Since the discovery of the gain-of-function mutation JAK2 V617F in 2005, an increasing number of mutations have been described in myeloproliferative neoplasms (MPN) and related myeloid malignancies. 1–4 For instance, use of array-based techniques such as comparative genomic hybridization and single-nucleotide polymorphism (SNP) analysis led to the identification of genes involved in epigenetic regulation such as tet-oncogene family member 2 (TET2), additional sex combs like 1 (ASXL1) and enhancer of zeste 2 (EZH2). 5–7 Inactivating mutations in TET2, ASXL1 and EZH2 are considered to promote myeloid tumorigenesis because of epigenetic modulation of target genes. More recently, a whole-genome sequencing study in acute myeloid leukemia (AML) uncovered recurrent mutations in 22% of AML patients in another epigenetic regulator, the DNA methyltransferase 3A gene DNMT3A. 8 In this study, DNMT3A mutations were associated with poor outcome and, thus, are of clinical relevance. However, exact mechanisms of action of DNMT3A mutations are still unclear because global methylation patterns and 5-methylcytosine content in AML genomes were not significantly altered. Nevertheless, mutations of TET2, ASXL1, EZH2 and DNMT3A occur in a significant number of patients with myeloid malignancies underlining the pathogenic relevance of epigenetic changes. On the basis of these findings we sought to explore the mutation frequency of DNMT3A in a series of well-characterized ‘classic’MPN cases. In our study, 115 chronic-(n= 80) or blast-phase (n= 35) MPN cases were screened for DNMT3A mutations using direct DNA sequencing of all coding exons (2–23): essential thrombocythemia (ET), n= 30; polycythemia vera (PV), n= 30; primary myelofibrosis, n= 16; post-ET MF, n= 2; post-PV MF, n= 2; AML secondary to MPN (sAML), n= 35. In chronic-phase MPN, DNA from peripheral blood granulocytes was analyzed, whereas mononuclear cell DNA from bone marrow (n= 9) or peripheral blood (n= 26) was used in sAML cases. Mutation data on JAK2 V617F, MPL W515L, TET2 (exon 3–11), IDH1/2 (exons 4), ASXL1 (exon 12), CBL (exon 8–9) and EZH2 (exon 2–20) were available in all cases. Technical details on mutation assays are available on request. In addition, all cases were analyzed on Affymetrix (Santa Clara, CA, USA) 250K SNP arrays allowing for genome-wide screening of copy number alterations and uniparental disomies at high resolution. All patients gave written informed consent for molecular analyses. The study was approved by the Ethics Committee of the University Hospital of Ulm. In total, 12 DNMT3A sequence variants were identified in our study, resulting in an overall frequency of 10%(12/115). Sequence variants were most frequently detected in sAML (17%, 6/35) and MF (15%, 3/20), followed by PV (7%, 2/30) and ET (3%, 1/30); of note, MF was secondary to ET or PV in 2/3 cases. All sequence variants were heterozygous, and none of the patients harbored more than one DNMT3A alteration. DNMT3A alterations consisted of eight nucleotide substitutions and four frameshift deletions; two nucleotide substitutions resulted in direct stop codons (E523n and E477n), two represented missense alterations (N501S and W860R) and four affected the amino acid residue R882 recurrently (R882H, n= 3; R882C n= 1).

The alterations E477n (n= 1) and R882H/C (n= 4) were already described by Ley et al. 8 as somatic nonsense and missense mutations, respectively. In our study, somatic origin could be proven in five cases, including two known (E477n and R882H) and three novel (W305fs, R488fs and E523n) variants in …