Internal tandem duplication mutations of the FLT3 gene (FLT3/ITD mutations) are the most frequent molecular abnormality in acute myeloid leukemia (AML) and are associated with a worse overall survival. While the normal FLT3 receptor is expressed in early hematopoietic progenitor cells, it has not been determined whether FLT3 mutations are present in the leukemic stem cells. In this study, we sorted primary AML samples into stem cell-enriched CD34+/CD38− fractions and then analyzed the sorted and unsorted cells for FLT3 mutant to wild type ratio. In each case, the FLT3 mutant to wild type ratio was not changed by selection of CD34+/CD38− cells, implying that the mutations are present in the leukemic stem cells. We injected NOD-SCID mice, in parallel groups of 4, with 106 unsorted and 5 x 103 CD34+/CD38− sorted cells from 6 FLT3/ITD AML samples. After 14 weeks, the level of human leukemic engraftment was determined using flow cytometric analysis of human CD45 compared with murine CD45 expression. Three of the 6 samples engrafted in the mice, with human CD45 positive cells ranging from 4 to 97% of total nucleated cells. In all three of the sample groups that successfully engrafted, the FLT3/ITD mutation was present at the same or higher mutant-to-wild type allelic ratio seen in the cells used to inject the mice. We then used a small molecule FLT3 inhibitor, CEP-701, to treat mice injected with CD34+/CD38- FLT3/ITD cells. 14 weeks following injection, 7 of 10 vehicle control mice had an easily detectable (by flow cytometry and colony assay) population of cells that stained for human CD45. Meanwhile, no human CD45+ cells were discernible in the marrows from the CEP-701 treated mice. Colony assays on these marrow samples yielded CFU-Ls from all vehicle-treated animals, with a frequency of 2.1 (range 0.94–3.62) CFU-L/103 human CD45+ cells plated. PCR amplification of genomic DNA obtained from the marrows of engrafted mice revealed the expected FLT3/ITD mutation in 7 out of 10 vehicle-treated mice. The mutation was also detectable by PCR only in 2 of the 8 mice treated with CEP-701. This indicated that these mice were indeed engrafted, but the level of engraftment was too low to be detected by the FACS analysis or colony assays. Quantitative PCR performed on genomic DNA from the marrow samples of the CEP-701 treated mice confirmed that the DNA from the marrows of these two mice contained FLT3/ITD DNA at a level of 0.8% and 0.03%, respectively. The other 6 CEP-701-treated mice had no detectable FLT3/ITD DNA. Finally, as a control experiment, we injected each of 10 mice with 6 x 106 normal donor human CD34+ cells. 5 of the mice were treated with CEP-701, while the other 5 mice were treated with vehicle only. After 14 weeks, analysis of human CD45 expression and CFU-GM/BFU-E colony assays was performed on the marrow of these mice. The assay results show that CEP-701 treatment had no significant effect on normal hematopoietic engraftment. This data provides evidence that 1) Leukemic stem cells harboring FLT3/ITD mutations are dependent on FLT3 signaling for survival and engraftment capability in NOD-SCID mice; 2) FLT3/ITD mutations are present in leukemia stem cells; and 3) FLT3 inhibitors may have activity against leukemia stem cells harboring these mutations.