Dipyridamole Potentiates Antipurine Antifolate Activity in the Presence of Hypoxanthine in Tumor Cells but not in Normal Tissues in Vitro1

The cytotoxicity of the antifolate inhibitors of de novo purine biosynthesis, lometrexol (LTX) and LY309887, can be abolished by hypoxanthine (HPX) salvage. The nucleoside transport inhibitor, dipyridamole (DP) can prevent HPX rescue from LTX growth inhibition in a cell linespecific manner. The studies described here have shown that, excluding colon and hematological malignancies, DP prevents HPX rescue from LTX growth inhibition in approximately one-third of cell lines with otherwise limited tissue specificity. The clinical dose-limiting toxicities of antipurine antifolates are to the bone marrow and gastrointestinal tract. In vitro models of these normal tissues were established, and the effect of DP on HPX rescue from LY309887 treatment was studied. Growth inhibition assays are not feasible in these primary cultures; therefore, an alternative assay, ccllular ATP depletion, was validated in four tumor cell lines as a marker of de novo and salvage purine synthesis. In LY309887-treated cells, DP prevented HPX-mediated maintenance of ATP levels only in cell lines in which DP inhibited HPX rescue from antifolate cytotoxicity. Hence, ATP depletion is a reliable indicator of sensitivity of HPX transport to DP when direct cell growth measurement is impractical. In primary cultures of human hematopoetic progenitor cells and mouse small intestine, coincubation with HPX prevented LY309887-mediated ATP depletion, which was Received 6/5/98; revised 8/31/98; accepted 9/1/98. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. I Supported by the North of England Cancer Research Campaign and Cancer Research Campaign, United Kingdom. 2 To whom requests for reprints should be addressed, at Cancer Research Unit, Medical School, University of Newcastle upon Tyne, Framlington Place, Newcastle upon Tyne, NE2 4HH, United Kingdom. Phone: 44 (0)191 222 8239: Fax: 44 (0)191 222 7556. not blocked by DP. These data suggest that DP would not prevent HPX rescue from antipurine antifolate growth inhibition in sensitive normal tissues, whereas activity against certain solid human tumors would be maintained. INTRODUCTION The antifolate LTX3 (6R-5, l0-dideaza-5,6,7,8-tetrahydrofolate) inhibits GARFT (EC 2. 1.2. 1), the first folate-dependent enzyme of de novo purine synthesis ( 1 ). LTX was found to have potent antitumor activity in a number of murine and human xenograft solid tumor models (2), and Phase I studies with a range of schedules were initiated. However, regardless of schedule, there were delayed and cumulative side effects, particularly, myehosuppression and gastrointestinal toxicities (3), although these may be ameliorated by folate supplementation (4). The second-generation GARET inhibitor LY309887 (6R-2’,5 ‘-thienyh-5,10-dideazatetrahydrofolic acid) is 9-fold more potent than LTX as an inhibitor of recombinant monofunctional human GARET (5). In growth inhibition assays, LY309887 was approximately 3-fold more potent than LTX against CCRF-CEM cells, exhibited equal or superior in s’is’o antitumor activity in four murine and five human xenograft models, and was clearly superior to LTX against two pancreatic xenograft tumors (6). Biochemical and pharmacological differences between these compounds, such as differential affinities for folate receptor isoforms, suggest that LY309887 may have a greater antitumor efficacy and less toxicity than LTX in cancer patients (5). Resistance to inhibitors of de novo nucleotide biosynthesis such as LTX and LY309887 may arise as a result of a number of mechanisms, including the salvage of extracellular nucleosides and nucheobases (7). In cancer cells, the activities of enzymes involved in purine and pyrimidine salvage are higher than those of the rate-limiting enzymes of de novo nucleotide synthesis (8). Moreover, the ability of cells to overcome the cytotoxicity of antimetabolites by salvage processes appears to increase in parallel with malignancy (9). Release of nucleic acids from dead cells and their subsequent degradation may exacerbate this mode of resistance by producing local high concentrations of salvageable nucleosides and bases in tumors. Thus, circumvention of the inhibition of de novo nucheotide synthesis by efficient salvage processes in tumor cells may reduce or abolish the antitumor effect of an antimetabohite. As expected from their mode of action, cells treated with GARET inhibitors have reduced ATP and GTP pools. Coincubation with 3 The abbreviations used are: LTX, lometrexol; GARET. glycinamide ribonucleotide transformylase; HPX, hypoxanthine: DP, dipyridamole: NBTI, nitrobenzylthioinosine; MTX, methotrexate: SRB, sulforhodaminc B. Research. on April 26, 2017. © 1998 American Association for Cancer clincancerres.aacrjournals.org Downloaded from 2896 Selective Potentiation of Antifohates by Dipyridamohe HPX repletes these poo1s ( I , 10) and abolishes the growth inhibitory and cytotoxic effects (10, 1 1). Inhibition of purine nucheoside and base salvage should, therefore, increase the antiproliferative effects of the antipurine antifolates. Carrier-mediated