A Pharmacokinetic and Pharmacogenetic Analysis of Osteosarcoma Patients Treated With High‐Dose Methotrexate: Data From the OS2006/Sarcoma‐09 Trial

Abstract

Growing evidence suggests that polymorphisms of genes coding for transporters or enzymes may partially explain the large between subject variability reported for methotrexate (MTX) pharmacokinetics (PK). This prospective study aimed to develop a population PK-pharmacogenetic model to evaluate the part of between-subject variability due to single-nucleotide polymorphisms (SNPs) in transporters and enzyme genes implicated in MTX distribution and elimination. MTX concentrations and 54 SNPs (located in ABCB1, ABCC1, ABCC2, ABCC3, ABCC4, ABCG2, SLC19A1, SLCO1B1, and UGT1A1 genes) were analyzed in patients treated with MTX included in the OS2006/sarcoma-09 trial (a multicenter, open-label, phase III trial, ClinicalTrials.gov. Identifier: NCT00470223). PK data were analyzed using the nonlinear mixed-effect modeling software program Monolix. The influence of each SNP was evaluated using a stepwise procedure under additive, recessive, or dominant genetic model. The likelihood ratio test was used to test the effect of each SNP on PK parameters. Overall, 187 patients with 7898 MTX blood concentrations were included in the PK-pharmacogenetic analysis. A 2-compartment model adequately described the data. Although high-dose MTX dosing recommendations in pediatric patients are currently based on body surface area, body weight was more predictive of clearance between-subject variability than body surface area. The most significant polymorphism associated with MTX clearance was rs13120400 (on the ABCG2 gene) under the recessive genetic model (P < .0001). GG genotype carriers for rs13120400 appeared to have a moderate decrease in MTX exposure compared to AA or GA carriers.

Systematic Review of Pharmacogenetic Factors That Influence High-Dose Methotrexate Pharmacokinetics in Pediatric Malignancies

“…The variant A allele was significantly associated with decreased MTX clearance in two studies [ 20 , 26 ], while the same directionality was observed in another study, without reaching significance [ 25 ], and no association was identified in the other three studies [ 23 , 24 , 27 ]. In addition, three SNPs (rs12505410, rs13120400, and rs13137622) were significantly associated with MTX PK in a single study [ 22 ]. Nevertheless, since this is the only article in which they were studied, additional studies would be needed to confirm those associations.…”

Analysis of Drug Metabolizing Gene Panel in Osteosarcoma Patients Identifies Association Between Variants in SULT1E1, CYP2B6 and CYP4F8 and Methotrexate Levels and Toxicities

“…Potential contributors to contrasting findings are differences in outcomes, different populations with different genetic backgrounds, and different HD-MTX treatment regimens (2 – 5 g/m² in hematological malignancies compared to 12 g/m² in osteosarcoma). A prospective candidate gene study investigating the relation of genetic variants and HD-MTX pharmacokinetics in osteosarcoma patients identified statistically significant associations between MTX clearance and variants in ABCG2 and UGT1A ( Lui et al., 2018 ). ABCG2 was not included in this study but the identified variant in UGT1A (rs4148324) was in high LD (r² > 0.90) with three variants in the present study (rs887829, rs111741722, rs10929302), but these variants showed no association with MTX 48 h plasma levels (using both an additive and dominant model, data not shown).…”

Effect of SLCO1B1 Polymorphisms on High‐Dose Methotrexate Clearance in Children and Young Adults With Leukemia and Lymphoblastic Lymphoma

“…We performed population PK analysis with a nonlinear mixed‐effects model implemented in NONMEM version VII 28 using a typical pediatric population PK modeling approach, similar to two recent HD MTX studies 29,30 . We chose the base model by comparing several compartment PK models with either a proportional residual error model or a combined additive and proportional residual error model, assuming lognormal distribution for random effects PK parameters.…”

“…27 Population PK analysis We performed population PK analysis with a nonlinear mixed-effects model implemented in NONMEM version VII 28 using a typical pediatric population PK modeling approach, similar to two recent HD MTX studies. 29,30 We chose the base model by comparing several compartment PK models with either a proportional residual error model or a combined additive and proportional residual error model, assuming lognormal distribution for random effects PK parameters. For estimations, we used the stochastic approximation expectation maximization, which was followed by importance sampling to obtain the object function and standard errors, because stochastic approximation expectation maximization could better handle more complex PK/ pharmacodynamic problems with many parameters than the first order conditional estimation with interaction.…”