Virtual screening approaches for the identification of novel non-steroidal mineralocorticoid receptor ligands

Program: Abstracts - Orals, Featured Poster Presentations, and Posters
Session: MON 355-388-Sex Hormone Receptor Action & Reaction
Monday, June 17, 2013: 1:45 PM-3:45 PM
Expo Halls ABC (Moscone Center)

Poster Board MON-386
Carlos F Lagos*, Andrea Vecchiola, Cristobal A Fuentes, Fidel Allende, Alejandra Tapia-Castillo, Carolina Valdivia, Sandra Solari, Rene Baudrand, Carmen Campino, Gareth Ivor Owen, Cristian A Carvajal and Carlos F Lagos
Pontificia Universidad Catolica de Chile, Santiago, Chile
Mineralocorticoid Receptor (MR) antagonists such as Spironolactone and Eplerenone are widely used for treating a variety of cardiovascular disorders, however, concerns of side effects due to cross reactivity with other steroid hormone receptors such as the Glucocorticoid (GR), Estrogen (ER), Androgen (AR) and Progesterone (PR), highlights the need for the search of novel non-steroidal MR modulators.

Aim.To identify novel non-steroidal potential MR antagonists using virtual screening techniques.

Methods.Crystal structures of the ligand binding domains of AR, GR, PR, ER and MR in complex with ligands were retrieved from the PDB Databank and structurally aligned. Shape-based queries and structure-based pharmacophore models were generated with vROCS and LigandScout respectively. An in-house database of compounds of ~ 3,000,000 compounds was compiled from libraries available at ZINC database. The libraries were filtered to eliminate duplicates and ADME/Tox constraints using FILTER, and the conformers generated by OMEGA. Virtual screening of the conformer database was performed over 3 MR complexes using FRED and ranked according to the average Chemgauss3 scoring function. Shape filtering criteria was based on the Tanimoto-Combo Score.

Results.From the initial database of 3,633,079 compounds compiled, a library of 3,358,681 filtered compounds and 478,988,388 conformers was obtained. The top 4000 hits from docking in the MR-complexes structures were further filtered with the shape-based query of the other nuclear receptors to refine the search, and the best 400 hits binding mode analyzed. The combined Chemgauss3 scores range from -91.6 to -31.3. The 8.9% of the top scoring solutions were able to bind to the three MR structures. The combined shared feature pharmacophore model obtained with the AR, GR, PR, and ER structures contains 3 hydrophobic characteristics, and 3 H-bond donor acceptors. An area under the curve of 0.86 indicates that query is predictive and well able to separate the actives from the decoys. A final selection of 40 diverse compounds was identified as putative selective MR ligands and selected to further biological assays.

Conclusion. A combined ligand and structure-based virtual screening approach to identify putative non-steroidal ligands selective for the MR was developed. The methods allow us to identified novel non-steroidal scaffolds with desirable physicochemical properties and predicted to have better affinity to MR than for other NR.

Nothing to Disclose: CFL, AV, CAF, FA, AT, CV, SS, RB, CC, GIO, CAC, CFL

*Please take note of The Endocrine Society's News Embargo Policy at

Sources of Research Support: FONDEF CAi101050, FONDECYT 1100356 & IMII P09/016-F grants. CFL thanks OpenEye Scientific Software for academic license of their products (FILTER, OMEGA, vROCS & FRED), and the DTP/NCI for kindly providing the compounds screened. CFL and CAC are CONICYT fellows.