Role of proton-coupled folate transporter in pemetrexed resistance of mesothelioma: clinical evidence and new pharmacological tools.

BACKGROUND: Thymidylate synthase (TS) has a predictive role in pemetrexed treatment of mesothelioma; however, additional chemoresistance mechanisms are poorly understood. Here, we explored the role of the reduced-folate carrier (RFC/SLC19A1) and proton-coupled folate transporter (PCFT/SLC46A1) in antifolate resistance in mesothelioma. PATIENTS AND METHODS: PCFT, RFC and TS RNA and PCFT protein levels were determined by quantitative RT-PCR of frozen tissues and immunohistochemistry of tissue-microarrays, respectively, in two cohorts of pemetrexed-treated patients. Data were analyzed by t-test, Fisher's/log-rank test and Cox proportional models. The contribution of PCFT expression and PCFT-promoter methylation to pemetrexed activity were evaluated in mesothelioma cells and spheroids, through 5-aza-2'-deoxycytidine-mediated demethylation and siRNA-knockdown. RESULTS: Pemetrexed-treated patients with low PCFT had significantly lower rates of disease control, and shorter overall survival (OS), in both the test (N = 73, 11.3 versus 20.1 months, P = 0.01) and validation (N = 51, 12.6 versus 30.3 months, P = 0.02) cohorts. Multivariate analysis confirmed PCFT-independent prognostic role. Low-PCFT protein levels were also associated with shorter OS. Patients with both low-PCFT and high-TS levels had the worst prognosis (OS, 5.5 months), whereas associations were neither found for RFC nor in pemetrexed-untreated patients. PCFT silencing reduced pemetrexed sensitivity, whereas 5-aza-2'-deoxycytidine overcame resistance. CONCLUSIONS: These findings identify for the first time PCFT as a novel mesothelioma prognostic biomarker, prompting prospective trials for its validation. Moreover, preclinical data suggest that targeting PCFT-promoter methylation might eradicate pemetrexed-resistant cells characterized by low-PCFT expression.

The GATA site-dependent hemogen promoter is transcriptionally regulated by GATA1 in hematopoietic and leukemia cells.

Hemgn (a gene symbol for hemogen in mouse, EDAG in human and RP59 in rat) encodes a nuclear protein that is highly expressed in hematopoietic tissues and acute leukemia. To characterize its regulatory mechanisms, we examined the activities of a Hemgn promoter containing 2975 bp of 5' flanking sequence and 196 bp of 5' untranslated region (5' UTR) sequence both in vitro and in vivo: this promoter is preferentially activated in a hematopoietic cell line, not in nonhematopoietic cell lines, and is sufficient to drive the transcription of a lacZ transgene in hematopoietic tissues in transgenic mice. Mutagenesis analyses showed that the 5' UTR including two highly conserved GATA boxes is critical for the promoter activity. GATA1, not GATA2, binds to the GATA binding sites and transactivates the Hemgn promoter in a dose-dependent manner. Furthermore, the expression of human hemogen (EDAG) transcripts were closely correlated with levels of GATA1 transcripts in primary acute myeloid leukemia specimens. This study suggests that the Hemgn promoter contains critical regulatory elements for its transcription in hematopoietic tissues and Hemgn is a direct target of GATA1 in leukemia cells.

Molecular and cellular biology of the human reduced folate carrier.

The natural folates are water-soluble members of the B class of vitamins that are essential for cell proliferation and tissue regeneration. Since mammalian cells cannot synthesize folates de novo, tightly regulated and sophisticated cellular uptake processes have evolved to sustain sufficient levels of intracellular tetrahydrofolate cofactors to support the biosynthesis of purines, pyrimidines, serine, and methione. Membrane transport is also a critical determinant of the antitumor activity of antifolate therapeutics (methotrexate, Tomudex) used in cancer chemotherapy, and impaired uptake of antifolates is a frequent mode of drug resistance. The reduced folate carrier is the major transport system for folates and classical antifolates in mammalian cells and tissues. This review summarizes the remarkable advances in the cellular and molecular biology of the human reduced folate carrier over the past decade, relating to its molecular structure and transport function, mechanisms of transcriptional and posttranscriptional regulation, and its critical role in antifolate response and resistance. Many key in vitro findings have now begun to be extended to studies of reduced folate carrier levels and function in patient specimens, paving the way for translating basic laboratory studies in cultured cells to improvements in human health and treatment of disease. The results of research into the human reduced folate carrier should clarify the roles of changes in expression and function of this system that accompany nutritional folate deficiency and human disease, and may lead to improved therapeutic strategies for enhancing drug response and circumventing resistance in cancer patients undergoing chemotherapy with antifolates.

Identification of a highly glycosylated methotrexate membrane carrier in K562 human erythroleukemia cells up-regulated for tetrahydrofolate cofactor and methotrexate transport.

