Purpose: CD47 is highly expressed on a variety of tumor cells. The interaction of CD47 with SIRPα, a protein on phagocytic cells, transmits a "don't eat me" signal that negatively regulates phagocytosis. CD47-SIRPα antagonists enable phagocytosis by disrupting the inhibitory signal and can synergize with Fc-mediated pro-phagocytic signals for potent elimination of tumor cells. A potential limitation of therapeutic CD47-SIRPα antagonists is that expression of CD47 on normal cells may create sites of toxicity or an "antigen sink." To overcome these limitations and address selective tumor targeting, we developed SIRPabodies to improve the therapeutic benefits of CD47-SIRPα blockade specifically towards tumor. Experimental Design: SIRPabodies were generated by grafting the wildtype SIRPα either to the N-terminus or the C-terminus of the heavy chain of rituximab. Selective tumor binding was tested using CFSE-labeled human primary CLL cells in the presence of 20-fold excess of human RBCs. NSG mice were transplanted with Raji-luciferase cells and were assigned to controls versus SIRPabody treatment. Cynomolgus non-human primates were administered a single intravenous infusion of SIRPabody at 3, 10, or 30 mg/kg. Results: SIRPabodies selectively bound to dual antigen-expressing tumor cells in the presence of a large antigen sink. SIRPabody reduced tumor burden and extended survival in mouse xenograft lymphoma models. SIRPabody caused no significant toxicity in non-human primates. Conclusions: These findings establish SIRPabodies as a promising approach to deliver the therapeutic benefit of CD47-SIRPα blockade specifically towards tumor cells. SIRPabodies may be applied to additional cancer types by grafting SIRPα onto other tumor-specific therapeutic antibodies.
- Received October 14, 2015.
- Revision received March 30, 2016.
- Accepted April 5, 2016.
- Copyright ©2016, American Association for Cancer Research.