Impact of Genetic Variants of RFC1, DHFR and MTHFR in Osteosarcoma Patients Treated with High-Dose Methotrexate

ORIGINAL ARTICLE Impact of genetic variants of RFC1, DHFR and MTHFR in osteosarcoma patients treated with high-dose methotrexate

S Jabeen1,2, L Holmboe3, GIG Alnæs2,4, AM Andersen5, KS Hall6 and VN Kristensen1,2,4

Osteosarcoma patients are commonly treated with high doses of methotrexate (MTX). MTX is an analog of folate, which is essential for DNA synthesis. Genetic polymorphism at single nucleotide can be indicative to the prognostic outcome in patients. Germ-line variants in candidate genes, coding for enzymes active in the metabolism of MTX, were studied in 62 osteosarcoma patients. Patients harboring the GG genotype in reduced folate carrier 1 (RFC1) rs1051266 had signiﬁcantly better survival in comparison with patients having the AA genotype (P = 0.046). These patients also had a lower frequency of metastasis (15%, P = 0.029). Also patients homozygous for the G allele of rs1053129 in the dihydrofolate reductase (DHFR) gene were more likely to have a metastasis (45%, P = 0.005), and the methylenetetetrahydrofolate reductase (MTHFR) 677C allele was associated with higher degree of liver toxicity (88%, P = 0.007). The study suggests that germ-line variants in the MTX metabolic pathway are associated with survival and side effects in patients treated with MTX.

The Pharmacogenomics Journal (2015) 15, 385–390; doi:10.1038/tpj.2015.11; published online 17 March 2015

INTRODUCTION Intracellular MTX is converted to polyanionic polyglutamates Methotrexate (MTX) is a folate analog used in the treatment of (PGMTX), which creates a gradient for continual cellular uptake of 11,12 autoimmune diseases and cancers such as osteosarcoma. Bone folates and folate analogs. Accumulated intracellular MTX and sarcomas such as osteosarcoma constitute approximately 0.2% of PGMTX interact with dihydrofolate reductase (DHFR), the enzyme all cancers. Of the approximately 40 new bone sarcomas that converts dihydrofolate (DHF) to tetrahydrofolate (THF), which 13,14 diagnosed each year in Norway 12–15 will be osteosarcoma.1 is essential for de novo purine synthesis. Osteosarcoma patients show a bimodal age distribution with a Treatment response mainly depends on the route of adminis- first peak during adolescence and a second peak in older tration, dose and duration of treatment. Response may also be adulthood.2 The second incidence peak in the elderly typically influenced by the activity of the proteins related to MTX uptake, appears in previously irradiated areas or is secondary to Paget’s and genetic variation in the genes coding for these proteins may disease.3 The current standard treatment for osteosarcoma is neo- affect their production or function. Therefore, we analyzed single- adjuvant chemotherapy given before surgery with a high-dose nucleotide polymorphisms (SNPs) and repeat polymorphisms in methotrexate (HDMTX) and subsequent leucovorin rescue, dox- key metabolic genes of the MTX pathway: RFC1, the major influx orubicin and/or cisplatin followed by surgical resection. A major pump; DHFR, the main target of MTX and PGMTX; methylene- challenge of HDMTX therapy is the great interpatient variability in tetetrahydrofolate reductase (MTHFR), en enzyme from the folate clinical response. Bielack et al.4 observed that over 40% of the metabolism pathway and ATP-binding cassette, sub-family G, patients showed poor response to chemotherapy-related treat- member 2 (ABCG2), one of the major MTX efflux pumps. ment, in a study where a good response was defined as less than 10% viable tumor. The wide spectrum of side effects during HDMTX treatment may include gastrointestinal disturbances, MATERIALS AND METHODS alopecia, elevated levels of liver enzymes, bone-marrow suppres- Patient material 5–7 sion and renal toxicity. Serum concentrations of MTX are Serum samples were available for 62 HDMTX-treated osteosarcoma monitored and used to adjust leucovorin rescue to prevent patients with detailed pharmacokinetic and genetic data (see below). toxicity. The patients were treated at The Norwegian Radium Hospital between MTX interacts with a number of enzymes of the folate metabolic 1994 and 2003. pathway. Folate is essential for the synthesis of DNA nucleotides, All treatment protocols in this study were multi-drug regimens most notably thymine, but also the purine bases. MTX enters cells consisting of HDMTX, cisplatin, doxorubicin and ifosfamide, and in some through three possible routes. The major route is via reduced patients, additional etoposide was administered. folate carrier 1 (RFC1).8–10 Reduced or altered expression of the proteins related to MTX uptake, commonly seen in malignant cells, Approvals may be another cause of altered transport kinetics of MTX into the The study protocol was approved by the Regional Ethics Committee. cells.9 Written informed consent was obtained from all participants. Studies on

