Targeting of a Novel MNK-eIF4E-b-Catenin Axis in Blast Crisis Chronic Myelogenous Leukemia Inhibits Leukemia Stem Cell Function

Abstract 963

Blast crisis (BC) chronic myeloid leukemia (CML) is characterized by expansion of a granulocyte macrophage progenitor-like population (GMPs) that has acquired self-renewal capacity, a feature not seen in normal or chronic phase (CP) GMPs. The ability to self-renew is thought to be mediated by b-catenin activation, and may contribute to disease persistence, as well as act as a reservoir for resistance. The mechanisms contributing to b-catenin activation remain obscure, and will need to be identified to improve the control of BC. In this study, we investigated the role of the mRNA translation machinery in mediating b-catenin-mediated self-renewal, since our prior work had implicated aberrant mRNA translation in drug-resistance and BC pathophysiology (Ly et al. Cancer Research 2003; Prabhu et al. Oncogene, 2007; Zhang et al. MCB, 2008). Using immunofluorescence (IF), we first confirmed that BC GMPs have activated nuclear b-catenin compared to GMPs isolated from normal cord blood, and that this was associated with increased eIF4E expression and phosphorylation at Ser209. Next, using biochemical and genetic approaches in CML cell lines (K562 and KCL22), we demonstrated that eIF4E overexpression was sufficient to increase b-catenin activity (as measured by IF for nuclear b-catenin, b-catenin reporter assays, and expression of b-catenin-regulated genes). By expressing phospho-mutant forms of eIF4E (S209A, S209D), we also found that the increase in b-catenin transcriptional activity is dependent on phosphorylation of at Ser209. In line with these observations, siRNA-mediated knockdown or pharmacologic (CGP57380) inhibition of the MNK1/2 kinases (which mediate in vivo eIF4E phosphorylation) prevented the increased b-catenin activity induced by eIF4E overexpression. Mechanistically, we found that eIF4E activated b-catenin signaling via a two-step mechanism. First, eIF4E overexpression increased total cell b-catenin. Second, eIF4E phosphorylation facilitated b-catenin nuclear translocation. The latter step was associated with increased b-catenin phosphorylation at Ser552, a site known to be involved in nuclear translocation, and directly regulated by AKT. Consistent with this model, siRNA-mediated knockdown or small molecule inhibition of AKT (AKT Inhibitor IV) prevented eIF4E-mediated increases in b-catenin transcriptional activity. The importance of eIF4E phosphorylation on b-catenin activation and the self-renewal capacity of primary BC GMPs cells was assessed next. First, we showed that treatment with CGP57380, but not imatinib or dasatinib, inhibited eIF4E phosphorylation, as well as prevented accumulation of active nuclear b-catenin in BC GMPs. Next, we evaluated the effect of MNK1/2 inhibition on the stem cell function of BC cells using both in vitro and in vivo assays. In an in vitro serial replating assay, we showed that CGP57380 impaired the ability of CD34+ BC cells (including those carrying T315I mutation), but not normal CD34+ cells, to serially replate for more than 8 weeks in methylcellulose. Interestingly, treatment with either imatinib or dasatinib only partially impaired the ability of BC CML to serially replate. Next, we found that in vitro treatment of BC CD34+ CML cells, but not normal cord blood CD34+ cells, with CGP57380 retarded their ability to engraft NSG mice. Finally, we developed an in vivo serial transplantation assay for assessing the leukemia stem cell (LSC) function of patient-derived BC GMPs. Here, we injected either BC GMPs or BC CD34+ CML cells intrafemorally into 8- to 10-week old sublethally irradiated NSG mice. Following engraftment, mice were treated with vehicle, CGP53780 (40 mg/kg/d), or dasatinib (5mg/kg/d) for three consecutive weeks. Following treatment, human CD34+ cells were isolated from the mice, and transplanted into a second recipient mouse. At 16 weeks, we found that in vivo treatment with CGP57380, but not dasatinib, prevented BC cells from serially transplanting NSG mice. In summary, our results demonstrate that: 1. eIF4E is overexpressed and phosphorylated at Ser209 in BC, but not normal, GMPs; 2. eIF4E phosphorylation activates b-catenin signalling in BC GMPs; 3. MNK inhibition prevents eIF4E phosphorylation and b-catenin signalling in BC GMPs; and 4. MNK inhibition prevents BC GMPs from functioning as LSCs. Our studies suggest that pharmacologic inhibition of the MNK1/2 kinases may be therapeutically useful in BC CML.