A K562 human erythroleukemia line (designated K562.4CF) was selected for increased tetrahydrofolate cofactor transport in a growth-limiting concentration (0.4 nM) of (6R,S)-5-formyltetrahydrofolate. K562.4CF cells exhibited elevated methotrexate uptake relative to parental cells, attributable to a 10-fold increased influx Vmax. The rate of methotrexate efflux in K562.4CF cells was somewhat increased (55%) as well. The transport system in K562.4CF cells had similar and high apparent binding affinities for methotrexate and 5-formyltetrahydrofolate and a markedly reduced affinity for folic acid, properties typically associated with the "classical" methotrexate/tetrahydrofolate cofactor transporter in tumor cells. Methotrexate uptake in K562.4CF cells decreased substantially under nonselective conditions; high levels of transport were restored in 0.4 nM 5-formyltetrahydrofolate. Treatment of parental and K562.4CF cells with N-hydroxysuccinimide methotrexate inhibited methotrexate influx. N-Hydroxysuccinimide-[3H]methotrexate (700 nM) radiolabeled a broadly migrating band at Mr 76,000-85,000. Incorporation from N-hydroxysuccinimide-[3H]methotrexate into this band was increased 7-fold in K562.4CF over parental cells and was blocked by unlabeled methotrexate, (6S)-5-formyltetrahydrofolate, or, to a lesser extent, folic acid. Whereas incubation with endoglycosidase F had no effect on the electrophoretic migration of the labeled protein, treatment with endoglycosidase F and glycopeptidase F, or endo-beta-galactosidase, reduced the apparent molecular weight to Mr approximately 52,000 or approximately 58,000, respectively. These results suggest that the high-affinity transporter in K562.4CF cells is an N-linked glycoprotein containing internal beta-galactosidic linkages in, or immediately after, unbranched poly-N-acetyllactosamine sequences. Differences in the level of glycosylation may, in part, account for the disparity in the apparent sizes of the homologous folate transport proteins from human and murine cells.

Role of the cellular oxidation-reduction state in methotrexate binding to dihydrofolate reductase and dissociation induced by reduced folates.

5-Formyltetrahydrofolate promotes the net dissociation of methotrexate bound to dihydrofolate reductase in the Ehrlich ascites tumor (L. H. Matherly et al., Cancer Res., 43: 2694-2699, 1983). Treatment of Ehrlich tumor cells with glucose or inhibitors of electron transfer stabilized the association of the antifolate with dihydrofolate reductase as reflected by a 2-fold increased fraction of dihydrofolate reductase-bound methotrexate and an abolition of the 5-formyltetrahydrofolate-induced dissociation of the inhibitor-enzyme complex. Glucose and azide were also found to increase the intracellular ratio of reduced nicotinamide adenine dinucleotide phosphate (NADPH) to oxidized nicotinamide adenine dinucleotide phosphate (NADP+) in the tumor approximately 8- and 11-fold, respectively. However, other agents which enhanced the association between methotrexate and its target enzyme were less effective in increasing the intracellular level of NADPH relative to NADP+. Micromolar concentrations of NADPH promoted methotrexate binding to the purified Ehrlich tumor dihydrofolate reductase. Bound methotrexate could be dissociated from the purified enzyme by 5-methyltetrahydrofolate but less readily by 5-formyltetrahydrofolate and only in the presence of reduced levels of NADPH relative to NADP+. The tetraglutamate derivative of 5-methyltetrahydrofolate was even more effective than the underivatized compound in dissociating methotrexate from dihydrofolate reductase. These findings suggest a critical role for the cellular oxidation-reduction state in determining the affinity of dihydrofolate reductase for methotrexate and thus the cellular sensitivity to the antifolate. In addition, the data are consistent with the possibility that dihydrofolate reductase is a key locus for intracellular competitive interactions between reduced folates and methotrexate during leucovorin rescue from the pharmacological effects of the antifolate.

Role of methotrexate polyglutamylation and cellular energy metabolism in inhibition of methotrexate binding to dihydrofolate reductase by 5-formyltetrahydrofolate in Ehrlich ascites tumor cells in vitro.

