Methylation of eukaryotic DNA is a major mechanism by which epigenetic changes regulate gene expression. Imprinting patterns associated with ontogenetic changes and epigenetic silencing are cell-type specific and can lead to profound functional consequences. Aside from physiologic methylation, altered methylation patterns have been found in malignant cells and during aging. It is likely that specific methylation patterns critical for self-renewal/pluripotence exist in hematopoietic stem cells. Due to technical limitations, most studies focused on a narrow empiric selection of promoters. However, new technologies have been developed to assess the “methylome” of a given cell/tissue. We stipulated that if we applied such a global technology to normal and malignant hematopoietic stem cells, we would discover a physiologic methylation pattern distinct from stem cell-derived hematologic malignancies. Here, genome-wide DNA methylation was determined using ChIP-DSL promoter array technology. This entailed using a 5-methyl-cytosine mAb to immunoprecipitate methylated chromatin fragments, which were annealed to a pool of 40K paired oligo probes corresponding directly to 20,000 human promoter regions, PCR amplified, and used to probe the promoter arrays to identify methylated promoters allowing for analysis of the global methylome. Previously, we validated the results of this technology through analysis of individual promoters by bisulfite sequencing. First, the array method was applied to analysis of highly purified CD34 cells from controls to establish a normal CD34-specific methylation pattern and compared to that of CD34 blasts derived from AML and CD34-expressing AML cell lines (KG1 and TF1) and more differentiated AML cell lines (K562, HL60, U937). Globally, methylation in AML patients (2.1% of all promoters studied) and malignant cell lines (2.5%) was greater than CD34 cells (0.6%). Overall, in comparison to CD34 cells, AML promoters demonstrated 1.4% and 0.9% hyper- and hypomethylation, respectively. Using ≥2.5 fold enrichment as a parameter, we identified 64 (0.33%) promoters concordantly methylated in normal CD34 cells, constituting a cellular ”methylation signature”. The methylomes of AML and CD34 cells showed concordant methylation levels in 26 (0.13%) promoters. Methylation profiles of the AML patients differed greatly; only 24 promoters demonstrated similar enrichment, suggesting that AML epigenotypes are variable, consistent with the heterogeneity of phenotype and clinical features. Examples of these hypermethylated promoters in AML include tumor suppressor genes (ETF1, Loc284948), apoptotic and stress-response genes (MAP2K7, PRKC1), and various loci previously demonstrated to play a role in AML (MAN1A1, PH2k-08, BCL11A, PAK1). The methylome of cell lines was remarkably dissimilar to normal CD34 cells, AML blasts and each other. Of the distinctively hypermethylated promoters in primary AML cells only 54% demonstrated concordant hypermethylation in KG1, 38% in TF1, 66% in U937, 43% in K5262 and 72% in HL60. Primary AML cells, KG1 and TF1 cells shared hypermethylation at 34% of loci. To our knowledge, our studies represent the first comprehensive report of global promoter methylation profiles of CD34 cells. Our results indicate the need for study of primary cells in study of disease-associated methylation patterns and point towards a variety of genes aberrantly silenced upon AML transformation.
Read full abstract