CSF1R Inhibition Targets AML Cells By Depleting Supportive Microenvironmental Signal from CD14+ Monocytes

Introduction: CSF1R (M-CSFR) is a receptor tyrosine kinase that, upon binding with CSF1 ligand (M-CSF), activates downstream survival, proliferation, and differentiation pathways in myeloid lineage cells. Expression and activity of CSF1R has been shown to be important for sustaining tumor-associated macrophages (TAMs) in a variety of solid tumor types, and these TAMs have been shown to provide a supportive microenvironment that maintains tumor cell viability and growth. Monocyte and macrophage lineage cells have also been implicated in providing a supportive microenvironment in certain hematologic malignancies, such as nurse-like cells in CLL, but have not been extensively explored in AML. Using an ex vivo functional screening platform applied directly to primary specimens from AML patients, we have identified recurrent sensitivity to CSF1R inhibition or silencing due to a depletion of CSF1R-expressing monocytic-lineage cells in the AML microenvironment.

Methods: To identify new therapies for AML patients, our lab has screened hundreds of AML patient specimens against panels of targeted small-molecule inhibitors or siRNA and evaluated the impact on cell viability. These panels have included two small-molecule inhibitors with exquisite specificity for CSF1R (GW-2580 and ARRY-382). We have also used a panel of recombinant growth factors and cytokines, including CSF1, to measure their impact on stimulating AML cell growth.

Results: We found that 20-30% of AML specimens showed sensitivity to inhibition or silencing of CSF1R, with good concordance between CSF1R siRNA, GW-2580, and ARRY-382. Exhaustive analysis of the clinical, demographic, and genetic features of these patients failed to reveal correlation between CSF1R and any prominent disease subsets. Samples sensitive to CSF1R silencing or inhibition are significantly more likely than non-sensitive samples to respond to CSF1 growth stimulation, indicating a ligand-dependent mechanism of CSF1R signaling. Additionally, we examined CSF1R and CSF1 expression in discrete hematopoietic cell populations by flow cytometry in our specimens, both with and without CSF1R inhibition. Interestingly, we found that CSF1R was predominantly expressed a subpopulation of CD14-expressing monocytes, not tumor cells, and this population disappeared after exposure to the CSF1R inhibitor. Moreover, we identified a correlation between sensitivity to CSF1R inhibition and the presence of this CD14/CSF1R-expressing cell population.

Conclusions: These results suggest that CD14/CSF1R-expressing cells support AML blasts through a ligand-dependent mechanism, and depleting these cells eliminates the supportive microenvironment, resulting in leukemia cell death. Therefore, we propose using CSF1R inhibitors, such as ARRY-382, as a promising new line of therapy to target AML by disrupting the tumor microenvironment.