5-Formyltetrahydrofolate was found to reverse the binding of methotrexate to dihydrofolate reductase in the Ehrlich ascites tumor in vitro. When cells pretreated with methotrexate were resuspended in methotrexate-free buffer containing 5-formyltetrahydrofolate (or 5-methyltetrahydrofolate), net dissociation of the antifolate from the enzyme was observed. Methotrexate associated with the enzyme under these conditions was below the enzyme binding capacity. However, glucose or azide increased the fraction of dihydrofolate reductase associated with methotrexate and abolished the effect of tetrahydrofolates on this intracellular component. Addition of 5-fluoro-2'-deoxyuridine had no effect on this response to the reduced folate, thereby precluding a direct role for the thymidylate synthase-dependent generation of dihydrofolate in this dissociation of methotrexate from dihydrofolate reductase. Enzyme-bound methotrexate could also be reduced by exposure to 5-formyltetrahydrofolate prior to uptake and efflux of free methotrexate. When cells were incubated under conditions which favored formation of methotrexate polyglutamate derivatives, subsequent treatment with 5-formyltetrahydrofolate had no effect on the binding of the conjugated antifolate to dihydrofolate reductase. These findings support a role for dihydrofolate reductase as a locus for competitive binding interactions between reduced folates and methotrexate that may be a basis for the ability of 5-formyltetrahydrofolate to prevent the biochemical effects of this antifolate. These data suggest that the presence of methotrexate polyglutamate derivatives and cellular energy metabolism may be critical determinants of the responsiveness of methotrexate-treated cells to reduced folates and may play important roles in the selectivity of 5-formyltetrahydrofolate rescue.

Antifolate polyglutamylation and competitive drug displacement at dihydrofolate reductase as important elements in leucovorin rescue in L1210 cells.

Previous studies from this laboratory have shown that the addition of leucovorin to tumor cells dissociates methotrexate, but not methotrexate polyglutamates, from dihydrofolate reductase (L. H. Matherly, D. W. Fry, and I. D. Goldman, Cancer Res., 43: 2694-2699, 1983). To further assess the importance of these interactions to leucovorin rescue, antifolate growth inhibition toward L1210 cells in the presence of leucovorin was correlated with the metabolism of (6S)-5-formyl tetrahydrofolate to dihydrofolate as a measure of dihydrofolate reductase activity. Growth inhibition (greater than 95%) by methotrexate (5-10 microM) following its intracellular polyglutamylation during a 3-h preexposure, or by continuous treatment with high levels of the lipophilic antifolate, trimetrexate (1 microM), was only slightly diminished by 10 microM leucovorin (15-25%). High-pressure liquid chromatographic analyses of the derivatives formed from radiolabeled (6S)-5-formyl tetrahydrofolate under these conditions showed an incomplete conversion to dihydrofolate and metabolism to predominantly 10-formyl tetrahydrofolate. Neither of the antifolates interfered appreciably with the metabolism of the folate derivatives to polyglutamates. Growth inhibition in the presence of leucovorin correlated with the accumulation of dihydrofolate (1.5-2.2 nmol) from radiolabeled (6S)-5-formyl tetrahydrofolate, reflecting continued suppression of dihydrofolate reductase activity at these drug concentrations. With lower equitoxic levels of the trimetrexate (7.5 nM), the provision of leucovorin allowed for a restoration of cell growth to a level greater than 90% of control. Under these conditions, control levels of dihydrofolate (0.2 nmol) were formed from radiolabeled cofactor, consistent with sustained dihydrofolate reductase activity. These findings support a role for the activation of dihydrofolate reductase as an important component of the reversal of the effects of diaminoantifolates by leucovorin, presumably by a competitive displacement of drug from the enzyme. Since no displacement occurs in cells which have accumulated methotrexate polyglutamates, or in the presence of high levels of trimetrexate, it appears that the concentration of unbound drug within cells is a significant determinant of the extent of this competitive binding interaction. From these considerations, the high levels of methotrexate polyglutamates that accumulate in sensitive tumors relative to bone marrow and gastrointestinal cells would appear to represent an important factor for the selectivity of leucovorin rescue in vivo.

In vitro formation of polyglutamyl derivatives of methotrexate and 7-hydroxymethotrexate in human lymphoblastic leukemia cells.

The intracellular synthesis of polyglutamyl derivatives of both methotrexate (4-amino-N-10-methylpteroylglutamic acid) and 7-hydroxymethotrexate, the primary plasma metabolite of methotrexate in humans, was evaluated in a methotrexate-sensitive, acute lymphoblastic leukemia cell line, MOLT 4. These studies were performed using a highly specific ion-pairing high-pressure liquid chromatography method which permits the simultaneous determination of methotrexate, 7-hydroxymethotrexate, and their corresponding polyglutamyl derivatives. When MOLT 4 cells were exposed to 1 microM methotrexate, the monoglutamate attained a steady state after 30 min, and polyglutamyl derivatives having from one to 4 additional glutamyl residues were observed over 4 hr. Four additional metabolites were also detected upon incubation with 1 microM 7-hydroxymethotrexate. On the basis of the retention times for these compounds relative to methotrexate polyglutamyl standards and since these metabolites reverted to 7-hydroxymethotrexate upon treatment with a preparation of hog kidney conjugase, they were identified as polyglutamyl derivatives of 7-hydroxymethotrexate. The identification of 7-hydroxymethotrexate polyglutamyl derivatives in vitro raises the possibility of an important new dimension in the pharmacological action of methotrexate. We investigated the effect of extracellular 7-hydroxymethotrexate on net methotrexate uptake and metabolism when cells were exposed simultaneously to 1 microM [3H]-methotrexate and unlabeled 7-hydroxymethotrexate. A decrease in the levels of both intracellular methotrexate and the corresponding polyglutamyl derivatives was noted for cells treated with 1 or 10 microM 7-hydroxymethotrexate. However, no appreciable effect of 7-hydroxymethotrexate on the amount of polyglutamyl derivatives formed relative to the total intracellular antifolate was noted. These studies show that in MOLT 4 cells (a) both methotrexate and 7-hydroxymethotrexate are rapidly converted to polyglutamyl derivatives, and (b) 7-hydroxymethotrexate interferes with net methotrexate accumulation and metabolism when present simultaneously in the extracellular medium. These results, moreover, suggest a potential role for 7-hydroxymethotrexate in modulating the biochemical effects of methotrexate in vivo.

