DOT1L As a Therapeutic Target for the Treatment of DNMT3A-Mutant Acute Myeloid Leukemia

Mutations of the DNA methyltransferase, DNMT3A, occur in approximately 20% of adult patients with acute myeloid leukemia (AML), and portend a poor prognosis. The most common of these mutations results in a dominant negative loss of function. Our lab observed that upon conditional inactivation of Dnmt3a in the murine hematopoietic system, Dnmt3a^–/– hematopoietic stem cells (HSCs) expanded dramatically while their differentiation was inhibited, consistent with a pre-leukemic state. The likely mechanism by which Dnmt3a loss contributes to leukemogenesis is altered DNA methylation and the attendant gene expression changes, however our current understanding is incomplete.

In analyses of gene expression data, we observed that murine Dnmt3a^–/– HSCs markedly overexpress the histone 3, lysine 79 (H3K79) methyltransferase, Dot1l. This is of interest given the known functional interplay between DNA methylation and histone modifications. Additionally, DOT1L plays a critical role in leukemia with MLL-rearrangements, lesions that essentially never occur concomitantly with DNMT3A mutations in AML. The mutual exclusion of these lesions combined with the observed overexpression of Dot1l in our murine model, led us to postulate that MLL-rearrangements and DNMT3A mutations are distinct epigenetic aberrations that converge on a common mechanism resulting in dysregulated gene expression, specifically mediated by H3K79 methylation (H3K79me). Therefore, in the pathogenesis of DNMT3A-mutant AML, like in MLL-rearranged leukemia, DOT1L-induced H3K79me may play a central role, and may represent a viable therapeutic target.

Throughout the genome of normal HSCs, expansive regions with low DNA methylation (canyons) exist. These canyons span conserved domains frequently containing transcription factors. In our Dnmt3a^-/- model, canyon borders, particularly flanking genes frequently dysregulated in human leukemia such as HOX genes, are highly prone to DNA methylation loss when Dnmt3a is deleted, resulting in canyon expansion. However, not all canyons expand with Dnmt3a loss. We found a close association between canyon behavior and the associated histone marks, with expanding canyons characterized by a lack of the repressive histone mark, H3K27me. To determine if in Dnmt3a-mediated malignant hematopoiesis, H3K79me also correlates with altered DNA methylation, we performed chromatin immunoprecipitation followed by deep sequencing (ChIP-seq) for H3K79 di-methylation (H3K79me2) and aligned these data with whole genome DNA methylation data. This revealed that H3K79me2 specifically coats canyons that lose methylation with Dnmt3a loss, including the HoxA and HoxB clusters, but is not present at canyons without methylation loss. This strong correlation between H3K79me and DNA hypomethylation with Dnmt3a loss suggests a functional interaction. To examine whether this also occurred in human samples with DNMT3A mutations, we analyzed TCGA data, which confirmed many canyon borders as regions with marked DNA methylation loss. Further, many canyon-associated genes, including HOX genes are significantly changed in human DNMT3A-mutant AML.

To explore the role of H3K79me, and specifically of DOT1L in human DNMT3A-mutant AML, we utilized the DNMT3A-mutant human AML cell lines OCIAML2 and OCIAML3. These cell lines were found to have increased total H3K79me compared to DNMT3A-wild type controls, consistent with the increased Dot1l expression in Dnmt3a^–/– HSCs. We then tested the in vitro efficacy of two selective DOT1L inhibitors, SYC-522 (Anglin. J Med Chem. 2011) and the Epizyme compound, EPZ004777 (Daigle. Cancer Cell. 2011), against DNMT3A-mutant cells. Both compounds led to a dose- and time-dependent inhibition of proliferation and induction of apoptosis in the DNMT3A-mutant cell lines at concentrations comparable to those used for MLL-rearranged cell lines. With treatment, DNMT3A-mutant cells also had evidence of induction of differentiation with increased expression of the mature monocyte marker, CD14. Importantly, oncogenic HOX genes overexpressed in DNMT3A-mutant AML were repressed in a time-dependent fashion with DOT1L inhibitor treatment.

In conclusion, our data suggest that DOT1L may be a novel, immediately actionable therapeutic target for the treatment of DNMT3A-mutant AML.