Abnormal epigenetic regulation has been implicated in oncogenesis. Mutations in the DNA methyltransferase 3A (DNMT3A) gene were recently demonstrated in acute myeloid leukemia (AML) as a candidate for the initiating of lesions in AML with adverse clinical outcome.
Using direct sequencing, identification of the DNMT3A mutations was done in 320 AML patients. Additionally, we analyzed NPM1, FLT3-ITD, FLT3-D835, IDH1 and IDH2 mutations by PCR and DNA sequencing. The retrospective analysis of diagnostic bone marrow samples was performed in AML patients treated in our institution. In all AML patients double induction therapy containing of “7+3” therapy followed by 4 cycles of high dose AraC consolidation was implicated in accordance with therapeutic standards of our institution. We identified DNMT3A mutations in 22% of AML patients. The most common of mutations affect amino acid R822 in exon 23. The DNMT3A mutation was highly enriched in the group of patients with intermediate-risk profile (35%, P<0.01). Unlikely FLT3, DNMT3A mutations were absent in all AML patients with favorable-risk group (P<0.001). AML patients with DNMT3A mutations were older (P=0.05), had higher WBC and platelet counts (P=0.02 for both comparison) and higher relapse rate (P=0.01). The median overall survival among AML patients with DNMT3A mutations was significantly shorter than among patients without such mutation (13.3 months versus 31.3 months, P<0.01). Occurrence of DNMT3A mutations was associated with presence of other common mutations, such as NPM1, FLT3-ITD, FLT3-D835, IDH1 and IDH2. Correlations of the DNMT3A mutations with NPM1 and FLT3-ITD mutations were significant (P<0.001).
Our results indicate that DNMT3A mutations are highly recurrent in AML patients with intermediate-risk profile. DNMT3A mutations are highly correlated with NPM1/FLT3-ITD mutations and are associated with an unfavorable prognosis. The discovery of recurrent mutations in DNMT3A gene may provide a new prognostic marker for the risk stratification for AML patients.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.