Although the majority of children with acute lymphoblastic leukemia (ALL) are cured with combination chemotherapy containing methotrexate (MTX), drug resistance contributes to treatment failure for a substantial fraction of patients. The primary transporter for folates and MTX is the reduced folate carrier (RFC). Impaired drug transport is a documented mechanism of MTX resistance in patients with ALL; however, to the authors' knowledge it is not known whether inactivating RFC mutations are a contributing factor.
The authors devised a genomic polymerase chain reaction-single strand conformational polymorphism assay followed by sequencing and screened the entire RFC coding region for sequence alterations in DNA from 246 leukemia specimens from patients with diverse ethnic variation, 24 at the time of recurrence and the rest at the time of diagnosis. This cohort was comprised of 203 B-precursor ALL specimens (82.5%), 32 T-lineage ALL specimens (13%), and 11 acute myeloblastic leukemia specimens (4.5%).
Of 246 DNA samples, only 3 diagnosis B-precursor ALL specimens (1.2%) were found to harbor alterations in the RFC gene, including heterozygous single nucleotide changes resulting in D56H and D522N substitutions in the first extracellular loop and the C-terminus of this transporter, respectively. The third sample had a sequence alteration in exon 3 that could not be identified because of the lack of availability of DNA.
Reduced folates are essential cofactors that function as one-carbon donors necessary for the biosynthesis of purines, thymidine, and glycine.1 Unlike prokaryotes, animal cells are devoid of de novo biosynthesis of folates, including the major circulatory form, 5-methyl-tetrahydrofolate (5-CH3-THF).1 Because the diffusion of folate anions is an ineffective process, particularly at the nanomolar levels present in the blood, mammalian cells meet their absolute folate requirement through facilitated transport processes from exogenous sources.2, 3 The major route for the delivery of folates and antifolates, including methotrexate (MTX), in mammalian cells is the reduced folate carrier (RFC).2, 3 RFC is a 591-amino-acid transporter with 12 transmembrane domains (TMDs).4, 5 RFC has a high affinity (Km = 0.3–10.0 μM) for reduced folates and various hydrophilic antifoltes, including MTX.2 RFC functions as a classic facilitative transporter through a bidirectional anion-exchange mechanism with intracellular organic phosphates.6–8
Defective transport as a result of qualitative and/or quantitative alterations in cultured human and animal cells has been documented recently as a frequent mechanism of antifolate drug resistance.9–24 Using antifolate-resistant human and murine leukemia cell lines that display defective drug uptake, along with others, we recently have showed that the predominant mechanisms underlying these transport alterations are inactivating RFC mutations12–22 and loss of RFC gene expression due to alterations in the function of transcription factors.23, 24 Thus, RFC mutations have been identified that alter or disrupt RFC expression, resulting in impaired antifolate transport. It is noteworthy that these mutations cluster primarily in TMD1 of the human and mouse RFC, resulting in a major loss of antifolate uptake, while preserving sufficient reduced folate uptake to sustain cellular growth.
The treatment of mice bearing L1210 leukemia cells with MTX results in antifolate-resistance, frequently because of impaired drug transport.25 Consistently, although the majority of children with acute lymphoblastic leukemia (ALL) are cured with modern combination chemotherapy containing MTX, intrinsic or acquired drug resistance contribute to treatment failure for a substantial minority of such patients.26 Resistance to MTX, which plays a prominent role in drug therapy for patients with ALL, can result from a variety of mechanisms, including defective antifolate uptake through alterations in RFC expression and/or function.27–30 However, the exact molecular mechanisms underlying these qualitative and/or quantitative alterations in the RFC gene in ALL patients remain unknown to date. Toward this end, in this study, we devised a genomic polymerase chain reaction-single strand conformational polymorphism (PCR-SSCP) assay followed by DNA sequencing for the rapid screening of the entire RFC coding region for sequence variations using minute amounts of genomic DNA isolated from ALL specimens. In this report, we show for the first time that mutations in the RFC gene occur at a very low frequency in childhood leukemia specimens. We conclude that RFC mutations do not play any important role in intrinsic or acquired resistance to MTX in childhood leukemia.
