The Role of -Folate Receptor-Mediated Transport in the Antitumor Activity of Antifolate Drugs

Purpose: Raltitrexed, pemetrexed, lometrexol, and ZD9331 are antifolate drugs transported into cells via the ubiquitously expressed reduced-folate carrier. They display also high affinity for the -folate receptor ( -FR), a low capacity folate transporter that is highly overexpressed in some epithelial tumors. The role of -FR in the activity of the antifolates has been evaluated in two -FR-overexpressing cell lines grown in a physiological concentration of folate (20 nM R,S-Leucovorin). Experimental Design and Results: A431-FBP cells (transfected with the -FR) were 3–5-fold more sensitive to the antifolates than A431 cells. KB cells (constitutive -FR overexpression) were less sensitive to the drugs when coexposed to 1 M folic acid to competitively inhibit binding to the -FR. Raltitrexed, pemetrexed, and lometrexol are polyglutamated in cells leading to drug retention, e.g., the raltitrexed 4and 24-h IC50s in A431 cells were 0.6 and 0.008 M, respectively, compared with 0.003 M for 72-h continuous exposure. A431-FBP cells were 3-fold more sensitive to raltitrexed and pemetrexed at all exposure times. ZD9331 is not polyglutamated, and the 4and 24-h IC50s in A431 cells were >100 and 100 M, respectively, reducing to 2 and 0.1 M, respectively, in A431-FBP cells. The ZD9331 4and 24-h IC50s in KB cells were 20 and 1 M, respectively, and reversible by coaddition of 1 M folic acid. An in situ thymidylate synthase assay demonstrated continued thymidylate synthase inhibition after ZD9331-treated A431-FBP and KB, but not A431, cells were placed in drug-free medium for 16 h. A model is proposed in which the antifolates accumulate in the -FR/endosomal apparatus, leading to slow release into the cytoplasm. In particular, this leads to cellular retention of the nonpolyglutamatable ZD9331. Conclusions: Antifolate drugs, particularly ZD9331, have the potential for increased efficacy in tumors that highly overexpress the -FR. INTRODUCTION Antifolate drugs have been developed that act on a number of folate-dependent enzymes and include methotrexate (MTX) that primarily targets dihydrofolate reductase; CB3717, raltitrexed (Tomudex; ZD1694), and ZD9331 that target thymidylate synthase (TS); Alimta (pemetrexed; MTA; LY231514) that primarily inhibits TS but may have other relevant targets; and lometrexol (5,8-dideazafolate) that inhibits glycinamide ribonucleotidetide formyltranserase (Refs. 1–8; Fig. 1). With the exception of CB3717, these drugs are efficiently transported into cells via the reduced-folate carrier (RFC; Refs. 7 and 9–14). The RFC is a low affinity, high capacity system that bidirectionally transfers folates/antifolates across the plasma membrane via an energy-dependent, carrier-mediated process (15– 17). Although the RFC is ubiquitously expressed (18), another folate transporter, the -folate receptor ( -FR) displays a more restricted range of tissue expression but is nevertheless proposed to function as a high affinity, low capacity folate/antifolate transporter (11, 16, 19, 20). The low capacity is ascribed to the receptor-mediated endocytotic mechanism that requires recycling of the receptor back to the cell surface, a process that has been reported to take between 30 min and 5 h (21–24). The -FR is highly overexpressed in some solid epithelial tumors, such as ovarian carcinoma and mesothelioma (25–31). Furthermore, its relatively low expression in most normal tissues (exceptions are placenta, proximal tubule of the kidney, choroid plexus, and glandular epithelia, particularly breast; Refs. 25, 26, and 32) is leading to the development of diagnostic agents and therapy targeted at -FR-overexpressing tumors (33– 36). Some of these therapies are folic acid drug conjugates or novel antifolates that rely on -FR functional activity to deliver the agent into the cells. Several antifolate drugs have high affinity for the -FR, similar to that of folic acid or the biologically active reduced folate cofactors (8, 11, 14, 19). Functionality of the receptor in terms of transport of antifolates was demonstrated using cell lines deficient in RFC function but overexpressing the -FR (11, 37). Uncertainty has surrounded the question of the relevance of -FR-mediated transport of antifolates when the RFC is coexpressed. This is partly attributable to the fact that in vitro models have relied on culturing cells in medium containing a low concentration of folate ( 1 nM) either to maintain or increase -FR expression (11, 37–39). This may result in reduced competition between folates and antifolates for binding to the -FR or, if intracellular folates are concomitantly low, reduced competition for enzymes relevant to drug action, e.g., folylpolyglutamate synthetase and/or TS. Nevertheless, studies using these model systems demonstrated that raltitrexed, pemetrexed, ZD9331, lometrexol, and particularly CB3717 could be transported into mouse and human cells via the -FR. However, it was concluded that, possibly with the exception of CB3717, drug uptake was predominantly via the RFC. Received 9/5/03; revised 10/28/03; accepted 10/31/03. Grant support: Grant from Cancer Research United Kingdom. 