Background: Acute myeloid leukemia (AML) is a type of cancer that consists of a group of hematological malignancies with high heterogeneity. AML patients with DNMT3A mutations often have a poor prognosis. Metabolic alterations have long been recognized in cancer cells. Metabolic homeostasis is a fundamental property of cells that becomes dysregulated in cancer to meet the altered, often heightened, demand for metabolism for increased growth and proliferation. Oncogenic mutations can directly change the cellular metabolism, priming cells for malignancy. It is well known that 2-hydroxyglutarate (2-HG) is an oncometabolite resulting from mutations of the isocitrate dehydrogenase 1 and 2 (IDH1and IDH2) genes and isa strong prognostic predictor independent of other clinical and molecular features. Alpha-ketoglutarate-dependent dioxygenase encodedby the TET2 gene could catalyze 5-methylcytosine (5-mC) to 5-hydroxymethylcytosine (5-hmC), thereby regulating DNA methylation and inducing hematopoietic malignancies. Methods: Metabolomic profiles of all serum samples were achieved using the GC-TOF MS platform. CCK-8 assays were used to examine cell viability and cytotoxic activity of 2-ketoisocaproic acid (2-KA) to leukemia cells. Methyl-Cellulose colony-forming cell assays and flow cytometry were performed to assess cell proliferation and cell differentiation. The pIpC was administrated in the Dnmt3aR878H/WTMx1Cre+mice at the time of weaning (4 weeks old) to activate transgene expression in vivo. The expressions of genes/proteins were detected by RNA sequencing, quantitative RT-PCR and western blot. Results: We report a metabolomics study with a total of 171 newly diagnosed AML patients and identified a distinct leucine metabolic signature in AML patients with DNMT3A mutations. The prognostic value of 2-KA was demonstrated in cytogenetically normal AML patients, and a low 2-KA level predicted a poor overall patient survival. We further compared the gene expression patterns of the blast cells in bone marrow (BM) between AML groups with and without a DNMT3A mutation; these patterns correlated well with those of the leucine metabolic pathways and exhibited enhanced ACAA1 and decreased ALDH7A1 gene expression in DNMT3A-mutated leukemic cells. In vitro results demonstrated that a high 2-KA level could inhibit the proliferation of DNMT3A-mutated cells but did not affect the differentiation of these leukemic cells. Conclusions:Our data suggest that 2-KA isa unique feature of AML with a DNMT3A mutation and might be a metabolic marker of early-stage hematopoietic malignances.This study also suggests that ACAA1 and ALDH7A1 were aberrantly expressed in DNMT3A mutant leukemic cells and may be potential targetsfor AML therapy. We also hypothesized that a combination of the metabolic inhibitor and chemotherapeutic agent may be a possible treatment strategy for DNMT3A-mutated AML. Disclosures No relevant conflicts of interest to declare.