Computational Evolution of Anti-PD-1 Antibodies Induces Structural Refolding for High-Affinity Interactions.

Area of Science:

• Immunology
• Computational Biology
• Structural Biology

Background:

• Checkpoint inhibitors targeting the PD-1/PD-L1 axis are crucial immunotherapies.
• The dynamic and flexible nature of PD-1 presents challenges for antibody engineering.

Purpose of the Study:

• To evolve pembrolizumab variants with enhanced binding affinity for PD-1 using computational approaches.
• To demonstrate the utility of integrated computational methods for targeting flexible protein interfaces.

Main Methods:

• Computational saturation mutagenesis
• AlphaFold prediction
• Molecular dynamics (MD) simulations
• Analysis of antibody-protein structural interactions

Main Results:

• Seven engineered antibodies demonstrated improved binding through additional salt bridges and hydrophobic contacts.
• One variant, m7p.5, exhibited biphasic kinetics and high-affinity binding (K_D,apparent = 62 pM).
• Observed structural changes included an α-helix to loop transition in the antibody heavy chain and a significant PD-1 loop shift.

Conclusions:

• Computational evolution can generate high-affinity antibodies for intrinsically flexible targets, overcoming limitations of traditional design.
• Integrated computational approaches, including MD simulations, offer a cost-effective method for discovering novel picomolar affinity antibodies.
• This study presents a valuable tool for AI-driven antibody generation against PD-1 and other protein-protein interactions.