A Network of Activated Cancer Signaling Pathways in a High Risk Subgroup of Acute Myeloid Leukemia Targetable By SRC-Kinase Inhibitors

Resistance to therapy including potent and selective targeted agents remains the major clinical challenge in AML. In fact, allogeneic stem cell transplantation remains the best curative treatment option for AML patients with high-risk features. This unmet clinical need may be addressed by novel approaches based on targeting networks of activated cancer signaling pathways (NACSPs) instead of only individual pathways. Signaling pathway activation in AML is not only related to class I mutations such as FLT3-ITD, but also to class II driver mutations. This is best exemplified by AML subtypes harboring non-random chromosomal aberrations, encoding driver mutations able to induce and maintain leukemia. Despite the contribution of class I mutations to a poor prognosis, targeting these lesions does not decisively contribute to the cure of AML in the great majority of the cases. AML with the translocation t(6;9)(p23;q34) encoding the related DEK/NUP214 fusion protein is a high risk group of AML patients characterized by young age and presence of FLT3-ITD in ~ 75% of the cases, which responds only transiently to FLT3-ITD inhibitors. Using this AML subtype as a model for high risk disease, we investigated the NACSPs activated by a class II driver mutation. In these patients FLT3-ITD represents the only established recurrent genetic aberration at diagnosis in addition to the t(6;9). We have shown that the driver mutation DEK/NUP214 transforms very immature hematopoietic stem cells with the contribution of activated STAT5, present also in FLT3-ITD-negative patients. Furthermore it has been reported that AKT/mTOR is activated in DEK/NUP214-positive cells.

Here we investigated whether these signaling pathways are components of a leukemogenic NACSP and are therapeutically/clinically significant. We used different inhibitors to target either selectively or to determine a inhibition pattern of a.) PI3K/AKT/mTOR signaling (BKM120, BEZ2315, RAD001, Torin1 and AZD1208) at different levels; b.) receptor tyrosine kinases (RTK- ruxolitinib, sorafenib); and c.) members of the SRC kinase family (dasatinib, ponatinib, PF114, PP2). We employed four different models of t(6;9)-positive AML: U937 cells stably transfected with DEK/NUP214, t(6;9)-positive FKH1 cells, syngeneic DEK/NUP214-driven murine AML cells and primary t(6,9)-positive AML cells. Here we show that i.) STAT5 and AKT/mTOR activation was genetically determined by the t(6;9)-DEK/NUP214; ii.) STAT5 and AKT/mTOR activation were independent of JAK2 and PI3K activation, respectively; iii) selective inhibition of the AKT/mTOR cascades strongly increased STAT5 activation; iv.) both signaling pathways form a NACP, with activated members of the SRC kinase family (SKF - LYN and SRC) as a central node; v.) the NACSP was effectively targeted by inhibitors of the SRC-kinase activity (SKI) such as dasatinib, ponatinib, PF114 and the selective SKF inhibitor PP2, resulting in cell growth arrest and induction of apoptosis in t(6;9)-positive leukemic cells; vi.) SKI not only inhibited SKF/STAT5 but also the AKT/mTOR cascade; vii.) this NACP was independent of the activation of RTKs such as PDGFR, KDR, c-KIT, FLT3 a.o, as part of the target profile of many SKIs used above, as shown by the lack of activity of inhibitors such as ruxolitinib, sorafenib or ibrutinib; viii.) addition of AKT/mTOR inhibitors strongly increased effects of low dose dasatinib or ponatinib in primary t(6;9) leukaemic cells.

Our findings implicate the t(6;9)-DEK/NUP214 oncogene as a central inductor of an NACSP including SFK, AKT/mTOR and STAT5 which is independent of the recurrent FLT3-ITD signaling in these patients. This NACSP is able to maintain the leukemia in the presence of effective inhibition of FLT3-ITD signaling by Sorafenib, AC220 and other inhibitors in clinical use. Furthermore, the increase of STAT5 activation upon AKT/mTOR inhibition, suggests that the use of such a treatment would not lead to the eradication of the disease, because of the role of activated STAT5 in the maintenance of leukemic stem cells. On the other hand, SKIs target the entire NACSP. In fact, hitting the central node of the NACSP abolished the activation of both STAT5 and AKT/mTOR.

Taken together these data establish SKIs as a valid therapeutic concept not only in t(6;9)-positive AMLs but also in all other AML subtypes characterised by the same NACSP.