1Division of Medicine, Department for Clinical Molecular Biology and Lab Science (EpiGen), Akershus University Hospital, Lørenskog, Norway; 2KG Jebsen Center for Breast Cancer Research, Department of Genetics, Institute for Clinical Medicine, Faculty of Medicine, University of Oslo, Oslo, Norway; 3Department of Nuclear Medicine, Oslo University Hospital Ullevål, Oslo, Norway; 4Division of Cancer Surgery and Transplantation, Department of Genetics, Institute for Cancer Research, OUS HF, Radiumhospitalet, Oslo, Norway; 5Department of Pharmacology, Oslo University Hospital, Rikshospitalet, Oslo, Norway and 6Department of Oncology, Oslo University Hospital, Radiumhospitalet, Oslo, Norway. Correspondence: Professor V Kristensen, Clinical Molecular Biology, University of Oslo, P. B. 28, Lorenskog 1478, Norway. E-mail: [email protected] Received 22 August 2014; revised 13 January 2015; accepted 28 January 2015; published online 17 March 2015 SNPs and methotrexate treatment of osteosarcoma S Jabeen et al 386 human participants were approved according to the Declaration of Haploview was also used to estimate Hardy–Weinberg equilibrium for Helsinki (http://www.wma.net/e/policy/b3.htm). the genotypes, as a control for technical genotyping errors and biases in the selection of study populations. MTX treatment Genotype frequencies of SNPs and haplotypes were compared to binary variables with the Pearson’s chi square (χ2) test implemented in Patients were treated with MTX as a 4-h infusion with leucovorin rescue Haploview.17 The resulting associations were permuted 1000 times to started 24 h after initiation of MTX as described by Holmboe et al.15 obtain a measure of significance corrected for multiple testing. The results Evaluation of efficacy and toxicity was assessed by case note review were also verified by Pearson χ2 analysis in SPSS. We used statistical from three independent investigators (KRH, JRFC and DPJ), blinded for packages R version 2.5 (r-project.org) and SPSS (v13. CA, USA), unless genotyping results. otherwise stated. Since in the studied cases the actual underlying mode of inheritance of the causal allele was unknown, the co-dominant model, a DNA isolation single two degrees of freedom test, was shown to have a good overall DNA was isolated from serum samples using the QIAamp® DNA Mini kit performance19 and was applied here. ANOVA was applied to examine (Qiagen, Hilden, Germany) following suppliers specifications. The provided continuous versus categorical variables for clinical metabolic data versus spin-columns were loaded with 200 μl proteinase-treated serum, and DNA SNP frequencies. After Bonferroni correction for six tests (since two pairs of was cleaned with ethanol before eluted in 200 μl DNase-free water. SNPs were in LD), a P-value of 0.008 was considered as significant and presented in the paper. For survival analysis Kaplan–Meier analysis was Genotyping of SNP used to test association between three possible genotypes for each SNP and patient survival times. Log-rank test with P-value under 0.05 was SNPs: RCF1 rs1051266, DHFR rs1053129, DHFR rs1650723, DHFR rs1677666, considered as significant. MTHFR rs1801133, MTHFR rs1801131, ABCG2 rs2231135 and ABCG2 rs2231142 were genotyped on the Nanogen Molecular Biology Workstation 16 platform (Nanogen, San Diego, CA, USA) as described in Nordgard et al. RESULTS The accuracy of genotyping on Nanogen has previously been evaluated, and, with regard to both success and error rate, found comparable to the Polymorphism in RFC1 showed association with survival other platforms assayed, such as TaqMan, SNPstream and MassArray3. The major influx transporter of MTX, RFC1, is encoded by the fi Further, the genotypes obtained with Nanogen were veri ed by solute carrier family 19, member 1. Downregulation of RFC1 has sequencing and a perfect match (100%) to the genotype calls was been reported to be associated with impaired MTX transport and observed. resistance.20–22 We genotyped rs1051266 in exon 2, a well-studied, non-synonymous polymorphism (80G4A) that results in substitu- Metabolic data tion of a histidine for an arginine at residue 27 of the protein MTX and its major extracellular metabolite 7-OH-MTX were measured (Arg27His). A log-rank test for Kaplan–Meier analysis for GG versus in 8 serum samples per HDMTX cycle in 62 consecutive osteosarcoma AA showed that the patients homozygous for the G allele of RFC1 patients. Alanine aminotransferase (ALAT), alkaline phosphatase, gamma- fi glutamyltransferase and total bilirubin were measured in the same serum rs1051266 had good prognosis, measured as disease-speci c samples by methods in routine clinical practice. Details are described by survival (P = 0.046). The analysis was verified by Breslow test with Holmboe et al.15 Unless specified, the phenotype data used are related to P = 0.050, and Tarone-ware test with P = 0.047 (Supplementary the first cycle of MTX treatment. Survival analysis and metastasis were Table 1). On the other hand, the AA genotype at the same locus followed up until 138 months (median 88 months) or death. scored higher frequency of metastasis in osteosarcoma patients (Figures 1a–c) (Table 1). Most patients homozygotes for the A Statistical analysis allele showed poor prognosis. Patients carrying the AG genotype Haplotypes (for DHFR and MDHFR) were inferred and visualized using showed intermediate survival and those who harbored GG Haploview v3.32.17 Haploview estimates haplotypes using the accelerated showed the best outcome (Figures 1a and b). Results from 18 EM algorithm. A haplotype block was defined as by Gabriel et al., and is Pearson correlation analysis show that among the osteosarcoma ’ fi ’ based on the 95% |D | con dence interval. D is a common measure of patients who developed metastasis, 80% of them were homo- linkage disequilibrium (LD), and estimates the likelihood of a recombina- tion event to have occurred between two adjacent loci I a population. D’ zygous for A allele at the rs1051266 in exon 2 of RFC1 SNP. In scored 1.0 for LD between DHFRrs1677666 and DHFRrs1650723, and LD contrast, 30% of G-allele carriers developed metastasis, and only between MTHFR rs1801133 (MTHFR677) and MTHFR rs1801131 15% of the patients homozygous for GG developed metastatic (MTHFR1298). This was accounted for in the multiple testing correction. tumor (P = 0.029) (Table 2) (Figure 1c).