MATERIALS AND METHODS
[3′,5′,7′-3H]MTX (23.4 Ci/mmol) obtained from Moravek Biochemicals (Brea, CA) was purified prior to use by thin-layer chromatography and was stored at − 80 °C.31 MTX, folic acid, and (d,l) leucovorin (calcium salt) were purchased from Sigma Chemical Company (St. Louis, MO). G-418 was obtained from GIBCO Chem. Co., Grand Island, NY).
Cell Lines and Tissue Culturing
Human CCRF-CEM leukemia cells, their MTX transport-defective CEM/MTX and CEM/T sublines,32 MTX transport-null murine leukemia L1210/MTXRA cells,11, 12 along with human breast carcinoma ZR-75-1 cells and their MTX transport-impaired MTXR-ZR-75 subline,33 were grown in RPMI-1640 medium (containing 2.3 μM folic acid; Biological Industries, Beth Haemek, Israel) supplemented with 10% fetal calf serum (GIBCO), 2 mM glutamine, 100 U/mL penicillin G (Sigma Chemical Company), and 100 μg/mL streptomycin sulfate (Sigma Chemical Company). The growth medium for CEM/MTX, CEM/T, and L1210/MTXRA cells contained 1 μM MTX.
The leucovorin and folic acid growth requirements as well as the sensitivity to MTX of human leukemia CEM cells and the various clonal transfectants were determined as described previously.16, 17 The 50% inhibitory concentration (IC50) was defined as the drug dose at which cell growth was inhibited by 50% compared with untreated controls. EC50 was defined as the folate cofactor concentration necessary to produce 50% of maximal cell growth.
Genomic PCR-SSCP Assay and DNA Sequencing
Genomic PCR-SSCP analysis of the 5 coding exons of the human RFC gene was performed using 10 oligonucleotide primer pairs (Table 1), as described previously.16, 17 PCR products that showed an altered RFC electrophoretic mobility pattern were fractionated on 1.5% agarose gels, purified using a DNA purification kit (Qiagen, Valencia, CA), and sequenced. At least three independent PCR reactions and DNA sequence determinations were performed (using different DNA preparations) to determine a definitive mutation.
Table 1. Oligonucleotide Primer Pairs Targeting the Entire Human Reduced Folate Carrier Coding Region Used for Genomic Polymerase Chain Reaction Analysis
Because exons 3 and 6 were too large to be covered by a single polymerase chain reaction-single strand conformational polymorphism fragment, they were divided into four fragments and two fragments, respectively, as denoted by Roman numerals.
After genomic PCR, the products were gel-purified (QIAquick Gel Extraction; Qiagen), incubated overnight at 4 °C in 10 mL 5 × reaction buffer (Promega, Madison, WI) containing Easy-pGEM-T plasmid and T4 ligase according to the instructions of the manufacturer. The recombinant plasmid was transformed into competent DH5α E. coli. The recombinant plasmid was purified using the Wizard Plus SV Minipreps DNA purification system (Promega) and subsequently was analyzed by PCR-SSCP and/or DNA sequencing.
