Pharmacophore-Based Virtual Screening for Identification of Adenosine Deaminase Inhibitors

Adenosine deaminase (ADA) is a key enzyme in purine metabolism, and abnormal ADA activity has been associated with various diseases, including severe combined immunodeficiency, cancer, neurodegenerative disorders, and liver diseases.

This study employed a pharmacophore-based virtual screening strategy to identify novel ADA inhibitors from an in-house library.

Two crystal structures of ADA co-crystallized with erythro-9-(2-hydroxy-3-nonyl) adenine (EHNA) and pentostatin were used to develop a pharmacophore model. The validated model was used to screen an in-house library. The resulting hits were further evaluated using molecular docking with AutoDock Vina. The compound with the highest binding affinity in the docking study was subsequently assessed by molecular dynamics (MD) simulation and in silico absorption, distribution, metabolism, excretion, and toxicity (ADMET) analysis.

The validated pharmacophore model comprised one hydrogen bond donor, one hydrogen bond acceptor, and one aromatic ring. When evaluated against the DUDE-Z database set, the model demonstrated acceptable sensitivity and specificity, with an enrichment factor of 23. Screening of the in-house library identified four promising hits. Compound 154 was the most notable hit because of its potent binding affinity and favorable interactions with key amino acids in the ADA active site. MD simulations showed that the ADA-compound 154 complex remained stable throughout the 150 ns simulation. MM-GBSA energy analysis indicated favorable binding driven by van der Waals and electrostatic interactions. Per-residue decomposition analysis identified critical residues contributing to complex stabilization. Compared with reference ADA inhibitors, compound 154 exhibited similar drug-likeness and pharmacokinetic properties, although further toxicity optimization may be required.

These results suggest that compound 154 is a promising lead for developing new ADA inhibitors and underscore the effectiveness of computational methods in drug discovery.