isolated chronic lymphocytic leukemia cells and acute myelogenous leukemia (AML) cells.The cs transporter mediates cellular uptake of cladribine and fludarabine 27 and thus differs from the csg process, which does not accept adenosine (nor, presumably, adenosine analogs). The perme- ant selectivity of the cs process has not been fully character- ized and the protein responsible has not been identified.

ROLE OF NUCLEOSIDE TRANSPORT IN CYTOTOXICITY AND ANTICANCER NUCLEOSIDE RESISTANCE

The cellular targets of nucleoside chemotherapy have been the focus of intense study, and mechanisms of action of the anticancer nucleosides include incorporat ion into DNA and RNA, DNA strand breakage and perturbat ion of intracellular nucleot ide pools. 28 Because the pharmacologic targets of most nucleoside analogs are intracellular, permeat ion through the plasma membrane is an obligatory first step in

© 1998 Harcourt Brace & Co. Ltd Drug Resistance Updates (I 998) I, 310-324 !

the manifestation of CytOtOXiCity. 7 However, as nucleoside analogs are generally hydrophllic, diffusion through the plasma membrane is slow, and transporter-mediated uptake is the major route of drug influx. Inefficient cellular uptake is, therefore, a potential mechanism of resistance to anti- cancer nucleosides. 25

In the discussion that follows, we have interpreted earlier studies of NT processes in the context of current knowledge of the molecular biology of NT proteins.Thus far, the molec- ular identities of the NT proteins responsible for mediating the es, el, cit and c/f processes in various cell types have proven to be members of the ENT or CNT protein families. In the examples given be low for which the molecular iden- tity of the NT process has not yet been defined, the process is designated by its trivial name (e.g. es or et3 linked wi th the name of the presumptive NT protein (e.g. hENT1 or rCNT1). Thus, the NBMPR-sensitive equilibrative process of murine leukemia cells is te rmed ' es /mENTl ' and is a predicted pro- tein, whose existence remains to be established. In addition, the transporter families (lENT vs CNT) to which the cib, cs

and csg proteins belong is unknown.

Transport-related resistance to nucleoside analogs: cultured cancer ce l l l i ne s The first NT-deficient cell line was the AE1 clone, generated from $49 murine T-cell lymphoma cells by chemical mutage- nization and subsequent selection in media containing toxic concentrat ions of adenosine. 29 31 AE1 cells exhibit greatly reduced uptake of physiological nucleosides and high-level resistance to cytotoxic nucleosides, including cytarabine, 5- fluoro-2'-deoxyaaridine, 5-fluorouridine 31 and gemcitabine. 32 AE1 cells lack es/mENTl-mediated transport activity and NBMPR-binding sites, suggesting a mutational loss of func- tional transporter protein. The first human transport-related resistant cell line, ARAC-SC, was obtained from a hypoxan- thine-guanine phosphoribosyltransferase deficient clonal derivative of the human T-lymphoblast CCRF-CEM cell line. 33,34 The parental CCRF-CEM cells exhibit primarily hENTl-mediated transport activity. After chemical mutagen- esis, cells were selected for cytarabine resistance.The result- ing ARAC-SC cells exhibit cross-resistance to a spect rum of cytotoxic nucleosides due to reduced uptake. 33'34 ARAC-8C cells, like AEI cells, appear to be resistant by virtue of an absence of functional hENT1 protein since they completely lack NBMPR-sensitive NT activity and high-affinity NBMPR- binding sites.

Transport-related resistance has also been observed in the absence of chemical mutagenesis. 33 Exposure of murine ery- throleukemia cells to increasing concentrat ions of periodate- oxidized adenosine yielded cells wi th genetically stable high-level resistance and a 20-fold decrease in NBMPR-bind- ing sites, suggesting a loss of the es/mENT1 transport process. In human HCT-8 colon cancer cells, 36 700-fold resis- tance to increasing concentrat ions of 5-fluoro-2'-deoxyuri- dine (FdUrd) was obtained and there was no measurable uptake of FdUrd or binding of NBMPR to FdUrd-resistant cells. Each of these studies demonstrates that the absence of the es process, presumably due to altered ENT1 product ion and/or function, confers high level-resistance to cytotoxic nucleosides.

Transport-related resistance to nucleoside analogs: c l i n i c a l evidence In vitro studies have demonstrated that NT-deficient cells are highly resistant to cytotoxic nucleosides. However, the importance of NT deficiency in clinical drug resistance is uncertain, in large part due to the difficulty of performing transport studies on malignant cells derived from clinical specimens, and the problems associated wi th quantifying NT proteins in malignant clones admixed with normal cells. Nonetheless, there is compell ing evidence that clinical resis- tance to cytarabine can be mediated by NT deficiency (reviewed below), although NT-based clinical resistance to o ther anticancer nucleosides has yet to be proven.

Transport of anticancer nucleoside drugs The anticancer nucleosides in routine clinical use (Fig. 1) include cytarabine (ara-C), cladribine (2-CdA), fludarabine (F- ara-A) and gemcitabine (dFdC). Deoxycoformycin has largely been supplanted by the arrival of the newer purine nucleo- side analogs, and will not be discussed further in this review.

Cytarabine Cytarabine (ara-C, 1-[~-D-arabinofuranosyl cytosine) has a major role in the curative therapy of acute leukemia. Given as a single agent at standard doses, cytarabine induces hema- tological remissions in about 30% of patients withAML?7 A 7- day infusion of cytarabine in combinat ion with three daily doses of an anthracycline (doxorubicin or idarubicin) is the standard induction therapy for AML, and cytarabine is also a componen t of the consolidation, maintenance and intensifi- cation regimens. Combinations including cytarabine are also employed for the blastic phase of chronic myelogenous leukemia (CML) and for non-Hodgkin's lymphomas? 8

The incorporat ion of cytarabine into DNA is closely linked to its cytotoxic effect on leukemic cells. 39 DNA incor- porat ion requires intracellular accumulation of cytarabine 5'- t r iphosphate (ara-CTP), which is the net result of several processes, including membrane transport, anabolism to ara- CTP and degradation of ara-CTP to inactive metaboli tes.The relationship be tween ara-CTP accumulation and treatment efficacy has been clearly demonstrated. The magnitude of radiolabelled cytarabine retention by bone mar row mononu- d e a r cells derived from AML patients correlates positively wi th clinical response to cytarabine-based induct ion chemotherapy. 4° In a study wi th AML patients treated with cytarabine and anthracycline, 41 freshly harvested myeloblasts were classified as 'high' o r ' l ow ' retent ion based on intracellu- lar quantities of ara-CTP; the median duration of clinical remission, once attained, was two-fold greater in the high- retent ion group. Mediated inward transport of cytarabine is the major determinant of ara-CTP accumulation at low cytarahine concentrat ions (< 1 ~tM), whereas phosphoryla- tion capacity predominates at concentrat ions above 10 ~tM;42 44 cytarahlne given during low or standard dose treat- ment regimens (e.g, 100-200 mg/m2/day) produces plasma levels be low 1 gM. 4~

The efficiency of cytarabine uptake by leukemic blast cells has been related to clinical ou tcome in AML patients receiving standard dose cytarabine. Nueleoside transport parameters were measured in blasts from 15 AML patients

Drug Resistance Updates (I 998) I, 310-324 © 1998 Harcourt Brace & Co. Ltd

Table 2 Kinetic parameters of nucleoside drug uptake

Drug Nucleoside Species transport process

Cell type Krn (~tM)

Vm~x Reference

Cytarabine es/hENT I Human

Fludarabine cif/mCNT2 Mouse Cladribine es/hENT I Human

Gemcitabine

AML blasts

LI210 CCRF-CEM

ei/hENT2 Human K562

c//TmCNT2 Murine MA27

255_+ I I

13.4 5.1

10.3

4.9

es/hENT I Human CCRF-CEM 329 _+ 91

T-cell lymphoblasts

ei/hENT2 Human HeLa + 832 _+ 204

100 nM NBMPR

cit/hCNT I Human hCNT I - 18.3 _+ 7.2 transfected HeLa cells

cif/hCNT2 Human hCNT2-transfected No measurable transport

0.84 _+ 0.44 49 (pmol/s/I 0 ~ cells)

57 1.43 62

(pmol/uL cell water/s)

0.50 62 (pmol/uL cell water/s) 0.29 62 (pmol/uL cell water/s) 17.0 + 4.3 32 (pmol/s/I 0' cells) 4.2 _+ 1.3 32 (pmol/s/I 06 cells) 0.94 _+ 0.22 32 (pmol/s/106 cells) N.A. 32

cib Human HeLa cells CaCo2 colon Rate insufficient Rate insufficient cancer cells + for accurate for accurate dilazep kinetics kinetics

32

treated with standard induction therapy containing cytara- bine? 6 Of the seven patients who failed therapy, three exhibited the lowest rates of cytarabine uptake and NBMPR- binding-site numbers. Another report 47 described a patient with T-cell acute lymphoblastic leukemia in first relapse, whose blasts prior to therapy had (i) rapid eytarabine uptake and ara-CTP accumulation, and (ii) high numbers of NBMPR- binding sites. Mthough this patient's leukemia initially responded to standard doses of cytarabine, the NBMPR-bind- ing-site number and ara-CTP accumulation rate had decreased by approximately 75% at the time of disease relapse, suggesting that cytarabine chemotherapy had selected for variant cells with reduced es/hENT1 activity.