Role for cytosolic folate-binding proteins in the compartmentation of endogenous tetrahydrofolates and the 5-formyl tetrahydrofolate-mediated enhancement of 5-fluoro-2'-deoxyuridine antitumor activity in vitro.

A major portion of the intracellular folates in L1210 cells grown in (6R,S)-5-formyltetrahydrofolate (leucovorin) was bound to cytosolic proteins when cell extracts were fractionated by rapid gel filtration or adsorption with activated charcoal. Only low levels of intracellular folates were associated with mitochondria (less than 5%). Protein-bound folates comprised 37-100% of the cytosolic cofactors following growth in 2-600 nM 5-formyltetrahydrofolate. Total intracellular folates increased in proportion to the changes in media folate concentration; however, binding was saturable. The maximum level of protein-bound folates in L1210 cells was 66 pmol/mg protein. Protein-bound folates were also detected in HT29 human colon adenocarcinoma cells grown in 5-formyltetrahydrofolate (maximum, 11 pmol/mg protein). For both lines, folate binding was specific for the tetrahydrofolate and 5,10-methylenetetrahydrofolate pool, and, to a lesser extent, 5-methyltetrahydrofolate. Extremely low levels of protein-bound 5-formyl-, 10-formyl-, and 5,10-methenyltetrahydrofolates were measured, even though considerable amounts were detected intracellularly. Pentaglutamyl folates were the predominant cofactor forms in L1210 cells; conversely, the tetraglutamates were the most abundant protein-bound folate derivatives. Increasing media concentrations of 5-formyltetrahydrofolate potentiated 5-fluoro-2'-deoxyuridine cytotoxicity. For L1210 cells, essentially all of the intracellular tetrahydrofolate and 5,10-methylenetetrahydrofolate fraction was protein bound over the concentration range of 5-formyltetrahydrofolate which maximally augmented fluoropyrimidine cytotoxicity. The relative changes in the 50% inhibitory concentrations for fluorodeoxyuridine directly approximated the increases in the levels of protein-bound tetrahydrofolates in L1210 cells. There was no direct relationship between the levels of unbound folates and fluorodeoxyuridine cytotoxicity. Similar results were obtained with HT29 cells. The major folate-binding protein in L1210 cells eluted during Sephacryl S-300 chromatography with a molecular weight of approximately 200,000; a small amount of a higher molecular weight folate-binding protein (Mr 450,000) was also detected. These findings support the concept of a compartmentation of endogenous folates involving specific binding to cytosolic proteins. These associations may regulate reduced folate availability for metabolic processes, and also mediate utilization of 5,10-methylenetetrahydrofolate for ternary complex formation with thymidylate synthase in cells treated with fluoropyrimidines. In this fashion, the levels of protein-bound tetrahydrofolates could represent an additional, previously unrecognized, determinant of fluoropyrimidine pharmacological activity toward mammalian cells.

Enhanced polyglutamylation of aminopterin relative to methotrexate in the Ehrlich ascites tumor cell in vitro.

The polyglutamylation of aminopterin and methotrexate (N10-methylaminopterin) was compared in the Ehrlich ascites tumor in vitro. Three poly-gamma-glutamyl conjugates of methotrexate and aminopterin were detected, although at an equal (1 microM) extracellular drug concentration, the net accumulation of aminopterin polyglutamates exceeded that for the methotrexate polyglutamyl derivatives by a factor of 9. When compensation was made for transport differences between these compounds by adjusting the extracellular drug concentrations to achieve equivalent intracellular monoglutamyl substrate levels, the polyglutamylation of aminopterin was still 2.8-fold greater than that for methotrexate, suggesting that aminopterin is a better substrate for the folylpolyglutamate synthetase as well as the transport carrier. An additional metabolite of aminopterin was detected within seconds following drug exposure. This derivative did not bind tightly to dihydrofolate reductase, yet it was rapidly converted to a polyglutamate. The formation of both aminopterin polyglutamates and these novel derivatives was enhanced by increases in the free intracellular level of aminopterin. Aminopterin polyglutamates were bound tightly to dihydrofolate reductase and were retained intracellularly relative to unaltered aminopterin when Ehrlich cells containing these forms were suspended in drug-free medium. These findings support a role for the polyglutamylation of aminopterin as a critical element in drug action and as a factor in addition to membrane transport in the disparate antifolate potencies of aminopterin and methotrexate.