equilibrative transport is widely responsible for uptake of salvageable nucleosides (12). DP is an inhibitor of both NBTI-sensitive (es) and -insensitive (ei) equilibrative nucleoside transport. DP has been found to potentiate MIX cytotoxicity in lymphocytes (13) and in HCT116 colon carcinoma cell line (14) by prevention of thymidine salvage. In A549 (lung carcinoma) cells, the activity of the thymidylate synthase inhibitor CB37 17 was increased by inhibition of thymidine uptake and deoxyuridine efflux (15). Less is known about nucleobase transport and its inhibition, although modulation of antimetabohite cytotoxicity via nucleobase transport has been observed. For example, potentiation of MIX cytotoxicity by DP in an ovarian carcinoma cell line was demonstrated as a result of the inhibition of both HPX and thymidine uptake (16). More recently, Turner et a!. (17) investigated prevention of HPX rescue from LIX cytotoxicity by DP and demonstrated that cells differed in their response to the triple combination of LIX, HPX, and DP. In some cell lines, DP prevented HPX rescue from LIX growth inhibition, and these cells were described as having DP-sensitive (ds) HPX salvage. In contrast, DP was unable to prevent HPX salvage in other cells, in which HPX salvage was termed DP insensitive (di). In total, Turner et a!. (17) studied nine cell lines of human and rodent origin, and it appeared that sensitivity of HPX salvage to DP had some tissue specificity. For example, the colon carcinoma and leukemia cell lines investigated had di HPX salvage and remained viable in the presence of LIX, HPX, and DP, whereas certain lung, breast, and ovarian cell lines were sensitive to the triple combination. Because colorectal and leukemia cell lines represent the malignant counterparts of the normal tissues associated with antipurine antifolate dose-limiting toxicities, it was postulated that the combination of HPX and DP might protect the bone marrow and gut from LIX toxicity in VitO. In contrast, tumors of the lung, breast, and ovary would remain sensitive. Here, we report an investigation of the effect of DP on HPX rescue from LIX growth inhibition in an extended panel of human tumor cell lines representative of common human malignancies. In addition, the growth-inhibitory properties of LIX and the new clinical candidate, LY309887, were compared in A549, CORL-23, MCF7, and I-47D cell lines. Using cellular AlP depletion to measure inhibition of de novo purine synthesis, HPX salvage, and the inhibition of salvage, we studied the effect of the triple combination of antipurine antifolates using the above four cell lines and primary cultures of human hematopoietic progenitor cells and mouse intestinal epithehium. The results contained suggest that the combination of an antipurine antifohate, HPX, and DP may protect hematopoietic and gastrointestinal stem cells, while retaining activity against certain common human malignancies. MATERIALS AND METHODS Materials. LIX and LY309887 disodium salts were gifts from Eli Lilly and Co. (Indianapolis, IN) and were dissolved in distilled water. DP and HPX (Sigma Chemical Co., Poole, United Kingdom) were dissolved in 100% DMSO and 0. 1 si NaOH, respectively. Drug stocks were stored at 4#{176}C in the dark and were not used if they were over 4 weeks old. All other chemicals were purchased from Sigma unless otherwise mdicated. Human Cell Lines. The following cell lines were used: MCF7 and I-47D, breast adenocarcinoma; 0AW42, ovarian epithelial tumor; SK-OV-3, ovarian adenocarcinoma; HX147 and CORL-23, non-small cell lung carcinoma; RT1 12, epithelial bladder carcinoma; RT4, transitional-cell bladder papilloma; K-562, chronic myelogenous leukemia; MOLI-4, acute hymphoblastic leukemia; and SW620, colon adenocarcinoma. Cells were adapted to growth in RPM! 1640 (Sigma) that was supplemented with 1000 units/mI penicillin, 100 p.g/ml streptomycm (Life Technologies, Inc., Paisley, United Kingdom), and 10% FCS (Sigma) that had been dialyzed for 24 h in PBS containing 1 glliter activated charcoal to remove any endogenous HPX. All other medium constituents were obtained from Life Technologies, Inc. Cell lines were obtained from the European Collection of Animal Cell Cultures or American Type Culture Collection except the CORL-23 cell line, which was a gift from Dr. P. Twentyman (Medical Research Council Clinical Oncology and Radiotherapeutics Unit, Cambridge, United Kingdom). Cells were found to be free of Mycoplasma infection when they were tested at monthly intervals using a Hoechst 33258 DNA fluorescence technique (18). Growth Inhibition Assays. Cells were maintained as exponentially growing cultures. MOLI-4 and K-562 cells were seeded at between 2 X l0 and 4 X l0 cells/ml in 24-well plates (Nunc-Life A/S. Roskilde, Denmark) in medium containing LIX, HPX, and DP as indicated in “Results.” After determining the day 0 cell count (Coulter Counter model Zl; Coulter Electronics), plates were incubated at 37#{176}C in a 5% CO2 enriched humidified atmosphere for 72 or 96 h to allow a minimum of three cell population doubhings in the control wells. Cells were then fixed with 0.5 ml Carnoy’s fixative per well (3:1 methanol:acetic acid) and