In oilier studies of isolated leukemic blasts, ",48 the mean K and V values of cytarabine influx were approximately 260 ~tM, while the Vma ~ values of cytarabine influx varied 80-fold (Table 2). Approximately 80% of cytarabine entered leukemic blasts by the es/hENTl-mediated process over a wide range of cytarabine concentrations, and a positive correlation was shown between NBMPR-binding site numbers and cytarabine influx in both normal and leukemic leukocytes? ~ The abtm- dance of es/hENT1 in plasma membranes of leukemic blasts has also been determined by binding of SAENTA-fluorescein, an impermeant fluorescent analog of NBMPR, and is corre- lated with in vitro sensitivity to cytarabine. 49,5°

High-dose cytarabine regimens (e.g. 3 g/m2/day) produce remissions in some patients refractory to conventional-dose cytarabine and generate plasma levels above 50 ~tM, 51 thereby overcoming the rate-limiting effect of es/hENTl-mediated transport on formation of ara-CTP Two randomized studies have shown benefit to high-dose cytarabine-based induction by demonstrating that, although remission rates are similar to standard-dose therapy, superior disease-free survival rates are seen in the high-dose groups. 52,53

In summary, cytarabine uptake in human leukemic cells is mediated by the es/hENT1 process, and its abundance in plasma membranes is a determinant of clinical efficacy of standard dose cytarabine therapy in adult AML. Studies are in progress to more precisely determine cytarabine transporta- bility by the other NT processes of human cells. Cytarabine appears to be a poor permeant for recombinant hCNT1 (the pyrimidine-nucleoside selective concentrative transporter) based on the high concentrations required to inhibit uridine transport in transiently transfected cultured cells (Coe I et al, unpublished results).

Fludarabine Fludarabine (F-ara-A, 9---D-arabinosyl-2-fluoroadenine, Fludara®) is the most active agent in the treatment of chronic lymphocytic leukemia, and also has activity in

© 1998 Harcourt Brace & Co. Ltd Drug Resistance Updates (1998) 1,310-324

indolent non-Hodgkin's lymphoma, prolymphocyt ic leukemia, and Waldenstr6m's macroglobulinemia. Fludara- bine has little solid tumor activity. 28 Myelosuppression is the dose-limiting toxicity and profound cell-mediated immuno- deficiency characterized by reduced CD4+ counts commonly results in opportunistic infections¢ 4 Following intravenous administration, fludarabine 5 '-monophosphate is rapidly dephosphorylated extracellularly to fludarabine. Fludarabine is transported into cells and is rephosphorylated by deoxycy- tidine kinase and thence to fludarabine 5'-triphosphate, which induces toxicity through incorporation into DNA and/or RNA and inhibition of ribonucleotide reductase, DNA polymerase-o~, DNA primase, and/or DNA ligase I. 28,5s

Cellular differences in mediated transport of fludarabine have been proposed as a mechanism of selectivity of fludara- bine therapy, since there is less in vivo accumulation of flu- darabine 5'-triphosphate in murine gastrointestinal mucosa and bone marrow compared to that in P388 murine leukemia cells, s~ Evidence support ing a role for differential uptake of fludarabine in its tissue selectivity was obtained in the L1210 murine leukemia cell line, 56 in which fludarabine accumulation was eight-fold more rapid than that observed in mouse intestinal crypt epithelial cells, and the influx step, rather than the subsequent phosphorylat ion step, was rate- limiting for routine gastrointestinal epithelial cells, s6

A kinetic study of fludarabine uptake by L1210 cells 57 found evidence for transport via two separate processes, wi th K m values, respectively, of 190 btM and 13 gM. However, because subsequent studies 58,59 revealed that L1210 cells pos- sess three distinct NT processes (es, ei, cOO , it is likely that the repor ted low-affinity transport of fludarabine in L1210 cells was due to the combined operat ion of the es/mENT1

and ei/mENT2 processes and the high-affinity transport rep- resented the operat ion of the clf/mCNT26° transport process (Table 2). In lymphoblasts harvested from acute lymphoblas- tic leukemia (ALL) patients, fludarabine sensitivity in vitro correlated wi th es/hENT1 abundance in the plasma mem- brane, as determined by 5'-S-(2-aminoethyl)-N6-(4-nitroben- zyl)-5'-thioadenosine (SAENTA) - f luorescein binding, s° In summary, fludarabine appears to be a substrate for es/hENT1, ei/hENT2 c/f/hCNT2 and cs ~7 and its tissue selec- tivity may be related to differences in abundance of func- tional NTs.

Cladribine Cladribine (CdA, 2-chloro-2'-deoxyadenosine, Leustatin®) is a purine nucleoside analog that is resistant to degradation by adenosine deaminase. Cladribine achieves durable complete remissions in the majority of patients wi th hairy cell leukemia, and short4ived remissions in chronic lymphocytic leukemia and non-Hodgkin's lymphoma; cladribine has mini- mal solid tumor activity. The major toxicity of standard doses is myelosuppression and cell-mediated immunosuppression. Cladribine, after permeat ion across the plasma membrane, is conver ted to the 5 ' -monophosphate by deoxycytidine kinase. Inhibition of DNA synthesis and repair, inhibition of r ibonuclotide reductase, and imbalances in deoxyribonu- cleotide pools contr ibute to cytotoxicity. 28

Cladribine uptake is evidently mediated by both es/hENT1 and ei/hENT2-mediated processes, as shown by

NBMPR and dipyridamole protect ion studies in cladribine- exposed human hematological cancer cell lines. 61 Transport kinetics of radiolabelled cladribine have been determined for es/hENT1, ei/hENT2, and ci f /mCNT2 in several leukemic cell lines, and the degree of cladribine cytotoxicity was directly correlated wi th the efficiency of transport 6~ (Table 2). In freshly harvested ALL lymphoblasts, in vitro sensitivity to cladribine correlated with es/hENT1 abundance in plasma membranes determined by SAENTA - f luorescein binding. 5°

Uptake of cladribine by the recombinant c~ type trans- porter of rats (rCNT2) has been demonstrated in the Xenopus oocyte system. TM Additionally, cladribine inhibited inosine uptake (IC50 13 btM) by recombinant rCNT2 in a HeLa cell expression system, t° However, the transportability of cladrib- ine via recombinant hCNT2 has not yet been determined.

In summary, cladribine uptake in human ceils is mediated by es/hENT1, ei/hENT2 and cs, ~7 and inward transport of cladribine appears to be an important determinant of in vitro cytotoxicity. Studies to directly determine the transportabil- ity of cladribine by the recombinant human NT proteins are in progress.

Gemcitabine Gemcitabine (dFdC, 2',2'-dlfluorodeoxycytidine, Gemzar®) is a pyrimidine analog of deoxycytidine in which two fluorine atoms are present at the 2'-position of the deoxyribose ring (Fig. 1). The dosedimiting toxicity is myelosuppression. Gemcitabine is unique among nucleoside drugs in that it has considerable activity against solid tumors, including non-small cell lung, breast, bladder, ovarian, head and neck cancers2 8

Gemcitabine is first conver ted intracellularly to dFdCMP by deoxycytidine kinase, and subsequently phosphorylated to the 5'-diphosphate and 5'-triphosphate derivatives by pyrimidine monophospha te and diphosphate kinases. 63 dFdCDP inhibits r ibonucleotide reductase, while dFdCTP is incorporated into DNA and RNA. 64 Although the relative con- tributions of these effects to cytotoxicity are not known, gemcitabine exhibits the proper ty of self-potention in that dFdCTP inhibits deoxycytidine monophospha te deaminase, thereby decreasing its o w n catabolic degradation. 63 Plasma membrane transport of gemcitabine was initially studied in Chinese hamster ovary cells, where the rates of inward trans- port of gemcitabine and cytarabine were similar and not responsible for the observed differences in cytotoxicity of gemcitabine and cytarabine; rather, the differences were attributed to gemcitabine's higher affinity for deoxycytidine kinase and the longer intracellular retention of dFdCTP than ara-CTR 65

The transportability of gemcitabine, and its importance in gemcitabine cytotoxicity, has been examined in detail in studies wi th both native and recombinant NTs.32 A deficiency in NT, produced either pharmacologically or genetically, con- fers two- to three-log protect ion of cultured cells from gem- citabine cytotoxicity. Kinetic studies wi th a panel of murine and human cell lines wi th defined NT processes demon- strated that gemcitabine uptake is mediated by the es, el, cit and cib processes, but not by the c/f process (Table 2); in these studies, the molecular identity of the responsible trans- por ter proteins (i.e. hENT1, hENT2, hCNT1, hCNT2) was known for all but the cib-mediated process. The most

W Drug Resistance Updates (I 998) I, 310-324 © 1998 Harcourt Brace & Co. Ltd

efficient processes were those mediated by hENT1 and hCNT1, although the rates of gemcitabine transport were approximately 10-fold slower than those of uridine. The hCNT2/cif transporter did not accept gemcitabine as a per- meant.