Interactions between 7-hydroxymethotrexate and methotrexate at the cellular level in the Ehrlich ascites tumor in vitro.

Studies were undertaken to characterize the cellular pharmacology of 7-hydroxymethotrexate (7-OH-MTX) in Ehrlich ascites tumor cells, compare it to that of methotrexate (MTX), and define the interactions between the parent compound and its catabolite. Transport of 7-OH-MTX is mediated by the MTX-tetrahydrofolate cofactor carrier, with a Km of 9 microM in comparison to the MTX Km of 5 microM. Both compounds mutually inhibit their influx and steady-state levels of free drug accumulated. While influx of 7-OH-MTX is slower than influx of MTX, 7-OH-MTX efflux is likewise slower, so that the steady-state level of 7-OH-MTX achieved is comparable to that of MTX. Influx of 7-OH-MTX is inhibited by extracellular 5-formyltetrahydrofolate and trans-stimulated in cells preloaded with this tetrahydrofolate cofactor. The energetics of 7-OH-MTX transport is similar to that of MTX in the influx and net transport are stimulated by sodium azide, while net transport is reduced by glucose. As observed for MTX, 7-OH-MTX transport is sensitive to the anionic composition of the extracellular compartment and was shown to be inhibited by organic and inorganic phosphates. 7-OH-MTX does not, alone, inhibit [3H]deoxyuridine incorporation into DNA at concentrations of up to 50 microM. However, the catabolite reduces MTX inhibition of deoxyuridine metabolism, presumably due to the reduction in the free level of intracellular MTX achieved. These findings support the possibility that when 7-OH-MTX accumulates to high levels relative to MTX in clinical regimens, it may modulate the pharmacological effects of MTX.

Cystathionine-beta-synthase cDNA transfection alters the sensitivity and metabolism of 1-beta-D-arabinofuranosylcytosine in CCRF-CEM leukemia cells in vitro and in vivo: a model of leukemia in Down syndrome.

The significantly higher event-free survival rates of Down syndrome (DS) children with acute myeloid leukemia compared with non-DS children is linked to increased sensitivity of DS myeloblasts to 1-beta-D-arabinofuranosylcytosine (ara-C) and the enhanced metabolism of ara-C to ara-C triphosphate (J. W. Taub et al., Blood, 87: 3395-3403, 1996). The cystathionine-beta-synthase (CBS) gene (localized to chromosome 21q22.3) may have downstream effects on reduced folate and S-adenosylmethionine pathways; ara-C metabolism and folate pools are linked by the known synergistic effect of sequential methotrexate and ara-C therapy. We have shown that relative CBS transcripts were significantly higher in DS compared with non-DS myeloblasts, and CBS transcript levels correlated with in vitro ara-C sensitivity (J. W. Taub et al., Blood, 94: 1393-1400, 1999). A leukemia cell line model to study the relationship of the CBS gene and ara-C metabolism/sensitivity was developed by transfecting CBS-null CCRF-CEM cells with the CBS cDNA. CBS-transfected cells were a median 15-fold more sensitive in vitro to ara-C compared with wild-type cells and generated 8.5-fold higher [3H]ara-C triphosphate levels after in vitro incubation with [3H]ara-C. Severe combined immunodeficient mice implanted with CBS-transfected CEM cells demonstrated greater responsiveness to therapy, reflected in significantly prolonged survivals after ara-C administration compared with mice implanted with wild-type cells and treated with the same dosage schedule. The transfected cells also demonstrated increased in vitro and in vivo sensitivity to gemcitabine. Deoxycytidine kinase (dCK) activity was approximately 22-fold higher in transfected CEM cells compared with wild-type cells. However, levels of dCK transcripts on Northern blots and protein levels on Western blots were nearly identical between CBS-transfected and wild-type cells. Collectively, these results suggest a posttranscriptional regulation of dCK in CBS-overexpressing cells that contributes to increased ara-C phosphorylation and drug activity. Further elucidating the mechanisms of increased sensitivity of DS cells to ara-C related to the CBS gene may lead to the application of these novel approaches to acute myeloid leukemia therapy for non-DS patients.

Structure and organization of the human reduced folate carrier gene.

The human reduced folate carrier gene was found to contain 7 exons, including two alternative non-coding exons (exons 1 and 2), spanning approximately 29 kb. Two transcript variants involving exon 7 were detected in K562 cells by RT-PCR, distinguishable from the wild-type transcript by deletions of 625 bp (KS32) and 988 bp (KS1). The presence of consensus splice donor and acceptor elements in the deleted KS1 isoform suggested that this form was likely a splice variant; however, KS32 likely arose during reverse transcription rather than by alternative splicing.

