Baicalein (Bai) is one of the most important bioactive flavonoids isolated from the well-known traditional Chinese medicine called “Huang Qin”. Though it has broad therapeutic capability, the bioavailability is limited due to its poor solubility. In this study, we aimed to modulate its solubility through cocrystallization. Cocrystals of Bai with isoniazide (Inia), isonicotinamide (Inam), caffeine (Caf), and theophylline (Tph) were obtained. And different cocrystallization methods lead to different cocrystal phases for Inam and Tph. These cocrystals were characterized using powder X-ray diffraction, thermogravimetric analysis, differential scanning calorimetry, dynamic vapor sorption, and Fourier transform infrared spectroscopy. Among all the cocrystals studied, BaiCaf is found to be superior in powder dissolution and pharmacokinetic behavior. Area under the curve values of BaiCaf is improved by a factor of 4.1, and the bioavailability of baicalein is thus expected to be accordingly increased. Given that Caf is a central nervous system stimulant available in many prescription and nonprescription medications, BaiCaf can be a promising alternative solid form of Bai to be developed.

A machine-learning model trained on the whole Cambridge Structural Database was developed to assist high-throughput cocrystal screening. With only 2D structures taken as inputs, the probability of cocrystal formation is returned for two given molecules. All of the cocrystal records in the CSD were used as positive samples, while negative samples were constructed by randomly combining different molecules into chemical pairs. Our model showed a prediction ability comparable with that of a widely used ab initio method in a head-to-head comparison test. Both experimental and virtual cocrystal screening against captopril were conducted at the same time to further validate the model. Two cocrystals of captopril with L-proline and sarcosine were obtained and characterized by PXRD, DSC, and FT-IR. These two coformers were also successfully predicted by our model. These results suggest that the tool we developed can be used to effectively guide coformer selection in the discovery of new cocrystals.

Scheme S1. Library of coformers. Scheme S2. Polarizing microscopy pictures of (a) MD• 1-HNA, (b) MD• 6-HNA, and (c) MD• Sul.