BACKGROUND60% to 70% of Acute Myeloid Leukemia (AML) patients enter complete remission after induction regimen, but the majority relapse within 3 years due to the outgrowth of therapy resistant Leukemia Stem Cells (LSCs). Identification of novel treatment strategies effective against these cells thus represents an outstanding medical need. We developed a cell culture method, which transiently maintains LSC activity ex vivo (Pabst et al., Nature Methods, 2014) and enables chemical interrogation of cell types relevant for the progression of the disease. Overall, HSCs and LSCs share numerous biological traits, making specific LSC eradication challenging. However, striking differences in energy metabolism between normal and leukemic stem cells have recently been suggested. While HSCs appear to rely primarily on anaerobic glycolysis for energy production, LSCs seem to depend on mitochondrial oxidative phosphorylation for their survival. Targeting mitochondrial respiration could therefore represent an effective approach for the specific eradication of LSCs. AIMWe aimed to identify novel therapeutic targets for AMLs with poor treatment outcome. The study relied on the Leucegene approach that integrates results generated by RNA sequencing analysis of primary human AML specimens, detailed clinical and cytogenetic annotations provided by the Quebec leukemia cell bank and ex vivo responses of primary AML samples to various chemical compounds. Our study specifically focused on specimens originating from patients with poor (overall survival < 3 years) and good (overall survival ≥ 3 years) response to standard chemotherapy, and did not include cases of Acute Promyelocytic Leukemia (APL). RESULTSWe identified Mubritinib, previously described as an ERBB2 inhibitor, as a novel anti-leukemic agent, which selectively inhibits the viability of leukemic cells from therapy-resistant AML patients, but does not affect normal CD34+ cord blood cells. Exposure to Mubritinib triggered apoptotic cell death in a subset of AML samples with high mitochondrial function-related gene expression, high relapse rates, and short overall survival. Sensitivity to Mubritinib also strongly associated with the intermediate cytogenetic risk category, normal karyotype (NK), and NPM1, FLT3 (ITD) and DNMT3A mutations. Conversely, resistance to Mubritinib associated with favorable cytogenetic risk AMLs, Core Binging Factor (CBF) leukemias and KIT mutations. Mubritinib has been developed as an ERBB2 kinase inhibitor. Intriguingly, we found that ERBB2 is not expressed in Mubritinib-sensitive AML specimens, suggesting that the anti-leukemic activity of this compound is likely not mediated by ERBB2 inhibition. Using a combination of functional genomics and biochemical analyses, we demonstrated that Mubritinib directly inhibits the mitochondrial Electron Transport Chain (ETC) complex I, which leads to a decrease in oxidative phosphorylation activity and to induction of oxidative stress. The impact of Mubritinib on AML progression was explored using a syngeneic mouse model (MLL-AF9 tdTomato-positive leukemia). Recipients of MLL-AF9 cells treated with Mubritinib exhibited a 19-fold decrease in the number of tdTomato-positive cells in the bone marrow and a 42-fold decrease in the spleens compared to control mice. Short-term treatment also led to a 37% increase in the median overall survival of Mubritinib exposed recipients compared to vehicle treated mice. Importantly, and in agreement with our observation that Mubritinib treatment does not impede proliferation of normal hematopoietic CD34+ cells in vitro, Mubritinib treatment had no impact on the number of non-transduced (tdTomato negative) nucleated bone marrow cells of recipients. CONCLUSIONSWe uncovered the clinical, mutational, and transcriptional landscape of mitochondrial vulnerability in AML and identified Mubritinib as a novel ETC complex I inhibitor with therapeutic potential for approximately 30% of AML cases currently lacking effective treatment options. As Mubritinib completed a phase I clinical trial in the context of ERBB2-positive solid tumors, our work suggests an opportunity to re-purpose Mubritinib's usage for this genetically distinct subgroup of poor outcome AML patients. DisclosuresNo relevant conflicts of interest to declare.