The Role of Human Nucleoside Transporters in Cellular Uptake of 4 -Thio- -D-arabinofuranosylcytosine and -D-Arabinosylcytosine

4 -Thio-D-arabinofuranosyl cytosine (TaraC) is in phase I development for treatment of cancer. In human equilibrative nucleoside transporter (hENT) 1-containing CEM cells, initial rates of uptake (10 M; picomoles per microliter of cell water per second) of [H]TaraC and [H]1-D-arabinofuranosyl cytosine (araC) were low (0.007 003 and 0.034 0.003, respectively) compared with that of [H]uridine (0.317 0.048), a highactivity hENT1 permeant. In hENT1and hENT2-containing HeLa cells, initial rates of uptake (10 M; picomoles per cell per second) of [H]TaraC, [H]araC, and [H]deoxycytidine were low (0.30 0.003, 0.42 0.03, and 0.51 0.11, respectively) and mediated primarily by hENT1 ( 74, 65, and 61%, respectively). In HeLa cells with recombinant human concentrative nucleoside transporter (hCNT) 1 or hCNT3 and pharmacologically blocked hENT1 and hENT2, transport of 10 M [H]TaraC and [H]araC was not detected. The apparent affinities of recombinant transporters (produced in yeast) for a panel of cytosine-containing nucleosides yielded results that were consistent with the observed low-permeant activities of TaraC and araC for hENT1/2 and negligible permeant activities for hCNT1/ 2/3. During prolonged drug exposures of CEM cells with hENT1 activity, araC was more cytotoxic than TaraC, whereas coexposures with nitrobenzylthioinosine (to pharmacologically block hENT1) yielded identical cytotoxicities for araC and TaraC. The introduction by gene transfer of hENT2 and hCNT1 activities, respectively, into nucleoside transport-defective CEM cells increased sensitivity to both drugs moderately and slightly. These results demonstrated that nucleoside transport capacity (primarily via hENT1, to a lesser extent by hENT2 and possibly by hCNT1) is a determinant of pharmacological activity of both drugs. The synthesis of 4 -thio-D-arabinofuranosyl cytosine (TaraC) was reported more than 30 years ago (Whistler et al., 1971). In recent studies (Waud et al., 2003), TaraC and araC were evaluated for activity against a panel of human tumor cell lines. TaraC differs from araC in the replacement of the oxygen atom by a sulfur atom in the arabinose ring. Although araC was more cytotoxic than TaraC in cell culture studies, TaraC showed superior antitumor activity in human tumor systems in mice compared with that of araC (Waud et al., 2003). AraC, which has shown promising activity in human leukemia xenograft models, was poorly active against human solid tumor xenografts (Gourdeau et al., 2001a). Although previous reports indicated that araC is a better substrate than TaraC for deoxycytidine kinase (Parker et al., 2000; Someya et al., 2002), recent studies (Someya et al., 2003) suggest that the difference in phosphorylation rates at pharmacologically relevant concentrations is only 2to 3-fold. The prolonged retention time of TaraC-5 -monophosphate resulting from its high rate of phosphorylation by UMP/CMP kinase has been suggested as a more important determinant for its activity against solid tumors than its resistance to cytidine and dCMP deaminase activities (Parker et al., 2000; Someya et al., 2005). Preliminary studies indicated that the This work was supported by OSI Pharmaceuticals, a Canadian Cancer Society Grant to C.E.C. from the National Cancer Research Institute of Canada, and the Alberta Cancer Foundation of the Alberta Cancer Board. C.E.C. is Canada Research Chair in Oncology. J.D.Y. is a Heritage Scientist of the Alberta Heritage Foundation for Medical Research. J.Z. held studentships from the Canadian Institutes of Health Research and the Alberta Heritage Foundation for Medical Research. Portions of this work were presented in a preliminary report (Clarke et al., 2003). M.L.C. and V.L.D. are co-first authors. 1 Current affiliation: GTI Inc., Waukesha, WI. Article, publication date, and citation information can be found at http://molpharm.aspetjournals.org. doi:10.1124/mol.105.021543. ABBREVIATIONS: TaraC, 4 -thio-D-arabinofuranosylcytosine; araC, 1-D-arabinofuranosylcytosine; NBMPR, nitrobenzylmercaptopurine riboside; hENT, human equilibrative nucleoside transporter; hCNT, human concentrative nucleoside transporter. 