Disease Stage Strongly Modulates Responsiveness to Targeted Agents in Nf1 Mutant Mice with Myeloprolfierative Disease (MPD) and Acute Myeloid Leukemia (AML).

Hyperactive Ras is a biochemical abnormality in many cases of myelodysplastic syndrome (MDS), MPD, and AML. The genetic mechanisms that deregulate Ras signaling in myeloid malignancies include NRAS, KRAS2, and PTPN11 point mutations, FLT3 internal tandem duplications, and inactivation of the NF1 tumor suppressor. Inactivating Nf1 or expressing an oncogenic allele of Kras2 in murine hematopoietic cells induces MPD with 100% penetrance; however, this does not evolve to AML. We exploited retroviral insertional mutagenesis to induce mutations that might cooperate with Nf1 inactivation in myeloid leukemogensis. In these studies, Mx1-Cre, Nf1^flox/flox and control pups were injected with the MOL4070A retrovirus (Wolff, et al. J Virol 2003) and with pIpC to induce Cre recombinase expression and thereby inactivate Nf1. Approximately 25% of the Mx1-Cre, Nf1^flox/flox mice developed AML. The Ras effectors MEK and ERK were hyper-phosphorylated in these AMLs. We next compared the effects of CI-1040, a highly selective inhibitor of the MEK kinase, on colony growth in methylcellulose of Nf1 mutant cells from mice with MPD and AML. Importantly, there was no differential sensitivity of the MPD mutant cells to CI-1040 as CFU-GM colony formation from Mx1-Cre, Nf1^flox/flox and wild-type bone marrow was inhibited at the same drug dose. By contrast, colony formation of AML Mx1-Cre, Nf1^flox/flox cells was abrogated at much lower CI-1040 concentrations. A randomized trial of CI-1040 in Mx1-Cre, Nf1^flox/flox MPD model (n=10) at the maximal tolerated dose of 100mg/kg twice a day showed no improvement in leukocyte counts, splenomegaly, or survival in Nf1 mutant mice with MPD. We demonstrated transient inhibition of the ability of GM-CSF to phsophorylate ERK in primary bone marrow cells from CI-1040-treated mice. Whereas ERK phosphorylation was markedly decreased 2 hours following a CI-1040 dose, inhibition was marginal at 4 hours and absent by 8 hours. Based on the increased in vitro sensitivity of Mx1-Cre, Nf1^flox/flox AML cells to CI-1040, we performed a second randomized trial in 25 recipient mice that were transplanted with 4 independent leukemias. CI-1040 had dramatic effects in this setting. Whereas the leukocyte counts of vehicle-treated mice increased progressively, we observed a transient fall in leukocyte counts in all mice randomized to receive CI-1040. Treatment with CI-1040 was also associated with markedly prolonged survival (24 versus 7 days in the vehicle-treated cohort; odds ratio 3.5 95% CI 3.0–3.8). These data demonstrate that the biologic response to a molecularly-targeted inhibitor is strongly modulated by the genetic context in which a disease-initiating mutation occurs. Specifically, although the MPD induced by Nf1 inactivation was resistant to CI-1040, progression to AML was associated with enhanced sensitivity to this agent. Our data support further clinical development of MEK inhibitors with better pharmacologic characteristics as a rational strategy for treating advanced myeloid malignancies. This mouse model also provides a tractable system for identifying cooperating mutations that cause progression to AML and modulate drug sensitivity.