Week 2, May 2024: Computational Design of Antibodies

The use of bioinformatics and computational methods were of high interest in several papers published this past week. Protein language models, akin to natural language processing tools, predict antibody structures and interactions based on vast datasets of protein sequences. Machine learning and AI are also pivotal, enabling the prediction of antibody-antigen interactions and the optimization of antibody sequences for improved binding affinity. Molecular dynamics simulations further contribute by modeling the atomic-level behavior of antibodies, providing insights into their stability and function.

Researchers from Lawrence Livermore National Laboratory redesigned an antibody drug against Omicron BA.1 and BA.1.1 strains of SARS-CoV-2, while maintaining efficacy against the dominant Delta variant! Many antibody therapeutics lost potency when the SARS-CoV-2 Omicron variant emerged in 2021, such as Evusheld and cilgavimab. These results show that computational methods can optimize an antibody to target multiple escape variants, in addition to increasing potency via generative unconstrained intelligent drug engineering (GUIDE). This approach combines high-performance computing, simulation and machine learning to co-optimize binding affinity to multiple antigen targets.

Last week, researchers from University of Washington also computationally designed antibody nanoparticles that respond to changes in the environment! These non-porous, pH-responsive, octahedral nanoparticles have a targeting antibody on the two-fold symmetry axis, a designed trimer programmed to disassemble below a tunable pH transition point on the three-fold axis, and a designed tetramer on the four-fold symmetry axis. They can package protein and nucleic acid payloads and can incorporate almost any antibody, with potential for targeted delivery of biologics.

On a similar topic, a team from University of Colorado Boulder developed an integrated technology for quantitative wide mutational scanning of human antibody Fab libraries. Named MAGMA-seq, this integrated technology combines multiple antigens and multiple antibodies and determines quantitative biophysical parameters using deep sequencing. The team showed comprehensive mapping of potential antibody development pathways, sequence-binding relationships for multiple antibodies simultaneously, and identification of paratope sequence determinants for binding recognition for broadly neutralizing antibodies (bnAbs).