Figure 1. (a) Kaplan–Meier curve for genotypes homozygous for alleles G and A among patients with signiﬁcant difference in disease- dependent survival (P = 0.046). (b) Kaplan–Meier curve for all three genotypes (AA, AC and CC) in reduced folate carrier 1 (RFC1) rs1051266 inﬂuencing the survival among osteosarcoma patients is shown. GG genotype carriers showed a trend in association with best survival (n = 26), AG variants with medium survival (n = 27) and AA variants with poor survival (n = 5) (P = 0.121). (c) Genotype distribution between groups of patients who developed metastasis (blue) compared with metastasis-free patients (orange). For RCFI G80A, 57% of the variant genotype AA, 44% of AG and 15% GG developed metastasis.

The Pharmacogenomics Journal (2015), 385 – 390 © 2015 Macmillan Publishers Limited SNPs and methotrexate treatment of osteosarcoma S Jabeen et al 387 Polymorphism in 3′-untranslated region of DHFR in association frequency (G) allele to be associated with higher degree of with mucositis and metastasis mucositis, although this trend was not statistically significant DHFR, the main substrate of MTX in the cell, converts dihydro- (P = 0.066) (Table 3). folate into tetrahydrofolate. Polymorphisms in DHFR can alter the Variation in another SNP in the 3′-UTR of DHFR, rs1053129, was encoded enzyme’saffinity for folate-analog MTX, which may result associated with metastasis. Pearson correlation analysis showed in varied effectiveness of MTX in tumor cells.23–25 In this study, the that the G allele of rs1053129 was associated with higher rate of T allele rs1650723 in the 3′-untranslated region (UTR) of DHFR, was metastasis (Figure 2b) (Table 2). Among the patients with associated with mucositis (P = 0.005) (Figure 2a). Results from metastatic tumor, 45% carried the GG genotype, and 6% carried ANOVA show that after the first cycle of MTX treatment, 66% of the GT genotype (P = 0.005). Out of 15 T-allele carriers, 14 were found to be in the group with no metastasis, suggesting that this the patients harboring the allele T suffered from higher grade of rare allele might have a protective function against metastasis or mucositis (Figure 2a). The total frequency of the T allele is rather fi may contribute to better response to the treatment. One patient low and larger studies will be needed to con rm this dependency carrying the T allele had metastasis at the time of diagnosis. (total genotype frequency distribution = 90% CC, 8% CT and 2% However, the allele frequency of this SNP is also very low, TT). Variant rs1677666 in DHFR showed a similar trend for its lower requiring future validation.

