Genome Wide DNA Methylation Analysis Reveals Significant Differences Between Blasts of Patients with Acute Promyelocytic Leukemia and CD34⁺-Hematopoietic Stem Cells of Healthy Individuals

Acute promyelocytic leukemia (APL) accounts for approximately 5 % of all acute myeloid leukemias. Recent experiments have shown that aberrant DNA-promotor-methylation of tumor suppressor genes may be an important mechanism contributing to leukemogenesis and oncogenic transformation. Figueroa et al. recently identified characteristic methylation profiles in a large cohort of acute myeloid leukemia patients using HpaII tiny fragment enrichment by ligation-mediated PCR and array-based methylation analysis. However, as APL is a rare entity, little is still known about the role of deregulated DNA methylation in APL.

Bone marrow blast cells of 45 APL patients at initial diagnosis were analyzed (median age 47 years, gender: 28 female, 17 male). These samples have previously been characterized by high-density-SNP-array analysis (Nowak D et al., Abstract# 38779, ASH 2011), which allowed a segregation into defined molecular and clinical risk subgroups (samples with deletion of chromosome 7q: n = 6; samples with > 5 copy-number-alterations as detected by SNP array: n = 8; samples with white blood cell count at initial diagnosis > 10000/μl: n = 5; samples of patients who suffered relapse: n = 7; samples of patients with initial white blood cell count < 10000/μl: n=19). Furthermore, CD34⁺-cells from the bone marrow of 10 healthy donors (median age 32 years, gender: 5 female, 5 male) were analyzed as a healthy control group. DNA was extracted using the QIAGEN Allprep Kit® (Qiagen, Hilden, Germany). Genome wide DNA methylation analysis was performed using the HumanMethylation450 BeadChip (Illumina, San Diego, USA). Differential methylation of CpGs between subgroups was defined by a minimum mean methylation difference of 25% as expressed by the beta-value of the array data and statistical significance set at p ≤ 0.05 in a two-sided t-test. Analysis of array data was performed using Genome-Studio Software® (Illumina, San Diego, USA), SAS 9.2./JMP-Genomics 4.2® (SAS Software, Cary, USA) and Microsoft Excel 10.1® (Microsoft Software, Redmond, USA).

Using a detection p-value of p ≤ 0.01 for the beta-value of the array and excluding X-chromosomal CpGs, 472.261 CpGs were evaluable for analysis. Highly significant differential methylation profiles between APL-blasts and CD34⁺-cells of healthy donors could be detected for 18163 CpGs (3.8 %). CpGs differentially hypermethylated in APL versus healthy individual affected the coding regions of interesting candidate genes such as transcriptional regulators microRNA 125b1 (MIR125B1), wilms tumor gene 1 (WT1), WT1 antisense RNA (WIT1), v-ets erythroblastosis virus E26 oncogene homolog 1 (ETS1), chromodomain helicase DNA binding protein 5 (CHD5), deleted in colon cancer (DCC) and homeobox genes like iroquois homeobox 1 (IRX5) and NK6 homeobox 2 (NKX6–2). CpGs in the vicinity of genes involved in regulation of the epigenetic landscape such as jumonji, AT rich interactive domain 2 (JARID2) and DNA (cytosine-5-)-methyltransferase 3 alpha (DNMT3A) were significantly hypomethylated in APL as compared to the control group. The analysis of molecular and clinical subgroups revealed that methylation profiles frequently clustered according to the respective molecular lesions in the analyzed samples.

We present a comprehensive genome wide methylation analysis of different APL patient subgroups defined by molecular aberrations and clinical risk profiles. Thereby we have detected sets of significantly differentially methylated CpGs in all subgroups. Integration of this data with results from high density SNP array analysis and detailed clinical outcome data may allow to elucidate the pathomechanistic contribution of aberrant methylation in APL and possibly establish novel prognostic markers based on differential methylation.

No relevant conflicts of interest to declare.