The in-silico evaluation and synthesis of silicon-based analogues of the cholinesterase inhibitor Donepezil

Abstract

Alzheimer’s disease (AD) is a multi-factorial and irreversible progressive neurodegenerative disease of growing social and financial concern. Donepezil is one of the best therapeutic drugs for the alleviation of the symptoms associated with AD available today, due to its competitive inhibition against both acetylcholinesterase (AChE) activity and AChE induced Aβ aggregation. The use of silicon as an alternative to carbon in medicinal chemistry has become a useful tool to increase the lipophilicity and absorption of drug molecules. In many literature examples silicon has been shown to increase blood brain barrier (BBB) penetration of molecules. This study was undertaken to identify potential lead compounds based on the donepezil scaffold with incorporated silicon groups for increase lipophilicity and BBB penetration through in-silico methods.

The study utilized the Schrödinger suite to develop and validate a method for the screening of silylated derivatives of donepezil for inhibition of AChE. A typical workflow was first used as described in literature, whereby the prepared ligands where docked using Glide XP into the prepared protein as the first screening step and prime MM-GBSA as the second step. It was then discovered in literature that improved models could be produced by introducing ConfGen and macro model minimization into the workflow as provided in the Schrödinger suite. The new workflow model was then validated by screening a small library of known actives compounds (against AChE) in the new model and the scored ligands were analysed through the enrichment task within the Schrödinger suite. The enrichment function compares the scored ligands against a list of the known active compounds sorted from highest to lowest activity for comparison. Additional enrichment scores are then assigned to each ligand to assess which model performed well in selecting the higher activity molecules first over molecules with lower activity.

Several ligand scaffolds were identified which are predicted to have improved and comparable binding energies to donepezil (-72.990). The most promising in silico leads identified have the indanone ring system of donepezil replaced with a silylated acetophenone ring which is linked to the benzyl piperidine ring system by either an ether (-75.736) or β-keto (-71.922) linker. In both instances O-silylation of the indanone ring system showed the best binding scores. In addition, O-silylation of the indanone moiety and C-silylation of the benzyl moiety of donepezil also showed great promise with comparable scores to donepezil.

Access to several promising silylated analogues was attempted with limited success. The isolation of the acetophenone ether linker derivative (Z)-2-((1-benzylpiperidin-4-yl)methoxy)-1-phenylethenol was achieved. In addition, the formation of (Z)-2-(2-((1-benzylpiperidin-4-yl)methoxy)-1-hydroxyvinyl)phenol occurred, but as a mixture of inseparable products. The desired analogues were shown to be synthetically accessible, however, the synthetic procedures still require reaction optimisation to produce the ligands in sufficient quantity and quality for biological testing and further derivatisation.