NPM1c Alters FLT3-D835Y Localization and Signaling Thereby Inducing a Fulminant Myeloproliferative Disorder in Mice

Activating mutations in FLT3 and NPM1 are the most common alterations in AML and frequently coincidental. The NPM1 mutation (NPM1c) dislocates the protein to the cytoplasm due to a disruption of the nuclear localization signal. In FLT3, two types of mutations are present: Tandem duplications of the juxtamembraneous domain (ITD) and point mutations of the tyrosine kinase domain (TKD). In murine models, both NPM1c and FLT3-ITD induce a myeloproliferative disease (MPN), while FLT3-TKD expression alone is not sufficient to cause a myeloid disease in mice. Co-expression of NPM1c and FLT3-ITD rapidly induces an AML in C57Bl/6 mice. In the present study, we investigated the impact of NPM1c co-expression with the TKD mutation FLT3-D835Y and the molecular mechanism underlying the functional dependency of those two proteins.

For this purpose, we used a heterozygous conditional Npm1c knockin mouse model, where bone marrow cells were harvested, retrovirally infected with MiG- Flt3-D835Y and injected into lethally irradiated C57Bl/6 wt mice. Lymphatic/leukemic organs of moribund mice were analyzed for their immune phenotype by flow cytometry. To investigate the molecular function and mechanism leading to disease induction by these oncogenes, intracellular flow cytometry as well as immunofluorescence imaging of primary murine cells were performed to analyze signaling and localization differences. In order to examine the clinical relevance, AML patient samples harboring either NPM1c and FLT3-TKD mutations or FLT3-TKD mutation in combination with wildtype NPM1 were analyzed by immunohistochemistry and flow cytometry for intracellular signaling mediators.

Flt3-D835Y⁺ Npm1c⁺mice rapidly developed an MPN with a median latency of 33 days, whereas Flt3-D835Y⁺ Npm1 wt mice did not develop a leukemic disease. To investigate the underlying mechanism leading to MPN induction in Flt3-D835Y⁺ Npm1c⁺mice, we analyzed the activation of signaling transducer and activator of transcription (STAT) pathways in primary splenocytes of moribund mice. Strikingly, we were able to detect a profound activation of STAT5 in murine Flt3-D835Y⁺ Npm1c⁺ cells, which could not be seen in the Flt3-D835Y⁺ Npm1 wt cells. Next we asked for the clinical significance of these differential FLT3-TKD signaling capacities and performed immunohistochemistry stainings as well as intracellular flow cytometry of primary AML patient BM at initial diagnosis. In line with the findings in the murine model, immunohistochemic analyses showed significantly increased pSTAT5 signal in AML samples harboring both NPM1c and FLT3-TKD mutations, when compared to FLT3-TKD⁺ samples without NPM1 alteration. These results could be confirmed by intracellular pSTAT5 staining of primary human AML blasts, where STAT5 activation was significantly elevated in FLT3-TKD patients harboring an NPM1c mutation.

To investigate the molecular mechanism underlying these differential signaling capacities we performed immunofluorescence and flow cytometry analyses of murine leukemic cells and could show a profound localization shift of FLT3-TKD from the cell surface to the endoplasmic reticulum only in cells harboring the NPM1c mutation. In NPM1 wt cells FLT3-TKD could exclusively be detected at the cell surface. These results could be confirmed by immunoblotting analyses, where we detect an underglycosylated, intracelullar FLT3-TKD 130 kDa-form in NPM1c⁺ cells while it is found as a highly glycosylated, cell-surface bound 150 kDa form in cells lacking NPM1c.

In the present study, we could demonstrate that co-expression of Npm1c and Flt3-D835Y induces a rapid-onset MPN in C57Bl/6 mice. Interestingly, the presence of NPM1c also alters the cellular localization of FLT3-TKD from the cell surface to the endoplasmic reticulum, and thereby leads to STAT5 activation by FLT3-D835Y. We could further show that STAT5 activation not only occurs in murine cells but also in AML patients with FLT3-TKD in combination with NPM1c mutation. Thus, our data indicate that the NPM1c mutation shifts the signaling properties of FLT3-TKD towards those of the FLT3-ITD mutation, which may explain its occurrence in myeloid disorders and suggest that STAT5 might be a potential unique therapeutic target in FLT3-mutated AML.