Effect of polymorphisms in folate-related genes on in vitro methotrexate sensitivity in pediatric acute lymphoblastic leukemia

Reduced tetrahydrofolates (THFs) serve as cofactors that carry one-carbon moieties for (1) the synthesis of de novo purines and thymidylate (dTMP) and (2) remethylation of homocysteine to methionine (Figure 1). The antifolate methotrexate (MTX) and its active polyglutamate forms inhibit RNA and DNA synthesis in cells by inhibiting dihydrofolate reductase (DHFR) and thymidylate synthase (TS), respectively, 2 crucial enzymes in pyrimidine synthesis and folate recirculation (Figure 1). MTX is an important chemotherapeutic drug in the treatment of acute lymphoblastic leukemia (ALL) and other malignancies. Recent clinical and in vitro studies indicate that common polymorphisms in genes coding for enzymes involved in folate metabolism affect the sensitivity of patients to folate-based chemotherapeutic drugs such as MTX.^1-10  This may be a direct effect or indirect due to shifts in intracellular folate distribution.^9,11  We investigated whether common polymorphisms in folate-related genes influence (ex vivo) MTX sensitivity.

From newly diagnosed Western European pediatric patients with ALL, peripheral blood or bone marrow mononuclear cells were isolated with Ficoll Isopaque (> 80% leukemic cell enrichment) and MTX sensitivity was previously measured with the in situ thymidylate synthase inhibition assay (TSI).^12,13  In short, leukemia cells were incubated with different concentrations of MTX for 21 hours (TSI_cont) or for 3 hours followed by an 18-hour drug-free period (TSI_short), after which the degree of inhibition of TS was measured using [5-³H]-2′-deoxycytidine as substrate.¹²  TSI data were expressed as the concentration of MTX needed to inhibit 50% of the control TS activity (TSI₅₀). Due to limited cell numbers, of all 157 patients included in this study, the TSI_{50, cont} was measured in 86 patients and the TSI_{50, short} in 120 patients. DNA was obtained from unstained cytospin slides of 157 patients. Informed consent from each patient was obtained according to local procedures in accordance with the Declaration of Helsinki. Genomic DNA was isolated using the QIAmp DNA Mini Blood Kit (Qiagen, Vealo, Netherlands). PCR-RFLP was used to detect polymorphisms in methylenetetrahydrofolate reductase (MTHFR 677C>T, MTHFR 1298A>C),^14,15  methionine synthase (MTR 2756A>G),¹⁶  serine hydroxymethyl transferase (SHMT1 1420C>T),¹⁷  methylenetetrahydrofolate dehydrogenase (MTHFD1 1958G>A),¹⁸  and the reduced folate carrier (RFC 80G>A).¹⁹  Double (2R2R) or triple (3R3R) 28-bp tandem repeats in the promoter region of the thymidylate synthase gene (TS 2R3R) were visualized on agarose gel directly after the PCR reaction.²⁰  The polymorphism in the methionine synthase reductase gene (MTRR 66A>G) was determined with real-time PCR using hybridization probes and Lightcycler Technology (Roche, Mannheim, Germany). Because TSI data were not normally distributed, ln (natural logarithm)-transformed TSI values were used as outcome measure in univariate and multivariate regression analysis.

For each genetic variant, the median (50th percentile) and interquartile range (25th-75th percentile) for the TSI_{50, cont} and the TSI_{50, short} is presented in Tables 1 and 2. Univariate regression analysis revealed MTHFR 1298AC (P = .01), MTRR 66AG (P = .08), and MTRR 66GG (P = .09) as potential determinants of TSI_{50, cont}; MTHFR 1298AC (P = .03), MTRR 66AG (P < .001), and MTRR 66GG (P = .02) were determinants of TSI_{50, short}. Multiple regression analysis identified MTHFR 1298AC (P = .06), MTRR 66AG (P = .08), and SHMT1 1420TT (P = .04) as determinants of TSI_{50, cont}; determinants of TSI_{50, short} were MTHFR 1298AC (P = .01), MTRR 66AG (P = .001), and MTRR 66GG (P = .03).