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. Requests for reprints: Ann L. Jackman, Haddow Laboratories, Institute of Cancer Research, 15 Cotswold Road, Sutton, Surrey, SM2 5NG, United Kingdom. E-mail: [email protected]. 1080 Vol. 10, 1080–1089, February 1, 2004 Clinical Cancer Research Cancer Research. on October 31, 2017. © 2004 American Association for clincancerres.aacrjournals.org Downloaded from Recently, it has been speculated that -FR overexpression may play a role in the clinical activity of antifolate drugs. This is based partly on the in vitro evidence described above but partly on some circumstantial evidence arising from clinical trials with CB3717, raltitrexed, and pemetrexed in -FR-overexpressing cancers. In Phase I/II clinical studies of CB3717, an 18% response rate was observed in platinum-refractory ovarian cancer (40). Activity of CB3717, raltitrexed, and pemetrexed has been demonstrated in mesothelioma (40–42). A deeper understanding, therefore, of -FR-mediated uptake may lead to the improved use of antifolate drugs by, e.g., selectively targeting -FR-overexpressing tumors. The current study compares several antifolate drugs for their affinity for the human -FR relative to folic acid and for their activity in two human tumor cell lines (A431-FBP and KB) that highly overexpress the -FR. Importantly, the cells express functional RFC and can be grown in media containing a physiological folate concentration without down-regulating the -FR (36). The A431-FBP cell line has been transfected with the -FR so that direct comparisons can be made with its A431 neo-transfected isogenic partner (43). The KB cell line constitutively overexpresses the -FR (39, 44). We have shown that A431-FBP tumor cells display a 3–5-fold increased sensitivity, compared with A431 cells, to raltitrexed, ZD9331, pemetrexed, and lometrexol when exposure is continuous and cells are cultured in a physiological concentration of folate. Similar results were obtained with KB tumor cells in which sensitivity was compared with that in the presence of 1 M folic acid. This concentration of folic acid competitively inhibits binding to the -FR but not RFC, because the two transporters display high and very low affinity for folic acid, respectively (16, 17, 36). After short exposure to ZD9331, -FR-mediated uptake led to prolonged TS inhibition after extracellular drug removal and a marked increase in the growth inhibitory activity of ZD9331 that was not seen with the other antifolates. This has led to the hypothesis that entrapment of ZD9331 in the endosomal apparatus of -FR-overexpressing cells leads to the slow but continuous delivery of ZD9331 into the cytosol. The consequences of this effect are less apparent with polyglutamatable drugs that become trapped in cells via polyglutamation, irrespective of the transport route. Thus, it is hypothesized that in a clinical situation, these antifolates, particularly ZD9331, have the potential to localize more highly to tumors overexpressing the -FR. MATERIALS AND METHODS Compounds. Raltitexed, pemetrexed, and ZD9331 were synthesized at AstraZeneca plc (Alderley Park, Cheshire, United Kingdom). Lometrexol was a kind gift from Eli Lilly Pharmaceuticals (Indianapolis, IN). MTX and folic acid were purchased from Sigma (Poole, Dorset, United Kingdom). The chemical structures are given in Fig. 1. Stock solutions (10 mM; 1 ml) were prepared in 0.15 M NaHCO3 [one to three drops of 1 M NaOH were added to aid dissolution (pH 8.0)] and stored at 20 for not 4 months. Cell Culture. Human A431 epidermoid vulval (neotransfected) and A431-FBP cells (transfected with the human -isoform of the FR) were a generous gift from Dr. A. Tomassetti (Instituto Tumori, Milan, Italy). Details of these cell lines and the culture conditions have been published recently (36). Importantly, the medium (DMEM) was purchased without folic acid, because commercial media contain a supra-physiological concentration of folic acid (2–8 M). The medium was supplemented before use with 1 or 20 nM R,S-leucovorin (LV) and Fig. 1 Structures of antifolates. 1081 Clinical Cancer Research Cancer Research. on October 31, 2017. © 2004 American Association for clincancerres.aacrjournals.org Downloaded from referred to as low and physiological folate concentrations, respectively. Similarly, the culture conditions for human KB cells (folate-free RPMI medium supplemented with LV) are found in Theti et al. (36). Cell surface expression of -FR, as measured by the surface binding capacity of [H]folic acid, was 1 pmol/10 cells, 171 42 pmol/10 cells, and 91 17 pmol/10 cells for A431, A431-FBP, and KB cells, respectively, grown in 20 nM LV. In 1 nM LV, the values were 1 pmol/10 cells, 211 65 pmol/10 cells, and 123 43 pmol/10 cells (36). The population doubling times were 18, 21, and 15 h for A431, A431-FBP, and KB cells, respectively, irrespective of the folate concentration. Affinity of the Antifolates for the -FR. This method is essentially a whole cell [H]folic acid competitive binding assay reported by Westerhof et al. (45) adapted for use with adherent cell lines (36). Relative affinities