Strategies fo r m o d u l a t i o n o f nuc leos ide t r a n s p o r t p rocesses to e n h a n c e the rapeu t i c eff icacy Improvement in the therapeutic index of anticancer nucleo- sides, to achieve maximal ant±cancer activity with minimal toxicity to the patient, may be possible through manipula- tions based on understanding of NT processes. Clinical NT inhibitors are required for several of the strategies outlined below.At present, however, no ideal NT inhibitor is available for clinical use. NBMPR, the classic inhibitor of es/hENT1- mediated processes, produces significant hypotension when administered as the soluble prodrug NBMPR 5'-monophos- phate (Venner PM, et al, unpublished results), and all pub- lished clinical trials attempting to exploit NT inhibitors have used dipyridamole. 6~-73 Dipyridamole, although having the advantage of inhibiting both es/hENTl and ei/hENT2 processes, has the disadvantage of non-specificity (inhibition of phosphodiesterase, stimulation of prostacyclin biosynthe- sis), and binding to plasma o~-acid glycoprotein and albumin, 74 which limits achievable free serum concentrations 75 even when given at maximally tolerated doses by continuous cen- tral intravenous infusion. 69 Draflazine appears to be well tol- erated in the setting of acute ischemia for which it is being studied as a cardioprotective agent 76 and currently appears to be the most promising es/hENT1 transport inhibitor for fur- ther in vivo study. Like dipyridamole, draflazine is a less effec- tive inhibitor of ei/hENT2 than es/hENT1.4

Pharmacologic resistance by protection of normal tissues In the 'protection' strategy, transport inhibitors are used to induce pharmacologic resistance by reducing cellular uptake of nucleoside drugs by dose-limiting normal tissues, thereby increasing the maximum tolerated drug dosage. Proof of principle has been demonstrated in studies with rodents in which NBMPR 5'-monophosphate (NBMPR-P) reduced the toxicity of tubercidin (7-deazaadenosine, 4-

77 amino-7-(~-D-ribofuranosyl)pyrrolo [2,3-d] pyrimidine) and nebularine (9-[3-D-ribofuranosylpurine).V7.7a In L1 2 1 0 leukemia-bearing mice, NBMPR-P increased the therapeutic index of both nebularine 79 and flndarabine. 8° BIBW22BS, (4- [N-(2-hydroxy-2-methyl-propyl)-ethanolamino]-2,7-bis(cis -2,6-dimethyl-morpholino)-6-phenylpteridine), which is a potent inhibitor of equilibrative NT processes, did not impair gemcitabine-induced growth inhibition of human tumor xenografts grown subcutaneously in nude mice, but slightly reduced therapy associated weight loss. 8. The efficacy of such protection strategies may be due, in part, to higher lev- els of es/ENT1 activity in proliferating malignant cells rela- tive to their normal counterparts (Table 3). Human breast, liver, stomach and colorectal tumor tissues have more NBMPR-binding sites than the adjacent normal tissues, 82. and mitotic rates correlate with NBMPR-binding site numbers in human thymocytes 83 and human myeloblasts, s4 Thus, the transport inhibitors appear to reduce normal tissue drug influx below a threshold for cytotoxicity, while leaving

Table 3 Numbers of NBMPR-binding sites on representative human cells

Cell type Binding sites/cell Reference

Erythrocyte II 000 122 Bone marrow 21 000 85

mononuclear cells Peripheral blood I 123 _+ 553 83

lymphomcytes I 120 ± 610 46 Unstimulated peripheral 2203 83

blood lymphocytes Mitogen stimulated 68 024 83

peripheral blood

lymphocytes Normal CD5+/CD 19+ 553 _+ 178 100

lymphocytes Chronic lymphocytic 485 _+ 425 100

leukemia cells 1430 _+ 730 84 Bone marrow plasma 997 _+ 1096 92

cells from normal

donors Normal blood 2270 _+ 690 46

polymorphonuclear cells Leukemic blasts 7490 + 8080 123 Acute lymphocytic 2300-7400 46

leukemic blasts 3100 _+ 940 48

Acute myelogenous 3800-24 200 84 leukmic blasts

Thymocytes 26 068 ± 8776 83 Bone marrow plasma 1777 _+ 2181 92

cells from patients with myeloma

CI 80-13S 232 000 -+ 70 800 124 undifferentiated ovarian cancer

RC2a 100 780 ± 19 700 124

myelomonocytic leukemic blasts

HL-60 promyelocytic 170 000 85 leukemia

CCRF-CEM T 330 000 85 lymphoblast leukemia

N84 promyelocytic 535 000 23

leukemia BeWo choriocarcinoma 2 7000 000 125 MCF-7 breast cancer 245 000 126 HeLa cervical 150 000 127

carcinoma 410 000 125

malignant cells with sufficient transport capability to achieve cytotoxic intracellular drug concentrations.

Pharmacologic resistance to cytotoxic nucleosides has been achieved with human hematopoietic progenitors by treatment with transport inhibitors, suggesting that combina- tions of equilibrative NT inhibitors and ant±cancer nucleo- sides may be capable of selectively targeting malignant cells possessing inhibitor-insensitive NT activities, while reducing myelotoxicity. Progenitor cells derived from normal human

© 1998 Harcourt Brace & Co. Ltd Drug Resistance Updates (1998) 1,310-324

bone marrow were spared from in vitro tubercidin toxicity by pretreatment with transport-inhibitory concentrations of NBMPR, dilazep or dipyridamole and tolerated in vitro expo- sures to relatively high concentrations of the transport inhibitors (5-1 0 ~tM) without loss of viability. 85 Other studies have confirmed that hematopoietic progenitor cells possess predominantly es/hENT 1 activity. 86

In summary, inhibitors of equllibrative NT processes, most likely mediated by the ENT1 and ENT2 proteins, are capable of providing pharmacologic resistance to some nor- mal tissues while permitting cytotoxicity to malignant cells. Pursuit of such strategies in clinical trials will require effec- tive and well-tolerated transport inhibitors.

Nucleoside trapping Cells that possess both equJlibrative and concentrative NT processes may be rendered more sensitive, rather than resis- tant, to nucleoside drugs by the loss of equllibrative trans- port capability. For example, dipyridamole enhances the cytotoxicity of ara-A in L1210/C2 c e l l s , 87 which have both eqnilibrative NT processes and the purine-selective concen- trative process. 59 Although dipyridamole reduces initial rates of uptake of ara-A, the subsequent intracellular accumulation of ara-A is enhanced because of the almost complete inhibi- tion of ara-A efflux (via es/mENT1 and ei/mENT2) while inward transport of ara-A via the sodium-dependent concen- trative transporter continues. Ara-A cytotoxicity is signifi- candy enhanced in cells so treated. Nucleoside 'trapping' by this strategy could be exploited to increase the therapeutic index of anticancer nucleoside drugs in human cancers if (i) the malignant cells possess either cit/hCNT1 (sensitization to pyrimidine nucleoside analogs) or c/f/hCNT2 (sensitiza- tion to purine nucleoside analogs) transporters, and (ii) the dose-limiting normal tissues (in which the CNT proteins are presumed to be absent or non-functional) can be rendered pharmacologically resistant. The trapping strategy failed in a recent in vitro study with gemcitabine; 32 dipyridamole treat- ment of cultured human CaCo-2 cells, which possess both equllibrative and concentrative NT processes, had no effect on gemcitabine sensitivity.

