Directed evolution of drug-like Aβ conformation-specific antibodies

Alec A Desai^1,2Matt D Smith^1,2Jennifer M Zupancic^1,2Emily K Makowski^2,3Yulei Zhang^1,2Julia E Gerson⁴Shannon J Moore^5,6Alexandra Sutter^4,7Sean P Ferris⁸Magdalena Ivanova^4,7,9Geoffrey G Murphy^5,6,9Henry Paulson^4,6,9Peter Tessier^1,2,6,9*

• ¹Department of Chemical Engineering, University of Michigan, Ann Arbor, United States
• ²Biointerfaces Institute, University of Michigan, Ann Arbor, United States
• ³Pharmaceutical Sciences, University of Michigan, Ann Arbor, United States
• ⁴Department of Neurology, University of Michigan, Ann Arbor, United States
• ⁵Department of Molecular and Integrative Physiology, University of Michigan, Ann Arbor, United States
• ⁶Protein Folding Disease Initiative, University of Michigan, Ann Arbor, United States
• ⁷Biophysics Program, University of Michigan, Ann Arbor, United States
• ⁸Department of Pathology, University of Michigan, Ann Arbor, United States
• ⁹Michigan Alzheimer’s Disease Center, University of Michigan, Ann Arbor, United States

Monoclonal antibodies that recognize conformational epitopes in protein aggregates are important for research, diagnostic, and therapeutic applications related to neurodegenerative disorders such as Alzheimer's and Parkinson's disease. Unfortunately, it remains challenging to discover and engineer high-quality conformational antibodies that are specific for protein aggregates and possess optimal combinations of three key binding properties, namely high affinity, high conformational specificity, and low off-target binding. Here, we report a directed evolution approach for generating high-quality conformational antibodies against Alzheimer's Ab fibrils in the native IgG format. Our directed evolution approach uses targeted mutagenesis, yeast surface display, cell sorting, and deep sequencing to identify antibody candidates with optimized binding properties. Notably, we find that this approach yields robust isolation of IgGs with higher affinity, higher conformational specificity, and lower off-target binding than multiple clinical-stage Aβ antibodies, including aducanumab and crenezumab. This antibody engineering platform can be readily applied to generate conformational antibodies against diverse types of peptide and protein aggregates linked to human diseases.