counted. Adherent cells in log-phase growth were trypsinized; syringed gently with a 25-gauge needle; seeded in 100 p.1 of medium into 96-well plates (Nunc) at 2 X l0 -5 X l0 cehls/ well, depending on the growth characteristics of the cell line; and left to attach overnight. The medium was removed with a drawn-out glass Pasteur pipette, to reduce cell disruption, and replaced with 100 p.1 of medium containing LIX, HPX, and/or DP. Replicate wells were fixed with Camoy’s fixative to estimate the cell density at the start of the drug incubation. Plates were incubated at 37#{176}Cin a 5% CO, enriched humidified environment for between 72 and 144 h, depending on the doubling time of the cell line, to ensure at least three cell doublings in control wells. After incubation, plates were fixed with Carnoy’s, washed, air-dried, and stained with SRB, as described previously (19). The absorbance relative to an air blank was measured on a computer-interfaced Dynatech MR7000 96-well microtiter plate reader using a 570-nm filter. A comparison of the growth inhibition properties of LIX and LY309887 was conducted in exponentially growing I-47D, MCF7, CORL-23, and A549 cells, as described above. Research. on April 26, 2017. © 1998 American Association for Cancer clincancerres.aacrjournals.org Downloaded from Clinical Cancer Research 2897 Measurement of ATP in Human Tumor Cells. Exponentially growing T-47D, MCF7, CORL-23, and A549 cells were seeded at -3 X l0 cells/mi in 24-well plates and left to attach overnight. The medium was replaced with medium contaming LY309887, HPX, and/or DP, and after 24 h, cellular AlP was determined using a bioluminescent somatic cell assay kit purchased from Sigma. Cells were washed with Du1A (ICN Biomedical Inc., Irvine, United Kingdom) and permeabilized using somatic cell releasing reagent (Sigma). Equal volumes of AlP assay mix containing luciferase, luciferin, and MgSO4 and permeabilized cell extract were added to a disposable huminometer cuvette (Promega, Madison, WI), and the luminescence was immediately read on a Turner TD-20e luminometer. The ATP concentration was determined by comparison with an internal standard. Protein estimations were performed using a bicinchoninic acid-based method (Ref. 20; Pierce, Rockford, IL); absorptions were read at 575 nm on a Titertek Multiskan MCC/ 340 plate reader, and estimates were quantified using a standard curve constructed with BSA. Isolation and Culture of Intestinal Epithelial Cells. Mouse intestinal cell aggregates were obtained by adapting a method described previously by Evans et a!. (21) in studies with rats. Briefly, two 6-8-day-old BALB/c mice were killed by cervical dislocation. Small intestinal segments were diced, washed thoroughly in HBSS (Life Technologies, Inc.), and then incubated in a combination of type I dispase (0. 1 mg!ml; Boehringer Mannheim, Lewes, United Kingdom) and type XI crude collagenase (300 units/mi: Sigma) in HBSS for 20 mm at room temperature. The digested preparation was then allowed to undergo differential sedimentation in DMEM (Life Technologies, Inc.) with 2% (w/v) sorbitol (Sigma) and 5% FCS to yield cell aggregates of proliferative crypt epithelium in morphological contact with adjacent mesenchymal elements. Approximately 30 aggregates per well were seeded in DMEM supplemented with 10% dialyzed FCS, 20 ng/ml epidermal growth factor, 20 units/ liter insulin, 50 p.g/mh heparin (to reduce nonepithelial cell growth), 10 p.g/ml transferrin (all purchased from Sigma), 1000 units/mi penicillin, and 100 p.g/mI streptomycin (Life Technologies, Inc.) in two 24-well plates (Costar, Cambridge, MA). The plates had been previously coated with collagen by the addition of 250 p.1 of Vitrogen (Imperial Laboratories, Andover, United Kingdom) to each well, overnight incubation at 37#{176}C, removal of liquid, and air drying. After an attachment period of 48 h at 37#{176}Cin a humidified atmosphere of 7.5% CO2 in air, the medium was replaced with medium containing LY309887, HPX, and/or DP. After a further 24-h incubation, ATP and protein concentrations were determined as described above. To ensure that the cells being studied were epithelial cells, an epithelial cell detection kit (Baxter Healthcare Ltd., Newbury, United Kingdom) was used. This method involves immunocytochemical detection of cells expressing cytokeratins, abundant proteins within the cytoskeleton of epithehial cells (22). Aggregates in a representative well were washed with DuIA then carefully removed using a cell scraper (Costar), and 10 aggregates were then deposited on to a glass slide (BDH, Poole, United Kingdom) using a Shandon Cytospin 2 centrifuge and left to air dry overnight. Following cell permeabihization and fixation, cytokeratins were detected using the Fab fragment of antibody A45-BIB3 conjugated to alkaline phosphatase. A staining solution containing naphthol phosphate, sodium nitrite, and new fuchsin facilitated the formation of an insoluble red reaction product at the site of binding of the specific conjugate. HI-29 colon adenocarcinoma cells and CCRF-CEM lymphoblastic leukemia cells were prepared and examined in an identical manner to act as a positive and negative controls. respec-