HIGH-RESOLUTION PHARMACOGENETIC PROFILES OF GENES INVOLVED IN DRUG ABSORPTION, DISTRIBUTION, METABOLISM AND ELIMINATION IN ADULT PHILADELPHIA-POSITIVE ACUTE LYMPHOBLASTIC LEUKEMIA PATIENTS
I Iacobucci,1 A Lonetti,1 M Sazzini,2 S Formica,3 A Ferrari,1
C Papayannidis,1 P Garagnani,2 A Boattini,2 A Astolfi,3 S Paolini,1
D Cilloni,4 MC Abbenante,1 V Guadaguolo,1 A Vitale,5 F Pane,6
S Soverini,1 M Vignetti,5 R Foà,5 M Baccarani,1 G Martinelli1
Background. Inter-individual variations in genes encoding drug metabolizing enzymes and transporters have been demonstrated to influence the response to therapy. However so far, how these genetic variations interact to produce specific drug related phenotypes in Philadelphia-positive (Ph+) acute lymphoblastic leukemia (ALL) has not yet been investigated. Aim. In order to investigate potential genetic structure and related pharmacogenetic profiles, around 2000 variants in more than 200 genes involved in drug absorption, distribution, metabolism and elimination were genotyped and studied with a population genetics approach in 45 Ph+ ALL patients. Methods. The Drug Metabolizing Enzymes and Transporters (DMET™, Affymetrix) platform, covering more than 90% of the most biologically relevant drug absorption, distribution, metabolism and excretion (ADME) markers was used for successfully genotyping 1931 variants in Ph+ ALL patients treated with the tyrosine kinase inhibitor Dasatinib. A model-based clustering method for inferring population structure using genotype data was applied by means of the Structure software assuming a model in which there are K populations - each one of them being characterized by a set of allele frequencies at each locus - to which individuals are probabilistically assigned according to their genotypes. Distribution of the genetic variance observed among the identified leukemia sub-groups was investigated with a locus by locus Analysis of the Molecular Variance (AMOVA) by means of the Arlequin 3.01 package, exploiting information on genotypes allelic content and frequencies. Results. Three different sub-groups (G1, G2, G3), made up of 2, 12 and 31 patients respectively, were identified in the examined ALL sample, according to their different patterns of allele frequency. A statistical support for this finding was provided by AMOVA results which pointed out a substantial level of genetic differentiation among G1 and the other two sub-groups (Fst=0.099, P<0.001) and a milder but still remarkably significant differentiation between G2 and G3 (Fst=0.020, P<0.001). Fst values for each genotyped variant were also computed in order to single out those changes which were actually responsible for such statistical significances. As regards the comparison of allele frequencies among G1 and G2-G3, 56 variants, affecting the NAT1, NAT2, CYP1A2, CES2, CYP1A2, CDA, SLC22A1, CYP3A5, CYP2B6, CYP3A43, FMO2, UGT1A1, CYP3A7 and VKORC1 genes, showed very high (>0.3) and significant Fst values; whereas a total of 50 loci, located on the NAT2, VKORC1, CYP4F2, CYP2B6, UGT2B7 and CYP2D6 genes, showed moderate to high (>0.08) significant Fst values in the G2/G3 comparison. Conclusions. Differences of allele frequencies observed among the identified ALL sub-groups prove that an evident genetic structure is detectable in our sample by genotyping loci involved in drug metabolism.
Supported by: Fondazione GIMEMA Onlus, European LeukemiaNet, AIL, AIRC, Fondazione Del Monte di Bologna e Ravenna, FIRB 2006, Ateneo RFO grants, Project of integreted program (PIO), Programma di Ricerca Regione - Università 2007-2009.