Erlotinib, Gefitinib, and Vandetanib Inhibit Human Nucleoside Transporters and Protect Cancer Cells from Gemcitabine Cytotoxicity

Published OnlineFirst October 29, 2013; DOI: 10.1158/1078-0432.CCR-13-2293

Clinical Cancer Cancer Therapy: Preclinical Research

Erlotinib, Geﬁtinib, and Vandetanib Inhibit Human Nucleoside Transporters and Protect Cancer Cells from Gemcitabine Cytotoxicity

Vijaya L. Damaraju1, Tara Scriver1, Delores Mowles1, Michelle Kuzma1, Anderson J. Ryan3, Carol E. Cass1, and Michael B. Sawyer1,2

Abstract Purpose: Combinations of tyrosine kinase inhibitors (TKI) with gemcitabine have been attempted with little added benefit to patients. We hypothesized that TKIs designed to bind to ATP-binding pockets of growth factor receptors also bind to transporter proteins that recognize nucleosides. Experimental Design: TKI inhibition of uridine transport was studied with recombinant human (h) equilibrative (E) and concentrative (C) nucleoside transporters (hENT, hCNT) produced individually in yeast. TKIs effects on uridine transport, gemcitabine accumulation, regulation of hENT1 activity, and cell viability in the presence or absence of gemcitabine were evaluated in human pancreatic and lung cancer cell lines. Results: Erlotinib, gefitinib and vandetanib inhibited [3H]uridine transport in yeast and [3H]uridine and [3H]gemcitabine uptake in the four cell lines. Treatment of cell lines with erlotinib, gefitinib, or vandetanib for 24 hours reduced hENT1 activity which was reversed by subsequent incubation in drug-free media for 24 hours. Greater cytotoxicity was observed when gemcitabine was administered before erlotinib, gefitinib, or vandetanib than when administered together and synergy, evaluated using the CalcuSyn Software, was observed in three cell lines resulting in combination indices under 0.6 at 50% reduction of cell growth. Conclusions: Vandetanib inhibited hENT1, hENT2, hCNT1, hCNT2, and hCNT3, whereas erlotinib inhibited hENT1 and hCNT3 and gefitinib inhibited hENT1 and hCNT1. The potential for reduced accumulation of nucleoside chemotherapy drugs in tumor tissues due to inhibition of hENTs and/or hCNTs by TKIs indicates that pharmacokinetic properties of these agents must be considered when scheduling TKIs and nucleoside chemotherapy in combination. Clin Cancer Res; 20(1); 176–86. 2013 AACR.

Introduction ever, in one of the pivotal randomized studies, gefitinib was The U.S. Food and Drug Administration (FDA) approved combined at doses of 250 or 500 mg/day with gemcitabine/ the tyrosine kinase inhibitor (TKI) erlotinib for mainte- cisplatin with expectations that the two gefitinib combina- nance of patients with locally advanced or metastatic non– tions would be superior to gemcitabine/cisplatin alone in small cell lung cancer (NSCLC) and more recently as a first- NSCLC but there were no statistical differences between line treatment in patients with NSCLC with EGF receptor placebo and treatment arms in terms of progression-free (EGFR) exon 19 deletions or exon 21 (L858R) substitution survival (PFS; ref. 2). In a phase II trial in patients with mutations. Similarly, gefitinib has been approved for EGFR advanced urothelial carcinoma treated with gemcitabine/ mutation–positive NSCLC in more than 80 countries cisplatin and concurrent gefitinib, response rates were not worldwide. It showed impressive, single-agent activity and significantly different from gemcitabine/cisplatin alone (3). received accelerated FDA approval based on early phase II Gemcitabine and gefitinib were evaluated in patients with studies pending definitive randomized studies (1). How- advanced pancreatic cancer (4) and response rates were similar to those observed in a pivotal phase III study in patients with locally advanced or metastatic pancreatic Authors' Affiliations: 1Department of Oncology, University of Alberta; cancer with gemcitabine or gemcitabine and erlotinib with 2Department of Medical Oncology, Cross Cancer Institute, Edmonton, median survival of 5.9 to 6.2 months. Addition of gefitinib Alberta, Canada; and 3Department of Oncology, University of Oxford, Oxford, United Kingdom to gemcitabine in a different phase II trial in patients with urothelial carcinoma did not result in increased response Corresponding Author: Michael B. Sawyer, Department of Medical Oncol- ogy, Cross Cancer Institute, 11560 University Avenue NW, Edmonton, rates compared with the control arm of gemcitabine/cis- Alberta T6G 1Z2 Canada. Phone: 780-432-8248; Fax: 780-432-8888; platin (5). Vandetanib was initially evaluated in patients E-mail: [email protected] with advanced NSCLC and the combination of vandetanib doi: 10.1158/1078-0432.CCR-13-2293 with gemcitabine/cisplatin was not tolerated in these 2013 American Association for Cancer Research. patients (6).