A trapping strategy that requires only equilibrative trans- porters has also been explored, ss After short exposures to cytarabine, cultured ovarian carcinoma and promyelocytic leukemia cells that were subsequently incubated continu- ously with dipyridamole, which blocked outward transport via the equilibrative processes, experienced a lO-fold aug- mentation of cytotoxicity. Cladrihine may also be amenable to trapping strategies. Levels of cladribine phosphates pre- sent in human leukemic cells declined rapidly when the cells were in cladribine-free mediumfl 9 and in freshly harvested chronic lymphocytic leukemia (CLL) cells, NBMPR signifi- cantly decreased the rate of cladribine efflux, thereby increasing the drug content of NBMPR-treated cells over untreated control cells? °

The nuc leos ide t r a n s p o r t prof 'de and drug schedule The 'nucleoside transport profile' is the description of the numbers and types of functional NT proteins present in the plasma membrane of the cells or tissues of interest.There are marked differences in the kinetics of transport of anticancer

nucleoside analogs by the human NTs (Table 2), suggesting that it may be possible to rationally design chemotherapeutic schedules that take advantage of these differences. Were a given malignancy found to exhibit predominantly CNT pro- teins, one might target malignant cells by low-dose long-dura- tion exposure to a toxic nucleoside for which the transporter in question has high selectivity. Variation in the tissue distribution of NT proteins may contribute to the dif- fering toxicity profiles of continuous versus bolus infusions of gemcitabine, 91 cladribine and fludarabine, z8

Selective increases in nucleoside transport in cancer cells Mthough regulation of equilibrative NT processes is not well understood, it is clear that the number of es/ENT1 proteins (as determined by NBMPR-binding studies) varies with pro- liferation rate, and in response to cytotoxic and mitogenic stimuli. High cellular proliferation rates are associated with high levels of es/hENTl-mediated activity in human thymo- cytes, 83 leukemic blasts ~4 and myeloma cells? 2 These observa- tions suggest that proliferative stimulation of malignant cells will increase nucleoside drug uptake, thereby sensitizing oth- erwise resistant cells. Exposure of routine S 1 macrophages to colony stimulating factor-1 (CSF-1) stimulated adenosine uptake by increasing es/mENTl-mediated transport at the time of entry of the cells into the S phase of the cell cyclefl 3 while treatment of freshly isolated human myeloid leukemic blasts with granulocyte-macrophage CSF (GM-CSF) resulted in a four-fold increase in es/hENT1 abundance. 94 Unfor- tunately, the strategy of GM-CSF priming was shown to be ineffective in three randomized trials of treatment of AML with cytarabinc 95-97 as were attempts to use granulocyte CSF (G-CSF) to increase leukemia cell proliferation, thereby increasing cell sensitivity to cytarabine. 98,99

Nucleoside analog therapy itself may increase the number of NBMPR-binding sites, thereby acting in a self-potentiating fashion.Treatment of freshly isolated CLL ceils with fludara- bine or cladribine led to a five-fold increase in the binding of SAENTA-fluorescein as well as increases in proliferation and apoptotic rates. ~°8 Additionaily, induction of sodium-depen- dent NT processes has been demonstrated by agents that promote cellular differentiation, For example, in the human promyelocytic leukemia HL-60 line, chemical induction of differentiation ~°~-1°4 leads to increased rates of sodium-depen- dent nucleoside transport and decreased rates of NBMPR- sensitive nucleoside transport as well as decreased numbers of high-affinity NBMPR - binding sites.

Gene therapy strategies The application of gene therapy to the treatment of cancer has been limited by several technical hurdles, including the limited transfection efficiencies achievable in vivo.To circum- vent this problem, those strategies that have a significant 'bystander effect' are preferred. 1°5 For example, introduction of viral thymidine kinase followed by gancyclovir therapy leads to cytoxocity in transfected cells, I°6 and the release of activated (i.e. phosphorylated) gancyclovir during cell lysis is toxic to neighboring non-transfected cells. Increased accu- mulation of gancyclovir within a transfected cell would there- fore augment the bystander effect. Since gancyclovir requires

Drug Resistance Updates (I 998) I, 310-324 © 1998 Harcourt Brace & Co. Ltd

mediated uptake (both by nucleoside and nucleobase transporters1°7), introduction of a gene encoding a transporter wi th high affinity for gancyclovir together wi th the viral thymidine kinase gene might increase cytotoxicity.

Relationship between multidrug-resistant phenotypes and nucleoside transport Recent studies suggest that multidrug-resistant cell lines have increased nucleoside uptake capabilities. MCF-7 human breast cancer cells with different levels of P-glycoprotein-mediated multidrug resistance also exhibited different apparent K m and V a ~ values of adenosine transport, and the increase inVm= cor- related with an increase in NBMPR binding, a°8 The influx of cytarabine, and its subsequent accumulation as intracellular ara-CTE, was higher in a P-glycoprotein-containing subllne than in the parental P388 murine leukemia cells, due to an increase in NBMPR-insensitive uptake in the multidrug-resis- tant cells. 1°9 P-glycoprotein-containing and multidrug resis- tance protein (MRP)-containing cell lines, although resistant to P-glycoprotein and MRP substrates, are significantly more sen- sitive to gemcitabine than their parental cell linesY ° Although the mechanism(s) tmderlying these observations are unknown, these studies suggest that cytotoxic nucleosides may selectively target multidrug-resistant cells.

ROLE OF NUCLEOSIDE TRANSPORT IN RESISTANCE TO NON-NUCLEOSIDE ANTIMETABOLITES

Nucleoside salvage pathways are an important means of circumventing antimetabolite drug action. 11~-113 Although pur ine and pyrimidine nucleotides are essential for numer- ous biological processes, they can be synthesized endoge- nously via de novo synthetic pathways. However, some tissues lack these pathways and require salvage of exogenous nucleosides to mee t their metabolic requirements. The rela- tive contributions of de novo and salvage pathways vary be tween different tissues, and in some normal tissues (e.g. hematopoiet ic and intestinal epithelial cells), the rate of sal- vage synthesis exceeds the rate of de novo synthesis.

The non-nucleoside antimetabolite drugs in clinical anti- cancer therapy inhibit de novo synthesis of purine and/or pyrimidine nucleotides. Because salvage of nucleosides and nucleobases is enhanced in many human tumors, the salvage pathways can be utilized by cancer cells to replenish intra- cellular nucleotide pools, thereby overcoming antimetabo- lite cytotoxicity. TM Cells therefore become resistant to the cytotoxici ty of the antimetabolite drug in quest ion if they have access to substrates of the salvage synthesis pathways (nucleobases such as hypoxanthine or nucleosides), or to components of the de novo pathway distal to the inhibitory block.

Nucleic acid precursors are present in plasma and extra- cellular fluid, and participate in salvage synthesis pathways by means of functional nucleoside and nucleobase trans- porters, lnhihitors of equilibrative NT processes have been explored in both preclinical models and clinical trials as modulators of resistance to antimetabolite drugs. Dipyridamole has been most extensively studied, because of its ability to inhibit es/ENT1 and ei/ENT2-mediated processes, as well as nueleobase transport, ~1~,'6 although

other po ten t inhibitors of equilibrative transport (e.g. NBMPR, dilazep, draflazine) also have potential applications as modulators of salvage-dependent resistance to anti- metabolites.

5 -F luo ro t t r ac i l Dipyridamole modulat ion of 5-fluorouracil (5-FU) cytotoxic- ity has been previously reviewed. 117 5-FU treatment leads to inhibition of de novo product ion of dTMP and cytotoxicity by interference wi th DNA synthesis and repair. However, NT- mediated uptake of preformed thymidine can bypass dTMP depletion, thereby conferring resistance to 5-FU. Such resis- tance can be c i rcumvented in vitro by blocking salvage of exogenous thymidine - e.g. cancer cell lines exposed to 5-FU and inhibitors of eqnilibrative NTs (dipyridamole and NBMPR) show increased 5-fluoro-2'-deoxyuridine 5'- monophospha te (FdUMP) synthesis, decreased FdUrd efflux, prolonged intracelltflar half-life of FdUMP, and greater cyto- toxicity compared to controls exposed to 5-FU alone. "7,1~s

Two clinical trials in which 5-FU and dipyridamole were given by concurrent continuous intravenous infusion failed to modulate the clinical toxicity of 5-FU, and clinical responses were rare. 66,~7 Although the strategy of dipyri- damole modulat ion of 5-FU toxicity has been abandoned (in part due to the previously discussed pharmacologic limita~ tions of dipyridamole), clinical inhibitors of ENT proteins that are bet ter tolerated and more poten t might warrant fur- ther investigation. One such candidate inhibitor, the dipyri- damole derivative BIBW22BS, increased the antiproliferative effects of 5-FU two- to six-fold in colon cancer cell lines, but did not affect 5-FU-dependent growth inhibition in human tumor xenografts grown subcutaneously in nude mice. 81

N-phosphonacetyl-L-aspartate N-phosphonacetyl-L-aspartate (PALA), a transition state inhibitor of aspartate transcarbamylase, blocks de novo pyrimidine biosynthesis. ~19 Despite promising preclinical activity, PALA has generally demonstrated little efficacy as a single agent in human cancer.Trials of PALA combined with oral 72 or intravenous dipyridanlole 73 failed to improve anti- cancer efficacy, although gastrointestinal toxicity was higher than with PALA monotherapy. Further study of NT modula- tion of PALA toxicity does not seem warranted.