SNPs in MTHFR in association with liver toxicity Table 1. Clinical outcome after 138 months (median = 88 months) Methylenetetetrahydrofolate reductase (NAD(P)H) (MTHFR) cata- lyzes the 5,10-methylenetetrahydrofolate to 5- methyltetrahydro- Gene ID SNP ID Genotype Deceased Survived Total X2 folate, the substrate for conversion of homocysteine to methionine. MTHFR is essential for de novo purine synthesis, a AA vs AG Overall pathway target of MTX. Two variants of this gene have been studied in relation to disease development and prognosis: MTHFR RFC1 rs1051266 GG 4 22 26 0.046 0.121 1298 A4C or Glu429Ala (rs1801131) and MTHFR 677 C4Tor AA 4 3 7 26–29 AG 12 15 27 Ala222Val (rs1801133). Total 20 40 60 The patients with genotype CC at MTHFR 677 locus had signiﬁcantly higher liver toxicity, measured as ALAT score. We Abbreviations: RFC1, reduced folate carrier 1; SNP, single-nucleotide polymorphism. Genotype counts in RFC1rs1051266 variants in relation to divided the severity of liver toxicity (our continuous variable) into survival. more severe (scores 3 and 4 for higher grades) and less-severe groups (ALAT scores 1 and 2), and applied ANOVA to study the distribution of ALAT score between the genotypes at this locus in the patients. Patients harboring the CC genotype for this variant Table 2. Metastasis after 138 months (median = 88 months) comprised 65% of the group with ALAT scores 3 and 4, compared with 39% of the group with ALAT scores 1 and 2. Among CT Gene ID SNP ID Genotype Counts for Counts for no Total X2 carriers, 23% had ALAT scores 3 and 4 (compared with 46% with metastasis metastasis ALAT scores 1 and 2), and among those with the TT genotype, 11% had ALAT scores 3 and 4 (compared with 14% with ALAT DHFR rs1053129 GG 18 12 40 0.005 scores 1 and 2) (P = 0.007) (Figure 2c) (Table 3). TT 0 0 0 GT 1 15 16 An interesting dose-dependent interaction was observed Total 19 37 56 between the genotype in MTHFR 1298C and the increasing RFC1 rs1051266 GG 4 22 26 0.029 number of MTX treatment cycle by ANOVA. While at cycles 1–3 AA 4 3 7 the distribution of all three genotypes was not signiﬁcantly AG 12 15 27 different, with the increasing number of cycles, there was a trend Total 20 40 60 in decrease in number of CC and AC carriers (Supplementary Abbreviations: DHFR, dihydrofolate reductase; RFC1, reduced folate Figure 1A) (Table 3). The number of AA carriers either increased or carrier 1; SNP, single-nucleotide polymorphism. Genotype counts in remained stable with the increasing number of cycles. Among the RFC1rs1051266 variants and DHFRrs 1053129 in relation to metastasis. patients who received most cycles (10 cycles), 92% were carriers of allele A at MTHFR1298 SNP (Supplementary Figure 1A).

Figure 2. Histograms representing distribution of genotypes among the osteosarcoma patients in relation to the methotrexate (MTX) toxicity and metastasis in the correspondent genes. The ‘grade’ represents degree of severity of the symptoms ranging between 0 and 3 for mucositis and 1 to 4 for liver toxicity. Higher grade numbers correspond to increased severity. (a) Dihydrofolate reductase (DHFR) rs1650723 variants in relation with severity of mucositis. (b) DHFR rs1053129 variants on basis of metastatic growth of tumor. (c) Methylenetetetrahydrofolate reductase (MTHFR) 677 relative to liver toxicity. CC variants was associated with elevated alanine aminotransferase (ALAT) score, while CT and TT variants were associated with a decline or no change. SNP, single-nucleotide polymorphism.

Table 3. Toxicity and MTX treatment cycles in relation with genotype frequencies of SNPs in MTX metabolic genes

Gene ABCG2 DHFRa MTHFRb RFC1

SNP ID rs2231135 rs2231142 rs1057129 rs1650723 rs1677666 rs1801133 rs1801131 rs1051266

Genotpe counts TT CT CC CC AC AA GG GT TT CC CT TT CC CG GG CC CT TT AA AC CC GG AG AA 48 9 0 52 4 1 40 16 0 45 4 2 41 6 5 33 21 8 34 17 11 26 27 7 Total cases 56 57 56 51 52 52 62 60

Mucositis CTC score 03150 33002923 1700 510500 2340 310330 3900 411411 X2 0.01 0.356 0.356 0.005 0.066 0.722 0.613 0.296

ALAT score 1 11 4 2 2 092 3 12 5 1 4 10 3 3 X2 0.907 0.583 0.583 0.532 0.110 0.007 0.336 0.561 Abbreviations: ABCG2, ATP-binding cassette, sub-family G, member 2; ALAT, alanine aminotransferase (measure of liver toxicity); CTC, common toxicity criteria; DHFR, dihydrofolate reductase; LD, linkage disequilibrium; MTHFR, methylenetetetrahydrofolate reductase; MTX, methotrexate; RFC1, reduced folate carrier 1; SNP, single-nucleotide polymorphism. Genotype counts of SNPs under study, Chi-square table in relation with liver toxicity, mucositis and MTX treatment cycles. ars1650723 and rs1677666 are in LD. brs1801133 and rs1801131 are in LD. Bonferroni corrected X2 should be 0.008 (0.05/6).

