The Re-Activation of FoxO3 Transcription Factor in Acute Myeloid Leukaemia Is Responsible for the Induction of a Quiescent Status of Leukaemic Progenitor Cells.

Abstract

The FoxO transcription factor promotes apoptosis and triggers cell cycle inhibition through multiple mechanisms. It is inactivated by PI3K/Akt induced phosphorylation thus resulting in nuclear exclusion and degradation. Moreover, several data demonstrated an abnormal activation of PI3K/Akt pathway in acute myeloid leukemia (AML) with contrasting prognostic significance. The aim of this study was to clarify the role of FoxO3 in AML and to investigate alternative pathways eventually responsible for FoxO3 inactivation. Importantly, we explore the effects of FoxO3 re-activation in CD34+ AML cells. Cell cycle was analyzed by FACS in CD34+ cell population as well as the levels of CD47, which has been demonstrated to increased during progression through the cell cycle and stem cell mobilization. Protein amount and localization were analyzed by Western blot and immunofluorescence, the DNA binding activity was measured by EMSA. 35 BM samples from AML patients at diagnosis and in 20 healthy donors were analyzed. Furthermore Spred1, known to be a FoxO3 target gene was quantified by RQ-PCR. We previously described the absence of Spred1 in AML patients and demonstrated that it promotes growth arrest and apoptosis in haematopoietic cells. Finally BM cells were incubated with a PI3K inhibitor LY294002 and the IKK inhibitor PS1145, alone and in combination. Moreover, the t(8;21) positive Kasumi cell line was transfected with pECE-FoxO3 to evaluate FoxO3 effects on cell growth and apoptosis. We found that, while FoxO3 in control cells is localized in both nucleus (mean value of intensity of 21,4 ±2) and cytoplasm (14,6±1,7), it is completely cytoplasmatic in AML cells (18,1±4,6 in cytoplasm vs 8,2±4 in the nucleus) and enters the nucleus after chemotherapy or in vitro incubation with LY29400. Moreover, FoxO3 DNA binding activity in AML patients is completely absent at diagnosis and restored after therapy. Also the mRNA of Spred1 is rather undetectable at diagnosis (2−ΔΔCt= 0,009±0,3) and shows normal levels during remission (2−ΔΔCt = 2±1,5) or after LY29400 incubation (2−ΔΔCt =0,8±0,3). In addition LY294002 and PS1145 treatment results in FoxO3 partial nuclear re-localization while their association induces a complete nuclear shuttle suggesting that both pathways could be implicated in FoxO3 inactivation. The restoring of FoxO3 function by LY29400 in CD34+ cells induces quiescence of this progenitor cell compartment as demonstrated by the comparison of cell cycle kinetics and the decreased expression of CD47 (p=0,01). Finally, FoxO3 overexpression in transfected cells results in a block of proliferation rate (66% of inhibition compared to empty vector transfected cells) and apoptosis induction (35% vs 12%). Taken together these data suggest that FoxO3 inactivation may be crucial for the leukemic progression and demonstrate that also IKK pathway contributes to this effect, providing the rationale for a therapeutic strategy. On the other hand, the re-activation of FoxO3 induces quiescence of the stem cell compartment so providing a mechanism of escape from chemotherapy induced apoptosis.