Site-Directed Mutagenesis and RFC cDNA Transfections
The various point mutations and the single nucleotide polymorphic variations were inserted into a native RFC cDNA clone (hRFC1) by site-directed mutagenesis (QuikChange; Stratagene, La Jolla, CA), as detailed previously.16 Mouse L1210/MTXRA leukemia cells that lacked functional RFC transport activity11, 12 were electroporated with 40 μg of nonlinearized pcDNA3.1 (+) harboring the native or mutant RFC cDNAs, as described previously.16, 23
Northern Blot Analysis
Cytoplasmic RNA was isolated from parental mouse MTXrA cells and their different human RFC cDNA transfectants harboring various RFC mutations. Northern blot hybridization was then performed with a [32P]labeled human RFC cDNA probe, as described previously.16
[3H]MTX Transport Studies
Initial rates of [3H]MTX uptake were determined as described previously.16, 17
Leukemia specimens, including B-precursor and T-cell ALL lymphoblasts as well as acute myeloblastic leukemia (AML) cells, were obtained from patients with newly diagnosed or recurrent disease from the Departments of Pediatric Hematology-Oncology at the following medical centers: The Cancer Institute of New Jersey (New Brunswick, NJ), Wayne State University School of Medicine (Detroit, MI), University Hospital VU (Amsterdam, The Netherlands), Rambam Medical Center (Haifa, Israel), and Sheba Medical Center (Tel-Hashomer, Israel). All patient samples were obtained after informed consent was granted. Leukemia cells were separated from bone marrow by standard Ficoll-Hypaque density centrifugation. Specimens were documented for patient age, gender, race, presenting leukocyte count, percent blast cells in bone marrow, and immunophenotype. Genomic DNA was isolated from 3–5 × 106 leukemic cells using the DNA isolation kit of Puregene (Gentra Systems, Minneapolis, MN) or Qiagen DNeasy.
Identification by Genomic PCR-SSCP of RFC Mutations in MTX-Resistant Tumor Cell Lines
To screen for RFC sequence variations in MTX-resistant tumor cells, we developed a genomic PCR-SSCP assay. Genomic DNA is extracted first, and PCR is performed with specific oligonucleotide primers spanning the entire RFC coding region, which is comprised of exons 2–6 (Table 1). Then, single-stranded DNA products are examined for aberrant electrophoretic mobility on native polyacrylamide gels capable of detecting a single nucleotide change in a fragment of up to ≈ 300 nucleotides. DNA products with putative genetic alteration(s) are sequenced, and the role of the observed sequence variation on MTX resistance, folate growth requirement, and MTX transport is determined by transfection of the mutant RFC cDNA into transport-null cells. We showed recently that the severe transport impairment of human leukemia CEM/MTX cells32 is due to a single G→A mutation at nucleotide 133 in exon 2, thus resulting in an E45K mutation in TMD1 of the RFC (Table 2).14 Furthermore, an additional C→T mutation at nucleotide 258 in exon 3 also was identified in these CEM-MTX cells; however, this was a silent change, thus retaining wild-type leucine 86 (Table 2). Hence, first, we examined the ability of the genomic PCR-SSCP assay to detect these as well as other mutations from the MTX transport-defective cell lines CEM/T32 and from breast carcinoma MTXR-ZR-75 cells.33 Whereas parental CEM cells contained 3 distinct bands in exon 2 (Fig. 1A, arrows) and 2 discrete bands in exon 3 (Fig. 1B, arrows), CEM/MTX cells contained an additional band in exon 2 (Fig. 1A, arrowhead) and 2 additional bands in exon 3 (Fig. 1B, arrowheads). Furthermore, although it has been shown that human leukemia CEM/T cells and MTXR-ZR-75 breast carcinoma cells are highly MTX transport defective, the molecular basis was never identified.32, 33 Whereas parental CEM cells contained 3 distinct bands in exon 2 (Fig. 2C) and a single band in exon 4 (Fig. 1D), CEM/T cells had a single, aberrantly migrating band in exon 2 (Fig. 1C, arrowhead) and two bands in exon 4 (Fig. 1D, arrowhead), suggesting 2 distinct mutations in both exon 2 and exon 4. Similarly, a single band with an aberrant mobility was present in exon 4 in MTXR-ZR-75 cells (Fig. 1E, lane h, star and arrowhead) compared with parental CEM cells (Fig. 1E, lane g). Sequencing of DNA from CEM/T cells identified a G→A mutation at nucleotide 130 in exon 2, resulting in a G44R mutation in TMD1, as well as a G insertion between nucleotides 1045 and 1046 in exon 4, thus yielding a premature translation termination at amino acid 393 (Table 2). Furthermore, MTXR-ZR-75 cells had a 5-base-pair deletion of nucleotides 1061–1065 that also resulted in a premature translation termination at amino acid 361 (Table 2). These results establish that the genomic PCR-SSCP assay devised for the current study can detect various RFC mutations reliably.
