A molecular immunology approach to antibody humanization and functional optimization

Abstract

We introduce a new method of humanization based on a novel and immunologically relevant metric of antibody humanness, termed human string content (HSC), that quantifies a sequence at the level of potential MHC/T-cell epitopes. Use of this quantity rather than global identity as an optimization goal enables the sampling of human diversity from distinct human germline sequences across the framework and CDR regions, and allows for the generation of multiple diverse candidate sequences. As a result engineering is carried out at finer sequence resolution relative to standard CDR grafting methods, providing for the optimization of antibody properties beyond immunogenicity such as antigen affinity and solution behavior. We have applied this method to the humanization of four antibodies with different antigen specificities. The resulting variable domains differ fundamentally from CDR-grafted antibodies in that they are immunologically more human and their humanness is derived from several discrete germline sequences. Furthermore, these antibodies bind their respective antigens better than or comparable to those of the parent antibodies without the need for affinity maturation.