Effects of the loss of capacity for N-glycosylation on the transport activity and cellular localization of the human reduced folate carrier.

The role of N-glycosylation in reduced folate carrier (RFC) transport and membrane targeting was examined in transport-deficient K562 (K500E) cells transfected with human RFC cDNAs. Treatment of cells expressing wild-type RFC with tunicamycin (0-3 microg) resulted in a progressive shift of the approximately 85 kDa RFC on western blots to 65 kDa. At 3 microg/ml tunicamycin, the nearly complete loss of glycosylated RFC was accompanied by a approximately 25% decreased rate of methotrexate uptake. A deglycosylated RFC cDNA construct in which asparagine-58 was replaced by glutamine (Gln58-RFC) was expressed in K500E cells as a 65 kDa protein and restored transport capacity for methotrexate and (6S)5-formyl tetrahydrofolate. With both wild-type and Gln58-RFC constructs, expression of cDNA-encoded RFC protein far exceeded relative levels of RFC uptake. Wild-type and Gln58-RFCs containing a hemagglutinin (HA) epitope at the carboxyl terminus were similarly functional and, by immunofluorescence staining with rhodamine-conjugated anti-HA antibody, were localized to plasma membranes. Collectively, our results demonstrate that N-glycosylation of human RFC plays no significant role in either transport function or membrane targeting. The discrepancy between the stoichiometries of RFC expression and transport activity for both wild-type RFC and Gln58-RFC implies that identical regulatory controls and/or non-RFC transport components are necessary to completely restore transport function in the transfected cells.

Role of the E45K-reduced folate carrier gene mutation in methotrexate resistance in human leukemia cells.

Resistance to the antifolate methotrexate (MTX) can cause treatment failure in childhood acute lymphoblastic leukemia (ALL). This may result from defective MTX accumulation due to alterations in the human reduced folate carrier (hRFC) gene. We have identified an hRFC gene point mutation in a transport-defective CCRF-CEM human T-ALL cell line resulting in a lysine to glutamic acid substitution at codon 45 (E45K), which has been identified in other antifolate-resistant sublines (JBC 273:30 189, 1998; JBC 275:30 855, 2000). To characterize the role of this mutation in MTX resistance, transfection experiments were performed using hRFC-null CCRF-CEM cells. E45K transfectants demonstrated an initial rate of MTX influx that was approximately 0.5-fold that of CCRF-CEM cells, despite marked protein overexpression. Cytotoxicity studies revealed partial reversal of MTX and raltitrexed resistance in E45K transfectants, while trimetrexate resistance was significantly increased. Kinetic analysis indicated only minor differences in MTX kinetics between wild-type and E45K hRFCs, however, K(i)s for folic acid and 5-formyltetrahydrofolate were markedly reduced for E45K hRFC. This was paralleled by increased folic acid transport and reduced synthesis of MTX polyglutamates. Collectively, the results demonstrate that expression of E45K hRFC leads to increased MTX resistance due to decreased membrane transport and, secondarily, from alterations in binding affinities and transport of folate substrates. However, despite these findings, we could find no evidence of this mutation in 121 childhood ALL samples, suggesting that it does not contribute to clinical MTX resistance in this disease.

Reduced folate carrier gene expression in childhood acute lymphoblastic leukemia: relationship to immunophenotype and ploidy.

Reduced folate carrier (RFC) transcripts in human leukemias were measured by a competitive PCR assay. Total RNAs were reverse transcribed and amplified in the presence of competitive templates for RFC and beta-actin. RFC transcripts were normalized to transcripts for beta-actin. In a series of K562 sublines, a approximately 30-fold range of RFC transcripts measured by PCR assay closely agreed with results of Northern analysis and varied in proportion to RFC protein on Western blots and [3H]methotrexate transport. RFC transcripts varied over a 88-fold range in 49 specimens from 48 children with acute lymphoblastic leukemia (ALL). Median RFC transcripts were similar for 15 T-cell and 33 B-precursor ALL samples (RFC/beta-actin = 6.13 x 10(-3) and 7.92 x 10(-3), respectively) and for 41 diagnostic (7.20 x 10(-3)) and 8 relapse (5.58 x 10(-3)) samples. Whereas PCR measurements of RFC transcripts approximated changes in methotrexate transport in B-precursor ALL blasts (n = 10), for T-ALL blasts (n = 12) there was no apparent relationship between these parameters. For hyperdiploid B-precursor blasts (n = 11) with greater than 52 chromosomes and three to five copies of chromosome 21, the median RFC transcript level was approximately 3-fold higher than that for diploid B-precursor blasts. RFC transcripts were also elevated for two of three B-precursor specimens with acquired trisomy 21. Our results suggest that RFC gene expression is far more predictive of methotrexate uptake capacity in B-precursor than T-ALL and that increased copies of chromosome 21 in B-precursor ALL blasts are generally associated with increased RFC transcripts. Hence, the good prognosis for children with hyperdiploid B-precursor ALL treated with antimetabolite-based chemotherapy and the high levels of methotrexate and methotrexate polyglutamates accumulated may, in part, reflect elevated RFC gene expression and capacities for methotrexate transport.

