Structural basis for computational screening of non-steroidal androgen receptor ligands

Deep structural and chemical understanding of the protein target and computational methods for detection of receptor-selective ligands are important for the early drug discovery in the steroid receptor field.

This review focuses on the use of currently available structural information of the androgen receptor (AR) and known AR ligands to make computational strategies for the discovery of AR ligands in order to offer new chemical platforms for drug development.

AR is a challenging target for drug discovery and modeling even if there is a wealth of experimental data available. First, only the active structure of AR is currently known, which hampers the design of AR antagonists. Second, the structural similarity between the ligand-binding sites of AR and its mutated forms and closely related steroid receptors (SRs) such as progesterone receptors presents challenges for the development of drugs with receptor-selective function.

Research indicates that a very small chemical change in the structure of a non-steroidal ligand can cause a complete change in its activity. One source of this effect arises from binding to similar binding sites in related SRs and other proteins in the signaling pathway. Currently, computational methods are not able to predict the subtle differences between AR ligand activities but modeling does offer the possibility of generating new lead structures that might have the desired properties.