The antifolate drugs Methotrexate and tr imetrexate inhibit dihydrofolate reduc- tase and lead to impaired de novo synthesis of purines and pyrimidine nucleotides. Doseqimiting toxicities are myelo- suppresion and mucositis. Resistance of cancer cells to antifolate cytotoxicity can arise through increased capacity for salvage of nucleosides and/or nucleobases; such resis- tance can be c i rcumvented by inhibition of cellular uptake. For example, dipyridamole potentiates the growth-inhibitory effects of methotrexate against HCT 1 16 human colon can- cer cells through inhibition of thymidine transport.lz° Similar results are seen in routine sarcoma cells 1~1 and in human embryo fibroblast cells, 112 where dipyridamole increases sensitivity to methotrexate by approximately 10-fold (an effect which can be reversed by excess hypoxanthine plus thymidine).

© 1998 Harcourt Brace & Co. Ltd Drug Resistance Updates (I 998) I, 310-324

Unfortunately, dipyridamole also increases methotrcxate toxicity to normal cells. 7° Methotrexate toxicity (moderate to severe myelosuppression and/or mucositis) was greater in patients treated with a bolus injection of methotrexate 24 h after initiation of a high-dose dipyridamole infusion than in patients treated wi th methotrexate alone. Pharmacokinetic studies showed no change in methotrexate elimination. Dipyridamole potentiat ion of methotrexate toxicity appears to require high-dose dipyridamole (given by continuous infu- sion), since modulat ion of methotrexate toxicity was not observed wi th oral dipyridamole and methotrexate therapy. 71

The myelosuppression induced by antifolates is largely due to toxicity to cells in late stages of hematopoietic develop- ment, while early myeloid progenitor ceils are relatively resis- tant. In a study of the role of nucleoside salvage in protecting hematopoiet ic stem cells from the cytotoxicity of trimetrex- ate, 86 human progenitor cells were relatively resistant to trime- trexate toxicity because of their ability to salvage thymidine and hypoxanthine from serum. Exposure to NBMPR-P greatly increased trimetrexate toxicity to progenitors and stem cells, suggesting that severe myelotoxicity could limit the use of inhibitors of the ENT proteins in combination with trimetrex- ate. In contrast, the interaction of lometrexol (a novel antifo- late which inhibits de novo purine biosynthesis) and dipyridamole is more complex. Hematologic cell lines are res- cued from lometrexol cytotoxicity by hypoxanthine, even in the presence of dipyridamole, unlike some human carcinoma cell lines (HeLa,A549)/16 Thus, the combination of lometrexol, dipyridamole and hypoxanthine may be selectively toxic to some malignant cells while sparing bone marrow.

CONCLUSIONS

Nucleoside transport plays a vital physiological role in nucle- oside salvage, and is a major determinant of cytotoxicity for anticancer nucleoside analogs. Understanding of NT processes has been greatly aided by the recent cloning of cDNAs encoding four human NT proteins. Acquired resis- tance to nucleoside analogs can occur as a result of ineffi- cient nucleoside uptake, both in vitro and in vivo, although the contribution of transport deficiency in clinical resistance to nucleosides requires further study.There are several strate- gies wi th the potential to increase the therapeutic index of nucleoside analog drugs, by way of normal tissue protection, nucleoside trapping, selectively promot ing drug uptake in malignant cells, rational drug scheduling to exploit a tumor 's NT profile, and gene therapy. Strategies have also been con- sidered of modulat ion of antimetabolite chemotherapy by inhibition of nucleoside salvage.

F ~ DIRECTIONS

Although there is strong evidence for a role of membrane transport in the pharmacologic action of nucleoside anti- cancer drugs, this information has yet to be exploited to improve their efficacy. With the recent cloning of four human NTs, molecular and immunologic probes are becom- ing available to determine the tissue and tumor distribution of these proteins, and will al low a greater understanding of their role in normal physiology and in their relationship to

malignant progression. Such probes may allow characteriza- tion of the NT profile of a given cell or tumor, which may in turn allow rational drug selection or predict ion of response to nucleoside anticancer therapy. Furthermore, a molecular understanding of human NTs might allow the design of spe- cific inhibitors, as well as more tumor-specific nucleoside analogs. These approaches have the potential to overcome clinical nucleoside drug resistance and to increase the thera- peutic indices of nucleoside anticancer drugs.

Acknowledgments

We thank Joan Thibault, Denise Brooks, Linda Harris mad Xiao Fang for assistance in preparing this manuscript.This work was supported by the Alberta Cancer Board, the Medical Research Council of Canada, the National Cancer Institute of Canada, Medical Research Council and the Wellcome Trust, UK. CEC is a Terry Fox Cancer Research Scientist of the National Cancer Institute of Canada and JDY is a Heritage Medical Scientist of the Alberta Heritage Foundation for Medical Research.

References

I. Griffiths M, Beaumont N,Yao SY et al. Cloning of a human nucleoside transporter implicated in the cellular uptake of adenosine and chemotherapeutic drugs. Nature Med 1997; 3(I): 89-93.

2. Ritzei MWL,Yao SYM, Huang MY et al. Molecular cloning and functional expression of cDNAs encoding a human Na+- nucleoside cotransporter (hCNTI).Am J Physiol 1997; 272(2 4 I-2): C707-C714.

3. Crawford CR, Patel DH, Naeve C, Belt JA. Cloning of the human equilibrative, nitrobenzylmercaptopurine riboside (NBMPR)- insensitive nudeoside transporter ei by functional expression in a transport-deficient cell line.J Biol Chem 1998; 273(9): 5288-5293.

4. Griffiths M,Yao SYM,Abidi F et al. Molecular cloning and characterization of a nitrobenzylthioinosine- insensitive (ei)

equilibrative nucleoside transporter from human placenta 8iochem J 1997; 328(Pt 3): 739-743.

5. Wang J, Su S-F, Dresser MJ, Schnaner ME,Washington CB, Giacomini KM. Na+-dependent purine nucleoside transporter from human kidney: cloning and functional characterization.Am J Physiol 1997; 273(42): F1058-FI065.

6. Ritzel MWL,Yao SYM, NgAML et al. Molecular cloning, functional expression and chromosomal localization of a cDNA encoding a human Na+/nucleoside cotransporter (hCNT2) selective for purine nucleosides and uridine. Mol Membr Biol 1998:Accepted for publication.

7. Cass CE. Nudeoside transport. In: Georgopapadakou NH, ed. DrugTransport in Antimicrobial Therapy and Anticancer Therapy. NewYork: Marcel Dekker, 1995:403-45 I.

8. Griffith DA, Jarvis SM. Nucleoside and nucleobase transport systems of mammalian cells. Biochim Biophys Acta Rev Biomembr 1996; 1286(3): 153-181.

l Drug Resistance Updates (I 998) I, 310-324 © 1998 Harcourt Brace & Co. Ltd

9. Buolamwini JK. Nucleoside transport inhibitors: structure-activity relationships and potential therapeutic applications. Curr Med Chem 1997; 4(I): 35-66.

I 0. Wang J, Schaner ME,Thomassen S, Su SF, Piquette-Miller M, Giacomini KM. Functional and molecular characteristics of Na(+)- dependent nucleoside transporters. Pharm Res 1997; 14(I I): 1524-1532.

I I. Che M, Gatmaitan Z, Arias IM. Ectonucleotidases, purine nucleoside transporter, and function of the bile canalicular plasma membrane of the hepatocyte. FASEB J 1997; I I (2): I 0 I-108.

12. Thorn JA, Jarvis SM.Adenosine transporters. Gen Pharmacol 1996; 27(4): 613-620.

13. Yao SY, NgAM, MuzykaWR et al. Molecular cloning and functional characterization of nitrobenzylthioinosine (NBMPR)-sensitive (es) and NBMPR-insensitive (el) equilibrative nucleoside transporter proteins (rENT I and rENT2) from rat tissues.J Biol Chem 1997; 272(45): 28423-28430.

14. Yao SY, NgAM, Ritzel MW, Cass CE,Young JD.Transport of adenosine by recombinant purine- and pyrimidine-selective sodium/nucleoside cotransporters from rat jejunum expressed in Xenopus laevis oocytes. Mol Pharm 1996; 50(6): 1529-1535.

15. Che M, Ortiz DF, Arias IM. Primary structure and functional expression of a cDNA encoding the bile canalicular, purine- specific Na+-nucleoside cotransporter.J Biol Chem 1995; 270(23): 13596-13599.

16. Huang QQ,Yao SY, Ritzel MW et al. Cloning and functional expression of a complementary DNA encoding a mammalian nucleoside transport protein.J Biol Chem 1994; 269(27): 17757-17760.

17. Kwong FY, Wu JS, Shi MM et al. Enzymic cleavage as a probe of the molecular structures of mammalian equilibrative nucleoside transporters. J Biol Chem 1993; 268(29): 22127-22134.

18. Coe IR, Griffiths M,Young JD, Baldwin SA, Cass CE.Assignment of the human equilibrative nucleoside transporter (hENTI) to 6p2 I. I-p21.2. Genomics 1997; 45(2): 459-460.

