Identification of A Pharmacogenomic Profile Associated with High Sensitivity and Low Toxicity to a Combination of Gemtuzumab Ozogamicin Plus Fludarabine, Cytarabine, Idarubicin Regimen In CD33-Positive Acute Myeloid Leukemia (AML) Patients

Human genome sequence variation, in the form of single nucleotide polymorphisms (SNPs) as well as more complex structural variation (such as insertions, duplications and deletions) in drug-metabolizing enzyme and transporter genes (DMETs), affects each individual's response to anticancer agents and thus the likelihood of experiencing an adverse drug reaction. Therefore, understanding relationships between genetic variation and differential drug responses is a challenge to improve both the safety and efficacy of drugs. Accordingly, in order to investigate a potential association between SNPs and clinical response and toxicity, we comprehensively assessed the allele frequencies of 1931 genetic variations of 225 absorption, distribution, metabolism and excretion genes in 59 AML patients younger than 65 years, using the new Affymetrix drug-metabolizing enzyme and transport (DMET Plus) genotyping platform (Affymetrix, Inc. Santa Clara, CA, USA). All patients were enrolled in a phase III multicenter clinical trial combining low dose of Gemtuzumab-ozogamicin (GO) with FLAI regimen (Fludarabine, Cytarabine, Idarubicin) as Induction chemotherapy i (eudract: 2007–005248-26; ClinicalTrials.gov NCT00909168). The induction regimen (GO-FLAI) included fludarabine (25 mg/sqm) and Ara-C (2 g/sqm) on days 1–5, idarubicin (10 mg/sqm) on days 1, 3, and 5 and GO (3 mg/sqm) on day 6. Hematopoietic stem cell transplant was planned for all high risk AML patients in first complete remission (CR) after consolidation with intermediate doses of Ara-C and idarubicin (ID-AC and IDA). Cytogenetic, multidrug-resistance phenotype, FLT3 and NPM mutation status, WT1 quantitative expression analyses, were performed at diagnosis in all patients. Furthermore, high-resolution SNP array analysis by GeneChip® Human Mapping 250K NspI and Genome-Wide Human SNP 6.0 (Affymetrix) was also performed. DNA processing and genotype identifications were performed according to the Affymetrix DMET platform's instructions. All statistical analyses were performed using the R package 2.11.1. Allele frequencies estimation and principal component analyses (PCAs) were performed using the adegenet library, whereas an ad hoc R function was developed to compute two-side Fisher's exact test p-values for each SNP in order to test their association with clinical response and toxicity.Results of potential interest were limited to those in which the p-value was <0.01. The median SNP call rate was 99.48% (range, from 98.05 – 100%). Two samples were run in duplicate and results with "passed call rate" were compared across all the 1931 polymorphic sites, showing a repeatability of 99.99%. In an initial screening procedure, we tested the association among SNPs and response to the induction cycle (FLAI + Gemtuzumab-Ozogamicin). We compared the genotyping profiling of 50 patients in complete (81%) and partial (5%) remission versus that of patients (14%) with no response finding out that 2 SNPs in the alcohol dehydrogenase enzyme (ADH1A) were highly associated with response to therapy. Since genetic polymorphisms may influence the toxicity of chemotherapy drugs, we stratified SNPs according to liver toxicity and 2 genes (SLC22A5 and ABCC4) were found to be associated with a grade I/II liver toxicity. Moreover, SNPs in SLC7A8 and SULT2B1 genes were found to be strongly associated (p <0.001) with fever to gemtuzumab-ozogamicin and finally SNPs in UGT2B15, CYP2B6 and PPARD (peroxisome proliferator-activated receptor delta) were found to be associated with a higher level of expression of P-glycoprotein (defined as a mean fluorescence index > 6). In conclusion, we performed an exploratory pharmacogenomic study that related SNPs in multidrug enzymes and transporters genes with the efficacy and toxicity of a combination of Gemtuzumab-ozogamicin (GO) with FLAI regimen used in a clinical trial of AML patients younger than 65 years. We identified for the first time a pharmacogenomic panel made up of 1 gene (ADH1A) associated with clinical outcome and 4 genes associated with toxicity. Findings obtained from this exploratory study are being tested in a larger clinical sample sets to confirm the predictive value of the SNP panel. Supported by European LeukemiaNet, AIL, AIRC, FIRB 2006, PRIN 2008, Ateneo RFO grants, Project of integreted program (PIO), Programma di Ricerca Regione – Università 2007 – 2009.