The Pharmacological Inhibition of KDM1A Displays Preclinical Efficacy in AML and MDS By Inducing Myelomonocytic Differentiation

Background: Histone methylation is one of the major systems of epigenetics and reversibly regulated by lysine (K) specific methyltransferases (KMTs) and demethylases (KDMs). Dysregulation of KMTs such as EZH2 and MLL play a key role in the pathogenesis of acute myeloid leukemia (AML) and myelodysplastic syndromes (MDS) while the precise function of KDMs remains unclear.KDM1A is the first reported histone demethylase, which mainly catalyzes demethylation of mono- and di-methylated lysine 4 of histone 3 (H3K4me1 and me2 respectively). According to recent reports, the inhibition of KDM1A either alone or in combination of all-trans retinoic acid is effective for AML expressing MLL-AF9 or for several types of AML respectively, suggesting KDM1A could be a therapeutic target of AML. However, other reports argue that KDM1A is essential in hematopoiesis, raising concern that KDM1A-targeted therapy could lead to severe hematological toxicity. Here, we try to clarify what types of leukemia can be ameliorated by the pharmacological inhibition of KDM1A with a possible therapeutic window and what are functional and molecular mechanisms utilizing highly selective KDM1A inhibitors we have newly designed.

Results: First we validated the effect of our novel inhibitors on murine leukemia cells harboring MLL-AF9. In accordance with a previous report, our novel KDM1A inhibitors suppressed cell proliferation, diminished clonogenic capacity and induced G1-S cell cycle arrest and myelomonocytic differentiation but not apoptosis in quite low concentration that clonogenicity of normal murine bone marrow cells was spared. Next we examined the effect of these drugs on diverse types of human myeloid leukemia cells and found that our drugs were particularly effective in erythroid leukemia cells (HEL), megakaryocytic leukemia cells (CMK11-5), and a blastic subline from a MDS patient with complex karyotype (MDS-L). MDS-L cells were changed phenotypically and morphologically towards myelomonocytic differentiation such as the increase of CD11b expression level and the induction of neutrophil-like cells. HEL and CMK cells which are negative for myelomonocytic markers also gained CD11b expression and decreased erythroid markers such as CD235a and CD71. These data suggest that the inhibition of KDM1A induces myeloid differentiation across various types of leukemia.

To investigate an underlying molecular basis for the cell fate conversion in HEL and myeloid differentiation in MDS-L by the inhibition of KDM1A, we performed gene expression profiling and analyzed a change of gene signatures. The expression pattern of transcriptional factors was changed from the erythroid signature (e.g. GATA1 and TAL1) to the myeloid signature (e.g. SPI1 and CEBPA) in HEL. Gene Set Enrichment Analysis (GSEA) showed that the myeloid differentiation-associated gene signature was positively enriched and the leukemia stem cell-associated gene signature was negatively enriched in both HEL and MDS-L.

Finally, we investigated the effect of our KDM1A inhibitors on primary human samples such as AML with MLL-AF9 and MDS in the phase of overt leukemia (MDS/AML) with complex karyotype. The colony formation capacity was clearly impaired in relatively low concentration that normal colonies were spared. We transplanted primary MDS/AML cells with complex karyotype to immunodeficient mice and treated with a KDM1A inhibitor or vehicles after confirming the engraftment. In one MDS/AML case, all mice treated with vehicles (n=4) died of anemia and the increase of human leukemia cells within four months while 3 of 4 mice treated with a KDM1A inhibitor have survived for more than six months. Also in another case, both of two vehicle-treated mice died while all drug-treated mice had survived for more than six months. At day 200 after transplantation, we sacrificed all survived mice treated with a KDM1A inhibitor and found that human blasts were displaced from the bone marrow of treated mice. Those data suggest that our KDM1A inhibitor is effective in vivo and have a possibility for the clinical application.

Conclusion: Our study suggests that KDM1A involves with myeloid differentiation and the leukemia stem cell signature and that the pharmacological inhibition of KDM1A by highly selective inhibitors is a promising way to ameliorate AML with poor prognosis such as erythroleukemia and MDS/AML with complex karyotype, without impairing normal hematopoiesis.