Table 2. Summary of Human Reduced Folate Carrier Mutations Identified by the Single-Strand Conformational Polymorphism Assay in Antifolate-Resistant Tumor Cell Lines and Leukemia Specimens
Screening for RFC Mutations in DNA from 246 Human Leukemia Specimens
We screened DNA from 246 leukemia specimens obtained from patients of diverse ethnic variation, including 222 specimens from patients at the time of diagnosis and the remaining specimens at the time of disease recurrence. This cohort comprised of 203 specimens of B-precursor ALL (82.5%), 32 specimens of T-lineage ALL (13%), and 11 specimens of AML (4.5%). Of 246 DNA samples, only 3 diagnosis B-precursor ALL specimens (1.2%) harbored alterations in the RFC gene. One specimen (Patient M1) from a patient with B-precursor ALL at the time of diagnosis (i.e., prior to chemotherapy) involved a G→C shift at nucleotide 167 in exon 2 (Fig. 2A,B, arrowhead and star, respectively), resulting in a heterozygous D56H substitution predicted to reside in the first extracellular loop (Table 2). Furthermore, this same RFC allele contained an A nucleotide at position 80 and, thus, was homozygous for His27 (Table 2). The second mutation, which also was identified in a patient who was diagnosed with B-precursor ALL (Patient J7) involved a G→A shift at nucleotide 1564 in exon 6 (Fig. 2C, arrowhead), resulting in a heterozygous D522N mutation in the C terminus (Table 2). The third mutation (patient M4) was a definitive alteration in exon 3 (Fig. 2D, arrowhead) that we were unable to characterize due to the lack of DNA availability.
The Impact of RFC Mutations from Patients with Leukemia on MTX Resistance and Folate Growth Requirement
To study the effect of the D56H mutation identified in the diagnosis of Patient M1 with B-ALL, as well as the R27/H27 polymorphism, on MTX resistance and folate growth requirement, human RFC (hRFC) cDNAs harboring either or both of these genetic alterations were transfected stably into transport-null mouse leukemia MTXrA cells.11, 12 Only clonal transfectants that expressed similar hRFC mRNA levels were studied. The loss of folate transport activity in the recipient mouse MTXrA cells increased their folic acid and leucovorin growth requirements by 22-fold and 62-fold, respectively, relative to wild type RFC (i.e., R27) (Table 3). MTXrA cells transfected with the H27/H56 mutant present in Patient M1 retained only 1.7-fold resistance to MTX relative to transfectant cells that expressed native RFC (R27) (Table 3). In contrast, transfectant cells that expressed the R27/H56 RFC resumed complete sensitivity to MTX (Table 3). It is most noteworthy that H27/H56 transfectant cells displayed a 2–3-fold decrease in both folic acid and leucovorin growth requirements relative to the wild type RFC cDNA transfectant (Table 3). In contrast, transfection of either the polymorphic H27 or the mutant R27/H56 RFC cDNAs resulted in the complete resumption of MTX sensitivity and parental folate growth requirement (Table 3). These results suggest that the mutant H27/H56 RFC identified in Patient M1 with B-ALL imparts a low (but statistically significant) level of resistance to MTX but, more noteworthy, confers a prominent decrease in the folate growth requirement (Table 3). Consistently, transfection of an E45K mutant hRFC displaying a 30-fold increase in the transport affinity for folic acid14 also resulted in an ≈ 5-fold decrease in the folic acid growth requirement accompanied by low-level resistance to MTX (Table 3).
