Acute lymphoblastic leukemia represents 80 % of all leukemias in the pediatric population. During maintenance treatment, which is the longest treatment phase, patients receive 6-mercaptopurine daily and methotrexate weekly. The doses are calculated based on body surface area and adjusted accordingly to leukocyte levels and occurrence of severe adverse events. However, there is large interindividual variation in the response to the thus prescribed therapy. To explain these interindividual variations we conducted a study on Slovenian pediatric population undergoing thiopurine therapy, investigating genetic factors which might influence the safety profile and efficacy of thiopurine-based treatment. Candidate genes included in the study were selected from those encoding proteins in 6-mercaptopurine metabolism, methotrexate transmembrane transport, folate, and methionine cycles. The course of maintenance therapy was evaluated by calculating the ratio between administered and calculated dose of 6-mercaptopurine, both daily and cumulatively through the course of whole maintenance treatment. Variants in ABCC2 (rs717620) and MTHFD1 (rs2236225, dominant model) were significantly associated with higher doses than calculated, both on daily and cumulative basis (before Bonferroni correction was applied). ABCC2 (rs717620) remained statistically significant after we categorized the cumulative ratio to three distinct groups based on clinical utility of these values (poor metabolizer phenotype bellow 0,9; fast metabolizer phenotype above 1,1: normal metabolizer phenotype between 0,9 and 1,1). We did not detect the most common clinically relevant TPMT variants (TPMT*2, TPMT*3A, TPMT*3C). Using next-generation sequencing, we did however find a synonymous variant that had previously been associated with increased enzymatic activity in the pharmacogenomic PharmGKB database (rs17839843). Additionally, we described 4 intronic variants in the TPMT gene that had previously not been reported in the PharmGKB database and were annotated as deleterious by at least 2 prediction models, included in the bioinformatic prediction tool PredictSNP (rs114611199, rs146407521, rs2518469, rs1024721139). To our knowledge, we were the first to genotype samples from a Slovenian population for the presence of variants in exons 1 and 3 of NUDT15. We detected 4 variants (rs45465203, rs116855232, rs61746486, rs116855232), among which intronic variant rs45465203 and missense variant rs116855232 appear to have the biggest potential to influence enzymatic activity. The frequencies of variant alleles rs61746486 and rs116855232 are 10-times higher than those reported in the literature for the European population, suggesting the need for further research on a larger Slovenian sample. Among non-genetic factors, the categorization into risk groups is associated with the course of maintenance thiopurine therapy, with the standard and medium-risk groups receiving doses closer to the predicted ones compared to the high-risk group. From data of 15 patients, included in TPMT sequencing analysis, we developed a preliminary model for the prediction of ratio between administered and predicted cumulative doses of 6-mercaptopurine during maintenance therapy using AIC model selection method. It includes the following variables: stratification to intermediate and high risk groups, MTHFD1 genotype (dominant model) and the adjusted sum of TPMT variant alleles, which, in spite of its limitations, confirms most of study conclusions. Observations made in this thesis can serve as guidelines for future research on a larger sample.
|
