Poster - #UCLA-69 - Keywords: IL-2 Cytokines Cancer Autoimmunity Computational

Multivalency enhances the specificity of Fc-cytokine fusions

Brian Orcutt-Jahns, Peter Emmel, Scott Taylor, Aaron Meyer
Department of Bioengineering, University of California, Los Angeles Visterra, Inc., Waltham, MA Department of Bioinformatics, University of California, Los Angeles, United States of America Jonsson Comprehensive Cancer Center, University of California, Los Angeles, United States of America Eli and Edythe Broad Center of Regenerative Medicine and Stem Cell Research, University of California, Los Angeles, United States of America

ABSTRACT:
Interleukin (IL)-2 has potential as a therapy in cancer and autoimmunity but is limited in effectiveness by its specificity toward desired immune populations. IL-2 muteins may have improved cell type selectivity through altered receptor-ligand binding kinetics. Here, we analyze the response of immune cells to a panel of IL-2 muteins in monomeric and dimeric Fc fusions using tensor factorization. We find that dimeric muteins have considerably altered selectivity profiles. We then dissect the mechanism of altered specificity in dimeric ligands using a multivalent binding model and show that the enhanced selectivity by dimeric muteins arises due to multivalent ligands’ avidity for cells based on the abundance of their target receptors. Finally, we utilize this model to inform the design of tetravalent IL-2 muteins with potentially enhanced selectivity for regulatory T cells beyond what is achievable using mono- or bivalent variants. In total, we show that multivalent cytokines have unique selectivity profiles, that modeling can help to aid in their design, and that tensor factorization provides an effective approach to visualize ligand responses across diverse cell populations.