MTHFR c.665C>T guided fluoropyrimidine therapy in cancer: gender-dependent effect on dose requirements

The fluoropyrimidine derivatives 5-fluorouracil (5-FU) and capecitabine (CAP) are widely used for the treatment of solid tumors, including gastrointestinal, breast, head and neck cancer [1]. Fluoropyrimidine metabolism involves a cascade of different enzymes, including dihydropyrimidine dehydrogenase (DPD) enzyme that catabolizes fluoropyrimidines to inactive metabolites, thymidylate synthase (TS) enzyme that methylates deoxyuridine monophosphate (dUMP) to form thymidine monophosphate (dTMP) and MTHFR enzyme that converts 5,10-Methyltetrahydrofolate (5,10-MTHF) to 5-MTHF. Fluoropyrimidine metabolism finally leads to the generation of the active nucleotide analogues 5-fluoro-2′-deoxyuridine-5′-monophosphate (FdUMP) and 5-fluorouridine-5′-triphosphate (FUTP) that are incorporated into both DNA and RNA, exerting thus the fluoropyrimidine antineoplasmic action. FdUMP forms a ternary complex with 5,10-MTHF and TS; this complex restrains the synthesis of DNA. Imbalance of fluoropyrimidine nucleotide analogues ratio is associated with fluoropyrimidine response and incidence of adverse events (AEs) that can be life-threatening or fatal and represent a major cause of reduced dosage, delayed drug administration and/or therapy discontinuation [1].

Substantial inter-individual variation in fluoropyrimidine-related AEs is partly explained by genetic factors. A detailed graphic representation of the candidate genes involved in fluoropyrimidine pharmacokinetics is described by Thorn et al. [1]. Currently, DPYD is the only pharmacogene applied in fluoropyrimidine clinical setting [2]. MTHFR is also a potential PGx marker for fluoropyrimidine drug response since it reduces the amount of 5,10-MTHF available for binding to FdUMP and TS. Among MTHFR gene polymorphisms, rs1801133 (c.665C>T, historically referred to as C677T, p. Ala222Val), is a common functional polymorphism which promotes the thermolability and diminishes the enzyme activity [3]. Therefore, the MTHFR c.665C>T polymorphism enhances the cytotoxic potential of fluoropyrimidines and may be closely related to dose requirements and the efficacy of fluoropyrimidine treatment.

In the present study, we have analyzed the potential association of MTHFR c.665C>T polymorphism with the clinical response of Greek cancer patients to 5-FU/CAP regiments in terms of 5-FU/CAP induced AEs and 5-FU/CAP dose reduction and delayed drug administration or therapy discontinuation. Patient cohort consists of 313 cancer patients treated with fluoropyrimidines, as described elsewhere in detail [4]. In brief, eligible were patients treated with 5-FU or CAP in monotherapy or in combination with other antineoplastic drugs. Toxicity was recorded by grade according to the common terminology criteria for adverse events v5.0. All patients were evaluated by the same team of oncologists who were responsible for clinical decisions on chemotherapeutic regiment, dosages, timeline of drug administration and therapy discontinuation [4]. In the same cohort, we have previously shown a gender driven association of TYMS-TSER 3R/2R polymorphism with incidence of AEs in female cancer patients [4]. Therefore, in the present study, we have further seeked for potential gene-gender and gene-gene interactions both in pooled sample of fluoropyrimidine treated cancer patients as well as in sub-groups after stratifying patients according to the type of cancer. All patients participated after being informed about the study by their attending clinician and giving written consent. The protocol of the study was approved by the Scientific Council and the Ethics Committee of the Academic General Hospital of Alexandroupolis (Greece) and was conducted according to the Declaration of Helsinki.

Main characteristics of patient cohort in total and stratified by gender are presented on Table 1. In stratification of patients in sub-groups for dose reduction (44 cases) and for having delayed drug administration or therapy discontinuation (59 cases), cases vs. controls were age-matched and no differences were observed in body weight, height or body surface area (BSA). As expected, cases in both dose reduction and delayed drug administration or therapy discontinuation groups had increased frequency of AE, lower dose intensity and greater % dose reduction (p<0.001 for all comparisons). Dose reduction cases had a 30% range of % dose reduction (minimum 20, maximum 50), whereas delayed drug administration or therapy discontinuation cases had a range of 10 cycles until therapy modification (minimum 1, maximum 11).

In patient cohort, the frequencies of MTHFR c.665C>T CC, CT and TT genotypes were 46.0, 45.4 and 8.6%, respectively, whereas the frequency of T allele was 31.3%. Genotype and allele frequencies were in Hardy-Weinberg equilibrium (χ²=0.94, p=0.33).

When comparing frequencies of MTHFR c.665C>T genotypes and alleles between AE cases and AE controls, no differences were observed (p=0.354 and 0.186, respectively) (Table 2). Similarly, no differences were observed when dominant and recessive models of inheritance were applied (p=0.301 and p=0.210, respectively, data not shown). When patients were stratified by gender, gene*gender interaction was not identified (p=0.175 and 0.097, for genotypes and alleles, respectively, in males; p=0.906 and 0.907, for genotypes and alleles, respectively, in females) (Table 2).

