Using Chemical Biology to Modulate Antibody Activity

Abstract

Monoclonal antibodies have shown promising results as therapeutic agents, and yet they can also be associated with adverse side effects due to activity outside the disease site. Aiming to reduce these side effects, we have explored the possibility of a tunable antibody, whose activity can be manipulated via the addition of a small molecule. Previously, we incorporated a single cavity-forming mutation (tryptophan to glycine) into an antibody, and observed reduced antigen-binding activity that could be restored by addition of a complementary ligand (indole) — albeit with binding affinity too low for potential therapeutic applications. Here, I describe a novel computational strategy for enumerating larger cavities in a fluorescein-binding single-chain variable fragment (scFv), leading to a designed variant with three large-to-small mutations (triple mutant) at the domain-domain interface with reduced antigen-binding. Through a complementary virtual screen, we identified a rescuing small molecule (JK43) that enhances binding affinity for antigen. Thorough characterization of this system shows that the loss of activity upon mutation was due to loss of stability and domain dissociation; conversely, addition of JK43 restores stability of the antibody fragment, induces domain re-association, and rescues antigen binding. Beyond this initial model system, I will also describe the transferability of this design (triple mutant and JK43) from the fluorescein-binding scFv onto an unrelated scFv that shares the same three residues used in this design. We hypothesize that this design will also prove transferable onto the many therapeutic antibodies that also share these three residues, including Ipilimumab (anti-CTLA-4), Atezolimumab (anti-PD-L1), Nivolumab (anti-PD-1) and Adalimumab (anti-TNF-α).