Abstract
Abstract 1375
Introduction:
Acute myeloid leukemia (AML) is a highly heterogeneous disease, with diverse genetic and epigenetic variables determining sensitivity to current standard therapy. A disruption of the normal DNA methylation pattern, which can result in altered gene and microRNA (miRNAs) expression, has been observed in different AML subtypes. Hydroxymethylation of 5-methylcytosine (5-mC) has recently been described as an intermediate key step in the process of DNA demethylation. Nonetheless, the correlation of DNA methylation and hydroxymethylation levels with clinical and biological characteristics and clinical outcome in AML is mostly unknown.
Aim:
To investigate the prognostic impact of overall methylation and hydroxymethylation levels in patients with intermediate-cytogenetic risk AML (IR-AML) and to identify miRNAs correlated with methylation and hydroxymethylation in these patients.
Patients and Methods:
We have analyzed 86 IR-AML patients (median age, 53 [range, 17–74]; 52% males) who received intensive therapy from 1994 to 2009 in a single institution. The level of overall methylation and hydroxymethylation in total DNA was estimated after determining the percentage of 5-mC and hydroxymethylcytosine (5-hmC), using anti-5-mC and anti-5-hmC monoclonal antibodies (MethylFlash Methylated or Hydroxymethylated DNA Quantification Kit, Epigentek). The expression of 670 mature miRNAs was analyzed using TaqMan Human MicroRNA Arrays (Applied Biosystems). The statistical analysis was performed with SPSS version 15.0.1 and R software version 2.9.0. MaxStat package of R was used to determine the optimal cutoffs.
Results:
The univariate analysis for overall methylation showed that patients with lower levels of methylation (cutoff < percentile 75) had shorter overall survival (OS) than those with higher 5-mC levels (5-year OS: 30±6% vs 52±11%; p=0.03) and a trend for shorter leukemia-free survival (LFS)(p=0.06). Overall methylation levels did not show any correlation with clinical features at diagnosis or with gene mutations, including DNMT3A. Concerning hydroxymethylation, patients with lower 5-hmC levels had a worse prognosis than those with higher 5-hmC levels, with a lower complete response rate (79% vs. 96%; p=0.04), shorter OS (5-yr OS: 21± 6% vs. 55± 6%; p=0.008), and shorter LFS (5-yr LFS: 24±8% vs. 52 ±10%; p=0.03). Interestingly, when analyzed as a continuous variable, 5-hmC levels retained their prognostic value as a marker of response rate (T-test, p=0.007), OS (Cox, p=0.015), and LFS (Cox, p=0.041). Moreover, 5-hmC levels were inversely correlated with FLT3-ITD (p=0.001) and the FLT3-ITD/FLT3wild-type ratio (Pearson correlation:-0.6; p=0.01). The multivariate analyses, including the main clinical and biological variables, identified older age, wild-type NPM1, FLT3-ITD, and lower 5-hmC levels (HR=3.072; 95% CI: 1.096–3.917; p=0.025) as independent prognostic markers of shorter OS and wild-type NPM1, FLT3-ITD, and lower 5-hmC levels (HR=2.002, 95% CI: 1.032–3.881, p=0.040) as independent prognostic markers of shorter LFS. Of note, lower 5-hmC levels retained their value as a marker of worse prognosis in the subgroup of IR-AML patients with unfavorable molecular markers (wild-type NPM1 and CEBPA and/or FLT3-ITD; p= 0.037).
Finally, we have identified a 3-miRNA signature (miR-378*, p=0.02; miR-493, p=0.02; and miR-181, p=0.02) associated with global methylation levels, and a 12-miRNA signature associated with hydroxymethylation, including miR-183* (p=0.001), miR-125a-3p (p= 0.01), miR-586 (p=0.02), and miR-513–3p (p=0.02).
Conclusions:
Hydroxymethylation levels appear as an independent prognostic factor in IR-AML and maintain their prognostic value in the subset of patients with unfavorable molecular markers. Moreover, methylation and hydroxymethylation are associated with specific miRNA profiles. Further studies are warranted to confirm the clinical impact of these findings and to clarify the underlying molecular mechanisms.
Disclosures:
No relevant conflicts of interest to declare.
Microsphere-Based Multiplex Analysis of DNA Methylation in Acute Myeloid Leukemia