Interestingly, the cases carrying the MTHFR 1298C allele were not Several studies have shown that RFC180A is associated with significantly associated with ALAT score per se in our study poor prognosis and a higher risk of proliferation in certain cancers. (P = 0.33) (Table 3). For instance, Laverdière et al.32 showed that among the acute lymphoblastic leukemia (ALL) children, carriers of the RFC1 80 A Variations in ABCG2 affects MTX efficacy and toxicity variant had significantly worse prognosis compared with 33 ABCG2, also known as breast cancer resistance protein, is an integral patients harboring the G allele. Gregers et al. observed that plasma membrane protein belonging to the ABC super family of ALL childhood patients harboring RFC1 80 AA suffered from higher degree of bone morrow toxicity compared with other membrane transporter. We tested for association between variation 34 in rs2231135, located in the 5′-UTR of ACCG2, and mucositis after variants. Wang et al. concluded that RFC1 80 homozygote AA fi was significantly associated with increased risk of esophageal the rst cycle of MTX treatment. We divided the patients under 35 observation for mucositis into four groups: from no symptoms, cancer in a Chinese population. Ando et al. observed that the mild, medium and severe as 0, 1, 2 and 3 grade, respectively. We level of PGMTX (PG-5) in red blood cells was lower in rheumatoid observed through ANOVA that an increase in genotype frequency arthritis with 80 A allele. The common feature in these studies of CT carriers from lower to higher grade of mucositis. We had nine appears to be that the lower the concentration or retention of PGMTX in tumor cells is, the less the efficacy of the MTX and worse patients with the genotype CT, four of them suffered from severe 36 mucositis. We had 48 cases that were homozygous for TT, 12 of patient response. Leyva- Vãzquez et al. concluded in their paper that ALL patients with RFC1 80 A allelic polymorphism have higher them developed severe toxicity (25%). The fraction of CT genotype 37 carriers of grade 2 and 3 sufferers was much higher (44%) risk of relapse and showed poor outcome. Drozdzik et al. (P = 0.009) (Supplementary Figure 1B). Again, these are effects of observed that among their rheumatoid arthritis patients, the very rare alleles and need to be confirmed in larger studies. probability of remission of the disease symptoms was significantly higher in allele A carriers. Our results show that RFC1 80AA was associated with poor survival and metastasis in the osteosarcoma DISCUSSION patients. Patients with genotype RFC1 80GG showed better Patients with osteosarcoma who are treated with MTX regimes survival, as there were comparably fewer cases that developed have a variety of outcomes. The possible functional effects of metastasis in this study. We conclude that bearing allele G of the studied here SNPs are either impaired uptake of MTX due to rs1051266 polymorphism in RFC1 may contribute to increased polymorphic variation of RFC1, or altered substrate affinity efficacy of MTX treatment and may contribute to improved resulting from varied activity or amounts of the enzyme DHFR, outcome in osteosarcoma patients. shorter MTX intracellular retention time due to variation in Regarding the metabolism of MTX, variation in 3′-UTR of DHFR folypolyglutamate synthase or overexpressed efflux proteins such can have an effect on termination codon, polyadenylation signal, as ABCG2 or ABCC1–5.13,14,22,30 A recent study by Patiño-García translation rate and/or, secondary structure and stability of mRNA et al.31 showed that expression levels of RFC1 and DHFR were and protein, hence altering DHFR’s enzymatic properties.38 Our lower in osteosarcoma cells in patients where the disease had study shows that the T allele at the two regions of 3′-UTR of DHFR progressed. This deregulation of expression may contribute to may impact the MTX metabolic pattern in osteosarcoma patients: decrease efficacy of MTX, thus poor prognosis. In this association DHFR rs1053129 T allele may have a protective role against study, we have applied the co-dominance genetic model as this metastasis of tumor cells, and DHFR rs1650723 T allele may have a model is considered as the most powerful one (over additive, role in inducement of digestive tract toxicity. recessive or dominant) to detect associations when the inheri- Regarding MTHFR variations, D’ Angelo et al.39 observed that tance model is not known.19 patients with ALL who were homozygous for 677 T had a 12-fold

