ChEMBL web services: streamlining access to drug discovery data and utilities

A PLS study on the psychotropic activity for a series of cannabinoid compounds

INTRODUCTION

The use of the Cannabis sativa plant by man has been common for centuries due to its numerous therapeutic properties resulting from the compounds present in it, called cannabinoids. However, the use of these compounds as drugs is still limited due to the psychotropic effects caused by them. The proteins that act as receptors of cannabinoid compounds were identified and characterized, being called CB1 and CB2 receptors. There is a series of 50 cannabinoid compounds that was studied through quantum and chemometric methods in order to obtain a mathematical model that could relate the structure of these compounds to their psychotropic activity. That model proved to be effective by predicting the psychoactivity of the 50 compounds from the series and elucidating relevant characteristics that imply in psychoactivity. However, most of these 50 compounds do not have experimental data of biological activity with CB1 and CB2 receptors.

OBJECTIVES

This study aims to generate QSAR models in order to predict the biological activity of the 50 cannabinoid compounds and then relate the predicted biological activity values to the already known psychoactivity.

METHODS

Another series of cannabinoid compounds was selected to generate and validate QSAR models, aiming to predict the biological activity of the 50 cannabinoid compounds with both CB1 and CB2 receptors.

RESULTS

The PLS-CB1 and PLS-CB2 QSAR models were generated and validated in this work, proving to be highly predictive, and the biological activities (pK ) of the 50 cannabinoid compounds were predicted by them. It is important to highlight compounds Ic14, Ic18, and Ic19 (psychotropic inactive) which presented higher predicted pK values than the main cannabinoid compounds (Δ9-THC and Δ8-THC). Also, compound Ic21 stood out as the highest value of the predicted biological activities in the interaction with the CB2 receptor.

CONCLUSION

The generated PLS models and the predicted pK_i values of the 50 cannabinoid compounds can provide valuable information in the drug design of new cannabinoid compounds that can interact with CB1 and CB2 receptors in a therapeutic way with no psychotropic effects.