Metabolic Profiling Defines A New Characterization Of Acute Myeloid Leukemia And Identifies Npm1-Mutated Cases As A Distinct Subgroup

Genomic and functional alterations of enzymatic activity drive cancer metabolic reprogramming along with mutations of tumor suppressors and oncogenes. IDH1/2 lesions represent a paradigmatic example in acute myeloid leukemia (AML). The study aimed to stratify AML patients based on their metabolic landscape and to map potential connections between their metabolic and genomic profile. The genomic landscape of 166 AML patients was obtained by whole exome sequencing. Variants were called by MuTect and Varscan2. Six additional cases were analyzed by targeted sequencing (SOPHiA GENETICS). Metabolites were quantified by mass spectrometry of bone marrow cells (35 CD34+ and 15 CD33+ AML from the above-mentioned cohort) and compared with CD34+ cord blood and CD33+ healthy blood cells (n=21 each, Metabolon) by Welch9s t-test. In AML, 17% of somatic variants targeted metabolism-related genes: 38% were enzymes (according to Recon2), including 3% of electron transport chain genes encoded by mitochondrial DNA (COX1-3, ND1-6), 3% were AML-related genes with a known involvement in cell metabolism (e.g. IDH1/2, MYC, KRAS) and 59% were metabolic regulators, defined by gene ontology annotation. Ninety-one% of patients carried at least one mutation in a metabolism-related gene, with 43% of variants rated as damaging. The most represented pathways were lipid, carbohydrate, nucleotide (AK9, H6PD), amino acids (IDO2) and glucose metabolism (PKM, HK3). Principal component analysis of metabolic data showed a distinct profile between AML and healthy cells, with a predictive accuracy of 86% and 95% for CD34+ and CD33+ cells, respectively. Conversely, few differences were observed between CD34+ and CD33+ AML. Unsupervised hierarchical clustering clearly defined 3 AML clusters (C1-3). Moreover, 3 subgroups could be identified in C3 without ambiguous assignments. C1 was enriched for NPM1-mutated (mut) cases (83%, 33%, 27% in C1, C2 and C3, respectively, p=0.03). NPM1-mutated AML were distinguished by a 12-metabolites signature. They showed increased levels of spermidine, cytidine 2′ or 3′-monophosphate (P), thymidine 3′-monoP, uridine-29,39-cyclic monoP and decrease of inosine 59-monoP, suggesting altered polyamine, pyrimidine and purine metabolism. Moreover, NPM1-mut AML had reduced levels of intermediates involved in acyl carnitine, lysophospholipid, phosphatidylethanolamine and sphingolipid metabolism. Overall, mutations of metabolism-related genes are common in AML. We defined a metabolic-based classification of AML and identified a new metabolic signature based on 12 metabolites that distinguish NPM1-mut AML from wild-type cases and healthy CD34+/CD33+ cells. Major alterations in the nucleotide and polyamine pathways suggest novel potential therapeutic approaches. Supported by: EHA research fellowship award, AIRC, FP7-NGS-PTL, Fondazione del Monte. Citation Format: Giorgia Simonetti, Antonella Padella, Eugenio Fonzi, Martina Pazzaglia, Margherita Perricone, Maria Chiara Fontana, Samantha Bruno, Maria Teresa Bochicchio, Eugenia Franchini, Jacopo Nanni, Giovanni Marconi, Italo F. do Valle, Rossella De Tommaso, Anna Ferrari, Enrica Imbrogno, Claudio Cerchione, Cristina Papayannidis, Emanuela Ottaviani, Daniel Remondini, Giovanni Martinelli. Metabolic profiling defines a new characterization of acute myeloid leukemia and identifies NPM1-mutated cases as a distinct subgroup [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 5279.