Author(s)

Yining Zhang

Corresponding Author

Yining Zhang

Abstract

Nonsteroidal anti-inflammatory drugs (NSAIDs) are widely used in veterinary medicine for managing pain and inflammation in cats and dogs. However, their clinical application is often limited by severe adverse effects, including gastrointestinal (GI) and renal toxicity, due to species-specific metabolic differences. This study proposes an AlphaFold-driven structural optimization strategy for designing feline- and canine-specific COX-2 inhibitors to enhance drug safety and selectivity. Leveraging AlphaFold v2.3, we predicted species-specific structural variations in cyclooxygenase (COX-1/COX-2) enzymes from Felis catus and Canis lupus familiaris, identifying critical active-site residues such as the hydrophobic pocket formed by Leu355 in cats and the Arg120/Ser353 hydrogen-bond network in dogs. Virtual screening of a 10,000-compound library, guided by molecular docking simulations, yielded two candidate molecules: Cat-COXi-1 (targeting feline COX-2) and Dog-COXi-2 (targeting canine COX-2). In vitro validation demonstrated that Cat-COXi-1 achieved a 3-fold improvement in COX-2 inhibition (IC₅₀ = 0.8 μM) while reducing GI-related off-target effects by 60% (MUC1 expression downregulation <12%). Dog-COXi-2 exhibited prolonged pharmacokinetics (half-life = 8 hours) and no significant renal toxicity markers. This work pioneers the application of AlphaFold in veterinary pharmacology, revealing mechanistic insights into species-specific drug-target interactions and establishing a framework for precision anti-inflammatory therapies in companion animals. The findings underscore the potential of structure-based design to balance efficacy and safety in veterinary medicine, addressing long-standing challenges in NSAID toxicity.

Keywords

AlphaFold Protein Structure Prediction, Species-Specific Drug Design, COX-2 Inhibition, Veterinary Pharmacology