Azacitidine in Combination with the Selective FLT3 Kinase Inhibitor Crenolanib Effectively Disrupts Stromal Protection of CD34+ Leukemia-Initiating Cells (LIC) in FLT3-ITD+ Acute Myeloid Leukemia (AML)

Background: Effectively targeting the oncogenic mutation FLT3-ITD remains a crucial goal in acute myeloid leukemia (AML) therapy. Thus far, tyrosine kinase inhibitors (TKI) have not been able to eradicate the earliest leukemia-initiating cells (LIC) in FLT3-ITD⁺ AML, which are thought to be responsible for the frequent relapses seen in this disease. We have previously shown that LIC in FLT3-ITD⁺ AML persist during treatment with first-generation TKI despite effective inhibition of FLT3 phosphorylation owing to their selective protection by niche cells (Parmar et al, Cancer Res 2011). Hence, novel strategies to disrupt the protective interaction of stroma with FLT3-ITD⁺ LIC are urgently needed. Here, we asked whether stromal resistance of FLT3-ITD⁺ LIC can be overcome by the next generation TKI crenolanib alone or in combination with the hypomethylating agent azacitidine (AZA).

Methods: The efficacy of crenolanib alone or in combination with AZA was assessed in the human FLT3-ITD⁺ cell lines MV4-11 and MOLM13 and FLT3-ITD transfected BaF3 cells as well as in primary human FLT3-ITD⁺ AML bone marrow samples. Cells were cultured in suspension or on the mesenchymal stromal cell line EL08-1D2, which mimics the bone marrow niche and maintains LIC in vitro (Parmar et al, Cancer Res 2011). Cultures were treated with DMSO, crenolanib and/or AZA for defined periods. Apoptosis, cell cycle and differentiation of AML cells were analyzed by flow cytometry. Clonogenic capacity and frequency of primitive FLT3-ITD⁺ stem/progenitors were probed by standard CFU and LTC assays. Engraftment potential of FLT3-ITD⁺ patient-derived xenograft (PDX) AML cells after treatment with crenolani, AZA or the combination thereof was assessed in the NSG xenograft model (Vick et al, PlosOne 2015). Treatment-induced alterations in FLT3-ITD downstream signaling were investigated by western blots.

Results: Crenolanib effectively inhibited FLT3 and STAT5 phosphorylation in FLT3-ITD+ cells in suspension as well as in EL08-1D2 supported co-cultures whereas AZA had no effect on FLT3 signaling pathways. Monotherapy with crenolanib but not AZA effectively induced apoptosis and inhibited growth of FLT3-ITD⁺ cell lines. Primary CD34⁺ FLT3-ITD⁺ progenitor cells were also highly susceptible to inhibition by crenolanib as a single agent. However, crenolanib was completely unable to eradicate primitive CD34⁺ FLT3-ITD⁺ LIC when protected by niche cells as assessed by CFU and LTC assays as well as xenotransplantation into NSG mice. In contrast, the combination of crenolanib and AZA resulted in efficient apoptosis and dramatically impaired clonogenic capacity of FLT3-ITD⁺ LIC even in the presence of stroma. Pretreatment of EL08-1D2 cells with AZA before co-culture with AML cells did not influence stromal protection against crenolanib. Furthermore, soluble stromal factors did not account for TKI resistance. Successful targeting of stromally protected FLT3-ITD+ LIC by the combination of AZA and crenolanib is currently being confirmed by xenotransplantation in NSG mice.

Conclusion: The combination of azacitidine and the selective next generation FLT3 kinase inhibitor crenolanib is a novel promising treatment regimen to overcome niche protection and selectively target LIC in FLT3-ITD+ AML. We hypothesize that the mechanism of action involves loss of quiescence in FLT3-ITD+ LIC leading to increased susceptibility towards TKI as well as induction of differentiation by AZA. In depth analysis of involved signaling pathways and phenotypic alterations in LIC are ongoing.