19. Sundaram M,Yao SYM, Cass CE, Baldwin SA,Young JD. Chimaeric constructs between human and rat equilibrative nucleoside transporters (hENTI and rENTI) reveal hENTI structural domains interacting with coronary vasoactive drugs.J Biol Chem 1998: 273:21519-21525.

20. Boleti H, Coe IR, Baldwin SA,Young JD, Cass CE. Molecular identification of the equilibrative NBMPR-sensitive (es) nucleoside transporter and demonstration of an equilibrative NBMPR- insensitive (eO transport activity in human erythroleukemia (K562) cells. Neuropharm 1997; 36(9): 1167-1179.

2 I. Vijayalakshmi D, Belt JA. Sodium-dependent nucieoside transport in mouse intestinal epithelial cells.Two transport systems with differing substrate specificities.J Biol Chem 1988; 263(36): 19419-19423.

22. Huang QQ, Harvey CM, Paterson AR, Cass CE,Young JD. Functional expression of Na(+)-dependent nucleoside transport systems of rat intestine in isolated oocytes of Xenopus laevis. Demonstration that rat jejunum expresses the purine-selective system N I (cif) and a second, novel system N3 having broad specificity for purine and pyrimidine nucleosides.J Biol Chem 1993; 268(27): 20613-20619.

23. Roovers KI, Meckling-Gill KA. Characterization of equifibrative and concentrative Na+-dependent (clio nucleoside transport in acute promyelocytic leukemia NB4 cells.J Cell Physiol 1996; 166(3): 593-600.

24. Belt JA, Harper EH, Byl JA, Noel LD. Sodium-dependent nucleoside transport inhuman myeloid leukemic cell lines and freshly isolated myeloblasts. Proc Annu MeetAm Assoc Canc Res 1992; 33: 20.

25. Belt JA, Marina NM, Phelps DA, Crawford CR. Nucleoside transport in normal and neoplastic cells.Adv Enzyme Regul 1993; 33: 235-252.

26. Flanagan SA, Meckling-Gill KA. Characterization of a novel Na+- dependent, guanosine-specific, nitrobenzylthioinosine-sensitive transporter in acute promyelocytic leukemia cells.J Biol Chem 1997; 272(29): 18026-18032.

27. Paterson AR, Gati WRVijayalakshmi D, et aL Inhibitor-sensitive, sodium-linked transport of nucleoside analogs in leukemia cells from patients. Proc Annu MeetAm Assoc Cancer Res 1993; 34:A84.

28. Cheson BD. Miscellaneous chemotherapeutic agents. In: Devita VT, ed. Cancer: Principles and Practice of Oncology, vol I, 5th edn. Philadelphia: Lippincott-Raven, 1997: 490-498.

29. Aronow B, Ullman B.Thymidine incorporation in nucleoside transport-deficient lymphoma cells.J Biol Chem 1985; 260(30): 16274-16278.

30. Cass CE, Kolassa N, UeharaY, Dahlig-Harley E, Harley ER, Paterson AR.Absence of binding sites for the transport inhibitor nitrobenzylthioinosine on nucleoside transport-deficient mouse lymphoma cells. Biochim Biophys Acta 1981; 649(3): 769-777.

3 I. Cohen A, UIIman B, Martin DW, Jr. Characterization of a mutant mouse lymphoma cell with deficient transport of purine and pyrimidine nucleosides. J Biol Chem 1979; 254( I ): I 12-116.

32. Mackey JR, Mani RS, Selner Met al. Functional nucleoside transporters are required for gemcitabine influx and cytotoxicity in cancer cell lines. Cancer Res 1998: 58: 4349-4357.

33. Ullman B, Coons T, Rockwell S, McCartan K. Genetic analysis of 2',3'-dideoxycytidine incorporation into cultured human T lymphoblasts. J Biol Chem 1988; 263(25): 12391-12396.

34. UIIman B. Dideoxycytidine metabolism in wild type and mutant CEM cells deficient in nucleoside transport or deoxycytidine kinase.Adv Exp Med Biol 1989; 253B: 415-420.

35. Hoffman J. Murine erythroleukemia cells resistant to periodate- oxidized adenosine have lowered levels of nucleoside transporter. Adv Exp Med Biol 199 I; 309A: 443-446.

36. Sobrero AF, Handschumacher RE, Bertino JR. Highly selective drug combinations for human colon cancer cells resistant in vitro to 5- fluoro-2'-deoxyuridine. Cancer Res 1985; 45(7): 3161-3163.

37. EIlison RR, Holland JF, Weil M, et aLArabinosyl cytosine: a useful agent in the treatment of acute leukemia in adults. Blood 1968; 32(4): 507-523.

38. Allegra CJ, Grem JL.Antimetabolites. In: DevitaVT, ed. Cancer: Principles and Practice of Oncology, vol I, 5th edn. Philadelphia: Lippincott-Raven, 1997: 490~98.

39. Kufe D, Spriggs D, Egan EM, Munroe D. Relationships among ara- CTP pools, formation of (ara-C) DNA, and cytotoxicity of human leukemic cells. Blood 1984; 64: 54-48.

40. Zittoun R,Jean-Pierre M, Delanian S, SubervilleA-M,Thevenin D. Prognostic value of in vitro uptake and retention of cytosine arabinoside in acute myelogenous leukemia. Semin Oncol 1987; 14: 269-275.

41. RustumYM, Preisler HD. Correlation between leukemic cell retention of I-B-D-arabinofuranosylcytosine 5'-triphosphate and response to therapy. Cancer Res 1979; 39: 42-49.

42. White JC, Rathmell JR Capizzi RL. Membrane transport influences the rate of accumulation of cytosine arabinoside in human leukemia cells.J Clin Invest 1987; 79(2): 380-387.

© 1998 Harcourt Brace & Co. Ltd Drug Resistance Updates (I 998) I, 310-324 m

43. Jamieson GP, Snook MB,Wiley JS. Saturation of intracellular cytosine arabinoside triphosphate accumulation in human leukemic blast cells. Leuk Res 1990; 14(5): 475-479.

44. Wiley JS, Jones sP, SawyerWH. Cytosine arabinoside transport by human leukaemic cells. EurJ Cancer Clin Oncol 1983; 19(8): 1067-1074.

45. Ho DHW, Frei III E. Clinical pharmacology of I-beta-D- arabinofuranosyl cytosine. Clin Pharmacol Therapeut 197 I; 12: 944-954.

46. Wiley JS,Jones SP, SawyerWH, Paterson AR. Cytosine arabinoside influx and nucleoside transport sites in acute leukemia.J Clin Invest 1982; 69(2): 479-489.

47. Wiley JS,Woodruff RK, Jamieson GP, Firkin FC, Saw'/erWH. Cytosine arablnoside in the treatment ofT-cell acute lymphoblastic leukemia.Aust NZ J Med 1987; 17(4): 379-386.

48. Wiley JS,Taupin J,Jamieson GP, Snook M, SawyerWH, Finch LR. Cytosine arabinoside transport and metabolism in acute leukemias and T cell lymphoblastic lymphoma.J Clin Invest 1985; 75(2): 632-642.

49. Gati WP, Paterson AR, Larratt LM,Turner AR, Belch AR, Sensitivity of acute leukemia cells to cytarabine is a correlate of cellular es nucleoside transporter site content measured by flow cytometry with SAENTA-fluorescein. Blood 1997;90(I):346-353.

50. Gati WP, Paterson ARP, Belch AR, et al. Es nucleoside transporter content of acute leukemia cells: role in cell sensitivity to cytarabine. Leuk Lymph 1998: in press.

5 I. Capizzi RL,Yang JL, Cheng E et aI.AIteration of the pharmacokinetics of high-dose ara-C by its metabolite, high ara-U in patients with acute leukemia.J Clin Oncol 1983; I (I 2): 763-77 I.

52. Weick JK, Kopecky KJ,Appelbaum FRet aI.A randomized investigation of high-dose versus standard-dose cytosine arabinoside with daunorubicin in patients with previously untreated acute myeloid leukemia: a Southwest Oncology Group study. Blood 1996; 88(8): 2841-2851.

53. Bishop JF, Matthews JP, Young GA et aI,A randomized study of high-dose cytarabine in induction in acute myeloid leukemia. Blood 1996;87(5): 1710-1717.

54. Ross SR, McTavish D, Faulds D. Fludarabine.A review of its pharmacological properties and therapeutic potential in malignancy. Drugs 1993; 45(5): 737-759.

55. AvramisVI, PlunkettW. Metabolism and therapeutic efficacy of 9- beta-D-arabinofuranosyl-2-fluoroadenine against murine leukemia P388. Cancer Res 1982; 42:2587-259 I.

56. Barrueco JR, Jacobsen DM, Chang CH, Brockman RW, Sirotnak FM. Proposed mechanism of therapeutic selectivity for 9-beta-D- arabinofuranosyl-2-fluoroadenine against murine leukemia based upon lower capacities for transport and phosphorylation in proliferative intestinal epithelium compared to tumor cells. Cancer Res 1987; 47: 700-706.

