Deep eutectic solvents incorporating cannabinoids: The influence of water on stability

Therapeutic Deep Eutectic Solvents (THEDES) are a novel class of green, biodegradable, and low-cost solvents formed by mixing active pharmaceutical ingredients with hydrogen bond donors and acceptors, leading to liquids with significantly reduced melting points. While THEDES formed by hydrophilic HBD/HBA mixtures can be gradually hydrated to modulate their properties, THEDES formed by hydrophobic components might only accept minute amounts of water. In between those extremes, there are THEDES formed by components with dissimilar properties (standard deviation of the logP >3), whose behavior upon water addition is less clear and remains at the mercy of trial-and-error approaches. Aiming to address this gap in knowledge, this manuscript describes the use of a machine learning approach capable of predicting the stability of DES upon the addition of water. For this project, we focused on a mixture of cannabidiol (CBD), ibuprofen, and choline chloride, predicted by our model to form a stable DES. Our model also predicted that water addition would gradually decrease the probability of formation, eventually destabilizing the hydrogen bond network. Predictions from our model were validated using a series of complementary experimental and computational techniques. Overall, the presented approach aims at providing a tool to predict how THEDES would interact with the aqueous within the context of their potential therapeutic applicability.