176 Clin Cancer Res; 20(1) January 1, 2014

EGFR and VEGFR TKIs Inhibit Human Nucleoside Transporters

Materials and Methods Translational Relevance Materials Tyrosine kinase inhibitors (TKI) such as gefitinib and Nitrobenzylmercaptopurine ribonucleoside (NBMPR), erlotinib have been combined with nucleoside chemo- dilazep, unlabeled nucleosides, and other chemicals were therapy with negative or disappointing results. We noted obtained from Sigma Chemical Company. Tritiated nucleo- that vandetanib and other TKIs share structural features sides were purchased from Moravek Biochemicals. Tissue with a potent inhibitor of human equilibrative nucleo- culture (96- and 12-well) plates and flasks were from VWR side transporter 1 (hENT1) nitrobenzylmercaptopurine International. Cell culture media and FBS were from Gibco ribonucleoside (NBMPR), and speculated that failures of BRL. Ecolite was from ICN Pharmaceuticals. Dojindo Cell combination of TKIs with nucleoside chemotherapy Counting Kit-8 (CCK-8) was from Dojindo Molecular may be due to TKIs inhibiting human nucleoside trans- Technologies, Inc. Erlotinib, gefitinib, and vandetanib were porters (hNT) blocking entry of nucleoside drugs. We from LC Laboratories. showed that gefitinib, erlotinib, and vandetanib inhibited hENT1, the hNT necessary for activity of many nucleoside drugs. Not only did TKIs inhibit hNTs direct- Cell culture ly, they also inhibited hENT1 activity indirectly by Human lung cancer cell lines A549, H292, and H1975 reducing hENT1 protein in plasma membranes. hNT and the human pancreatic cancer cell line AsPC-1 were inhibition may be responsible for failure to successfully obtained from American Type Culture Collection (ATCC). combine nucleosides and TKIs. This previously unno- Cell lines were sent to DDC Medical to verify their authen- ticed feature of EGFR TKIs likely explains why they have ticity and mycoplasma status. Results showed that cell lines failed in gastrointestinal cancer combination chemo- were 100% matched to the ATCC panel of markers and were therapy for which the backbone is either capecitabine free of mycoplasma. Cells were maintained in RPMI-1640 or gemcitabine, known nucleoside permeants of hNTs. medium supplemented with 10% FBS, 2 mmol/L L-gluta- mine, and 10% glucose. All cultures were kept at 37Cin5% CO2/95% air and subcultured at 2 to 3-day intervals to maintain exponential growth. Transport and cytotoxicity Combining conventional cytotoxic drugs with novel experiments were conducted with cells in exponential agents that target key signaling pathways that control cancer growth phase. cell survival, proliferation, and/or invasion should be a promising approach, and several clinical trials of TKIs (2– Choice of TKI concentrations to be studied 6) suggest that there may be unfavorable interactions Plasma concentrations of gefitinib, erlotinib, and van- between TKIs and nucleoside chemotherapy drugs. Although detanib are 0.4, 2.5 and 2.0 mmol/L, respectively (13– literature reports suggest that TKIs interfere with uptake of 15). Earlier studies have shown that gefitinib concentra- nucleoside chemotherapy drugs (7–10), detailed studies of tions in tumor tissues were 42-fold higher than plasma TKI effects on individual human nucleoside transporters concentrations (15). Similarly, vandetanib tumor con- (hNT) or a theory to explain why TKIs inhibit hNTs are centrations were 30-fold higher in tumor tissues than in lacking. Transporters for physiologic nucleosides and nucle- plasma (16). Given structural similarities among the oside analogs include human equilibrative nucleoside trans- three TKIs, we expected similar concentrating effects to porters 1–4 (hENT1-4) and human concentrative nucleoside be seen with erlotinib and assumed that gefitinib, erlo- transporters (hCNT1-3). Detailed summaries of hNTs’ roles tinib, and vandetanib concentrations were likely to be in transport of nucleoside drugs can be found in recent considerably higher in tumor tissues than in plasma (15, reviews (11, 12). 16). We therefore studied TKI concentrations up to 100 We hypothesized that poor clinical outcomes of com- mmol/L in cell lines and 500 mmol/L in yeast model bining erlotinib, gefitinib, or vandetanib with nucleoside systems. chemotherapy drugs (e.g., gemcitabine) were due to TKI inhibition of hNTs, resulting in lowered intracellular accu- Uridine transport in Saccharomyces cerevisiae mulation of nucleoside drug metabolites and reduced Saccharomyces cerevisiae yeast were separately transfor- treatment efficacy. To study potential interactions between med with plasmids (pYPhENT1, pYPhENT2, pYPhCNT1, TKIs and hNTs, we investigated TKI inhibition of uridine pYPhCNT2, or pYPhCNT3) encoding hNTs (hENT1, hENT2, transport in yeast cells producing each of recombinant hCNT1, hCNT2, or hCNT3, respectively) as described else- hNTs individually and TKI effects on uridine uptake, where (17, 18). Uptake of 1 mmol/L [3H]uridine (Moravek gemcitabine accumulation, and cytotoxicity in one pan- Biochemicals) into yeast was measured as previously creatic adenocarcinoma cell line AsPC-1 and three human described (18, 19) using the semiautomated cell harvester NSCLC cells that differ in their EGF receptor (EGFR) and (Micro96 HARVESTER; Skatron Instruments). Yeast were KRAS mutation status—i.e., A549 (EGFR wild-type and incubated at room temperature with graded concentrations KRAS mutant), H292 (EGFR wild-type and KRAS wild- (0–0.5 mmol/L) of test compounds in the presence of 1 type), and H1975 (EGFR L858R, T790M mutation and mmol/L [3H]uridine in yeast growth media at pH 7.4. Yeast wild-type KRAS). were incubated with 1 mmol/L [3H]uridine in the absence of

www.aacrjournals.org Clin Cancer Res; 20(1) January 1, 2014 177

Damaraju et al.

test compound and the resulting uninhibited (i.e., control) gefitinib, or vandetanib (A549, AsPC-1 cells) or 2.5 mmol/L values were used to determine "% Control" values. Uridine erlotinib, gefitinib, or vandetanib (H292, H1975 cells) and self-inhibition was used to determine maximum inhibition allowed to recover for 24 hours in drug-free media after of mediated transport. which uridine uptake in untreated and treated cells was Concentration–effect curves were subjected to nonlinear measured in drug-free media. regression analysis using Prism software (version 4.03; GraphPad Software Inc.) to obtain the concentration of Cell-surface staining and confocal microscopic test compound that inhibited uridine uptake by 50% rel- visualization of hENT1 sites on A549 cells with 50-S-[2- 0 ative to that of untreated cells (IC50 values). Each IC50 value (6-aminohexanamido)]ethyl-6-N-(4-nitrobenzyl)-5 - determination was conducted with nine concentrations and thioadenosine-fluorescein-5-yl isothiocyanate six replicates per concentration and experiments were SAHENTA-FITC repeated three times. Synthesis and use of a fluorescent probe for evaluation of cell surface hENT1 sites was described earlier (20). For Nucleoside transport inhibition in A549, H292, H1975, evaluation of cell-surface abundance of hENT1, A549 cells and AsPC-1 cells grown on cover slip dishes were treated for 0 or 24 h with or Cells (105/well) were seeded in 12-well plates and on the without 5 mM erlotinib, gefitinib or vandetanib and allowed third day, uptake of [3H]nucleosides was measured at room to recover for 0 or 24 h in drug-free media. Cells were (i) temperature in transport buffer (pH 7.4) containing 20 washed twice with PBS, (ii) stained with 100 nmol/L 50-S- 0 mmol/L Tris, 3 mmol/L K2HPO4, 1 mmol/L MgCl2, 1.4 [2-(6-aminohexanamido)]ethyl-6-N-(4-nitrobenzyl)-5 - mmol/L CaCl2, and 5 mmol/L glucose with 144 mmol/L thioadenosine-fluorescein-5-yl isothiocyanate (SAHENTA- NaCl hereafter termed transport buffer. For uridine uptake FITC) in sodium buffer for 30 minutes at room temperature, assays, cell growth medium was aspirated, cells were washed (iii) washed with PBS, and (iv) resuspended in a small with sodium or sodium-free buffer, [3H]uridine was added, volume of PBS for fluorescence visualization. Confocal and uptake was measured over fixed timepoints in the microscopic analysis was done using a Zeiss 710 LSM with presence or absence of known NT inhibitors or graded Zeiss Plan-Apochromat 40 /1.3 oil DIC M27 lens with an concentrations of TKIs. Effects of TKIs on kinetics of uridine argon laser 488 and a photomultiplier tube detector. Images uptake were determined in A549 cells at graded concentra- were collected as 12 bit with 1024 1024 dimensions with tions of [3H]uridine (0–1000 mmol/L) at 0, 25, 50, or 100 pixel dwell time of 1.27 msec and Zen 2011 software was mmol/L of either erlotinib, gefitinib, or vandetanib using 30- used for analysis of the images. second incubations (established in preliminary experiments to be from linear portions of time courses of Cytotoxicity assays [3H]uridine uptake in A549 cells, data not shown). At the Dojindo Cell Counting Kit-8 (CCK-8) was used to end of uptake intervals, permeant-containing solutions quantify drug induced cytotoxicity. Cells were seeded in were removed by aspiration and cells were quickly rinsed 96-wellplatesandallowedto attach for 24 hours. Cells twice with sodium buffer and solubilized with 5% TritonX- werethenexposedtogradedconcentrationsofgemcita- 100. Radioactivity in solubilized extracts was measured by bine, vandetanib, or gefitinib in the absence or presence liquid scintillation counting. Uptake values were expressed of 1 mmol/L NBMPR (inhibits hENT1) or 25 mmol/L as pmol/106 cells and graphs generated using the Prism dilazep (inhibits hENT1/2) for 72 hours after which they software. Each experiment was conducted two or three were treated with CCK-8 reagent for assessment of cyto- times with triplicate measurements. toxicity. For evaluation of in vitro synergy of combinations For [3H]uridine uptake and [3H]gemcitabine accumula- of gemcitabine with any of the TKIs, experiments were tion experiments, cells were exposed to either 10 mmol/L based on the individual drug’s IC50 value, with the high- 3 3 [ H]uridine for one minute or 1 mmol/L [ H]gemcitabine in est concentration being 8 IC50 values as described sodium-containing transport buffer for 60 minutes in earlier (21). For sequential treatments, cells were treated absence or presence of 25 mmol/L TKI (erlotinib, gefitinib, with either drug for 24 hours, followed by drug-free or vandetanib) or 100 mmol/L dilazep (an inhibitor of media for 24 hours and subsequent 72-hour treatment hENT1 and hENT2) and processed as described above for with the other drug; simultaneous treatments were per- uptake assays. formed for 72 hours after 48 hours in drug-free media. To determine a suitable timepoint for study of recovery of Drug synergy was determined by the isobologram and transport activity in cell lines treated with TKIs, a prelim- combination index methods (CI), derived from the medi- inary experiment with A549 cells was conducted in which an effect principle of Chou and Talalay (22) using the cells were treated with vandetanib for 0, 2, 4, 6, 12, and 24- CalcuSyn software (Biosoft). Using data from the growth hour exposures followed by 24-hour recovery in drug-free inhibitory experiments and computerized software, CI medium. Decreased uridine transport activities were values were generated over a range of fraction affected observed as early as 2 hours and the effect was maximal at (Fa) levels from 0.05–0.90 (5%–90% growth inhibition). 24 hours (data not shown). We used 24-hour exposures for A CI of 1 indicates an additive effect between two agents, subsequent experiments with all four cell lines. Cells were whereas a CI < 1or>1 indicates synergism or antagonism, treated for 0 or 24 hours with or without 5 mmol/L erlotinib, respectively.

