Impact of DNMT3A mutations on Clinical Response to the Hypomethylating Agent Decitabine in Older Patients (pts) with Acute Myeloid Leukemia (AML)

Abstract 944

Loss of function mutations of the DNA methyltransferase 3A (DNMT3A) gene have been associated with a negative impact on outcome in AML pts treated with conventional chemotherapy and/or stem cell transplantation. DNA methyltransferases (DNMTs) such as DNMT3A are responsible for catalyzing the addition of methyl groups to CpG dinucleotides. In AML, aberrant DNMT activity plays a role in epigenetic silencing of genes involved in hematopoiesis, and inhibition of DNMT activity by hypomethylating agents such as azacitidine and decitabine has been explored as a way to reverse hypermethylated gene silencing and induce clinical response. However, whether DNMT3A mutations influence response to hypomethylating agents is unknown. We studied the frequency and impact of DNMT3A mutations in a cohort of AML pts similarly treated with low dose decitabine. Genomic DNA from pretreatment bone marrow specimens from 46 pts (age 32–85) with untreated or relapsed AML who received decitabine 20mg/m²/d for 10 days every 28 days alone (n=39) or in combination with escalating doses of bortezomib (1.3mg/m²/d on days 5, 8, 12 & 15; n=7) were evaluated. DNMT3A was studied for mutations in the “hot spot” regions, and the mutational status of NPM1, CEBPA, IDH1/IDH2 and TET2 was assessed using polymerase chain reaction (PCR) and direct sequencing. Mutations within FLT3 were determined by quantitative fluorescence-based PCR capillary electrophoresis. Of the 46 pts, 17 had cytogenetically normal (CN) AML. Ten pts belonged to the Favorable genetic group as defined by the European LeukemiaNet (ELN) reporting system, nine pts fell into the ELN Intermediate-I, 11 were in the Intermediate-II, and 16 in the Adverse. Of the 46 pts, eight (17%) carried a DNMT3A mutation: six had codon R882 missense mutations, one nonsense (c.1729A>T; p.K576X) and one splice-site (c.2322+1G>A) mutation. DNMT3A mutations were distributed across all ELN genetic groups: three in the ELN Favorable, one in the ELN Intermediate-I, one in the ELN Intermediate-II, and three in the ELN Adverse group. Among the other molecular markers, only NPM1 mutations were significantly associated with DNMT3A mutations (P=.004). The complete remission (CR) rate for the entire cohort was 41% (19/46, 95% CI: 27% to 57%). Six of eight DNMT3A-mutated pts (75%) reached CR, compared to 13 of 38 with wild-type DNMT3A (34%; P=.05). All five pts with mutated DNMT3A and mutated NPM1 achieved CR, a significantly better response rate than that observed among the remaining pts (P=.008). The median overall survival of DNMT3A-mutated pts was longer than that of DNMT3A-wild-type pts (16.8 vs 11.0 months), but the difference did not reach statistical significance, likely due to the relatively small number of pts. Our group has shown that high levels of microRNA miR-29b, which targets and downregulates DNMT3A, are significantly associated with response to decitabine. (Blum W et al, PNAS 107:7473) Among the pts achieving CR, DNMT3A mutations were present in the pts with the lowest miR-29b expression levels (Wilcoxon's P=.02 for the comparison of miR-29b expression between CR pts with and without DNMT3A mutations). Indeed, the miR-29b expression levels in these three pts was similar to that of pts who did not achieve CR, thereby supporting an impact of the mutation independent from that of the miR. These initial findings, while based on a small sample size, suggest a potential link between DNMT3A mutations and response to decitabine. If confirmed in a larger cohort, screening for DNMT3A mutation status could be used for risk-adapted stratification of AML pts to decitabine-based regimens.

Supported by NIH/NCI grants NO1 CM62207; UO1 CA76576; K23CA120708 (WB); R01 CA102031 (GM), K12CA133250 (JCB, AW); AW is an ASH-AMFDP Scholar.