Single nucleotide polymorphisms in the human reduced folate carrier: characterization of a high-frequency G/A variant at position 80 and transport properties of the His(27) and Arg(27) carriers.

The presence of sequence variants in the human reduced folate carrier (hRFC) was assessed in leukemia blasts from children with acute lymphoblastic leukemia (ALL) and in normal peripheral blood specimens. A CATG frame shift insertion at position 191 was detected in 10-60% of hRFC transcripts from 10 of 16 ALL specimens, by RFLP analysis and direct sequencing of hRFC cDNAs. In genomic DNAs prepared from 105 leukemia (n = 54) and non-leukemia (n = 51) specimens, PCR amplifications and direct sequencing of exon 3 identified a high-frequency G to A single nucleotide polymorphism at position 80 that resulted in a change of arginine-27 to histidine-27. The allelic frequencies of G/A80 were nearly identical for the non-leukemia (42.2% CGC and 57.8% CAC) and leukemia (40.7% CGC and 59.3% CAC) genomic DNAs. In cDNAs prepared from 10 of these ALL patients, identical allelic frequencies (40 and 60%, respectively) were recorded. In up to 62 genomic DNAs, hRFC-coding exons 4-7 were PCR-amplified and sequenced. A high-abundance C/T696 polymorphism was detected with nearly identical frequencies for both alleles, and a heterozygous C/A1242 sequence variant was identified in two ALL specimens. Both C/T696 and C/A1242 were phenotypically silent. In transport assays with [(3)H]methotrexate and [(3)H]5-formyl tetrahydrofolate, nearly identical uptake rates were measured for the arginine-27- and histidine-27-hRFC proteins expressed in transport-impaired K562 cells. Although there were no significant differences between the kinetic parameters for methotrexate transport for the hRFC forms, minor (approximately 2-fold) differences were measured in the K(i)s for other substrates including Tomudex, 5,10-dideazatetrahydrofolate, GW1843U89, and 10-ethyl-10-deazaaminopterin and for 5-formyl tetrahydrofolate.

Synthesis and biological activity of N omega-hemiphthaloyl-alpha,omega- diaminoalkanoic acid analogues of aminopterin and 3',5-dichloroaminopterin.

Analogues of N alpha-(4-amino-4-deoxypteroyl)-N delta-(hemiphthaloyl)-L-ornithine (PT523) with 3',5'-dichloro substitution in the p-aminobenzoyl moiety or with one less or one more CH2 group in the amino acid moiety were synthesized and tested as inhibitors of dihydrofolate reductase (DHFR) activity and cell growth. Replacement of L-ornithine in PT523 by L-2,4-diaminobutanoic acid or L-lysine did not decrease binding to human recombinant DHFR but resulted in some loss of activity against SCC25 human and SCC VII murine squamous cell carcinoma and against MCF-7 human breast carcinoma in culture. PT523 was several times more potent than methotrexate (MTX), aminopterin (AMT), or trimetrexate (TMQ). 3',5'-Dichloro substitution did not decrease either DHFR binding or cytotoxicity. A new synthetic route to PT523 from 2,4-diamino-6-(hydroxymethyl)pteridine and methyl N alpha-(4-aminobenzoyl)-N delta-phthaloyl-L-ornithinate was investigated but was not found superior to previously described methods. In comparative experiments on the ability of PT523 and MTX to competitively inhibit the influx of (6R)-5,10-dideazatetra-hydrofolate (DDATHF, lometrexol), used here as a surrogate for MTX and reduced folates, the Ki of PT523 was lower than that of MTX in both wild-type CCRF-CEM human leukemic lymphoblasts and the transport- and polyglutamylation-defective subline CEM/MTX. The CCRF-CEM cells were 10-fold more sensitive to PT523 than to MTX, whereas the CEM/MTX cells were 240-fold more sensitive. However, in contrast to other MTX-resistant cells where collateral sensitivity to PT523 has been seen. CEM/MTX cells still showed substantial cross resistance to PT523 which may reflect an unusual heightened ability to utilize exogenous folic acid. The good correlation observed with both cell lines between the cytotoxicity of PT523 and MTX and the ability to inhibit DDATHF influx supported the view that PT523 and MTX share, at least in part, a common protein carrier for membrane transport.

Evidence for a localized conversion of endogenous tetrahydrofolate cofactors to dihydrofolate as an important element in antifolate action in murine leukemia cells.