178 Clin Cancer Res; 20(1) January 1, 2014 Clinical Cancer Research

EGFR and VEGFR TKIs Inhibit Human Nucleoside Transporters

Results by the three TKIs. Vandetanib inhibited hENT1, hENT2, Inhibition of [3H]uridine uptake by erlotinib, gefitinib, hCNT1, hCNT2, and hCNT3, whereas erlotinib inhibited and vandetanib in Saccharomyces cerevisiae and in cell hENT1 and hCNT3 and gefitinib inhibited hENT1 and lines hCNT1. Erlotinib, gefitinib, and vandetanib (chemical structures To assess TKI inhibition of hNT activity in A549, H292, are shown in Fig. 1A–C) were assessed for their relative H1975, and AsPC-1 cells, we first established the type of abilities to inhibit transport of [3H]uridine by each of the hNT activities present by conducting [3H]uridine uptake five recombinant hNTs produced in yeast in inhibition experiments in sodium-containing or sodium-free media experiments to determine IC50 values (inhibitor concentra- with or without 100 nmol/L NBMPR (inhibits hENT1), or tions that produced 50% inhibition of transport). Repre- 100 mmol/L dilazep (inhibits hENT1/2) or 1 mmol/L sentative concentration–effect curves for inhibition of excess uridine (inhibits hENT1/2, hCNT1/2/3) in each hENT1-mediated uridine transport by erlotinib, gefitinib, of the cell lines. Uridine uptake activities in A549, H292, or vandetanib are shown in Fig. 1D–F and for each, inhi- and H1975 cells are shown in Fig. 2A–C and the % bition of [3H]uridine transport was observed at micromolar inhibition was similar with the various NT inhibitors, concentrations. IC50 values obtained from such experi- indicating that uridine uptake was mediated primarily by ments with erlotinib, gefitinib, and vandetanib for each of hENT1. Similar results were seen with AsPC-1 cells (data the five recombinant NTs are shown in Table 1, where it is not shown). Since erlotinib, gefitinib, and vandetanib evident that hENT1 was inhibited by all three TKIs, whereas were identified as hENT1 inhibitors in yeast radiotracer hENT2 and hCNT1/2/3 were inhibited to different extents experiments, they were tested in the four cell lines to

AD Erlotinib 100

(% control) 25 H]Uridine uptake

Figure 1. Structures of erlotinib, 3 fi [ 0 ge tinib, and vandetanib and their -7 -6 -5 -4 -3 effects on [3H]uridine uptake by Log [Erlotinib] (mol/L) recombinant hENT1 in yeast cells. Structures of erlotinib, gefitinib, and vandetanib are shown in (A–C). Yeast cells were incubated with BE 1 mmol/L [3H]uridine for 10 minutes Gefitinib in the absence or presence of 100 increasing concentrations (0–300 mmol/L) of erlotinib, gefitinib, or 75 vandetanib (D–F). Shown are representative experiments 50 performed with six replicates per (% control) concentration and data are 25 H]Uridine uptake 3 expressed as mean SE values. [ 0 Uptake values represent % uridine -7 -6 -5 -4 -3 uptake in the presence of TKIs relative to that in its absence Log [Gefitinib] (mol/L) (control). Error bars are not shown where SE values are smaller than the size of the symbol. Each CF experiment was repeated three times. Vandetanib 100 75

(% control) 25 H]Uridine uptake 3 [ 0 -7 -6 -5 -4 -3 Log [Vandetanib] (mol/L)

www.aacrjournals.org Clin Cancer Res; 20(1) January 1, 2014 179

Damaraju et al.

Table 1. Summary of IC50 values for inhibition of uridine transport in yeast and cell lines

Geﬁtinib

Transporter (yeast) Erlotinib IC50 (mmol/L SE) Vandetanib hENT1 34 614 211 1 hENT2 >300 >300 89 17 hCNT1 160 20 37 11 64 17 hCNT2 >300 >300 82 4 hCNT3 11 1 >300 28 9 Cell lines (hENT1) A549 3.0 0.3 5.0 0.4 16 4 H292 6.0 3.0 6.0 0.4 37 7 H1975 6.0 0.2 2.0 0.6 33 8 ASPC-1 1.6 0.4 2.0 0.4 18 3

NOTE: Inhibition of [3H]uridine uptake by TKIs was assessed in yeast producing each of the ﬁve recombinant hNTs in concentration– effect experiments as described in Materials and Methods. Inhibition experiments were also conducted in cell lines all of which have

major intrinsic hENT1 activity and negligible or no CNT or hENT2 activities. IC50 values (mean SE) are listed.

