Detection of high-frequency and novel DNMT3A mutations in acute myeloid leukemia by high-resolution melting curve analysis.

DNA methyltransferase 3A (DNMT3A) is mutated in a subset of de novo acute myeloid leukemia patients and is associated with poor overall and event-free survival. Because routine Sanger sequencing of the 23 DNMT3A exons is impractical in clinical laboratories, we developed a high-throughput method using high-resolution melting (HRM) analysis, which identifies sequence variants by detecting subtle changes in the melting patterns of mutant DNA in comparison with WT sequences. DNA from 104 acute myeloid leukemia patients was tested for mutations in 12 exons encoding 3 major functional domains of DNMT3A: the PWWP (proline-tryptophan-tryptophan-proline) domain (exons 8 to 10), the ADD (ATM-DNMT3-DNMT3L) zinc finger, and the methyltransferase domains encoded by exons 15 to 23. HRM analysis identified 20 of 104 patient samples as variants, which we confirmed by Sanger sequencing. Codon 882 of exon 23 was mutated at the highest frequency with an occurrence rate of 11.5%. All HRM WT calls were confirmed to be devoid of mutations by Sanger sequencing. We also identified seven novel and previously unreported DNMT3A mutations. Structural modeling showed seven of the eight missense mutations detected in our study increased the free energy, destabilized protein, and altered solvent accessibility, suggesting their loss-of-function nature. These data demonstrate HRM analysis to be a higher throughput, sensitive, and efficient alternative to Sanger sequencing for detecting DNMT3A mutations in the clinical diagnostic laboratory.