57. Sirotnak FM, Barrueco JR. Membrane transport and the antineoplastic action of nucleoside analogues. Cancer Metastasis Rev 1987; 6(4): 459-480.

58. Vijayalakshmi D, Dagnino L, Belt JA, Gati WP, Cass CE, Paterson AR. L I 210/B23. I cells express equilibrative, inhibitor-sensitive nucleoside transport activity and lack two parental nucleoside transport activities. J Biol Chem 1992; 267(24): 1695 I - 16956.

59. Crawford CR, Ng CY, Noel LD, Belt JA. Nucleoside transport in LI 210 murine leukemia cells. Evidence for three transporters.J Biol Chem 1990; 265(I 7): 9732-9736.

60. Patel D, Crawford C, Naeve C, Belt J. Molecular cloning and characterization of a mouse purine-selective concentrative nucleoside transporter (Meeting abstract). Proc Annu Meet Am Assoc Cancer Res 1997; 38:A405.

6 I. Avery TL, Rehg JE, Lumm WC, Harwood FC, Santana VM, Blakley RL. Biochemical pharmacology of 2-chlorodeoxyadenosine in malignant human hematopoietic cell lines and therapeutic effects of 2-bromodeoxyadenosine in drug combinations in mice. Cancer Res 1989; 49(I 8): 4972-4978.

62. King KM, Cass CE. Membrane transport of 2-chloro-2'- deoxyadenosine and 2-chloro-2'-arabinofluoro- 2'- deoxyadenosine is required for cytotoxicity. Proc Annu Meet Am Assoc Cancer Res 1994; 35:A3436.

63. HeinemannV, XuYZ, Chubb S et al. Cellular elimination of 2',2 ~- difluorodeoxycytidine 5'-triphosphate: a mechanism of self- potentiation. Cancer Res 1992; 52(3): 533-539.

64. Baker CH, Banzon J, Bollinger JM et al. 21-Deoxy-2 '- methylenecytidine and 2'-deoxy-2', 2'-difluorocytidine 5'- diphosphates: potent mechanism-based inhibitors of ribonucleotide reductase.J Med Chem 199 I; 34(6): 1879-1884.

65. HeinemannV, Hertel LVV, Grindey GB, PlunkettW. Comparison of the cellular pharmacokinetics and toxicity of 2', 2 ~- difluorodeoxycytidine and I-beta-D-arabinofuranosylcytosine. Cancer Res 1988; 48(I 4): 4024-403 I.

66. Remick SC, Grem JL, Fischer PH et al. Phase I trial of 5-fluorouracil and dipyridamole administered by seventy-two-hour concurrent continuous infusion. Cancer Res 1990; 50(9): 2667-2672.

67. Bailey H,Wilding G,Tutsch KD et aI.A phase I trial of 5-fluorouracil, leucovorin, and dipyridamole given by concurrent 120-h continuous infusions. Cancer Chemother Pharmacol 1992; 30(4): 297-302.

68. Fischer PR, Willson JKV, Risueno C et al. Biochemical assessment of the effects of acivicin and dipyridamole given as a continuous 72-hour intravenous infusion. Cancer Res 1988; 48:5991-5996.

69. Willson JKV, Fischer PH,Tutsch K et al. Phase I clinical trial of a combination of dipyridamole and acivicin based upon inhibition of nucleoside salvage. Cancer Res 1988; 48: 5585-5590.

70. Willson JKV, Fisher PH, Remick SC, et al. Methotrexate and dipyridamole combination chemotherapy based upon inhibition of nucleoside salvage in humans. Cancer Res 1989; 49:1866-1870.

71. Higano CS, Livingston RB. Oral dipyridamole and methotrexate in human solid tumors: a toxicity trial. Cancer Chemother Pharmacol 1989; 23(4): 259-262.

72. Casper ES, Baselga J, SmartTB, Magill GB, Markman M, Ranhosky A.A phase II trial of PALA + dipyridamole in patients with advanced soft-tissue sarcoma. Cancer Chemother Pharmacol 1991;28(I): 51-54.

73. Markman M, Chan TC, Cleary S, Howell SB. Phase i trial of combination therapy of cancer with N-phosphonacetyl -L- aspartic acid and dipyridamole. Cancer Chemother Pharmacol 1987; 19(I):80-83.

74. Szebeni J,Weinstein JN. Dipyridamole binding to proteins in human plasma and tissue culture media.J Lab Clin Med 1991; 117(6): 485-492.

75. Curtin N J, Newell DR, Harris AL. Modulation of dipyridamole action by alpha I acid glycoprotein. Reduced potentiation of quinazoline antifolate (CB3717) cytotoxicity by dipyridamole. Biochem Pharmacol 1989; 38( 19): 3281-3288.

76. Andersen K, Dellborg M, Swedberg K. Nucleoside transport inhibition by draflazine in unstable coronary disease. Eur J Clin Pharm 1996; 51 ( I): 7-13.

Drug Resistance Updates (I 998) I, 310-324 © 1998 Harcourt Brace & Co. Ltd

77. LynchTRJakobs ES, Paran JH, PatersonAR.Treatment of mouse neoplasms with high doses of tubercidin. Cancer Res 1981; 41 (8):

3200-3204. 78. Paterson AR, Paran JH,Yang S, Lynch TR Protection of mice against

lethal dosages of nebularine by nitrobenzylthioinosine, an inhibitor of nucleoside transport. Cancer Res 1979; 39(9): 3607-361 I.

79. Lynch TP, Paran JH, Paterson AR.Therapy of mouse leukemia L 1210 with combinations of nebularine and nitrobenzylthioinosine 5'-monophosphate. Cancer Res 198 I; 41 (2): 560-565.

80. Adjei AA, Dagnino L,Wong MM, Paterson AR. Protection against fludarabine neurotoxicity in leukemic mice by the nucleoside transport inhibitor nitrobenzylthioinosine. Cancer Chemother Pharmacol 1992; 31 (I): 71-75.

81. JansenWJ, Pinedo HM, van derWilt CL, Feller N, Bamberger U, Boven E.The influence of BIBW22BS, a dipyridamole derivative, on the antiproliferative effects of 5-fluorouracil, methotrexate and gemcitabine in vitro and in human tumor xenografts. EurJ Cancer 1995; 31 A( 13-14): 2313-2319.

82. Goh LB, Mack P, Lee CW. Nitrobenzylthioinosine-binding protein overexpression in human breast, liver, stomach and colorectal tumour tissues.Anticancer Res 1995; 15(6B): 2575-2579.

83. Smith CL, Pilarski LM, Egerton ML, Wiley JS. Nucleoside transport and proliferative rate in human thymocytes and lymphocytes. Blood 1989; 74(6): 2038-2042.

84. Wiley JS, Snook MB,Jamieson GR Nucleoside transport in acute leukaemia and lymphoma: close relation to proliferative rate. Br J Haematol 1989; 71 (2): 203-207.

85. Cass CE, King KM, Montano JT, Janowska-WieczorekA.A comparison of the abilities of nitrobenzylthioinosine, dilazep, and dipyridamole to protect human hematopoietic cells from 7- deazaadenosine (tubercidin). Cancer Res 1992; 52(21): 5879-5886.

86. Allay JA, Spencer HT, Wilkinson SL, Belt JA, Blakley RL, Sorrentino BR Sensitization of hematopoietic stem and progenitor cells to trimetrexate using nucleoside transport inhibitors. Blood 1997; 90(9): 3546-3554.

87. Dagnino L, Paterson AR. Sodium-dependent and equilibrative nucleoside transport systems in L I 210 mouse leukemia cells: effect of inhibitors of equilibrative systems on the content and retention of nucleosides. Cancer Res 1990; 50(20): 6549-6553.

88. Chan TC.Augmentation of I-beta-D-arabinofuranosylcytosine cytotoxicity in human tumor cells by inhibiting drug efflux. Cancer Res 1989; 49(I 0): 2656-2660.

89. SantanaVM, Mirro JJ, Harwood FC, et aI.A phase I clinical trial of 2-chlorodeoxyadenosine in pediatric patients with acute leukemia. J Clin Oncol 199 I; 9(3): 416-422.

90. Alessi-Severini S, Gati WR Belch AR, Paterson AR. Intracellular pharmacol<inetics of 2-chlorodeoxyadenosine in leukemia cells from patients with chronic lymphocytic leukemia. Leukemia 1995; 9(10): 1674-1679.

91. Grunewald R,AbbruzzeseJL,Tarassoff R PlunkettW. Saturation of 2',2'-difluorodeoxycytidine 5'-triphosphate accumulation by mononuclear cells during a phase I trial of gemcitabine. Cancer Chemother Pharmacol 199 I; 27(4): 258-262.

92. Petersen AJ, Brown RD, Pope BB, et al. Multiple myeloma: Expression of nucleoside transporters on malignant plasma cells and their relationship to cellular proliferation. Leuk Lymphoma 1994; 13(5-6): 491-499.

