Abstract

DNA methylation changes are a common finding in leukemia, and hypermethylation of CpG island promoters is associated with aberrant gene silencing. Some abnormal cancer related methylation changes have been associated with clinical phenotype including pathologic features, prognosis, and treatment response. However, other DNA methylation changes do not appear to have phenotypic consequences and may reflect a stochastic event or a downstream event of tumorogenesis, such as the CpG island methylator phenotype (CIMP). In order to obtain a better understanding of the DNA methylation changes found in leukemia we analyzed 18 acute promyelocytic leukemia (APL) and 36 chronic myeloid leukemia (CML) patients. We specifically chose to study APL and CML as these leukemia are initiated by specific genetic translocation events, t(15:17) and t(9:22) respectively. To measure the DNA methylation status, we used the GoldenGate Assay for Methylation and BeadArray technology from Illumina, Inc. The Standard Methylation Cancer Panel I from Illumina interrogates 1505 CpG sites, selected from 807 genes (231 genes contain one CpG site per gene, 463 genes contain two CpG sites and 114 genes have three or more CpG sites). In our study we found 142 and 269 genes that were hypermethylated in CML and APL. 31 genes were uniquely hypermethylated in CML, 158 genes were hypermethylated only in APL, and 111 genes were hypermethylated in both leukemias. There was a unique pattern of hypermethylated genes in each cancer; such there was a high concordance of hypermethylated genes within each leukemia type. These data suggest that the epigenetic events were a result of the genetic translocation BCR/ABL or PML/RARĪ± (associated with chromosomal aberrations t(9:22) or t(15:17)) that initiates these leukemias. Analysis of the number of hypermethylated genes in these two leukemias showed a bimodal distribution suggestive of CIMP, however, closer examination showed that this bimodal distribution could be attributed to the two different types of leukemia. APL patients had mean of 280 genes hypermethylated while CML patients only had a mean of 193 genes hypermethylated. APL had a stronger methylator phenotype than CML for the loci studied, which underscores the possible relationship of CIMP to a genetic phenotype. Subset analysis of our CML samples by chronic phase (23 patients), accelerated phase (5 patients), and blast crisis (8 patients) revealed 42 genes that became hypermethylated with progression of CML. It is possible that hypermethylation of these genes are clinically important in the leukemia phenotype, and maybe targets for epigenetic therapy. We examined the DNA methylation changes induced by the DNA methylation inhibitor, azacitidine, in a patient with blast crisis CML and refractory to imatinib mesylate therapy. Azacitidine could reverse the aberrant hypermethylation associated with progression of CML to blast crisis and supports the use of this drug as an epigenetic therapy. Our data show that the majority of DNA hypermethylation events in leukemia are dependent on genetic events, but there is a subset of DNA hypermethylation events that are involved in the progression of leukemia and may be therapeutically reversed by DNA methylation inhibitors.

Author notes

Disclosure: Consultancy: Pharmion, MGI Phrama. Ownership Interests: TherEpi. Research Funding: Methylgene, Pharmion. Honoraria Information: Pharmion, MGI Pharma. Membership Information: MGI Pharma; Pharmion.