evaluate their effects on hENT1-mediated uptake of vandetanib for 0 or 24 hours with or without recovery in [3H]uridine and results from only A549 cells are shown drug-free media for 24 hours and changes in uptake of here in detail (Fig. 2D–F); results from the four cell lines [3H]uridine were measured to monitor changes in hENT- are summarized in Table 1. mediated activity (Fig. 3C and D). Results indicated a The nature of the interactions of TKIs with hENT1 was decrease in hENT1-mediated activity after exposure to erlo- examined in A549 cells by studying effects of fixed con- tinib, gefitinib, or vandetanib in all cell lines tested. This centrations of erlotinib, gefitinib, or vandetanib on kinet- decrease in NT activity was reversed after culturing cells in ics of [3H]uridine uptake. Analysis of results using Line- drug-free media for 24 hours. To determine whether Weaver Burk plot (Fig. 2G–I) showed the competitive reduced NT activity was due to altered protein levels, we nature of uridine uptake inhibition by TKIs, suggesting used SAHENTA-FITC (a nonpermeable fluorescent hENT1 that TKIs and uridine bind to the same or overlapping probe) to measure the abundance of hENT1 sites on cell sites on hENT1. surfaces as described earlier (20). A549 cells were treated with erlotinib, gefitinib, or vandetanib for 24 hours and, Erlotinib, gefitinib, and vandetanib inhibit uridine after recovery in drug-free media for another 24 hours, were uptake and gemcitabine accumulation in A549, H292, stained with 100 nmol/L SAHENTA-FITC and processed as H1975, AsPC-1 cells and reduce cell surface hENT1 described in the Materials and Methods section. Fig. 3E staining in A549 cells (panels 1, 2) shows unstained and cell-surface stained Because of the failure of combination therapies with TKIs hENT1 in untreated cells and decreased staining after treat- and nucleoside analog drugs such as gemcitabine, effects of ment with TKIs (panels 3, 5, 7) that recovered after culturing erlotinib, gefitinib, and vandetanib on uridine uptake and cells in drug-free media (panels 4, 6, 8). gemcitabine accumulation were examined in all cell lines. Short-term uptake of 1 mmol/L [3H]uridine (1 minute) and Cytotoxicity of gemcitabine, erlotinib, gefitinib, or long-term uptake of 10 mmol/L [3H]gemcitabine (1 hour), vandetanib to A549, H292, H1975, and AsPC-1 cells which assessed intracellular accumulation of gemcitabine Cytotoxicity of gemcitabine, erlotinib, gefitinib, or van- and its metabolites, were measured after incubation in the detanib was assessed in all four cell lines. Erlotinib was not absence or presence of 25 mmol/L erlotinib, gefitinib, or toxic to A549 cells and therefore was not used in subsequent vandetanib or 100 mmol/L dilazep, a potent inhibitor of experiments. Since gemcitabine, gefitinib, and vandetanib both hENT1 and hENT2-mediated activities (23). In were toxic to A549 cells, we assessed whether mediated [3H[uridine uptake inhibition studies, the antibodies cetux- transport by hENTs as observed previously with gemcita- imab and bevacizumab were tested at three different bine (24) contributed to toxicity with TKIs. Cytotoxicity amounts (10, 50, and 100 mg/mL). The antibodies did studies were conducted with gemcitabine, gefitinib, or not inhibit [3H[uridine uptake (data not shown). In con- vandetanib in absence and presence of dilazep, an estab- trast, the three TKIs inhibited uridine uptake (Fig. 3A) and lished inhibitor of hENT1/2. A549 cells were exposed for gemcitabine accumulation (Fig. 3B) in all cells tested. 72 hours to graded gemcitabine concentrations (0–100 To further study effects of erlotinib, gefitinib, or vande- mmol/L) in the absence or presence of 25 mmol/L dilazep. tanib on hENT1 activity, A549, AsPC-1, H292, and H1975 Protection against gemcitabine toxicity was observed (Fig. cells were exposed to 0 or 5 mmol/L erlotinib, gefitinib, or 4A) in the presence of dilazep confirming previous

180 Clin Cancer Res; 20(1) January 1, 2014 Clinical Cancer Research

EGFR and VEGFR TKIs Inhibit Human Nucleoside Transporters

A D G 0.5 100 100 0.4 cells/s) 75 75 6 0.3 50 50 0.2

(% control) 25 (% control) 25 0.1 H]Uridine uptake H]Uridine uptake 3 3 1/(pmol/10 [ 0 [ 0 0.0 -8 -7 -6 -5 -4 0.00 0.25 0.50 0.75 1.00 KCl NaCl Log [Erlotinib] (mol/L) 1/[Uridine] (mmol/L)

KCl + NBMPRKCl + DilazepKCl + Uridine B E H 1.00 100 100 0.75 75 75 cells/s) 6 0.50 50 50

(% control) 25 (% control) 25 0.25 H]Uridine uptake H]Uridine uptake 3 3 [ [

0 0 1/(pmol/10 0.00 -8 -7 -6 -5 -4 0.00 0.25 0.50 0.75 1.00 KCl NaCl m Log [Gefitinib] (mol/L) 1/[Uridine] ( mol/L)

KCl + NBMPRKCl + DilazepKCl + Uridine C F I 0.12 100 100 75 75

cells/s) 0.08 6 50 50 0.04 25 25 (% control) (% control) H]Uridine uptake H]Uridine uptake

3 0 3 [ [ 0 0.00 -8 -7 -6 -5 -4 1/(pmol/10 0.00 0.06 0.12 0.18 KCl NaCl Log [Vandetanib] (mol/L) 1/[Uridine] (mmol/L)

KCl + NBMPRKCl + DilazepKCl + Uridine

Figure 2. Effects of erlotinib, gefitinib, and vandetanib on [3H]uridine uptake in cultured cells. A–C, uptake of 10 mmol/L [3H]uridine in A549, H292, and H1975 cells in sodium-containing or sodium-free media without or with inhibitors of hENT activity. Effects of increasing concentrations of erlotinib, gefitinib, or vandetanib on uptake of 10 mmol/L [3H]uridine in A549 cells are shown in (D–F). Competition effects of 0 (&), 5 (~), 25 (!), or 100 (}) mmol/L of erlotinib, gefitinib, or vandetanib on uridine uptake rates are shown in (G–I). Values with mean SE are shown in each panel. Each experiment was repeated three times. observations in other human cancer cell lines (24) that NT- vandetanib in all cell lines tested. This situation, when mediated permeation is important for gemcitabine toxicity. translated clinically, would mean reduced effectiveness of Similar experiments conducted with gefitinib and vandeta- gemcitabine of simultaneous administration regimens or nib (Fig. 4B) in A549 cells showed no effects of dilazep on when TKIs are administered before gemcitabine. In contrast, gefitinib or vandetanib cytotoxicity. Erlotinib, gefitinib, and we should see equivalent or higher effectiveness when vandetanib toxicities (0–100 mmol/L) were assessed over 72 gemcitabine is administered before TKIs. We explored cyto- hours in all four cell lines and IC50 values (mean SE) are toxicity of simultaneous and sequential administration of summarized in Table 2. gemcitabine with various TKIs in the four cell lines as follows. Individual toxicities of gemcitabine, erlotinib, gefi- In vitro combination studies with gemcitabine and tinib, or vandetanib to A549, H292, H1975, and AsPC-1 erlotinib, gefitinib, or vandetanib cells are summarized in Table 2. For in vitro combination Results described in Fig. 3B showed reduced accumula- studies cells were either (i) pretreated with TKIs before tion of gemcitabine in presence of erlotinib, gefitinib, or gemcitabine, (ii) pretreated with gemcitabine before TKIs,

www.aacrjournals.org Clin Cancer Res; 20(1) January 1, 2014 181

Damaraju et al.