MTHFR c.665C>T genotypes and alleles were assessed as for their potential association with fluoropyrimidine dose reduction and delayed drug administration or therapy discontinuation. In pooled sample, when MTHFR c.665C>T genotypes were assessed as for their potential association with fluoropyrimidine dose reduction, no association was found (data not shown). Similarly, no association was present in the pooled sample when either dominant or recessive model of MTHFR c.665C>T genotypes was applied (Dominant model is displayed in Table 3). When patients were stratified by gender, MTHFR c.665T allele carriage (CT and TT genotypes) was associated with dose reduction in female patients (77.8% in female patients with dose reduction vs. 50.4% in female patients without dose reduction, p=0.029) (Table 3). Our study detected a statistically significant difference of 27.4% in the frequency of MTHFR CT+TT genotypes between female patients requiring dose reduction and those without with 60.6% power. No difference was present in male patients (p=0.953 and 0.682, for dominant model and alleles, respectively). In regression analysis adjusted for age of disease onset, BSA and AE incidence, MTHFR CT and TT genotypes were the sole determinants of increased need for fluoropyrimidine dose reduction in female patients (OR 5.050, 95% CI 1.346–18.948, p=0.016). In a similar linear model with percentage of dose reduction [(final dose-initial dose)*100/initial dose] as a dependent variable, AE incidence (β=3.774, 95% C.I. 1.314–6.235, p=0.003) and MTHFR CT and TT genotypes (β=3.318, 95% C.I. 1.056–5.580, p=0.004), were predictors of dose reduction only in female patients.

Based on our previous findings of the association of TYMS-TSER 2R allele carriage with AEs in female patients [4] and considering that AE incidence is a main clinical factor for fluoropyrimidine dose adjustments, TYMS*MTHFR interaction was inserted as independent variable in binary regression model. This gene-gene interaction increased need for fluoropyrimidine dose reduction in female patients (OR 1.816, 95% CI 1.104–2.988, p=0.038 after Bonferroni correction).

No associations were found of MTHFR c.665C>T polymorphism with delayed drug administration or therapy discontinuation both in pooled sample [distinct genotypes (p=0.482), alleles (p=0.265), dominant model (p=0.230), recessive model (p=0.639)] and in gender stratification analysis (p=0.619 and 0.362, for genotypes and alleles, respectively, in males; p=0.782 and 0.544, for genotypes and alleles, respectively, in females) (Table 4).

Dose intensity was also calculated in our patient cohort to assess the potential effect of MTHFR c.665C>T polymorphism in a variable depicting both delayed drug administration and dose reduction. The studied SNP was not associated with dose intensity either in pooled sample or in gender stratification analysis (data not shown).

DPYD polymorphisms are valuable predictors of the response to fluoropyrimidine chemotherapy [2]. Screening of additional DPYD polymorphisms could enhance the prevention of fluoropyrimidine toxicity. Towards this direction, implementation of other genetic variants, such as variants in TYMS and MTHFR, holds promise to improve fluoropyrimidine response in cancer patients. Studies assessing the effect of TYMS gene polymorphisms on CAP-5-FU response have been thoroughly discussed previously [4].

For MTHFR, results of several studies have shown an association of MTHFR polymorphisms with different outcomes of fluoropyrimidine response, mainly focusing on drug toxicity [5], [6], [7] or tumor recurrence and survival [8], [9], [10]. However, results are still inconclusive and are not yet supported from current meta-analyses [11, 12].

In the present study, we have found a gender-dependent association of MTHFR c.665C>T polymorphism with fluoropyrimidine dose reduction, whereas in pooled sample analyses MTHFR c.665T allele was not associated with fluoropyrimidine dose reduction or toxicity. Female carriers of MTHFR c.665T allele had significantly increased need (OR 3.392) for decreased dose requirements. Additionally, the gene*gene interaction of MTHFR c.665C>T and TYMS-TSER was also associated with fluoropyrimidine dose reduction in female cancer patients treated with 5-FU or CAP. A similar gender-driven association was present in a previous study of our team when assessing the association of TYMS-TSER polymorphism with fluoropyrimidine AE incidence in the same cancer patient cohort [4]. We thereby encourage gene*gender interaction analyses in cancer pharmacogenomic studies. Interestingly, we did not find an association of MTHFR c.665T allele with dose intensity of fluoropyrimidines. This is potentially attributed to the direct effect of MTHFR on dose requirements. Additionally, in our patient cohort a cancer heterogeneity is present both in terms of tumor site and fluoropyrimidine-based chemotherapeutic scheme.

MTHFR c.665C>T polymorphism is strongly associated with higher homocysteine levels. Gender dependent effects of MTHFR C677T polymorphism have been previously described in studies assessing the relation between the c.665C>T MTHFR polymorphism and fasting plasma total homocysteine concentrations [13, 14]. Both studies describe that male carriers of MTHFR c.665T allele have increased homocysteine levels. Interestingly, 5-FU activation interferes with folate–homocysteine cycle via TS inhibition [1, 15]. Thus, gender*MTHFR as well as TYMS*MTHFR interactions we have found in female patients in our study may be associated with homocysteine cycle. In the present manuscript, however, this assumption is not based on experimental data and should be therefore tested in different studies.

In conclusion, MTHFR c.665T allele is associated with fluoropyrimidine dose requirements in female cancer patients treated with 5-FU or CAP. We suggest that sub-group analyses are important and provide evidence for sex related pharmacogenomic markers that should be further investigated in larger patient cohorts.