93. Meckling-Gill KA, Guilbert L, Cass CE. CSF- I stimulates nucleoside transport in S I macrophages.J Cell Physiol 1993; 155(3): 530-538.

94. Wiley JS, Cebon JS, Jamieson GP et aI.Assessment of proliferative responses to granulocyte-macrophage colony-stimulating factor (GM-CSF) in acute myeloid leukaemia using a fluorescent ligand for the nucleoside transporter. Leukemia 1994; 8(I): 18 I - I 85.

95. Lowenberg B, Suciu S,Archimbaud E et al. Use of recombinant GM-CSF during and after remission induction chemotherapy in patients aged 61 years and older with acute myeloid leukemia:final report ofAML-I I, a phase III randomized study of the Leukemia Cooperative Group of European Organisation for the Research and Treatment of Cancer and the Dutch Belgian Hemato- Oncology Cooperative G~oup. Blood 1997; 90(8): 2952-2961.

96. Wilz F, Harousseau JL, Cahn JY, Abgrall JF. GM-CSF during and after remission induction treatment for elderly patients with acute myeloid leukemia. Blood 1994; 84(suppl I): 23 I A.

97. Lowenberg B, Boogaerts MA, Daenen SM et aI.The value of different modalities of GM-CSF applied during or after induction therapy of acute myeloid leukemia.J Clin Oncol 1998; in press.

98. Ohno R, Naoe T, Kanamaru A et aI.A double-blind controlled study of granulocyte colony-stimulating factor started two days before induction chemotherapy in refractory acute myeloid leukemia. Kohseisho Leukemia Study Group. Blood 1994; 83(8): 2086-2092.

99. Estey E,Thall P, Andreeff Me t aL Use of granulocyte colony- stimulating factor before, during, and after fludarabine plus cytarabine induction therapy of newly diagnosed acute myelogenous leukemia or myelodysplastic syndromes: comparison with fludarabine plus cytarabine without granulocyte colony-stimulating factor.J Clin Oncol 1994; 12(4): 671-678. '

100. Petersen AJ, Brown RD, Gibson Jet al. Nucleoside transporters, bcl-2 and apoptosis in CLL cells exposed to nucleoside analogues in vitro. Eur J Haematol 1996; 56(4): 213-220.

101. Chen SF, Cleaveland JS, Hollmann AB,Wiemann MC, Parks RE Jr, Stoeckler JD. Changes in nucleoside transport of HL-60 human promyelocytic cells during N, N-dimethylformamide induced differentiation. Cancer Res 1986;46(7): 3449-3455.

102. Sokoloski JA, Beardsley GP, Sartorelli AC. Induction of HL-60 leukemia cell differentiation by tetrahydrofolate inhibitors of de novo purine nucleotide biosynthesis. Cancer Chemother Pharmacol 199 I; 28: 39-44.

103. Lee CW, Sokoloski JA, Sartorelli AC, Handschumacher RE. Differentiation of HL-60 cells by dimethylsulfoxide activates a Na(+)-dependent nucleoside transport system. InVivo 1994;8(5): 795-80 I.

104. Lee CW, Sokoloski JA, Sartorelli AC, Handschumacher RE. Induction of the differentiation of HL-60 cells by phorbol 12- myristate 13-acetate activates a Na(+)-dependent uridine- transport system. Involvement of protein kinase C. Biochem J 199 I; 274(Pt I): 85-90.

105. Ishii-Morita H,Agbaria R, Mullen CA et al. Mechanism of 'bystander effect' killing in the herpes simplex thymidine kinase gene therapy model of cancer treatment. Gene Ther 1997; 4(3): 244-25 I.

106. Culver K, Ram Z,Wallbridge Set al. In vivo gene transfer with retroviral vector-producer cells for treatment of experimental brain tumors. Science 1992; 256: 1550-i 553.

© 1998 Harcourt Brace & Co. Ltd Drug Resistance Updates (I 998) I, 310-324 !

107. MahonyWB, Domin BA, ZimmermanTRGanciclovir permeation of the human erythrocyte membrane. Biochem Pharmacol 199 I; 41(2): 263-271.

108. Morgan PE Fine RL, Montgomery P, Marangos PJ. Multidrug resistance in MCF-7 human breast cancer cells is associated with increased expression of nucleoside transporters and altered uptake of adenosine. Cancer Chemother Pharmacol 199 I; 29(2): 127-132.

109. Higashigawa M, Ido M, Kuwabara H et al. Membrane transport of I-beta-D-arabinofuranosylcytosine and accumulation of I-beta-D- arabinofuranosylcytosine 5'-triphosphate in P388 murine leukemic cells resistant to vincristine. Leuk Res 1991; 15(4): 255-262.

I I 0. Bergman AM PH,Veerman G. Increased sensitivity to gemcitabine of MDR human cancer cells. ProcAnnu MeetAmAssoc Cancer Res 1998; 39:216.

I I I. Fox M, Boyle JM, Kinsella AR. Nucleoside salvage and resistance to antimetabolite anticancer agents. Br J Cancer 199 I; 64(3): 428-436.

112. Pickard M, KinsellaA. Influence of both salvage and DNA damage response pathways on resistance to chemotherapeutic antimetabolites. Biochem Pharmacol 1996; 52(3): 425-43 I.

I 13. KinsellaAR, Smith D, Pickard M. Resistance to chemotherapeutic antimetabolites: a function of salvage pathway involvement and cellular response to DNA damage. Br J Cancer 1997; 75(7): 935-945.

114. Weber G. Biochemical strategy of cancer cells and the design of chemotherapy: G.H.A. Clowes Memorial Lecture. Cancer Res 1983; 43(8): 3466-3492.

I 15. Plagemann PG,Wohlhueter RM,Woffendin C. Nucleoside and nucleobase transport in animal cells. Biochim Biophys Acta 1988; 947(3): 405-443.

II 6. Turner RN,Aherne GW, Curtin NJ. Selective potentiation of lometrexol growth inhibition by dipyridamole through cell- specific inhibition of hypoxanthine salvage. Br J Cancer 1997; 76(I 0): 1300-7.

117. Grem JL. Biochemical modulation of fluorouracil by dipyridamole: preclinical and clinical experience. Semin Oncol 1992; 19(2 Suppl

3): 56-65.

118. Grem JL, Fischer PH.Alteration of fluorouracil metabolism in human colon cancer cells by dipyridamole with a selective increase in fluorodeoxyuridine monophosphate levels. Cancer Res 1986;46(12 Pt I):6191-6199.

119. Collins KD, Stark GR.Aspartate transcarbamylase. Interaction with the transition state analogue N-(phosphonacetyl)-L- aspartate. J Biol Chem 197 I; 246(21 ): 6599-605.

120. Van MouwerikTJ, Pangallo CA,Willson JK, Fischer PH. Augmentation of methotrexate cytotoxicity in human colon cancer cells achieved through inhibition of thymidine salvage by dipyridamole. Biochem Pharmacol 1987; 36(6): 809-814.

121. Cabral S, Leis S, Borer L, Nembrot M, Mordoh J. Dipyridamole inhibits reversion by thymidine of methotrexate effect and increases drug uptake in Sarcoma 180 cells. Proc Natl Acad Sci USA 1984; 81 (10): 3200-3203.

122. Jarvis SM, Hammond JR, Paterson AR, Clanachan AS. Species differences in nucleoside transport.A study of uridine transport and nitrobenzylthioinosine binding by mammalian erythrocytes. Biochem J 1982; 208( I ): 83-88.

123. "fang JL, White JC, Capizzi RL. Modulation of the cellular pharmacokinetics of ara-CTP in human leukemic blasts by dipyridamole. Cancer Chemother Pharmacol 1992;29(3): 236-240.

124. Jamieson GR Snook MB, BradleyTR, Bertoncello I,Wiley JS. Transport and metabolism of I-beta-D-arabinofuranosylcytosine in human ovarian adenocarcinoma cetls. Cancer Res 1989; 49(2): 309-313.

125. Boumah CE, Hogue DL, Cass CE. Expression of high levels of nitrobenzylthioinosine-sensitive nucleoside transport in cultured human choriocarcinoma (BeWo) cells. Biochem J 1992; 288(Pt 3): 987-996.

126. Goh LB, Lee CW. Reduction of equilibrative nitrobenzylthioinosine-sensitive nucleoside transporter in tamoxifen-treated MCF-7 cells: an oestrogen-reversible phenomenon. Biochem J 1997; 327(Pt I): 31-36.

127. Paterson AR,Yang SE, Lau EY, Cass CE. Low specificity of the nucleoside transport mechanism of RPMI 6410 cells. Mo~ Pharmacol 1979; 16(3): 900-908.

E Drug Resistance Updates (I 998) I, 310-324 © 1998 Harcourt Brace & Co. Ltd