The Pharmacogenomics Journal (2015), 385 – 390 © 2015 Macmillan Publishers Limited SNPs and methotrexate treatment of osteosarcoma S Jabeen et al 389 31 increased risk of developing toxicity. Patiño-García et al. 7 Weinblatt ME, Kaplan H, Germain BF, Block S, Solomon SD, Merriman RC et al. observed that among their osteosarcoma patients population, Methotrexate in rheumatoid arthritis. A five-year prospective multicenter study. 677 T and 1298 T were more frequent in higher grade of Arthritis Rheum 1994; 37: 1492–1498. gastrointestinal and hematologic toxicities. Cáliz et al.40 also 8 Belkov VM, Krynetski EY, Schuetz JD, Yanishevski Y, Masson E, Mathew S et al. Reduced folate carrier expression in acute lymphoblastic leukemia: a mechanism concluded that for rheumatoid arthritis patients that C677T for ploidy but not lineage differences in methotrexate accumulation. Blood 1999; polymorphism harbors suffered increased MTX toxicity. Similar 93: 1643–1650. 41 results were reported by Yang et al. in ALL patients. There might 9 van der Heijden JW, Dijkmans BA, Scheper RJ, Jansen G. Drug Insight: resistance be many reasons for the contradictory findings in this publication to methotrexate and other disease-modifying antirheumatic drugs--from bench where the CC genotype appears to be associated with higher to bedside. Nat Clin Pract Rheumatol 2007; 3:26–34. toxicity: type of tumor, type of therapies and their combinations 10 Zhang L, Taub JW, Williamson M, Wong SC, Hukku B, Pullen J et al. Reduced folate ’ carrier gene expression in childhood acute lymphoblastic leukemia: relationship ratios and/or cohort s type, age, ethnicity, gender or size. 4 – These polymorphisms are of particular interest as they are to immunophenotype and ploidy. Clin Cancer Res 1998; : 2169 2177. 11 Goldman ID, Matherly LH. The cellular pharmacology of methotrexate. Pharmacol reported to change the activity of MTHFR (decrease in case of Ther 1985; 28:77–102. minor allelic form) and the levels of homocysteine (increase in 12 Green MR, Chowdhary S, Lombardi KM, Chalmers LM, Chamberlain M. Clinical 42,43 case of minor allele). MTHFR 1298C may result in lower activity utility and pharmacology of high-dose methotrexate in the treatment of primary of MTHFR thus low de novo purine synthesis, which might result in CNS lymphoma. Expert Rev Neurother 2006; 6:635–652. increased rate of cell death due to more active MTX thus 13 Bertino JR, Goker E, Gorlick R, Li WW, Banerjee D. Resistance mechanisms to increasing the effectiveness of the MTX treatment. The role of methotrexate in tumors. Stem Cells 1996; 14:5–9. these SNPs needs to be studied in further functional details. 14 Serra M, Reverter-Branch G, Maurici D, Benini S, Shen JN, Chano T et al. Analysis of dihydrofolate reductase and reduced folate carrier gene status in relation to methotrexate resistance in osteosarcoma cells. Ann Oncol 2004; 15:151–160. CONCLUSIONS 15 Holmboe L, Andersen AM, Morkrid L, Slordal L, Hall KS. High dose methotrexate chemotherapy: pharmacokinetics, folate and toxicity in osteosarcoma patients. We conclude that osteosarcoma patients carrying the RFC1 Br J Clin Pharmacol 2012; 73:106–114. rs1051266 GG genotype had better clinical outcome compared 16 Nordgard SH, Alnaes GI, Hihn B, Lingjaerde OC, Liestol K, Tsalenko A et al. Pathway to those with the AA genotype. The RFC1 rs1051266 A allele was based analysis of SNPs with relevance to 5-FU therapy: relation to intratumoral more frequent among patients with metastasis. The DHFR mRNA expression and survival. Int J Cancer 2008; 123:577–585. rs1053129 GG genotype was associated with a higher rate of 17 Barrett JC, Fry B, Maller J, Daly MJ. Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics 2005; 21: 263–265. metastasis among the patients. In this study, the MTHFR 677 C 18 Gabriel SB, Schaffner SF, Nguyen H, Moore JM, Roy J, Blumenstiel B et al. allele was significantly associated with increased liver toxicity. The structure of haplotype blocks in the human genome. Science 2002; 296: Osteosarcoma patients homozygous for RFC1 80 A might need 2225–2229. different combinations of chemotherapy compared to patients 19 Lettre G, Lange C, Hirschhorn JN. Genetic model testing and statistical power in with other genotypes. Many factors are involved in prognosis: population-based association studies of quantitative traits. Genet Epidemiol 2007; tumor site and size, primary metastases, response to chemother- 31:358–362. apy and surgical remission.9,44 Our cohort was small, and a larger 20 Gorlick R, Goker E, Trippett T, Steinherz P, Elisseyeff Y, Mazumdar M et al. fi fi Defective transport is a common mechanism of acquired methotrexate resistance study is needed to con rm our ndings. Genotyping from larger in acute lymphocytic leukemia and is associated with decreased reduced folate cohorts might help in the prediction of one of the factors related carrier expression. Blood 1997; 89: 1013–1018. to prognosis. 21 Wong SC, Zhang L, Proefke SA, Hukku B, Matherly LH. Gene amplification and increased expression of the reduced folate carrier in transport elevated K562 cells. Biochem Pharmacol 1998; 55: 1135–1138. CONFLICT OF INTEREST 22 Wong SC, Zhang L, Witt TL, Proefke SA, Bhushan A, Matherly LH. Impaired The authors declare no conflict of interest. membrane transport in methotrexate-resistant CCRF-CEM cells involves early translation termination and increased turnover of a mutant reduced folate carrier. J Biol Chem 1999; 274: 10388–10394. ACKNOWLEDGMENTS 23 Goto Y, Yue L, Yokoi A, Nishimura R, Uehara T, Koizumi S et al. A novel single- nucleotide polymorphism in the 3'-untranslated region of the human dihy- SJ is a PhD scholar from the Norwegian Health Authority grant nr HSE 2014061. This drofolate reductase gene with enhanced expression. Clin Cancer Res 2001; 7: study was financed by a grant from the Norwegian Research Council to VNK (PK01- 1952–1956. 2007-0356). We would like to thank David Quigley for the critical reading of the 24 Walling J. From methotrexate to pemetrexed and beyond. A review of the manuscript and to Silje Nord for the contribution to the statistical analysis and pharmacodynamic and clinical properties of antifolates. Invest New Drugs 2006; presentation. 24:37–77. 25 Matherly LH, Taub JW, Ravindranath Y, Proefke SA, Wong SC, Gimotty P et al. REFERENCES Elevated dihydrofolate reductase and impaired methotrexate transport as ele- ments in methotrexate resistance in childhood acute lymphoblastic leukemia. 1 Aksnes LH, Hall KS, Folleraas G, Stenwig A, Saeteri G et al. Management of high- Blood 1995; 85: 500–509. grade bone sarcomas over two decades: The Norwegian Radium Hospital 26 Etienne MC, Formento JL, Chazal M, Francoual M, Magne N, Formento P et al. experience. Acta Oncol 2006; 45:38–46. Methylenetetrahydrofolate reductase gene polymorphisms and response to 2 Ottaviani G, Jaffe N. The epidemiology of osteosarcoma. Cancer Treat Res 2009; fluorouracil-based treatment in advanced colorectal cancer patients. Pharmaco- 152:3–13. genetics 2004; 14:785–792. 3 McNairn JD, Damron TA, Landas SK, Ambrose JL, Shrimpton AE. Inheritance of 27 Lewis SJ, Ebrahim S, Davey SG. Meta-analysis of MTHFR 677C-4T polymorphism osteosarcoma and Paget's disease of bone: a familial loss of heterozygosity study. and coronary heart disease: does totality of evidence support causal role for J Mol Diagn 2001; 3: 171–177. homocysteine and preventive potential of folate? BMJ 2005; 331: 1053. 4 Bielack SS, Kempf-Bielack B, Delling G, Exner GU, Flege S, Helmke K et al. Prog- 28 Schwahn B, Rozen R. Polymorphisms in the methylenetetrahydrofolate reductase nostic factors in high-grade osteosarcoma of the extremities or trunk: an analysis gene: clinical consequences. Am J Pharmacogenomics 2001; 1: 189–201. of 1,702 patients treated on neoadjuvant cooperative osteosarcoma study group 29 Van der Linden IJ, Afman LA, Heil SG, Blom HJ. Genetic variation in genes protocols. J Clin Oncol 2002; 20:776–790. of folate metabolism and neural-tube defect risk. Proc Nutr Soc 2006; 65: 5 Treon SP, Chabner BA. Concepts in use of high-dose methotrexate therapy. Clin 204–215. Chem 1996; 42: 1322–1329. 30 Hagleitner MM, Coenen MJ, Schrauwen M, Vermeulen SH, de Bont ES et al. 6 Walker AM, Funch D, Dreyer NA, Tolman KG, Kremer JM, Alarcon GS et al. Association of a genetic variant in the ABCC2 gene with high methotrexate Determinants of serious liver disease among patients receiving low-dose meth- plasma concentrations in pediatric malignancies. Journal of Clinical Oncology otrexate for rheumatoid arthritis. Arthritis Rheum 1993; 36:329–335. 2010; 28: 9522.

