High-Resolution Genome Wide Copy Number Alteration (CNA) and Loss of Heterozigosity (LOH) Analysis in Chronic Myeloid Leukemia (CML) Shows That High and Intermediate Sokal Risk Pts (Pts) Have Multiple Losses Targeting Genes Involved in DNA Repair.

Abstract

Abstract 3262Poster Board III-1CML pts display a certain degree of clinical heterogeneity that is documented by the varying levels of response to tyrosine kinase inhibitor therapy and is best reflected by the Sokal risk score. Clinical differences must be a sign of some biological heterogeneity the basis of which, however, are still poorly understood. Today many high-throughput assays are available that allow to unravel the complexity of cancer cells in a genome-wide fashion. We have used Human 6.0 SNP Arrays (Affymetrix) to perform high-resolution (<1 kb) karyotyping of DNA samples from 73 newly diagnosed chronic phase CML pts. Median age was 55 years (range, 25–73 years); male to female ratio was 39/34; pts were almost equally distributed by Sokal risk score (low, n=23; intermediate, n= 23 high, n= 27). Of 189 genes known to be implicated in the cellular DNA repair pathways, 135 (71%) were found to map in regions affected by CNAs or copy-neutral LOH (uniparental disomy, UPD) in 44/73 (60%) pts. However, this was markedly more frequent in high and intermediate Sokal risk pts (20/27, 74% and 16/23, 69%, respectively) than in low Sokal risk pts (8/23, 33%), although neither the total number of detected regions of CNAs/UPD per sample nor the QC parameters differ significantly across different risk categories. Regions of CNA involving DNA repair genes ranged from 105 Kb to 1.1 Mb and were either focal lesions involving a part or the whole single gene (17% of cases), or more extensive losses/gains including 2 to 84 genes. Monoallelic deletions were much more frequent than amplifications. Regions of UPD involving DNA repair genes were much larger and ranged from 980 kb to 32 Mb.The pathways and genes most frequently affected by CNAs or UPD are listed in the Table below:Base Excision Repair (BER)MUTYHDNA glycosilase1p34.1loss/upd10 ptsPNKPPolynucleotide kinase19q13.33loss10 ptsNEIL1DNA glycosilase15q24.2loss7 ptsPOLBDNA polymerase beta8p11.21loss/upd6 ptsPCNASliding clamp for DNA polymerases20p12.3loss6 ptsMismatch Repair (MMR)PMS2L5Mut L homolog - mismatch and loop recognition7q11.23loss/upd11 ptsMSH2Mut S homolog - mismatch and loop recognition2p21loss8 ptsPOLD1DNA polymerase delta19q13.33loss8 ptsPOLA2DNA polymerase alpha, subunit 22p16.3loss/upd6 ptsPOLE2DNA polymerase epsilon, subunit 214q21–22loss5 ptsNucleotide Excision Repair (NER)ERCC15' incision subunit of TFIIH complex19q13.32loss11 ptsERCC25' to 3'DNA helicase of TFIIH complex19q13.32loss10 ptsXAB2Transcription-coupled NER factor19p13.2loss/upd9 ptsCDK7Kinase subunit of TFIIH complex5q13.2loss7 ptsRPA4Binds damaged DNA in preincision complexesXp21.33loss/upd7 ptsRPA2Binds damaged DNA in preincision complexes12q24.31loss/upd6 ptsHomologous Recombination (HR)RAD51CHomologous pairing17q23.2loss/upd7 ptsRAD52Accessory factor for recombination12p13.33loss7 ptsXRCC2DNA break and crosslink repair7q36.1loss5 ptsNon-Homologous End Joining (NHEJ)PRKDCDNA-dependent protein kinase, catalytic subunit8q11.21loss/upd5 ptsDCLRE1CArtemis nuclease10p13loss5 ptsREV7DNA polymerase zeta, subunit1p36.22gain5 ptsOther genes involved in DNA replication/repair/modification or chromatin remodelingCHAF1AChromatin assembly factor19p13.3loss15 ptsRECQL5DNA helicase17q25.1loss/upd11 ptsRAD9BPCNA-like DNA damage sensor12q24.11loss/upd10 ptsCHEK2DNA-damage checkpoint22q12.1loss/upd8 ptsRAD17RFC-like DNA damage sensor5q13.2loss6 ptsTREX1DNAase III exonuclease3p12.31loss/upd6 ptsFor some genes (e.g., RAD52), the monoallelic deletion we detected was found to translate into reduced mRNA expression, observation that was also independently confirmed in an additional group of high/intermediate versus low Sokal risk pts. In all the 44 pts, multiple pathways and multiple genes within the same pathway were affected, supporting the hypothesis that the lesions we detected might actually have consequences on DNA integrity despite the known partial functional redundancy of pathways and effectors. For many of the genes identified in this screen, activating or inactivating mutations are known to occur, and together with overexpression or haploinsufficiency, have been linked to a mutator phenotype in several malignant conditions. We are currently investigating whether this may be the case also in CML. Supported by European LeukemiaNet, AIL, AIRC, PRIN, Fondazione del Monte di Bologna e Ravenna. Disclosures:Baccarani:Novartis: Honoraria, Research Funding, Speakers Bureau; Bristol-Myers Squibb: Honoraria, Research Funding, Speakers Bureau.