(685a) Invited Talk: Computational and Experimental Design of Bispecific Protein Therapeutics

Bispecific and multispecific proteins, molecules capable of binding two or more targets, have a wide variety of therapeutic applications including monoclonal antibodies (Fab and Fc domain binding), receptor crosslinking on different cell types (T cell engagers), crosslinking targets on the same cell type (protein degraders), and forming AND/OR binding logic gates to multiple antigens to improve specificity or overcome heterogeneity. However, the complexity in the design of these agents, especially when considering polytherapy, makes their development challenging. Here, we use a combination of in vitro and in vivo experiments with computational modeling to highlight the trends in response to design features such as valency, binding affinity, linker flexibility, and antibody combinations for targets with varying expression levels. Specific examples for antibody drug conjugates and cross-linking bispecific molecules demonstrate the utility of these combined approaches. The use of computational modeling can help focus the experimental development of these agents, increasing the efficiency and effectiveness of novel compounds.