© 2015 Macmillan Publishers Limited The Pharmacogenomics Journal (2015), 385 – 390 SNPs and methotrexate treatment of osteosarcoma S Jabeen et al 390 31 Patiño-García A, Zalacaín M, Marrodán L, San-Julián M, Sierrasesúmaga L. Meth- 37 Drozdzik M, Rudas T, Pawlik A, Gornik W, Kurzawski M, Herczynska M. Reduced otrexate in pediatric osteosarcoma: response and toxicity in relation to genetic folate carrier-1 80G4A polymorphism affects methotrexate treatment outcome polymorphisms and dihydrofolate reductase and reduced folate carrier 1 in rheumatoid arthritis. Pharmacogenomics J 2007; 7:404–407. expression. J Pediatr 2009; 154: 688–693. 38 Chatterjee S, Pal JK. Role of 5'- and 3'-untranslated regions of mRNAs in human 32 Laverdiere C, Chiasson S, Costea I, Moghrabi A, Krajinovic M. Polymorphism G80A diseases. Biol Cell 2009; 101:251–262. in the reduced folate carrier gene and its relationship to methotrexate plasma 39 D'Angelo V, Ramaglia M, Iannotta A, Crisci S, Indolﬁ P, Francese M et al. Metho- levels and outcome of childhood acute lymphoblastic leukemia. Blood 2002; 100: trexate toxicity and efﬁcacy during the consolidation phase in paediatric acute 3832–3834. lymphoblastic leukaemia and MTHFR polymorphisms as pharmacogenetic 33 Gregers J, Christensen IJ, Dalhoff K, Lausen B, Schroeder H, Rosthoej S et al. determinants. Cancer Chemother Pharmacol 2011; 68: 1339–1346. The association of reduced folate carrier 80G4A polymorphism to outcome in 40 Caliz R, del AJ, Balsa A, Blanco F, Silva L, Sanmarti R et al. The C677T polymorphism childhood acute lymphoblastic leukemia interacts with chromosome 21 in the MTHFR gene is associated with the toxicity of methotrexate in a Spanish copy number. Blood 2010; 115: 4671–4677. rheumatoid arthritis population. Scand J Rheumatol 2012; 41:10–14. 34 Wang L, Chen W, Wang J, Tan Y, Zhou Y, Ding W et al. Reduced folate 41 Yang L, Hu X, Xu L. Impact of methylenetetrahydrofolate reductase (MTHFR) carrier gene G80A polymorphism is associated with an increased risk of polymorphisms on methotrexate-induced toxicities in acute lymphoblastic leu- gastroesophageal cancers in a Chinese population. Eur J Cancer 2006; 42: kemia: a meta-analysis. Tumour Biol 2012; 33:1445–1454. 3206–3211. 42 Frosst P, Blom HJ, Milos R, Goyette P, Sheppard CA, Matthews RG et al. 35 Ando Y, Shimada H, Matsumoto N, Hirota T, Oribe M, Otsuka E et al. Role of A candidate genetic risk factor for vascular disease: a common mutation in methotrexate polyglutamation and reduced folate carrier 1 (RFC1) gene poly- methylenetetrahydrofolate reductase. Nat Genet 1995; 10:111–113. morphisms in clinical assessment indexes. Drug Metab Pharmacokinet 2013; 28: 43 Weisberg I, Tran P, Christensen B, Sibani S, Rozen R. A second genetic poly- 442–445. morphism in methylenetetrahydrofolate reductase (MTHFR) associated with 36 Leyva-Vazquez MA, Organista-Nava J, Gomez-Gomez Y, Contreras-Quiroz A, decreased enzyme activity. Mol Genet Metab 1998; 64:169–172. Flores-Alfaro E, Illades-Aguiar B. Polymorphism G80A in the reduced folate carrier 44 Ingram SS, Seo PH, Martell RE, Clipp EC, Doyle ME, Montana GS et al. Compre- gene and its relationship to survival and risk of relapse in acute lymphoblastic hensive assessment of the elderly cancer patient: the feasibility of self-report leukemia. J Investig Med 2012; 60: 1064–1067. methodology. J Clin Oncol 2002; 20:770–775.

Supplementary Information accompanies the paper on the The Pharmacogenomics Journal website (http://www.nature.com/tpj)