0026-895X/06/7001-303–310$20.00 MOLECULAR PHARMACOLOGY Vol. 70, No. 1 Copyright © 2006 The American Society for Pharmacology and Experimental Therapeutics 21543/3122211 Mol Pharmacol 70:303–310, 2006 Printed in U.S.A. 303 at A PE T Jornals on O cber 3, 2017 m oharm .aspeurnals.org D ow nladed from mechanism of TaraC cytotoxicity differs from that of either araC or gemcitabine, a difluoro-2 -deoxycytidine analog with activity against solid tumors (Blajeski et al., 2002). TaraC was a better substrate for nuclear DNA polymerase and than either araC or gemcitabine, and its incorporation into DNA resulted in strong chain termination in comparison with that of araC and gemcitabine, which resulted in minimal termination (Richardson et al., 2004). In recent studies in which mice bearing human lung cancer xenografts (Calu-6, A549) were treated with graded doses of TaraC, TaraC was incorporated into internal nucleotide linkages in a manner that was doseand time-dependent, leading to the conclusion that DNA synthesis may play a role in TaraC cytotoxicity (Richardson et al., 2005). An area that has been neglected to date is the characterization of the uptake parameters of TaraC by plasma membrane nucleoside transporters. It is not known whether mediated uptake is required for TaraC cytotoxicity, as reported for araC (Ullman, 1989; Gati et al., 1997) and gemcitabine (Mackey et al., 1998a,b), or whether TaraC enters cells primarily by passive diffusion as reported for troxacitabine, a deoxycytidine derivative with an unusual dioxolane structure and non-natural L-configuration with activity against leukemic and solid tumor xenograft models (Gourdeau et al., 2001b). In humans, seven functionally distinct nucleoside transport processes have been characterized in molecular terms through isolation and functional expression of cDNAs encoding the transporter proteins in Xenopus laevis oocytes, mammalian cells, or yeast (Griffiths et al., 1997a,b; Ritzel et al., 1997, 1998, 2001; Baldwin et al., 2005). The human (h) nucleoside transporters exhibit different permeant selectivities, and only hENT1 is inhibited by nanomolar concentrations of nitrobenzylmercaptopurine riboside (NBMPR). The proteins (and their permeant selectivities) are as follows: hENT1 (accepts purine and pyrimidine nucleosides), hENT2 (accepts purine and pyrimidine nucleosides and nucleobases), hENT3 (accepts purine and pyrimidine nucleosides), hENT4 (accepts adenosine), hCNT1 (accepts pyrimidine nucleosides and adenosine), hCNT2 (accepts purine nucleosides and uridine), and hCNT3 (accepts purine and pyrimidine nucleosides). Six of the seven transporters (hENT1/2/4, hCNT1/2/3) are found in plasma membranes, whereas hENT3 is found in lysosomal membranes. This study was undertaken to characterize the transportability of TaraC by the human nucleoside transporters that are responsible for cellular uptake of exogenous nucleosides and to assess the importance of nucleoside transport in TaraC cytotoxicity by using cultured human cell lines with known nucleoside transporter phenotypes. The study was designed to compare TaraC and araC directly because of the structural similarities of the two compounds and the established role of araC in cancer chemotherapy. Transportability studies were conducted with well established model systems producing either native (cultured human cell lines) or recombinant human nucleoside transporters (yeast and Xenopus laevis oocyte expression systems) that enabled functional isolation of the transporter under investigation. Cytotoxicity studies were conducted in CEM cells with native hENT1 in the presence and absence of transport inhibitors or in a nucleoside transport-defective mutant line (CEM-ARAC) with either recombinant hENT2 or hCNT1 introduced by stable transfection. The goal was to determine whether nucleoside transporter-mediated accumulation of TaraC contributes to its unique pharmacological profile, which includes prolonged intracellular retention and schedule independence. Materials and Methods Materials. TaraC was provided by OSI Pharmaceuticals, Inc. (Boulder, CO). [5-H]AraC (20.3 Ci/mmol), [5-H]TaraC (8.2 Ci/ mmol), [5-H (N)]2 -deoxycytidine (27 Ci/mmol), and [5-H]uridine (40 Ci/mmol) were purchased from Moravek Biochemicals (Brea, CA). Research grade uridine, cytidine, 2 -deoxycytidine, 3 -deoxycytidine, 2 ,3 -dideoxycytidine, gemcitabine, and araC were from Sigma-Aldrich (St. Louis, MO). All other reagents used were analytical grade and from commercial sources. Cell Culture. The human CCRF-CEM leukemia, hereafter termed CEM, and HeLa cervical carcinoma cell lines were obtained, respectively, from William T. Beck (formerly at St. Jude Children’s Research Hospital, now at University of Illinois at Chicago, Chicago, IL) and the American Type Culture Collection (Manassas, VA). CEM/ARAC8C, a nucleoside transport-deficient derivative of CEM (Ullman et al., 1988), referred to as CEMARAC, was a gift from Dr. B. Ullman (Oregon Health and Science University, Portland, OR). CEM-ARAC/hCNT1 was derived from CEM-ARAC by stable transfection with a pcDNA3 plasmid that contained the coding sequence for hCNT1 (Lang et al., 2004). hENT2 stable transfectants (CEM-ARAC/hENT2) were