A E

100 A549 12 H292 H1975 AsPC-1 75

50 (% control)

H]Uridine uptake 25 3 [ 0

Control Dilazep Gefitinib Erolotinib Vandetanib

B 20 mm

A549 100 H292 34 H1975

50 (% control) 25 H]Gemcitabine uptake 3

[ 0

Control Dilazep Gefitinib Erolotinib Vandetanib

C 125 A549 AsPC-1 56 100

(% control) 25 H]Uridine uptake 3 [ 0

Control Erlotinib Gefitinib Vandetanib

24 h Recovered24 h Recovered24 h Recovered D 125 H292 78 H1975 100

50 (% control) 25 H]Uridine uptake 3 [ 0

Control Erlotinib Gefitinib Vandetanib

24 h Recovered24 h Recovered24 h Recovered

Figure 3. TKI effects on [3H]uridine uptake and [3H]gemcitabine accumulation in cultured cells. Incubations with 10 mmol/L [3H]uridine for 1 minute (A) or 1 mmol/L [3H]gemcitabine for 60 minutes (B) were conducted in the absence or presence of 25 mmol/L of erlotinib, geﬁtinib, or vandetanib or 100 mmol/L dilazep and cell-associated radioactivity was measured. Values plotted are % control values obtained in the absence of additives and average values from two or more experiments are shown in each panel. (Legend continued on following page.)

182 Clin Cancer Res; 20(1) January 1, 2014 Clinical Cancer Research

EGFR and VEGFR TKIs Inhibit Human Nucleoside Transporters

Figure 4. Effects of dilazep on gemcitabine and TKI cytotoxicity. A B A, A549 cells were exposed to graded concentrations (0–100 100 100 m mol/L) of gemcitabine in the 75 75 absence (*) or presence of 25 mmol/L dilazep (*) for 72 hours. 50 50 Cell viability was measured by CCK-8 assay kit as described in Viability 25 Viability 25 (% control) (% control) Materials and Methods. B, A549 cells were treated with geﬁtinib or 0 0 –10 –8 –6 –4 –5.5 –5.0 –4.5 –4.0 vandetanib in the absence (*~)or presence of 25 mmol/L dilazep Log [Gemcitabine] (mol/L) Log [TKI] (mol/L) (*~) for 72 hours. Plotted are % controls values relative to CD untreated cells and representative 0.2 experiments with mean SD are 0.3 shown in each panel. Each experiment was repeated three A549 H292 0.2 times. C–F, show isobologram 0.1 analysis of the sequential combination of gemcitabine 0.1 followed by TKIs in A549, H292,

H1975, and AsPC-1 cells. Gemcitabine ( m mol/L) 0.0 Gemcitabine ( m mol/L) 0.0 Individual doses of gemcitabine 0 50 100 150 200 0 400 800 1,200 and TKIs to achieve 90% (straight Vandetanib (mmol/L) Gefitinib (mmol/L) line) growth inhibition (Fa ¼ 0.90), 75% (hyphenated line) growth EF inhibition (Fa ¼ 0.75), and 50% (on 0.9 the axes, open circle) growth ¼ inhibition (Fa 0.50) were plotted 8 on x- and y-axes. CI values 0.6 calculated using Calcusyn H1975 AsPC-1 software are represented by points 4 above (indicate antagonism 0.3 between drugs) or below the lines (indicate synergy). Symbols Gemcitabine ( m mol/L) 0.0 Gemcitabine ( m mol/L) * * 0 indicate: ( ) for ED50, ( )for 0 100 200 300 400 0 3,000 6,000 ED75, and (&) for ED90, m m respectively. Erlotinib ( mol/L) Vandetanib ( mol/L)

or (iii) treated with both agents together as described in combinations involving TKIs and nucleoside chemothera- Materials and Methods. Isobologram and CI methods py drugs. developed by Chou and Talalay (22) were used to conﬁrm and quantify the synergism observed with the various Discussion combinations of gemcitabine and TKIs. Isobolograms were TKIs were designed to bind to the ATP regulatory pocket of constructed for Fa values of 0.50, 0.75, and 0.90, represent- growth factor receptors and compete with ATP at the binding ing 50%, 75%, and 90% growth inhibition, respectively, site. Since adenosine is the nucleoside moiety in ATP, we and are shown in Fig. 4C–F. CI values at Fa values of 0.5 for expected that TKIs may also bind to hNTs that recognize and all cell lines are summarized in Table 2. These results transport nucleosides and therefore interfere with nucleoside indicated that the sequence, gemcitabine followed by TKIs, chemotherapy. Interaction of TKIs with ATP-binding was synergistic, whereas the sequence TKI followed by domains of ATP-binding cassette (ABC) transporter-mediat- gemcitabine or simultaneous treatment ranged from addi- ed multidrug resistance (MDR) proteins in cancer cells (25– tive to antagonistic thus supporting our hypothesis that 27) and inhibition of P-gp activity (28) was shown earlier. inhibition of hNTs by TKIs lead to decreased effectiveness of Many TKIs modulate both P-gp and ABCG2 activities and

(Continued) C and D, effects of 24-hour treatment with TKIs on hENT1 activity in A549, AsPC-1 (C), and in H292, H1975 (D) before and after 24-hour recovery in TKI-free media; cells were pretreated with 5 mmol/L erlotinib, geﬁtinib, or vandetanib for 24 hours and uptake of 10 mmol/L [3H]uridine (2 minutes) was measured in untreated (control) and treated cells immediately or 24 hours after incubation in drug-free media (24-hour recovery). % Control values with mean SE are shown. Each experiment was repeated three times. E, confocal microscopic analysis of staining of hENT1 sites in A549 cells with SAHENTA-FITC in untreated-unstained (# 1), untreated-stained with SAHENTA-FITC (# 2), or treated-stained with SAHENTA-FITC for 24 hours with erlotinib, geﬁtinib, or vandetanib (# 3, 5, 7) followed by 24-hour recovery after TKI treatment (# 4, 6, 8) as described in Materials and Methods. Bright staining on membranes indicates binding of SAHENTA-FITC to hENT1 sites. Scale bar shown is 20 mm.

www.aacrjournals.org Clin Cancer Res; 20(1) January 1, 2014 183

Damaraju et al.

