Identification of Differential Methylation Markers Among Cytogenetic Risk Groups of Acute Myeloid Leukemia

Abstract

Abstract 1464Aberrant DNA methylation has been shown as an important mechanism in the progression from myelodysplasia (MDS) to acute myeloid leukemia (AML), leading to use of the demethylating agents, 5-azacitidine and decitabine, for treatment of both disorders. While these drugs produce responses, the ability to distinguish potential responders from potential non-responders remains limited. The purpose of this work was to bring new technology to study this problem. Recent studies demonstrated that promoter DNA methylation is not randomly distributed in AML blasts but rather is highly organized and associated with biologically distinct AML subtypes. Because cytogenetics at presentation is the most important prognostic factors in predicting response to therapy, remission duration and overall survival in AML, we aimed to identify differentially methylated genomic regions (DMR) in cytogenetically defined risk groups of AML. Published literature suggested that the new comprehensive high-throughput array-based relative methylation analysis (CHARM) has the highest sensitivity and specificity among all the array-based genome profiling methods and should be the most accurate means to identify methylation markers. It is a customized NimbleGen HD2 array of tiled 50mer-probes typically separated by 30–40 bases covering approximately 4.6 million CpG sites across the genome. This assay is highly quantitative for approximately 100,000 independent CpG sites.We performed methylation profiling with CHARM on 15 age-matched patients divided into 3 groups (n=5 in each): (1) high-risk AML, defined as patients with a complex or monosomal karyotype, inv(3)/t(3;3), t(6;9), or FLT3-ITD; (2) intermediate-risk AML, defined as patients with normal karyotype, trisomy 8, t(9;11), or others; (3) low-risk AML, defined as patients with t(8;21) or inv(16). Five age-matched normal individuals served as control. Randomly fractionated DNA was divided into two equal portions with and without McrBC treatment, which cleaves methylated DNA, then size-fractionated, purified and subject to whole-genome amplification prior to hybridization with the CHARM array. Data analysis was performed with R and Bioconductor. AML patients showed a very strong hypermethylation signature as compared with the normal control blood. A unique set of DMRs was identified which distinguishes between any two risk groups. The number of DMRs and those with p -values < 0.01 are shown in Table 1. There were fewer methylation discriminators between low- and mid-risk groups than between high-risk and the other risk groups. Figure 1 shows the comparison between high-risk group and other risk groups highlighting the 12 top DMRs with lowest p -values. In silico validation of the DMRs identified by CHARM verified previously reported aberrant hypermethylation of tumor suppressor genes like p15CDKN2B, discriminating normal from AML patients, CDH1 promoter hypermethylation in high-risk AML compared with mid-risk AML, and HOXB3 hypomethylation in mid-risk AML compared with both low-risk and high-risk AML, etc. Technical validation using quantitative bisulfite pyrosequencing on 8 genes, including DCC, DUOX2, NEFL, and PITX1, demonstrated an 87.5% concordant rate with CHARM. Testing of additional AML samples with validated markers are underway to confirm the top DMRs identified, which may serve as useful biomarkers to predict response to azacitidine and decitabine.Table 1Number of DMRsNumber with p-value < 0.01Normal blood VS at risk3768311Low-risk VS mid-risk156530Low-risk VS high-risk247573Mid-risk VS high-risk2651107 [Display omitted] Disclosures:No relevant conflicts of interest to declare.