A candidate gene approach to predict toxicity and pharmacokinetics of high-dose methotrexate

Abstract

Introduction: The simultaneous analysis of many single nucleotide polymorphisms (SNPs) is now possible through DNA microarrays. The GPCO (Groupe de Pharmacologie Clinique Oncologique) developed a SNP-chip from which it is possible to genotype SNPs of various genes involved in drug response. The wide variability in Methotrexate (MTX) toxicity may depend on variants in pharmacokinetic- and pharmacodynamic-related genes. Analyzing such SNPs simultaneously in patients represents an integrative approach that may help identifying relevant pharmacogenetic targets. Methods: We applied a candidate gene approach to select SNPs that may explain toxicity and clearance of MTX. Selected SNPs referred to genes encoding targets or enzymes of the intracellular metabolism of MTX (MTHFR, MTR, MTRR, SHMT1, ATIC, GART and RFC) and to genes encoding membrane transporters (RFC, MRP2, MRP4, BCRP, OATP1 and OATP3) involved in MTX distribution or elimination. Genotyping was performed on the genomic platform of Bordeaux II University by the custom 384-SNP Illumina(registered trademark) Golden Gate assay. The analysis was conducted in patients suffering from hematologic malignancies, included in a prospective population pharmacokinetic study of high dose MTX, in whom individual MTX clearance was available. Toxicities were recorded and graduated according to standard criteria. We compared the occurrence of severe toxicity among genotypes, by a chi-square test or a Fischer exact test. We compared the mean MTX clearance in each genotype by ANOVA. A Benjamini correction was applied to adjust the threshold of significance for multiple comparisons. Results: Twenty-one SNPs were genotyped in 51 patients. We found a trend for a higher risk to develop severe toxicity in patients homozygotes for the c.1286C minor allele in MTHFR gene (rs1801131) [OR = 8.73 (0.77; 462), P = 0.047 not adjusted, P = 0.494 adjusted). Furthermore, the patients homozygotes for the c.521C minor allele in SLCO1B1/OATP1 gene (rs4149056) had lower MTX clearance than patients carriers of at least a c.521T allele (CL = 6.06 (plus or minus)0.37 vs. 6.92 (plus or minus)2.00 L/h, P = 0.027 not adjusted, P = 0.382 adjusted). Conclusion: SNP microarrays provide rough information on possible links between genes and pharmacological response and could be used as preliminary screening tools. Simultaneous study of a very large number of SNPs requires the inclusion of larger populations to attain sufficient statistical power.