Table 2. Cytotoxicity of gemcitabine and TKIs to pancreatic and lung cancer cell lines

H292 H1975

Drug A549 (IC50 values SE) AsPC-1 Gemcitabine (nmol/L) 2.2 0.2 3.0 0.5 1.0 0.3 3.5 0.8 Erlotinib (mmol/L) >100 60.0 8.0 7.0 0.8 7.0 1.5 Geﬁtinib (mmol/L) 3.5 0.9 3.4 1.4 6.0 0.5 11.0 3.0 Vandetanib (mmol/L) 3.0 0.2 1.0 0.4 4.0 0.1 4.0 0.8

H292 H1975

Drug combination A549 CI values at Fa50 AsPC-1

Gemcitabine first 0.9 0.6 0.5 0.4 Vandetanib first 3 1.0 1.3 0.9 Simultaneously 0.3 1.1 0.8 0.6 Gemcitabine first 0.4 Gefitinib first 0.9 Simultaneously 3.0 Gemcitabine first 0.4 Erlotinib first 1.2 Simultaneously 1.4

NOTE: All four cell lines were treated with gemcitabine alone or in combination with TKIs as described in the Materials and Methods

section. IC50 values (mean SE) for gemcitabine and TKIs were derived from 72-hour exposure experiments and are listed above. Cytotoxicity of combinations of gemcitabine with TKIs either sequentially or simultaneously was determined in each cell line and CI values at Fa values of 50% are listed below.

Shown are CI values at Fa50 (50% cytotoxic effect) where CI values signiﬁcantly less than 1 indicate synergy; values close to 1, an additive effect; and values signiﬁcantly greater than 1, an antagonistic effect of the 2 agents.

enhance cytotoxic effects of multiple anticancer drugs by nucleoside drugs without any adverse effects on pharma- increasing accumulation of P-gp and ABCG2 substrates cokinetics of nucleoside drugs. (29, 30). Another group of potential target proteins are NTs. Results from cell surface staining of hENT1 levels by Inhibition of hENT1-mediated activity in K562 cells by p38 SAHENTA-FITC showed reduction in cell surface staining mitogen-activated protein kinase inhibitors (8) and of intensity thus suggesting that these TKIs also affected cell murine equilibrative NT 1 (mENT1; ref. 9) by imatinib were surface hENT1 expression in addition to direct inhibition of shown earlier. A more recent study showed inhibition of transport activity thereby causing reduction in uptake of cytarabine uptake by imatinib and nilotinib (31). nucleosides in the presence of TKIs. This effect was reversed In this study, we showed that hENT or hCNT mediated by culturing cells in absence of erlotinib, gefitinib, or [3H]uridine uptake was inhibited to varying extents by vandetanib for 24 hours. erlotinib, gefitinib, and vandetanib at concentrations that Pharmacologic evidence from several clinical trials with are readily achievable in patient plasma, tissue, and tumors. TKIs and nucleoside chemotherapy (2–6) suggests that Uridine uptake in A549 cells was inhibited in a competitive there may be unfavorable interactions between TKIs and manner that suggests binding of TKIs at the nucleoside nucleoside chemotherapy drugs. Changes in pharmaco- (permeant) binding site of hENT1. Although erlotinib, kinetics of nucleoside metabolites of capecitabine in gefitinib, and vandetanib interacted with the nucleoside- presence of erlotinib were shown in a phase Ib dose binding site of hENT1, they do not appear to be transported escalation study of erlotinib and capecitabine in patients on the basis of the results of indirect cytotoxicity experi- with colorectal cancer. Van Custem and colleagues (32) ments wherein dilazep protected cells against gemcitabine reported that the mean peak concentration (Cmax)for toxicity but not against TKI toxicity. Direct evidence for TKIs capecitabine was 27% lower and the mean area under the lack of transportability by hNTs requires use of [3H]TKIs. concentration curve (AUC) was 12% lower in the pres- Erlotinib, gefitinib, and vandetanib inhibited [3H]uridine ence than in the absence of erlotinib. They also noted uptake and accumulation of [3H]gemcitabine thus suggest- reduced levels of 50-deoxy-5-fluorouridine when erlotinib ing interference with nucleoside analog chemotherapy and capecitabine were administered together. In a study when combined with these TKIs. In contrast, EGFR and of erlotinib with capecitabine in patients with breast VEGFR antibodies cetuximab and bevacizumab had no cancer (33), the Cmax (ng/mL) of capecitabine when effect on [3H]uridine uptake in A549 cells (data not shown) administered with docetaxel was 6,274 ng/mL but when thus suggesting that antibodies can be combined with combined with erlotinib was only 3,934 ng/mL, thus

184 Clin Cancer Res; 20(1) January 1, 2014 Clinical Cancer Research

EGFR and VEGFR TKIs Inhibit Human Nucleoside Transporters

suggesting an effect of erlotinib on pharmacokinetics of In summary, we have shown that two EGFR TKIs and one capecitabine in plasma. Such interaction may also result VEGFR/EGFR TKI inhibit hENT1. Inhibition of hENT2, in lower clearance of nucleoside drugs—for example, hCNT1/2/3 are varied among the three TKIs, but all three Goss and colleagues showed that gemcitabine clearance TKIs inhibit at least one hCNT. Furthermore, these TKIs was significantly reduced in the presence of cediranib (P > caused a decrease in cell surface abundance of hENT1, a 0.02; ref. 34). The above studies clearly show adverse ubiquitous hNT necessary for activity of many nucleoside pharmacokinetic interactions between nucleoside drugs chemotherapy drugs. Clinical implications of this study are and TKIs that may reduce treatment efficacy when these that lower concentrations of nucleosides and their meta- agents are administered together. bolites in plasma and in tumor tissues would be expected In combination toxicity studies, cells showed greater sen- when TKIs and nucleoside drugs are administered together. sitivity to drug combinations when cells were exposed to This drug interaction between nucleoside drugs and TKIs gemcitabine for 24 hours followed by erlotinib, gefitinib, or will lead to reduced treatment efficacy when these agents are vandetanib as predicted from our uptake inhibition studies. administered together. To obtain the best clinical responses, There were mostly additive to antagonistic interactions if the nucleoside drugs should not be administered concurrently TKI was added first followed by gemcitabine or simulta- with EGFR and VEGFR TKIs but rather in sequence, with the neously with the exception of vandetanib in A549 and nucleoside drug (e.g., gemcitabine) followed by TKIs. We AsPC-1 in simultaneous addition. Synergism observed in suspect that many ATP-competitive TKIs have the potential the sequential schedule of nucleoside drug followed by TKIs to inhibit hNTs and that different classes of TKIs inhibit is supported by results from a recent phase III trial (FASTACT- different hNTs with different potencies. 2; ref. 35), wherein untreated stage IIIB/IV NSCLC patients were randomized and treated with gemcitabine plus plati- Disclosure of Potential Conflicts of Interest num followed by intercalated erlotinib or placebo every 4 V.L. Damaraju has ownership interest (including patents) in patent. C.E. Cass is employed (other than primary affiliation; e.g., consulting) as a sole weeks. Treatment with intercalated combination of erlotinib proprietor in CE Cass Consulting and has commercial research grant from and chemotherapy improved PFS versus chemotherapy Clavis Pharma ASA. M.B. Sawyer has ownership interest (including patents) alone as first-line treatment in patients with advanced in a patent. No potential conflicts of interest were disclosed by the other authors. NSCLC with known and unknown EGFR mutation status. in vitro Although it is difficult to extrapolate studies to Authors' Contributions the clinic especially with drugs that have such extensive Conception and design: V.L. Damaraju, C.E. Cass, M.B. Sawyer protein binding and accumulation in tumors, concentra- Development of methodology: V.L. Damaraju Acquisition of data (provided animals, acquired and managed patients, tions of erlotinib, gefitinib, or vandetanib achievable in provided facilities, etc.): T. Scriver, D. Mowles, M. Kuzma plasma are levels that would inhibit hNT activity and Analysis and interpretation of data (e.g., statistical analysis, biosta- tumor levels are at least 40-fold higher than plasma levels. tistics, computational analysis): V.L. Damaraju, T. Scriver, M.B. Sawyer Writing, review, and/or revision of the manuscript: V.L. Damaraju, A. In a recent study of neoadjuvant breast cancer patients, Ryan, C.E. Cass, M.B. Sawyer mean gefitinib plasma concentrations at steady state were Administrative, technical, or material support (i.e., reporting or orga- 0.18 mg/mL (0.403 mmol/L) and mean tumor levels were nizing data, constructing databases): T. Scriver, D. Mowles, M. Kuzma Study supervision: V.L. Damaraju, M.B. Sawyer 7.5 mg/g (17 mmol/L), approximately 42-fold higher (36). In two related studies, median plasma gefitinib levels Acknowledgments were 1,064 ng/mL on day 8 (37) and tumor levels were The authors thank the Cross Cancer Institute Cell Imaging Facility for shown to be approximately 40-fold higher than plasma assistance with confocal microscopy. levels (22.7 vs. 0.52 mmol/L; ref. 38). In a phase I study at the 300 mg/day dose of vandetanib, steady-state mean Grant Support vandetanib concentrations were approximately 1,000 ng/ This research was funded by the Alberta Cancer Board Research Initiative m Program (to C.E. Cass and M.B. Sawyer), the Canadian Cancer Society mL (2.1 mol/L; ref. 13). Vandetanib administered at Research Institute (to C.E. Cass) and AstraZeneca (to M.B. Sawyer). M.B. doses up to 300 mg/day resulted in mean steady-state Sawyer was a recipient of an American Society of Clinical Oncology Career plasma levels of 2.2 mmol/L (range of 1.6–6.3 mmol/L; Development Award during a portion of this work. The costs of publication of this article were defrayed in part by the ref. 39) and was shown to accumulate in tumor tissues to payment of page charges. This article must therefore be hereby marked much higher levels (16). At these concentrations, we can advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate expect adverse drug interactions between TKIs and nucle- this fact. oside chemotherapy drugs that are consistent with inhi- Received August 19, 2013; revised October 22, 2013; accepted October 23, bition of NTs by TKIs. 2013; published OnlineFirst October 29, 2013.

