Mechanisms of Relapse Following Targeted Therapy in An NRAS^G12V and Mll-AF9 Driven Mouse Model of AML

Abstract 2620

Acute Myeloid Leukemia (AML) is driven by genetic mutations that promote proliferation and prevent maturation of myeloid progenitors. While the majority of adults with AML achieve a remission with aggressive cytotoxic chemotherapy, a large proportion will ultimately die of relapsed/refractory disease. Given the limited efficacy and significant toxicity associated with current AML treatment strategies, new therapies are needed. While many genetic mutations have been shown to contribute to the development AML, the genetic heterogeneity seen in AML presents a challenge for the development of targeted therapies. The NRAS proto-oncogene and other key mediators of RAS signaling are frequently mutated in AML, making RAS-targeted therapies an attractive strategy for the treatment of AML. We previously developed an in vivo genetically-engineered mouse (GEM) model of AML driven by a tetracycline-repressible, constitutively-active form of NRAS, NRAS^G12V, and the leukemogenic fusion gene Mll-AF9. The leukemia cells in this model are “addicted” to NRAS^G12V, and inhibiting its expression with doxycycline (dox) results in rapid, complete remission of the AML. However, we have found that some mice relapse with NRAS^G12V-independent “dox-resistant” disease after continued suppression of NRAS^G12V expression, a phenomenon we might expect to occur in human AML after Ras signal pathway inhibition. To investigate the mechanisms of relapse in this model we generated two NRAS^G12V-independent (NRI) AML clones from a single primary NRAS^G12V-dependent (NRD) AML. We performed Affymetrix-based global gene expression analysis to compare the expression profiles of the primary NRD AML with the relapsed NRI AMLs. We identified 79 genes that were expressed at ≥ 2-fold higher levels in the NRD AML compared to both NRI AMLs. Among these were the putative tumor suppressor Cav2, and a negative regulator of Ras/Raf/Mek/Erk signaling, Dusp6. Down regulation of Dusp6 could enhance Raf/Mek/Erk signaling in the absence of NRAS^G12V, and thereby circumvent “addiction” to NRAS^G12V. Expression analysis also identified 20 genes that were expressed at ≥ 2-fold higher levels in both relapsed NRI AMLs compared to the primary NRD AML. Interestingly a Myc oncogene family member, Mycn (N-Myc), was expressed at > 60-fold higher levels in relapsed NRI AMLs compared to the primary NRD AML. Enforced expression of Mycn is sufficient to give rise to AML in a mouse model, and MYCN is widely expressed in human AML (Kawagoe et al. Cancer Res. 2007;67:10677), suggesting a critical role for Mycn in the development of relapsed NRI AML in this model. We are in the process of investigating whether enforced expression of Mycn and/or loss of Dusp6 are sufficient for the development of resistant NRI AML, and conversely if loss of Mycn and/or enforced expression of Dusp6 restore NRAS^G12V dependence in this model. We are also performing multi-parameter phosho-flow cytometery to further elucidate the mechanisms of AML resistance and relapse and to identify potentially “druggable” targets for AML. This represents an important step toward understanding the genetic determinates of treatment response and disease relapse with Ras pathway targeted therapies for AML, and thus provides a foundation for developing more effective and less toxic therapies for AML.