Pharmacogenomics of Methotrexate: Moving towards Individualized Therapy

Since its introduction in the late 1940s [1], the antifolate methotrexate (MTX) has become widely used in the treatment of malignancies and autoimmune diseases such as childhood acute lymphoblastic leukemia (ALL) and rheumatoid arthritis (RA) [2,3]. Still, a significant percentage of patients either respond poorly to the therapy or develop serious adverse effects [2]. Lack of insight into the pharmacological background for these clinical variations may lead to unwarranted treatment withdrawal rather than corrective individualized dose increments or reductions, respectively. Recently, diversity in treatment response has been linked to sequence variations (mainly SNPs) in genes involved in MTX absorption, metabolism, excretion, cellular transport, and targets or target pathways such as dihydrofolate reductase, thymidylate synthetase and purine de novo synthesis [2,4,5]. Such linkage has been explored for ALL [6], ovarian cancer [7], osteosarcoma [8], RA [9], psoriasis [10], inflammatory bowel disease [11] and prevention of graft-versus-host disease [12,13]. Unfortunately, the results of clinical pharmacogenetic studies have been diverse and generally inconclusive, not only because of insufficient statistical power, but also, and more importantly, due to an array of pharmacological and clinical confounders.

First, MTX is used in combination with other anticancer agents [4] or, in the treatment of RA, is co-administered with folate [14]. Furthermore, studies of the impact of pharmacogenetic variations are burdened by the wide differences in MTX dosages used (ranging from low dose oral 5–10 mg/m² to high dose i.v. 5.0 mg/m² or more) and routes of administration (oral, intramuscular, intravenous and intrathecal), which complicate comparison of pharmacogenetic results across studies. Thus, for the low oral doses used in RA, SNPs in the reduced folate carrier-1 (RFC-1 also known as SCL19A1) may influence the cellular influx [15], whereas for high-dose MTX used in ALL the passive diffusion of MTX over the cell membrane may become increasingly important, and hence diminish the impact of RFC-1 polymorphisms [16].

Second, the mechanism of action of MTX and thus the clinical importance of genetic variations may differ between diseases. In RA the anti-inflammatory effect of adenosine that accumulates after MTX administration may play a role [17], whereas depletion of nucleotide precursors through inhibition of purine and pyrimidine synthesis may be more important in ALL therapy [18].

Third, in oncology the efficacy of MTX and impact of genetic polymorphisms in MTX pathways may be modified by gene dosage differences in the malignant cells. Thus, the ability to form MTX polyglutamates, which influences MTX sensitivity, can be modified by chromosome 21 copy number [16]. Furthermore, since folate metabolism in itself may influence the risk of disease, the association of SNPs in MTX-related genes with treatment outcomes may not only reflect differential MTX sensitivity, but also differences in the pathogenesis and thus aggressiveness of the disease [19].

Fourth, as for folates, the pharmacology of MTX is complex, and numerous SNPs may influence the treatment response. Accordingly, it is unlikely that any single genetic variation will have more than a weak impact on treatment response. Some commercial pharmacogenetic SNP applications allow low-cost screening of multiple predefined genes (e.g., The DMET platform [Affymetrix, CA, USA] [20]), but they do not cover all genes relevant to MTX. Alternative available or rapidly developing techniques include conventional low-cost multiplex PCR methods, high-throughput customized SNP-chip arrays, or genome-wide sequencing approaches that will allow large-scale genetic profiling of MTX-treated patients. Which genes should then be explored? The most studied genes are those that influence MTX influx (e.g., RFC1 [SCL19A1]) and efflux (e.g., the ATP-binding-cassette proteins), MTX polyglutamation, which increases the affinity to the target enzymes and enhances intracellular retention (folylpolyglutamyl synthetase and hydrolase), some crucial target enzymes such as thymidylate synthetase and purine de novo synthesis genes, and genes that influence the tolerance to depletion of reduced folates (e.g., methylenetetrahydrofolate reductase) [2,4–13]. Although several studies have linked each of these to treatment efficacy and/or toxicity, the associations have often been weak or contradicted by other studies. Thus, generally multilocus risk profiling is more likely than using individual SNPs to predict diversity in treatment responses. However, large-scale studies including all the relevant SNPs and with validation of the identified risk profiles in independent patient populations are still lacking.

Finally, adding pharmacogenetic approaches to dose adjustment should not only be cost effective, but should also be superior to dose adjustment by toxicity [21] or by other clinical or paraclinical measures of disease activities such as inflammatory cytokines in RA, or minimal residual disease or MTX metabolite levels in ALL [22,23]. Adjustment of MTX doses to a targeted degree of myelotoxicity in childhood ALL leads to better cure rates than standardized dosing [24]. Accordingly, MTX pharmacogenetic variants should be regarded as supplementary information rather than a substitution of clinical dose adjustments. Thiopurine methyltransferase (TPMT), which methylates the anticancer agent 6-mercaptopurine, is an example of how pharmacogenetic profiling can add to the individualization of therapy compared with toxicity-based guidelines, in that for childhood ALL patients with TPMT low activity, the cure rate is independent of the degree of obtained myelosuppression, whereas for the TPMT high-activity patients dosing to toxicity significantly improves the cure rate [21].

In summary, the detailed mapping of the metabolism and target pathways of MTX renders it likely that multilocus genotype profiles will lead to clinically applicable individualized drug dosing, but large (optimally multicenter) clinical studies are needed to determine the true advantages of integrating MTX pharmacogenetics in individual patient care.

Financial & competing interests disclosure

This study has received financial support from The Danish Childhood Cancer Foundation, the Novo Nordic Foundation, the Home Secretary Research Grant for Individualized Therapy, and the Danish Research Council for Health and Disease. Kjeld Schmiegelow holds the Childhood Cancer Foundation Research Professorship in Pediatric Oncology. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

No writing assistance was utilized in the production of this manuscript.

Additional information

Funding

This study has received financial support from The Danish Childhood Cancer Foundation, the Novo Nordic Foundation, the Home Secretary Research Grant for Individualized Therapy, and the Danish Research Council for Health and Disease. Kjeld Schmiegelow holds the Childhood Cancer Foundation Research Professorship in Pediatric Oncology. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed. No writing assistance was utilized in the production of this manuscript.