References 1. Cohen MH, Williams GA, Sridhara R, Chen G, McGuinn WD Jr, Morse 2. Giaccone G, Herbst RS, Manegold C, Scagliotti G, Rosell R, Miller V, D, et al. United States Food and Drug Administration Drug Approval et al. Geﬁtinib in combination with gemcitabine and cisplatin in summary: Geﬁtinib (ZD1839; Iressa) tablets. Clin Cancer Res 2004; advanced non-small-cell lung cancer: a phase III trial–INTACT 1. 10:1212–8. J Clin Oncol 2004;22:777–84.

www.aacrjournals.org Clin Cancer Res; 20(1) January 1, 2014 185

Damaraju et al.

3. Philips GK, Halabi S, Sanford BL, Bajorin D, Small EJ. A phase II trial of 21. Chou TC. Drug combination studies and their synergy quantification cisplatin, fixed dose-rate gemcitabine and gefitinib for advanced using the Chou-Talalay method. Cancer Res 2010;70:440–6. urothelial tract carcinoma: results of the Cancer and Leukaemia Group 22. Chou TC, Talalay P. Quantitative analysis of dose-effect relationships: B 90102. BJU Int 2008;101:20–5. the combined effects of multiple drugs or enzyme inhibitors. Adv 4. Fountzilas G, Bobos M, Kalogera-Fountzila A, Xiros N, Murray S, Enzyme Regul 1984;22:27–55. Linardou H, et al. Gemcitabine combined with gefitinib in patients 23. Visser F, Vickers MF, Ng AM, Baldwin SA, Young JD, Cass CE. with inoperable or metastatic pancreatic cancer: a phase II Study of the Mutation of residue 33 of human equilibrative nucleoside transporters Hellenic Cooperative Oncology Group with biomarker evaluation. 1 and 2 alters sensitivity to inhibition of transport by dilazep and Cancer Invest 2008;26:784–93. dipyridamole. J Biol Chem 2002;277:395–401. 5. Philips GK, Halabi S, Sanford BL, Bajorin D, Small EJ. A phase II trial of 24. Mackey JR, Mani RS, Selner M, Mowles D, Young JD, Belt JA, et al. cisplatin (C), gemcitabine (G) and gefitinib for advanced urothelial tract Functional nucleoside transporters are required for gemcitabine influx carcinoma: results of Cancer and Leukemia Group B (CALGB) 90102. and manifestation of toxicity in cancer cell lines. Cancer Res 1998; Ann Oncol 2009;20:1074–9. 58:4349–57. 6. Blackhall FH, O'Brien M, Schmid P, Nicolson M, Taylor P, Milenkova T, 25. Burger H, van Tol H, Boersma AW, Brok M, Wiemer EA, Stoter G, et al. et al. A phase I study of Vandetanib in combination with vinorelbine/ Imatinib mesylate (STI571) is a substrate for the breast cancer resis- cisplatin or gemcitabine/cisplatin as first-line treatment for advanced tance protein (BCRP)/ABCG2 drug pump. Blood 2004;104:2940–2. non-small cell lung cancer. J Thorac Oncol 2010;5:1285–8. 26. Elkind NB, Szentpetery Z, Apati A, Ozvegy-Laczka C, Varady G, Ujhelly 7. Huang M, Wang Y, Cogut SB, Mitchell BS, Graves LM. Inhibition of O, et al. Multidrug transporter ABCG2 prevents tumor cell death nucleoside transport by protein kinase inhibitors. J Pharmacol Exp induced by the epidermal growth factor receptor inhibitor Iressa Ther 2003;304:753–60. (ZD1839, Gefitinib). Cancer Res 2005;65:1770–7. 8. Huang M, Wang Y, Collins M, Gu JJ, Mitchell BS, Graves LM. Inhibition 27. Erlichman C, Boerner SA, Hallgren CG, Spieker R, Wang XY, James of nucleoside transport by p38 MAPK inhibitors. J Biol Chem 2002; CD, et al. The HER tyrosine kinase inhibitor CI1033 enhances cyto- 277:28364–7. toxicity of 7-ethyl-10-hydroxycamptothecin and topotecan by inhibit- 9. Leisewitz AV, Zimmerman EI, Jones SZ, Yang J, Graves LM. Imatinib- ing breast cancer resistance protein-mediated drug efflux. Cancer Res resistant CML cells have low ENT activity but maintain sensitivity 2001;61:739–48. to gemcitabine. Nucleosides Nucleotides Nucleic Acids 2008;27: 28. Mi Y, Lou L. ZD6474 reverses multidrug resistance by directly inhibiting 779–86. the function of P-glycoprotein. Br J Cancer 2007;97:934–40. 10. Woodahl EL, Wang J, Heimfeld S, Ren AG, McCune JS. Imatinib 29. Dai CL, Tiwari AK, Wu CP, Su XD, Wang SR, Liu DG, et al. Lapatinib inhibition of fludarabine uptake in T-lymphocytes. Cancer Chemother (Tykerb, GW572016) reverses multidrug resistance in cancer cells by Pharmacol 2008;62:735–9. inhibiting the activity of ATP-binding cassette subfamily B member 1 11. Damaraju VL, Sawyer MB, Mackey JR, Young JD, Cass CE. Human and G member 2. Cancer Res 2008;68:7905–14. nucleoside transporters: biomarkers for response to nucleoside drugs. 30. Shi Z, Peng XX, Kim IW, Shukla S, Si QS, Robey RW, et al. Erlotinib Nucleosides Nucleotides Nucleic Acids 2009;28:450–63. (Tarceva, OSI-774) antagonizes ATP-binding cassette subfamily B 12. Zhang J, Visser F, King KM, Baldwin SA, Young JD, Cass CE. The role member 1 and ATP-binding cassette subfamily G member 2-mediated of nucleoside transporters in cancer chemotherapy with nucleoside drug resistance. Cancer Res 2007;67:11012–20. drugs. Cancer Metastasis Rev 2007;26:85–110. 31. Naud JS, Ghani K, de Campos-Lima PO, Caruso M. Nilotinib and 13. Holden SN, Eckhardt SG, Basser R, de Boer R, Rischin D, Green M, imatinib inhibit cytarabine cellular uptake: implications for combination et al. Clinical evaluation of ZD6474, an orally active inhibitor of VEGF therapy. Leuk Res 2012;36:1311–4. and EGF receptor signaling, in patients with solid, malignant tumors. 