Table 3. Methotrexate (MTX) Growth Inhibition and Folate Growth Requirement in Mouse MTXrA Leukemia Transfectants Harboring Either the Native Reduced Folate Carrier(RFC) (Arg27) or the Altered RFC cDNA
MTX concentration (nM) that inhibits cell growth by 50% after 72 hours of drug exposure (IC50).
Numbers in parentheses represent the fold resistance to MTX relative to native reduced folate carrier (i.e., transfectant/Arg27).
The leucovorin or folic acid concentration (nM) necessary to support 50% of maximal growth of control cells (EC50). Results are the mean ± standard deviation of five or six independent experiments.
Numbers in parentheses denote the fold change in folate growth requirement relative to native reduced folate carrier (i.e., transfectant/Arg27).
[3H]MTX Transport in Patient Lymphoblasts Lacking or Harboring RFC Mutations
The availability of sufficient blast cells from Patient J7 and from various patients who lacked RFC mutations allowed the direct measurement of [3H]MTX transport (Fig. 3). [3H]MTX transport in lymphoblasts that were isolated from 15 patients with B-ALL at the time of diagnosis varied over an ≈ 10-fold range, with an average of 0.051 pmol/107 cells per minute (Fig. 3). It is noteworthy that lymphoblasts from 3 patients with B-ALL, including Patient J7 (who harbored the D522N mutation) and Patients J6 and J7 (who lacked any RFC mutations), had [3H]MTX uptake that was 65–84% lower than the average (Fig. 3). Hence, decreased [3H]MTX uptake could be detected in leukemia specimens whether or not RFC mutations were present.
Polymorphisms in the RFC Gene
Recent studies have identified a high frequency His/Arg27 polymorphism in the RFC.34, 35 Therefore, we used the current SCCP assay to screen the entire RFC coding region for possible polymorphisms using 30 DNA samples from healthy individuals. Consistent with these previous studies, three polymorphic groups at nucleotide position 80 were found. One was homozygous for G80, thus encoding for R27; the second was homozygous for A80, thus displaying only the H27 RFC species; and the third group was heterozygous, thereby containing both the G80 and A80 alleles (G/A), resulting in the expression of two RFC species, one with R27 and another with H27 (Table 4). We also identified two novel, albeit silent, polymorphisms, including Pro232 and Ala324, resulting from C→T, and G→A variations at nucleotides 696 and 972, respectively (Fig. 4). At nucleotide 696, 85% of the population was comprised of the homozygous C form, 14.6% was comprised of the heterozygous C/T form, and the homozygous T form was extremely rare (0.4%) (Fig. 4A) (Table 4). Regarding the polymorphism at nucleotide 972, 98% of the population was the homozygous G form, whereas the heterozygous G/A form was rare and was found in only 2% of the population (Fig. 4B) (Table 4).
Table 4. Summary of Human Reduced Folate Carrier Polymorphisms and their Relative Frequencies
Nucleotide numbering starts at the first human reduced folate carrier codon.
To our knowledge the current study is the first documentation of RFC mutations in specimens from pediatric leukemias. Two definitive, single-amino-acid substitutions were identified in the RFC gene at the time of diagnosis in patients with B-precursor ALL. Blast cells from Patient M1 harbored an H27/H56 mutant RFC that, upon transfection into transport-null cells, appeared to confer low-level resistance to MTX. It is noteworthy that this H27/H56 mutant RFC imparted a prominent decrease in both folic acid and leucovorin growth requirements. These results are in accord with a recent suggestion that the ratio of MTX/5-CH3-THF uptake may be more valuable in predicting MTX resistance in leukemia patients; it was suggested that the poorer MTX transport, and/or the higher natural folate cofactor uptake is, the more MTX resistance is observed in leukemia patients.36 Another patient who was diagnosed with B-precursor ALL (Patient J7) harbored a C-terminal D522N mutation in the RFC, and freshly isolated lymphoblasts from this patient displayed [3H]MTX transport that was 60% lower than the average transport observed in multiple leukemia patients who did not have any RFC mutation. However, other patients, including Patients J6 and J11, who did not harbor any RFC mutations also had [3H]MTX uptake that was substantially lower than the average. Hence, decreased [3H]MTX uptake could be detected whether or not RFC mutations were present.
