A preliminary validation of the utility of the capsaicin molecule as a lead for the design of androgen receptor modulators with potential inhibitory activity

Abstract

Capsaicin, a naturally occurring molecule, has been shown to have an in vitro apoptotic effect through an anti-proliferative mode of action associated with oxidative stress induction on refractory prostate cell lines. This study used the capsaicin scaffold as a lead molecule for the design of high affinity analogs capable of similar antagonist interactions with the androgen receptor through a Virtual Screening and de novo approach. For the Virtual Screening process, conformational analysis was performed to determine the best capsaicin conformation with the highest affinity and lowest energy since this is predicted to bind optimally to the androgen receptor ligand binding pocket. Each conformer was considered separately and in each case, the total energy was calculated using Sybyl-X® Version 1.1 (Ash et al., 2010) and the ligand binding affinity (LBA) was calculated using X-Score® (Wang et al., 2002). The bioactive conformation of DHT, as extracted from PDB crystallographic deposition 2AMA (Pereira de Jésus-Tran et al., 2006) and the optimal capsaicin conformer (conformer 17), were read sequentially into LigandScout® (Wolber and Langer, 2005). The 2 molecules were superimposed and an average or consensus pharmacophore that incorporated the critical contact point forged between both small molecules was modelled. This consensus pharmacophore was then uploaded as a query onto the ZincPharmer® (Koes et al., 2015) database where hit structures having similar outer electronic configuration, physicochemical properties and spatial orientation were identified. Filters were imposed during the search process which ensured that all identified hits would be Rule of Three (Congreve et al., 2003) compliant making all hits suitable for use as lead molecules. The molecular weight was set out to be  300 while the number of rotatable bonds was targeted to be  3. Hit structures were identified and assembled into a single .sdf file using MONA® (Hilbig and Rarey, 2015) where Lipinski’s Rule of Five (Benet et al., 2016) was used for filtering. In the de novo approach, a two-dimensional (2-D) Topography Map displaying critical interactions between Conformer 17 and PDB crystallographic deposition 2AMA (Pereira de Jésus-Tran et al., 2006) in the Ligand Binding Domain was generated in Discovery Studio® Seeds are created by altering or eliminating the functional groups which are not critical to the molecule using Sybyl-X ® Version 1.1 (Ash et al., 2010). This was utilised by Ligbuilder® v1.2 (Yuan et al., 2011) using the GROW and LINK algorithms. The optimal molecules obtained from both processes will be further refined and validated to obtain clinical efficacy.