produced using procedures described previously (Lang et al., 2001, 2004). Cells were grown in RPMI 1640 medium supplemented with 10% fetal bovine serum (Invitrogen, Burlington, ON, Canada) as adherent (HeLa) or suspension cultures (CEM, CEM-ARAC, CEM-ARAC/hCNT1, and CEM-ARAC/hENT2). Cells were free of mycoplasma, maintained in the absence of antibiotics, incubated at 37°C in a humidified atmosphere (5% CO2), and subcultured at 2to 4-day intervals to maintain active proliferation. Cell numbers were determined with a Coulter Z2 electronic particle counter equipped with a size analyzer (Beckman Coulter Canada, Mississauga, ON, Canada). Transient Expression of hCNT1 and hCNT3 in HeLa Cells. The cDNAs encoding the hCNT1 and hCNT3 proteins were subcloned into the mammalian expression vector, pcDNA3, to produce pcDNA3-hCNT1 (Graham et al., 2000) and pcDNA3-hCNT3 (this work). The plasmids were separately transfected into actively proliferating HeLa cells as described previously (Mackey et al., 1998b; Graham et al., 2000). Nucleoside Uptake Assays in Cultured Cells. Nucleoside uptake assays were conducted at room temperature in transport buffer [20 mM Tris-HCl, 3 mM K2HPO4, 1 mM MgCl2(6H2O), 2 mM CaCl2, 5 mM glucose, and 130 mM NaCl, pH 7.4; 300 15 mOsM] in CEM or HeLa cells, respectively, as described for suspension cultures (Boleti et al., 1997) or adherent cultures (Graham et al., 2000). Cells were lysed with 5% Triton X-100 and mixed with Ecolite scintillation fluid to measure cell-associated radioactivity (Beckman LS 6500 scintillation counter; Beckman Coulter Canada). Transport Inhibition Assays in Yeast with Recombinant Transporters. Saccharomyces cerevisiae that were separately transformed with plasmids (pYPhENT1, pYPhENT2, pYPhCNT1, pYPhCNT2, pYPhCNT3) with cDNAs encoding, respectively, the hENT1, hENT2, hCNT1, hCNT2, or hCNT3 as described previously (Vickers et al., 2002; Visser et al., 2002; Zhang et al., 2003, 2005), were used to examine the ability of various cytosine-containing analogs to inhibit the uptake of 1 M [H]uridine. Transport experiments were conducted in 96-well plates with a semiautomated cell harvester (Micro96 HARVESTER; Skatron Instruments, Lier, Norway) as described previously (Zhang et al., 2003). Yeast were incubated at room temperature with graded concentrations of test compounds in the presence of 1 M [H]uridine for 20 min. Each 304 Clarke et al. at A PE T Jornals on O cber 3, 2017 m oharm .aspeurnals.org D ow nladed from experiment was repeated at least three times. Data were subjected to nonlinear regression analysis using Prism software (version 3.0; GraphPad Software Inc., San Diego, CA) to obtain IC50 values that were used to calculate Ki values (Cheng and Prusoff, 1973). Electrophysiological Studies in Frog Oocytes with Recombinant hCNT1. hCNT1 cDNA contained in the plasmid pGEM-HE (Ritzel et al., 1997) was linearized with NheI and transcribed with T3 or T7 polymerase using the mMESSAGE mMACHINE (Ambion, Austin, TX) transcription system. In vitro synthesized transcripts were injected into isolated mature stage VI oocytes from X. laevis, and oocyte membrane currents were measured using a GeneClamp 500B oocyte clamp (Molecular Devices, Sunnyvale, CA) in the twoelectrode, voltage-clamp mode as described previously (Smith et al., 2004, 2005). Mock-injected oocytes were injected with water alone. All experiments were performed at room temperature (20°C) and oocytes were discarded if membrane potentials were unstable or more positive than 30 mV. The membrane potential was clamped at a holding potential of 50 mV and either 100 M uridine or 1 mM TaraC, araC, or 2 -deoxycytidine was added. The transport medium contained 100 mM NaCl, 2 mM KCl, 1 mM CaCl2, 1 mM MgCl2, and 10 mM HEPES, pH 7.5. Current values are presented as means S.E. of three or more oocytes. Chemosensitivity Testing with Cultured CEM Cells. The relative cytotoxicities of TaraC and araC against CEM, CEM-ARAC, CEM-ARAC/hENT2, and CEM-ARAC/hCNT1 were assessed using the CellTiter 96 proliferation assay (Promega, Madison, WI) as described previously (Damaraju et al., 2005). This assay is based on the reduction of a tetrazolium compound to a soluble formazan derivative by the dehydrogenase enzymes of metabolically active cells. The absorbance (490 nm) is directly proportional to the number of living cells in culture. Cells were seeded in 96-well tissue culture plates (10 cells/well, six replicates/condition) and exposed to graded concentrations (0–10 nM) of TaraC or araC for 24 or 48 h. Chemosensitivity was expressed as the effective drug concentration at which cell proliferation was inhibited by 50% (EC50 values) and was determined from concentration-effect relationships using GraphPad