Previous studies have shown that lymphoblasts from children with recurrent B-precursor ALL are more MTX resistant compared with their diagnosis counterparts.37 Furthermore, several reports have demonstrated defective antifolate transport as a mechanism of MTX resistance in childhood ALL.26–30 Thus, because our results clearly show that RFC mutations are not the frequent mechanism of MTX resistance in childhood leukemia, several potential mechanisms may exist. First, defective MTX transport may result from RFC gene silencing due to alterations in transcription factor expression and function23, 24 as well as promoter methylation.38 Second, decreased formation of MTX polyglutamates due to diminished activity of the enzyme folylpolyglutamate synthetase (FPGS) has been documented in leukemia.37 Because polyglutamylation is the primary mechanism responsible for cellular retention of MTX, decreased activity of FPGS may contribute to antifolate resistance. Third, a high activity of the MTX polyglutamate breakdown enzyme folylpolyglutamate hydrolase (FPGH) also was reported in leukemia and, thus, was associated with decreased retention of MTX polyglutamates.37 However, recent studies have shown that transfection and consequent overexpression of FPGH in breast carcinoma and fibrosarcoma cells did not result in MTX resistance.39 Fourth, elevated mRNA levels of the key folate dependent enzymes dihydrofolate reductase and thymidylate synthase also were correlated with MTX resistance in leukemia.40 These results suggest that resistance to MTX in leukemia may be multifactorial and, thus, may require the simultaneous future examination of the qualitative and quantitative alterations in various proteins and enzymes involved in folate and antifolate transport and intracellular metabolism.
Recent studies have demonstrated both the occurrence of RFC mutations41 and the decreased expression of RFC protein in specimens from osteosarcoma patients.42 One of the frequent mutations that results in a serine-to-asparagine substitution at amino acid 46 was found in 9.2% of the osteosarcoma samples. This amino acid substitution was characterized previously in an MTX-resistant leukemia cell line and was found to inflict a marked loss of MTX transport while retaining sufficient folate uptake to support cellular proliferation.43 In contrast, in the current study, we found an extremely low frequency of RFC mutations in multiple leukemia specimens. Our current results are in excellent agreement with a recent preliminary report on leukemia and lymphoblastic lymphoma specimens.44 In the latter study, although only exon 2 of the RFC gene was examined in 40 patients with leukemia and lymphoblastic lymphoma, sequence variations associated only with the G/A polymorphism at nucleotide 80 were found. Thus, the consistent finding of the low occurrence of RFC mutations in leukemia specimens differs completely from the frequent inactivating RFC mutations observed in human13–17 and murine leukemia cell lines12, 18–22 after acquisition of in vitro resistance to MTX and other antifolates. One likely explanation for this discrepancy may be associated with the growth of the various tumor cell lines in culture media that contained nonphysiologic concentrations (2.3 μM) of a nonnaturally occurring, oxidized folate (i.e., folic acid); whereas 5-CH3-THF, the predominant reduced folate in the circulation, is present at much lower concentrations of ≈ 20–30 nM.1 One potentially important implication from the current clinical research is that future studies with antifolate-resistant model tumor cell lines may need to be established and maintained in medium containing natural reduced folates (e.g., 5-CH3THF or 5-formyl-tetrahydrofolate) at physiologic concentrations.
The authors thank Dr. G. Pals (Department of Anthropogenetics, Free University Hospital, Amsterdam, the Netherlands) for his assistance with the ExelGel analysis. They also thank Drs. J.W. Taub and Y. Ravindranath (Children's Hospital of Michigan, Detroit, MI) for providing some acute lymphoblastic leukemia specimens and Dr. S. Shah for isolating DNA from patient blasts for these studies.