32. Van Cutsem E, Verslype C, Beale P, Clarke S, Bugat R, Rakhit A, et al. A Ann Oncol 2005;16:1391–7. phase Ib dose-escalation study of erlotinib, capecitabine and oxaliplatin 14. Masago K, Togashi Y, Fukudo M, Terada T, Irisa K, Sakamori Y, et al. in metastatic colorectal cancer patients. Ann Oncol 2008;19:332–9. Plasma and pleural fluid pharmacokinetics of erlotinib and its active 33. Twelves C, Trigo JM, Jones R, De Rosa F, Rakhit A, Fettner S, et al. metabolite OSI-420 in patients with non-small-cell lung cancer with Erlotinib in combination with capecitabine and docetaxel in patients pleural effusion. Clin Lung Cancer 2011;12:307–12. with metastatic breast cancer: a dose-escalation study. Eur J Cancer 15. McKillop D, Partridge EA, Kemp JV, Spence MP, Kendrew J, Barnett S, 2008;44:419–26. et al. Tumor penetration of gefitinib (Iressa), an epidermal growth factor 34. Goss G, Shepherd FA, Laurie S, Gauthier I, Leighl N, Chen E, et al. A receptor tyrosine kinase inhibitor. Mol Cancer Ther 2005;4:641–9. phase I and pharmacokinetic study of daily oral cediranib, an inhibitor 16. Gustafson DL, Bradshaw-Pierce EL, Merz AL, Zirrolli JA. Tissue dis- of vascular endothelial growth factor tyrosine kinases, in combination tribution and metabolism of the tyrosine kinase inhibitor ZD6474 with cisplatin and gemcitabine in patients with advanced non-small cell (Zactima) in tumor-bearing nude mice following oral dosing. J Phar- lung cancer: a study of the National Cancer Institute of Canada Clinical macol Exp Ther 2006;318:872–80. Trials Group. Eur J Cancer 2009;45:782–8. 17. Vickers MF, Kumar R, Visser F, Zhang J, Charania J, Raborn RT, et al. 35. Wu YL, Lee JS, Thongprasert S, Yu CJ, Zhang L, Ladrera G, et al. Comparison of the interaction of uridine, cytidine, and other pyrimidine Intercalated combination of chemotherapy and erlotinib for patients nucleoside analogues with recombinant human equilibrative nucleo- with advanced stage non-small-cell lung cancer (FASTACT-2): a side transporter 2 (hENT2) produced in Saccharomyces cerevisiae. randomised, double-blind trial. Lancet Oncol 2013;14:777–86. Biochem Cell Biol 2002;80:639–44. 36. McKillop D, Raab G, Eidtmann H, Furnival A, Riva A, Forbes J, et al. 18. Zhang J, Visser F, Vickers MF, Lang T, Robins MJ, Nielsen LP, et al. Intratumoral and plasma concentrations of gefitinib in breast cancer Uridine binding motifs of human concentrative nucleoside transporters patients: Preliminary results from a presurgical investigatory study 1 and 3 produced in Saccharomyces cerevisiae. Mol Pharmacol 2003; (BCIRG 103) 2004. J Clin Oncol 22:14s, 2004 (suppl; abstr 581). 64:1512–20. 37. Nakamura Y, Sano K, Soda H, Takatani H, Fukuda M, Nagashima S, 19. Vickers MF, Zhang J, Visser F, Tackaberry T, Robins MJ, Nielsen LP, et al. Pharmacokinetics of gefitinib predicts antitumor activity for et al. Uridine recognition motifs of human equilibrative nucleoside advanced non-small cell lung cancer. J Thorac Oncol 2010;5:1404–9. transporters 1 and 2 produced in Saccharomyces cerevisiae. Nucleo- 38. Haura EB, Sommers E, Song L, Chiappori A, Becker A. A pilot study of sides Nucleotides Nucleic Acids 2004;23:361–73. preoperative gefitinib for early-stage lung cancer to assess intratumor 20. Robins MJ, Peng Y, Damaraju VL, Mowles D, Barron G, Tackaberry T, drug concentration and pathways mediating primary resistance. et al. Improved syntheses of 50-S-(2-aminoethyl)-6-N-(4-nitrobenzyl)- J Thorac Oncol 2010;5:1806–14. 50-thioadenosine (SAENTA), analogues, and fluorescent probe con- 39. Brave SR, Odedra R, James NH, Smith NR, Marshall GB, Acheson KL, jugates: analysis of cell-surface human equilibrative nucleoside trans- et al. Vandetanib inhibits both VEGFR-2 and EGFR signalling at porter 1 (hENT1) levels for prediction of the antitumor efficacy of clinically relevant drug levels in preclinical models of human cancer. gemcitabine. J Med Chem 2010;53:6040–53. Int J Oncol 2011;39:271–8.

186 Clin Cancer Res; 20(1) January 1, 2014 Clinical Cancer Research

Erlotinib, Gefitinib, and Vandetanib Inhibit Human Nucleoside Transporters and Protect Cancer Cells from Gemcitabine Cytotoxicity

Vijaya L. Damaraju, Tara Scriver, Delores Mowles, et al.

Clin Cancer Res 2014;20:176-186. Published OnlineFirst October 29, 2013.

Updated version Access the most recent version of this article at: doi:10.1158/1078-0432.CCR-13-2293

Cited articles This article cites 39 articles, 15 of which you can access for free at: http://clincancerres.aacrjournals.org/content/20/1/176.full#ref-list-1

Citing articles This article has been cited by 6 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/20/1/176.full#related-urls

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

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/20/1/176. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.