Preemptive optimization of a clinical antibody for broad neutralization of SARS-CoV-2 variants and robustness against viral escape.

Summary

Using deep mutational scanning and iterative computational design, the authors engineered AZD3152 into 3152-1142, restoring and broadening neutralization against current and anticipated SARS-CoV-2 variants, including XBB.1.5+F456L. This generalizable, preemptive optimization strategy integrates structure-based modeling, machine learning, and experimental validation to mitigate future viral escape.

Key Findings

• Deep mutational scanning identified key AZD3152 vulnerabilities at spike residues F456 and D420.
• Two rounds of structure- and ML-guided optimization produced 3152-1142 with ~100-fold improved potency against XBB.1.5+F456L and sustained activity across 24 variants.
• DMS confirmed no new susceptibility hotspots in 3152-1142, indicating improved robustness against future escape.
• The design co-optimized for 20 potential future escape variants, illustrating a preemptive strategy.

Clinical Implications

Could inform next-generation monoclonal antibody prophylaxis for immunocompromised patients by maintaining potency across emerging variants and reducing the risk of escape. Supports integrating DMS- and AI-guided updates into regulatory and clinical pipelines.

Limitations

• Predominantly in vitro neutralization without in vivo efficacy or clinical outcomes
• Pharmacokinetics, immunogenicity, and manufacturability of redesigned antibodies not reported

Future Directions

Translate preemptive optimization into clinical-grade candidates with in vivo efficacy and safety; extend the approach to other respiratory pathogens and polyclonal antibody cocktails.