Energy-guided combinatorial co-optimization of antibody affinity and stability

Abstract

Affinity maturation is an essential process in antibody engineering. Although powerful, it is iterative, time-consuming, and can result in trade-offs, where affinity is gained at the cost of important properties, such as specificity and stability. We present a scalable strategy, called LAffAb, that starts from a crystallographic structure of the antibody-antigen complex and introduces combinations of mutations to optimize its energy. A combinatorial library comprising 7,000 variants with up to nine mutations results in gains of up to 30-fold in affinity while co-optimizing stability. Surprisingly, the library does not converge on a single solution, instead favoring diverse variants with a high mutational load. Small-scale screening of 10 designs against a potential drug target results in an order-of-magnitude improvement in affinity while maintaining high developability. We also apply LAffAb to improve the developability of a therapeutic antibody without degrading affinity. We envision that LAffAb can be used to design stable, specific, and high-affinity binders and to improve our understanding of sequence, structure, and function relationships in antibodies.