Target specific virtual screening: Optimization of an estrogen receptor screening platform

In this work, we introduce a four-step scoring and filtering procedure, furnishing target specific virtual screening (TS-VS), which serves to minimize false positives resulting from conformational artifacts of the docking process and is optimized to converge on novel chemotypes of estrogen receptor alpha (ER(x). As a proof of concept, VS of a cornmercial compound database was undertaken (SPECs database release: Aug 2005, 202 054 compounds in total), resulting in the identification of both previously known and novel putative ER scaffolds. Application of distance constraints within TS-VS allowed facile identification of three novel active ligands with ER alpha binding affinities (IC50) of 1.4,mu M, 57 nM, and 53 nM. Importantly, they all exhibited ER(x over ER beta selectivity, with the most selective being 17-fold. The ligands also displayed low micomolar anti proliferative activity (7-15 mu M) in the human MCF-7 breast cancer cell line.