Non-Cycling Cells From Acute Myeloid Leukemia Patients Harbor the FLT3-ITD Mutation and Are Insensitive to TKI258, a Potent FLT3-Directed Inhibitor, in Vitro.

Abstract 479

The identification of activating mutations in the FLT3 gene and their impact on prognosis has been crucial to the rationale behind the development of FLT3 inhibitors. While it has been shown that some leukemic cells with high tumorigenic potential exist mostly in a dormant state, it is unclear if this quiescent/non-cycling leukemia-initiating fraction carries the FLT3-ITD mutation and if it is successfully targeted by FLT3 inhibitors. As a paradigm, quiescent Ph+ stem cells in CML have been shown to exhibit resistance to bcr-abl targeted inhibitors. Additionally, results from clinical trials suggest that FLT3 inhibitors reduce the leukemic blast count in peripheral blood but are less successful in the bone marrow where factors regulating hematopoietic stem cell quiescence are active. In order to investigate the non-cycling and cycling human leukemic cell boundary, we devised a biological model that allowed us to distinguish non-cycling AML cells from cycling AML progenitors in human FLT3-ITD positive AML samples. CD34+ cells were isolated from AML samples using magnetic cell sorting, labeled with the cell membrane dye PKH26 to enable tracking of cell division, and cultured on murine stroma for 12 days. Non-cycling AML cells were then separated from cycling cells by FACS sorting and were found to retain a CD34+ primitive phenotype in contrast to expanding leukemic blasts. Fluorescence in situ hybridization analyses revealed that non-cycling cells carried leukemic gene rearrangements (trisomy 8, trisomy 13, t[3;21]and t[16;16] in our cases), and were therefore part of the original leukemic clone. PCR for the FLT3-ITD region showed that in four out of five cases, the FLT3-ITD mutation was present in the non-cycling fraction. To examine the distribution of FLT3-ITD to FLT3 wild type (WT) bearing cells, non-cycling AML cells were FACS sorted, DNA extracted and the PCR products subsequently cloned. Bacterial colonies were sequenced and colony-PCR used to determine the ratio of FLT3-ITD to WT bearing colonies for each patient. These data indicated that at least 25% of non-cycling cells (range 25%-100%) harbored the FLT3-ITD mutation. We then assessed the impact of a potent FLT3-directed inhibitor, TKI258 (Novartis), on leukemic cell expansion and the viability of non-cycling cells. TKI258 has been found to induce apoptosis of FLT3-ITD bearing cells of the human acute monocytic leukemia MV4;11 cell line. In our present study, CD34+ AML blasts from the same FLT3-ITD positive patient samples were grown in vitro in the presence of 0 μM, 0.3 μM (IC50 dose) and 1.25 μM TKI258. In stromal cultures, TKI258 significantly reduced leukemic cell expansion (range 2.13 to 20 fold for untreated cultures and 0.07 to 2.27 fold for 1.25 μM TKI258 treated cultures at day 12, p ≤ 0.05). In methylcellulose colony assays, TKI258 exposure resulted in dose-dependent suppression of colony formation of CD34+ FLT3-ITD positive leukemia cells (60% to 81% reduction in the mean plating efficiency of CD34+ AML cells at 0.3 μM TKI258). Despite this striking anti-proliferative effect, the majority of non-cycling cells from AML patients showed resistance to TKI258 (five out of six cases). In these samples, FLT3-ITD positive non-cycling cells could still be detected after treatment with the equivalent highest clinical dose (1.25 μM) of TKI258. Moreover, at a functional level, limiting-dilution experiments on non-cycling AML cells pre-treated with TKI258 showed no impairment in a modified leukemic cobblestone assay at four weeks compared to untreated non-cycling cells. These results suggest that the majority of non-cycling AML cells that harbor FLT3-ITD are unaffected by a FLT3 inhibitor and may constitute an as yet untargeted disease reservoir. Only one FLT3-ITD AML case showed exquisite sensitivity to TKI258 with elimination of the non-cycling fraction observed from 0.3 μM of TKI258 upwards. Possible explanations for this may include specific mutant receptor sensitivities or generic multi-drug resistance mechanisms operating in dormant cells. Interestingly, TKI258 selectively eradicated an ‘intermediate' dividing progenitor population in two of the insensitive cases, an indication that leukemic progenitors may be rendered sensitive to FLT3 inhibition on transiting the dormancy-cycling boundary. Further studies are needed to determine if these findings are representative of a generation of FLT3 inhibitors or specific for TKI258.