This is the first report describing the influence on a large population of several genetic polymorphisms in folate-related genes on ex vivo MTX sensitivity. Our results show that patients with the MTRR 66 G-allele and the MTHFR 1298AC variant are less sensitive to MTX in both the TSI_{50, cont} and the TSI_{50, short}. Interestingly, both polymorphisms were also related to hematologic toxicity (manuscript in preparation). The MTHFR 677C>T polymorphism was not related to sensitivity in this study. In fact, published data on the role of MTHFR 677C>T are quite contradictory. Patients carrying the MTHFR 677C>T polymorphism are at a higher risk of developing MTX-related toxicity,^2,4,8  whereas in an earlier small pilot study (n = 6), greater MTX sensitivity was observed in the 2 MTHFR 677TT patients.⁶  The MTHFR 677T1298A haplotype or the MTHFD1 1958A variant, in interaction with the TS 3R repeat, resulted in a decreased event- and disease-free survival in pediatric patients with ALL treated with MTX.¹⁰  Sohn et al⁹  transfected human colon and breast cancer cells with either wild-type or mutant 677T human MTHFR cDNA. Only mutant cells had decreased MTHFR activity, changed intracellular folate distribution, increased TS activity, and decreased chemosensitivity toward MTX. Our study with clinical samples of childhood ALL could not confirm the influence of the MTHFR 677C>T polymorphism on ex vivo MTX sensitivity, although a trend toward decreased sensitivity was observed. Some of the discrepancies might be explained by the different conditions used for the in vitro assays. In our assay, high folate is present in the culture medium, possibly counteracting changed folate redistribution due to polymorphisms in folate genes. The clinical results are possibly explained by changed folate distribution. We have no explanation why only the heterozygous MTHFR 1298AC variant showed decreased MTX sensitivity whereas the homozygous (CC) variant did not.

Also, in other diseases in which MTX is being used, genetic polymorphisms play a role. MTHFR 1298A>C and not MTHFR 677C>T modulated the efficacy of MTX treatment in rheumatoid arthritis patients, whereas MTHFR 677C>T was related to toxicity.^3,5  Both TS 2R/3R^7,21  and RFC 80G>A²¹  polymorphisms influenced efficacy and not toxicity of low-dose MTX therapy in patients with rheumatoid arthritis, whereas both MTHFR polymorphisms showed no effect.⁷ 

Our study is the first to report on the MTRR 66A>G and SHMT1 1420C>T polymorphisms as pharmacogenetic determinants of MTX treatment efficacy. Further studies are warranted to evaluate the effect of these polymorphisms on enzyme activity and on intracellular folate redistribution. However, increased risk of neural tube defect pregnancy outcome and cardiovascular disease have been associated with the MTRR 66A>G polymorphism.²² SHMT1 1420CC individuals show increased homocysteine concentrations and lower folate status²²  and exhibit an increased ALL risk.¹⁷  TS competes with SHMT and MTHFR for methylenetetrahydrofolate.²³ MTHFR 677C>T and MTHFR 1298A>C reduce in vitro enzyme activity.^14,15 MTHFR 677C>T directs methylenetetrahydrofolate toward purine and pyrimidine synthesis at the expense of methyltetrahydrofolate, and increases catalytic activity of TS.^9,11  SHMT mediates the competition between TS and MTHFR for one-carbon units: increased SHMT expression favors DNA synthesis and inhibits homocysteine remethylation and S-adenosylmethionine (SAM) synthesis.²⁴  MTRR maintains adequate levels of methyl-cobalamin, the activated cofactor of MS, and thus maintains MS in its active state. Low MTRR activity prevents MTR-mediated remethylation of homocysteine and, because the MTHFR reaction is irreversible, results in trapping of methyl-tetrahydrofolate with concomitant inhibition of purine and thymidylate biosynthesis.²³  We have no explanation why SHMT1 TT patients only showed increased MTX resistance in the TSI_{50, cont}.

In contrast to the TSI_{50, short}, borderline significances were observed for the TSI_{50, cont} in our study. This may be due to the lower number of patients in which the TSI_{50, cont} was measured (n = 86) compared with the TSI_{50, short} (n = 120) with concomitant lower statistical power. More likely, however, accumulation of long-chain MTX polyglutamates is higher after continuous incubations (TSI_{50, cont}), resulting in an increased inhibition capacity of TS. As a consequence, TSI_{50, cont} is less sensitive to changes in folate redistribution due to polymorphisms in other folate-related enzymes. The TSI correlates with conventional cytotoxic assays in leukemic cell lines and the TSI_{50, short} gives a better reflection of clinical MTX sensitivity than the TSI_{50, cont}.¹² 

In conclusion, we show that the common polymorphisms in the folate-related genes MTHFR, MTRR, and SHMT1 are related to in vitro MTX sensitivity in pediatric patients with ALL. To date, the MTX dose applied is based mostly on the body surface area, while the patient's cellular folate status or pharmacogenetic background²⁵  is not yet routinely taken into account. Identifying predictors of MTX sensitivity may lead to the development of individualized treatment strategies with improved efficacy and reduced toxicity. In some countries, but not The Netherlands, the TPMT polymorphism is used to dose adjust the initial dose of 6-mercaptopurine. We feel that in a similar way polymorphisms in folate-related genes such as MTRR could be used to adjust the initial (not maintenance) MTX dose once it is firmly established that these polymorphisms influence efficacy and/or toxicity. In 500 healthy white subjects, we established high allele frequencies of MTRR 66A>G (55%), MTHFR 1298A>C (31%), and SHMT1 1420C>T (32%). Athough ethnic and geographic variations exist in these polymorphisms, high allele frequencies have also been reported in other ethnic groups and thus, they may significantly contribute to sensitivity and toxicity associated with MTX treatment.