AML-377 A “Designed” High-Throughput Drug Screening Strategy Identifies Aurora Kinase A Inhibitors as Promising Preclinical Candidates for the Treatment of NPM1-Mutated AML

Abstract

Nucleophosmin (NPM1)-mutated acute myeloid leukemia (AML) accounts for one-third of adult AML and displays distinctive biological and clinical features. Despite its chemo-sensitivity and a relatively favorable prognosis, this leukemia lacks any specific therapies. Because NPM1 mutant protein is not druggable, targeting its non-oncogene addictions and its specific vulnerabilities can represent an alternative option for identifying a tailored therapy for NPM1-mutated AML. We performed a high-throughput screening (HTS) of different chemical compounds and drug libraries to find drugs or compounds 'synthetically lethal' with NPM1 mutation and/or showing more selective activity in AML with NPM1 mutation. In collaboration with the Screening Unit of Leibniz-Forschungsinstitut für Molekulare Pharmakologie (FMP; (Berlin, Germany), we performed an HTS of 38720 drugs/compounds in a 384-well format. The assay readout was cellular metabolism/viability by CellTiter Blue assay at 48 hours. The screening was properly designed in 2 steps to increase efficiency, exclude compounds with general toxicity, and identify compounds ('hits') selectively active against the only two available NPM1-mutated AML cell lines, OCI-AML3 and IMS-M2, but not against wild-type controls. Sixty-two hit compounds that met qHTS threshold criteria were rescreened in a dose-response experiment. From this analysis, we obtained 37 compounds able to selectively kill IMS-M2 (IMS-M2 %viability <50 vs. OCI-AML2 %viability >75). We focused on Aurora kinase A inhibitors (AURKAi) because 4 of the selective compounds were directed against this kinase. In the validation assay, 2 compounds (MK-5108 and MK-8745) were confirmed as more active against NPM1-mutated than NPM1-wild-type AML cells by either CyQUANT or apoptosis assay. Strikingly, we confirmed selectivity for two additional AURKAi (Alisertib and ENMD-2076, both isoform-selective inhibitors of AURKA) that were not included in the screening library. The AURKAi were also effective in reducing tumor burden in NPM1-mutated PDX cell models and provided a survival advantage compared with vehicle-treated mice (60 vs. 32 days, P=0.005). We demonstrated that using HTS can considerably shorten drug discovery time scales, leading to the identification of promising lead compounds. Our study is the first to provide preclinical evidence of the antileukemic activity of AURKAi, with particular focus on NPM1-mutated AML.