Computational identification of potential COX-2 inhibitors from Devil’s Claw for anti-inflammation

Abstract

This study evaluated the anti-inflammatory potential of thirty-three phytoconstituents from Devil’s Claw, Harpagophytum ( H. ) procumbens, as inhibitors of human cyclooxygenase-2 (COX-2; PDB ID: 5F19 ). We employed a comprehensive in silico pipeline combining molecular docking, ADMET profiling, and 500 ns MD simulations, followed by MMPBSA analysis. Eight compounds displayed binding affinities (≤ -8.4 kcal/mol) comparable to the reference inhibitors, diclofenac and SC-558. ADMET and drug-likeness analyses identified flavonoids (fisetin, kaempferol, luteolin) as the most pharmacokinetically suitable candidates for oral development, combining high intestinal absorption (74-82%), no Lipinski violations, and low aggregation probabilities (all compounds ≤ 0.12; luteolin showed the highest aggregation probability at 0.122), though experimental confirmation remains necessary. The extended MD simulations indicated generally stable protein-ligand interactions for most complexes, with harpagoside showing notable conformational flexibility. MMPBSA analysis revealed favorable binding free energies for the glycosides (isoacteoside, ΔG TOTAL = -38.5 ± 6.1 kcal/mol, and harpagoside, ΔG TOTAL = -32.0 ± 4.3 kcal/mol), driven by electrostatic and hydrophobic forces. The findings from this study support the putative COX-2 inhibitory potential of H. procumbens . These outcomes are complementary rather than contradictory, flavonoids are prioritised as oral lead candidates on pharmacokinetic grounds, while glycosides represent high-affinity scaffold leads that warrant formulation-based development strategies. Both groups highlight H. procumbens as a promising source of plant-derived COX-2 inhibitors and identify candidate leads that warrant experimental validation for the development of safer, plant-based anti-inflammatory agents.