Cannabis is widely used as a therapeutic drug, especially by patients suffering from psychiatric and neurodegenerative diseases. However, the complex interplay between phytocannabinoids and their targets in the human receptome remains largely a mystery, and there have been few investigations into the relationship between the chemical composition of medical cannabis and the corresponding biological activity. In this study, we investigated 59 cannabis samples used by patients for medical reasons. The samples were subjected to extraction (microwave and supercritical carbon dioxide) and chemical analyses, and the resulting extracts were assayed in vitro using the CB1 and CB2 receptors. Using a partial least squares regression analysis, the chemical compositions of the extracts were then correlated to their corresponding cannabinoid receptor activities, thus generating predictive models that describe the receptor potency as a function of major phytocannabinoid content. Using the current dataset, meaningful models for CB1 and CB2 receptor agonism were obtained, and these reveal the insignificant relationships between the major phytocannabinoid content and receptor affinity for CB1 but good correlations between the two at CB2 receptors. These results also explain the anomalies between the receptor activities of pure phytocannabinoids and cannabis extracts. Furthermore, the models for CB1 and CB2 agonism in cannabis extracts predict the cannabinoid receptor activities of individual phytocannabinoids with reasonable accuracy. Here for the first time, we disclose a method to predict the relationship between the chemical composition, including phytocannabinoids, of cannabis extracts and cannabinoid receptor responses.
Keywords: cannabidiol; cannabinoid receptors; chemoinformatics; medical cannabis; partial least squares analysis; quantitative structure-activity relationship; tetrahydrocannabinol.