The folate antagonist methotrexate (MTX) is currently one of the most widely prescribed drugs for the treatment of rheumatoid arthritis (RA) (1, 2). Although MTX is among the best-tolerated disease-modifying antirheumatic drugs, a major drawback of MTX therapy is great interpatient variability in the clinical response and the unpredictable appearance of a large spectrum of side effects that include gastrointestinal disturbances, alopecia, elevation of liver enzyme levels, and bone marrow suppression (3, 4). Several well-controlled clinical trials have demonstrated that MTX decreases functional disability, with a maximum effect observable after 6 months of therapy (2, 3). However, recent findings from a large cohort of patients with RA have surprisingly demonstrated that the time to maximal MTX effects is longer than initially thought, thereby raising the concern that the currently recommended dosage of MTX may be suboptimal (5). A possible explanation proposed by the authors was the lack of efficient monitoring of the therapeutic effect of MTX and the difficulty for the physician to rapidly individualize the dose-maximizing response to MTX.
As already proposed by various investigators (6–9), the field of pharmacogenetics may fulfill part of the need for innovative markers that help predict MTX response, and data support the hypothesis that a C677T polymorphism in methylene tetrahydrofolate reductase may help identify patients with an increased likelihood of MTX-related adverse events (10–13). MTX enters cells through the reduced folate carrier (RFC-1) and is activated by folylpolyglutamate synthetase to MTX polyglutamates (MTXPGs) (14). This γ-linked sequential addition of glutamic acid residues enhances the intracellular retention of MTX and promotes the sustained inhibition of 5-aminoimidazole-4-carboxamide ribonucleotide (AICAR) transformylase (ATIC), the last enzyme in the de novo purine synthesis pathway (15, 16). This inhibition by MTXPGs promotes the accumulation of AICAR ribotide, a potent inhibitor of adenosine deaminase (17). The consequence is the buildup of adenosine, a potent antiinflammatory agent (18–20). Furthermore, MTXPGs are inhibitors of thymidylate synthase (TS) (21), which methylates deoxyuridine monophosphate to produce deoxythymidylate, the unique de novo source of thymidylate in the cell. Inhibition of TS by MTX causes cytotoxicity by deoxythymidine triphosphate pool depletion, leading to thymineless death (22).
It can be hypothesized that genetic polymorphisms in RFC-1, ATIC, and TS may account for part of the large interpatient variability in the therapeutic response to MTX. In fact, a G-to-A transition at position 80 of RFC-1 (G80A) was associated with clinical outcome in patients with acute lymphoblastic leukemia (23). In addition, an increasing number of variable-number 28-bp tandem repeats (TSER*2/*3 [2 or 3 28-bp tandem repeats]) in the 5′-untranslated promoter region of TS was associated with enhanced TS expression and a decreased response to MTX and 5-fluorouracil (24–26). However, to date, no polymorphisms in the de novo purine synthesis pathway (e.g., ATIC) have been associated with the therapeutic response to MTX.
We recently developed an analytical method for quantifying the concentration of MTXPG in red blood cells (RBCs) (27). In the present study, we applied this method to the routine monitoring of MTX therapy in patients with RA and investigated the contribution of 3 common polymorphisms (RFC-1 G80A, ATIC C347G, and TSER*2/*3) to the efficacy of MTX.
- Top of page
- PATIENTS AND METHODS
This report is the first to describe the contribution of MTXPGs and common polymorphisms in RFC-1, ATIC, and TS to the effects of MTX in patients with RA treated with low-dose MTX. Results of recent studies have suggested that the MTX dosage is suboptimal in RA, and that innovative approaches may be required to more rapidly maximize the effects of MTX (5).
The current study was cross-sectional rather than longitudinal but was useful in an initial examination combining the measurement of intracellular MTXPGs and certain MTX-related genetic polymorphisms. Because several investigators have advocated monitoring the therapeutic effect of MTX with measurement of MTXPGs in various diseases including RA (33–37), we cross-sectionally measured RBC MTXPG levels in a population of patients with RA who had been receiving MTX for at least 3 months. Because circulating erythrocytes lack folylpolyglutamate synthetase, MTXPGs in RBCs are representative of polyglutamation in bone marrow progenitors (38, 39) and, therefore, are representative of MTXPG levels in less accessible tissues such as lymphocytes. Our data revealed that increased RBC MTXPG concentrations were associated with an increased response to MTX, and we identified a therapeutic threshold of 60 nmoles/liter RBC MTXPGs to be associated with a 14-fold higher likelihood of a VAS score of ≤2 cm for the physician's assessment of patient's response to MTX (i.e., a good response to MTX). This is consistent with previous findings in the treatment of RA (33, 34), psoriasis (35), and cancer (36, 37), and is consistent with the notion that the quantification of RBC MTXPG can be useful for practicing physicians to achieve rapid, effective dosing of MTX.