are defined as the inverse molar ratio of compound required to inhibit [H]folic acid binding by 50%. The relative affinity of FA is set at 1 (100%). Compounds with a lower or higher affinity have values 1 and 1, respectively. Growth Inhibition Studies. Details of the growth inhibition assays are reported in Theti et al. (36). Briefly, A431 and A431-FBP cells were incubated for 72 and 96 h, respectively (approximately four control population doublings), with increasing concentrations of the antifolate drugs in 96-well plates before using an 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay as a measure of cell viability. KB cells were incubated for 72 h (approximately four control population doublings) with the drugs in 24-well plates before cell counting. Cell growth in the presence of the drugs is expressed as a percentage of control cell growth and the IC50 calculated (concentration of drug to inhibit cell growth by 50%). This was performed in media containing 1 or 20 nM LV as the folate source. In parallel, cells were coexposed to drug and 1 M folic acid (added 15 min before drug) to selectively inhibit -FRbut not RFC-mediated uptake. This, as might be expected, had no effect on the IC50s obtained for any of the antifolates in the -FR-negative A431 cells. Short exposure growth inhibition assays refer to experiments in which the exposure time to the antifolates ranges from 1 to 72 h, i.e., the cells were washed, and drug-free medium (DFM) was added for the remaining incubation period as described above, i.e., the end point is the same as for continuous exposure experiments. The IC50s obtained are plotted against exposure time. In Situ TS Assay. This assay measures the rate of [H] release (as [H]2O) from 5-[ H]dUrd over 1 h and is a semiquantitative measure of TS inhibition in cells. Details of this method, adapted to the cell lines described herein, have been published recently (36). TS inhibition was measured after 1-, 4-, 8-, and 16-h exposure to increasing concentrations of ZD9331, expressed as a percentage of control and the IC50 calculated. In parallel, cells were exposed to ZD9331 and 1 M folic acid to assess the contribution of -FR-mediated drug uptake to TS inhibition. In some experiments, the exposure to ZD9331 was fixed at 4 h, followed by removal of drug-containing medium and the addition of DFM for 4 or 16 h before the in situ TS assay was performed. RESULTS Affinity of Antifolates for the -FR Expressed by L1210-FBP and A431-FBP Cells Relative to Folic Acid. A description of antifolate relative binding affinities for the -FR expressed by mouse L1210-FBP cells has been published (11, 14), so this cell line was included as a control, and results were compared with A431-FBP -FR. Raltitrexed, ZD9331, and lometrexol displayed affinities for mouse -FR that were 61, 54, and 78% that of folic acid, respectively, whereas MTX displayed a considerably lower affinity ( 1%; Table 1). Only pemetrexed had a relative affinity higher than that of folic acid (120%). This compares with 150% for CB3717 (data not shown). Some differences in relative affinities were observed when human A431-FBP -FR-overexpressing cells were used, but generally, the same pattern was observed (Table 1). The published results of raltitrexed and lometrexol for -FR expressed by human KB cells are also similar, i.e., 31 and 70%, respectively, (11). One study reported that raltitrexed binds to KB -FR with an 100-fold lower affinity using a [I]-folic acid competitive binding assay (46). Inhibition of the Growth of the Human A431 and A431FBP Isogenic Pair of Cell Lines. The sensitivity of the A431 cells to the antifolates in a physiological folate concentration (20 nM LV) ranged from 3 nM IC50 (raltitrexed) to 90 nM IC50 (ZD9331; Table 2) and is similar to that reported for other non-FR-overexpressing tumor cell lines grown in standard commercial media (47). The results were similar in low folate conditions (1 nM LV), except that MTX was 8-fold more active (Table 2). This may be the result of a lower intracellular folate pool decreasing competition for binding to dihydrofolate reductase as reported for other cell lines (48, 49). In 20 nM LV, A431-FBP cells displayed a 3–5-fold higher sensitivity compared with A431 cells to all of the drugs, except for MTX, consistent with -FR-mediated transport contributing to drug uptake (Table 2). In 1 nM LV, A431-FBP cells were 14-fold more sensitive than A431 cells to raltitrexed and pemetrexed, suggesting that -FR-mediated uptake is more efficient in low folate compared with a physiological folate concentration. The similar sensitivity of both cell lines to MTX is consistent with its low affinity for the -FR relative to that of the Table 1 Relative affinities of the antifolates for the -FR expressed by L1210-FBP and A431-FBP cells The -FR affinity is expressed as the inverse molar ratio of compound required to displace [H]-folic acid from the -FR by 50%. Folic acid is designated as 1 (100%). Results are given as the mean SD of at least three experiments or as individual results. Values for CB3717 are 1.5 and 1.2 for L1210-FBP and A431-FBP, respectively (Ref. 36). Affinity relative to folic acid