Lysosomal Disruption Selectively Targets Leukemia Cells and Leukemia Stem Cells Through A Mechanism Related to Increased Reactive Oxygen Species Production

Abstract 61

To identify new therapeutic strategies for AML, we compiled and screened an in-house library of on-patent and off-patent drugs to identify agents cytotoxic to leukemia cells. From this screen, we identified mefloquine, an off-patent drug indicated for the treatment and prophylaxis of malaria. In secondary assays, mefloquine decreased the viability of 9/10 human and murine leukemia cell lines (EC50 3.25–8.0 μM). Moreover, it reduced the viability of 4/5 primary AML samples, but was not cytotoxic to normal hematopoietic cells (EC50>31 μM). Importantly, mefloquine reduced the clonogenic growth of primary AML samples, but not normal hematopoietic cells, and completely inhibited engraftment of primary AML cells into immune deficient mice. Finally, systemic treatment with oral mefloquine (50 mg/kg/day) decreased leukemic burden without evidence of toxicity in 4 mouse models of leukemia, including mice engrafted with primary AML cells. Thus, mefloquine effectively targets leukemic cells, including leukemia stem cells, at concentrations that appear pharmacologically achievable and are not toxic to normal hematopoietic cells.

To identify the mechanisms of mefloquine-mediated cell death in AML cells, we performed a binary drug combination screen, hypothesizing that drugs that synergized with mefloquine may share overlapping mechanism of action. From this combination screen of 550 drugs, we identified 18 that reproducibly synergized with mefloquine as measured by the Excess over Bliss additivism score, including 3 members of the artemisinin class of anti-malarials: artemisinin, artesunate and artenimol. Strikingly, 10/18 synergistic compounds, including the artemisinins, were known generators of reactive oxygen species (ROS). Therefore we tested mefloquine's ability to increase ROS in leukemic cells. Mefloquine increased ROS production in leukemia cells in a dose- and time-dependent manner. Co-treatment with ROS scavengers α-tocopherol and N-acetyl-cysteine abrogated mefloquine-induced ROS production and cell death, indicating that ROS production was functionally important for mefloquine-mediated cell death. Moreover, the artemisinins induced ROS as single agents, and synergistically increased ROS when combined with mefloquine.

To identify cellular target(s) of mefloquine's anti-leukemic effects, we performed a yeast genome-wide functional screen to identify heterozygous gene deletions that rendered yeast more sensitive to mefloquine. 21/37 genes whose depletion conferred >4-fold sensitivity to mefloquine were associated with function of the yeast vacuole, equivalent to the mammalian lysosome. Consistent with these data, fluorescent confocal microscopy demonstrated that mefloquine and artesunate disrupted lysosomes. Cell death after mefloquine and artesunate treatment was caspase-independent and associated with increased incorporation of monodancylcadaverin in autophagosomes, consistent with the effect of these drugs on the lysosomes. To further explore the anti-leukemic activity of lysosomal disruption, we evaluated the anti-leukemic effects of the known lysosomal disrupter L-leucine-leucine methyl ether (LeuLeuOMe). Similar to mefloquine and artesunate, LeuLeuOMe induced cell death in leukemia cells, increased ROS production, and disrupted the lysosomes. Highlighting the potential clinical utility of lysosomal disrupters for the treatment of leukemia, a patient with relapsed/refractory juvenile myelomonocytic leukemia self-administered artemisinin. The artemisinin cleared the circulating blasts from the circulating blasts and the patient proceeded to allotransplant.

Finally, to investigate the basis of leukemic cell hypersensitivity to lysosomal disruption, we assessed lysosomal characteristics of primary AML and normal hematopoietic cells. By gene expression analysis, AML patient samples had higher mRNA levels of the lysosomal cathepsins A, B, C, D, H, L, S and Z, compared to CD34+ normal hematopoietic cells, and cathepsins C, D and Z were significantly over-expressed in the LSC compartment, compared to normal HSCs.

In summary, our data demonstrate that lysosomal disruption preferentially targets AML cells and AML stem cells through a mechanism related to increased ROS production. Thus, this work highlights lysosomal disruption as a novel therapeutic strategy for AML.