There is growing evidence that a part of the large interpatient variability in response to xenobiotics is related to genetic polymorphisms (40). In the present study, we evaluated the contribution of 3 common polymorphisms in the folate (RFC-1 G80A), de novo purine (ATIC C347G), and pyrimidine (TSER*2/*3) synthesis pathways to the effects of MTX therapy. Recent evidence suggests that the G80A polymorphism in RFC-1 is associated with altered folate/antifolate levels and is modestly associated with the risk for neural tube defect (23, 41, 42). Data suggest that individuals carrying the homozygous mutant 80AA genotype tend to have higher plasma folate and MTX levels (23, 43) and higher RBC folate polyglutamate levels compared with those with the wild-type or heterozygous genotype (41). This latter finding is consistent with the observation that individuals with the RFC-1 homozygous mutant genotype (RFC-1 80AA) had a 2-fold higher frequency of MTXPG (>60 nmoles/liter) compared with those with the RFC-1 80GG and RFC-1 80GA genotypes. It is tempting to suggest that these higher levels may have contributed to the lower disease activity and improved scores for patient's assessment of disability (lower M-HAQ score) in individuals with the 80AA genotype, although the polymorphism could also directly impact disease activity through more subtle alteration in folate homeostasis (23, 44).
Investigators have previously demonstrated that inhibition of the de novo purine synthesis pathway is an important component of the mechanism of MTX (15, 17). MTXPGs are inhibitors of ATIC, a bifunctional enzyme that catalyzes the final steps in the de novo purine nucleotide biosynthetic pathway (45). The result is accumulation of AICAR and release of the antiinflammatory agent, adenosine (18–20). In the present study, we investigated the contribution of a threonine-to-serine substitution at position 116 of ATIC (C347G) to the effects of MTX, and our data suggest that patients carrying a homozygous variant genotype (347GG) may have an increased likelihood of response to MTX compared with those carrying a 347CC or 347CG genotype. These data are consistent with the hypothesis that MTX may produce part of its antiinflammatory effects through inhibition of ATIC. Whether the single-nucleotide polymorphisms may alter the enzymatic activity, thereby increasing the intracellular pools for the purine precursor AICAR, is not known.
Previous studies have demonstrated that TS levels increase by 10-fold 48 hours after activation of T lymphocytes (46), and evidence suggests that an increased number of tandem repeats in the TS promoter is associated with increased TS expression and a decreased response to 5-fluorouracil and MTX (24–26). In the present study, individuals homozygous for 2 tandem repeats had lower disease activity and improved response to MTX compared with patients with a third repeat. Therefore, our data suggest that inhibition of the pyrimidine synthesis is part of the mechanism of action of MTX in RA.
Because each of these common polymorphisms exhibits only a marginal phenotype, we calculated a pharmacogenetic index to demonstrate the additive association of these polymorphisms to MTX efficacy. Our data revealed that an increase in the number of variant homozygous genotypes was associated with an increased likelihood of response to MTX. Of course, it can be hypothesized that additional genetic markers in loci associated with MTX polyglutamation (e.g., folylpolyglutamate synthetase, γ-glutamyl hydrolase, multidrug resistance–associated protein, aldehyde oxidase) or MTX efficacy (e.g., dihydrofolate reductase, adenosine receptors) may also contribute to the efficacy of this drug.
Finally, the contribution of the total number of homozygous mutant genotypes to the effect of MTX was evident at low concentrations of MTXPGs, while increased MTXPG concentrations tended to overcome the interpatient variability of the effects of MTX. This latter observation is of interest and can have direct applications in clinical practice, because individuals with no homozygous mutant genotypes and low MTXPG levels may require more aggressive MTX treatment to maximize polyglutamation and achieve efficacy.