AbstractAbstract 1580Most adults with acute myeloid leukemia (AML) are not cured with current treatments due to primary chemo-resistance or relapse. Emerging evidence suggests that the pool of leukemic blasts is heterogeneous and disease persistence is due to a subset of leukemic (-stem) cells able to evade chemotherapy and sustain tumor growth. Cell surface marker expression has proven to be a valuable tool to isolate and study leukemic stem cells (LSC) which, similarly to normal hematopoiesis, are shown to reside in the CD34+/CD38- leukemic fraction. However, recent data indicate that the phenotype of LSC varies from patient to patient and it seems likely that no single phenotypic signature exists to uniformly identify LSC. Besides immunophenotype, isolating LSC on the basis of functional properties unique for these cells may enforce our understanding of AML biology and provide the basis to develop more effective therapies. Reactive oxygen species (ROS) regulate essential cellular functions such as self-renewal, proliferation and apoptosis. In normal neurogenesis and hematopoiesis, ROSlow states correlate with self-renewal and ROShigh is associated with differentiation. In malignant tissues, although cancer cells are commonly more oxidized than their normal counterparts, some cancer stem cells are shown to contain low ROS levels, a feature associated with increased resistance to therapy (Nature. 2009;458:780-783). We hypothesized that LSCs also reside in a less oxidized state than bulk leukemic cells, a condition which promotes self-renewal and confers resistance to chemotherapy. To validate this hypothesis, we evaluated the redox state of leukemic blasts isolated from bone marrow or peripheral blood from 21 AML and 2 high-risk MDS patients. Loading of cells with the fluorescent probe DCF-DA showed that primary AML specimens have a broad range of oxidative state, with cells clearly falling into ROShigh and ROSlow populations (ROSlow=11.5±9%). Phenotypic analyses of AML specimens with respect to primitive cell surface markers indicated that the ROSlow gate represented 18 ± 17% of the phenotypically primitive CD34+/CD38- cells and was significantly more enriched in CD34-/CD38- leukemic cells in comparison to ROShigh. We isolated ROSlow and ROShigh leukemic subsets by flow cytometric sorting on the basis of their DCF fluorescence from 11 AML patients' samples and analyzed them for stem cell properties and drug sensitivity. Importantly, we used the differential redox state and not phenotypic markers to isolate distinct leukemic subpopulations. Morphological evaluation of sorted CD45/SS blast gated, DCFlow and DCFhigh cells demonstrated that both subpopulations were leukemic. Comparative analysis of the cell cycle distribution after staining with Ki67 and 7AAD indicated in most cases that ROSlow cells are quiescent, in contrast to ROShigh and total blast cells which are more actively cycling. Despite their predominant quiescent state, ROSlow leukemic cells were able both to grow as colonies in CFU assays and also to engraft in NOD SCID mice in pilot experiments, suggesting the existence of both leukemic “progenitor” and “stem” cell types within the ROSlow leukemic fraction. Based on these data, we challenged primary AML specimens with conventional chemotherapy agents (daunorubicin and AraC). Intriguingly, ROSlow cells preferentially survived exposure to either antileukemic agent in vitro. Taken together, our data identify a novel quiescent “oxidative state – low” leukemic population from patients with AML/MDS at diagnosis, which displays stem cell properties and exhibits functional differences related to drug sensitivity. The detailed molecular and functional characterization of this novel leukemic population is the subject of our ongoing studies. Disclosures:No relevant conflicts of interest to declare.