The inhibition of de novo nucleotide, serine, and methionine biosynthesis in mammalian cells treated with antifolates has been attributed generally to a reduction in the levels of tetrahydrofolate cofactors. In L1210 leukemia cells grown in tritiated folic acid (1 microM), most of the endogenous radiolabeled folates were present as formyl-substituted tetrahydrofolates (60-73%, including 10- and 5-formyl and 5,10-methenyl tetrahydrofolate), with lower levels of tetrahydrofolate (including 5,10-methylene tetrahydrofolate), 5-methyl tetrahydrofolate, and non-metabolized folic acid. Trimetrexate (1 microM) caused an elevation of dihydrofolate levels within 5 min following drug addition, from approximately 1 to 20% of the total folates. Whereas total reduced folates were preserved, losses in the levels of individual forms ranged from minor changes in the formyl tetrahydrofolates (approx. 10% decrease), to significant losses in the levels of tetrahydrofolate (approx. 60%) and 5-methyl tetrahydrofolate (95%). Under these conditions, the incorporations of [3H]deoxyuridine into TMP and [14C]glycine into purines or of [14C]formate into biosynthetic products were inhibited (69-95%). The majority (59-100%) of the endogenous radiolabeled folates in L1210 cells grown in various concentrations (0.2 to 3 microM) of [3H]folic acid was bound to soluble intracellular proteins when cell-free extracts were fractionated by rapid gel filtration or charcoal adsorption. Total intracellular folate levels increased in proportion to the changes in medium folic acid concentration; however, cofactor binding was saturable. At low concentrations, below that which supported maximal growth (less than 0.75 microM), all of the intracellular folates were protein-bound; only when maximal growth was achieved, could unbound folates be detected. Incubation with trimetrexate (1 or 10 microM), methotrexate (10 microM), or calcium leuvovorin (50 microM) did not alter significantly the levels of total and protein-bound [3H]folates in cells grown in 1 microM [3H]folic acid. Under all conditions, formyl tetrahydrofolates were the major intracellular derivatives; however, these forms were poorly represented in the bound fraction. Conversely, all of the other intracellular folate forms were completely bound. Tetrahydrofolate was the predominant protein-bound derivative in control cells; in antifolate-treated cells, both bound tetrahydrofolate and 5-methyl tetrahydrofolate were largely replaced by protein-bound dihydrofolate. This interconversion in drug-treated cells was independent of (i) sustained levels of [3H]formyl tetrahydrofolates, or (ii) high extracellular concentrations of unlabeled calcium leucovorin (50 microM). Hence, protein-bound tetrahydrofolates must not only be substrates for enzyme mediated reactions (i.e. TMP synthesis) but also must slowly equilibrate with unbound cofactor. In this fashion, binding of endogenous folates to soluble proteins may function to "segregate' intracellular cofactor pools.(ABSTRACT TRUNCATED AT 400 WORDS)

Role of N-glycosylation in the structure and function of the methotrexate membrane transporter from CCRF-CEM human lymphoblastic leukemia cells.

The carrier protein for methotrexate and tetrahydrofolate cofactors (GP-MTX) in CCRF-CEM human lymphoblastic leukemia cells in a 117 kDa glycoprotein containing both N- and O-linked oligosaccharides (Matherly et al., J Biol Chem 267: 23253-23260, 1992). Tunicamycin, an inhibitor of N-glycosylation, was used to investigate the roles of asparagine-linked oligosaccharides in the structure, intracellular routing, and transport function of GP-MTX. Tunicamycin was growth inhibitory toward CCRF-CEM cells (IC50-0.80 micrograms/mL) and caused a potent suppression of [3H]mannose incorporation into nascent glycoproteins. From 1-3 micrograms/mL, inhibition of [3H]mannose incorporation was 66-87%, exceeding that for [35S]methionine incorporation by 2 to 4-fold. Tunicamycin (1 and 2 micrograms/mL) exposures decreased the median molecular masses of GP-MTX on immunoblots (to 82 and 67 kDa, respectively) and were accompanied by reduced maximal rates of methotrexate uptake (31 and 37%, respectively, of control levels). Conversely, the Ki values for methotrexate binding to the transporter were unaffected by tunicamycin treatments. The effects of tunicamycin on methotrexate influx closely correlated with lower levels of immunoreactive GP-MTX in plasma membranes and specific [3H]methotrexate binding to intact cells, suggesting that the transport effect was due to decreased numbers of carrier proteins at the membrane surface. The reduced molecular mass values for GP-MTX, which accompanied tunicamycin exposures, were further decreased (to 55 and 50 kDa at 1 and 2 micrograms/mL, respectively) by digestions with N-glycanase. Hence, despite the large loss of N-glycan from GP-MTX in tunicamycin-treated cells, residual core oligosaccharides remained. The sizes of hypoglycosylated GP-MTX following both treatments were similar to that of the functionally homologous methotrexate membrane carrier previously identified in L1210 murine leukemia cells.