Cocrystal Prediction of Bexarotene by Graph Convolution Network and Bioavailability Improvement.

in Pharmaceutics by Fu Xiao, Yinxiang Cheng, Jian-Rong Wang, Dingyan Wang, Yuanyuan Zhang, Kaixian Chen, Xuefeng Mei, Xiaomin Luo

TLDR

  • This study used a computer program to find new ways to make a drug called Bexarotene work better. The program looked at the molecules that the drug is made of and found ways to combine them with other molecules to make them work better together. The researchers then tested these new combinations and found that they worked better than the original drug. The study also looked at how well these new combinations worked in the body and found that they worked better than the original drug. This study shows that using computer programs to find new ways to make drugs work better is a good idea and could help people who have skin cancer.

Abstract

Bexarotene (BEX) was approved by the FDA in 1999 for the treatment of cutaneous T-cell lymphoma (CTCL). The poor aqueous solubility causes the low bioavailability of the drug and thereby limits the clinical application. In this study, we developed a GCN-based deep learning model (CocrystalGCN) for in-silico screening of the cocrystals of BEX. The results show that our model obtained high performance relative to baseline models. The top 30 of 109 coformer candidates were scored by CocrystalGCN and then validated experimentally. Finally, cocrystals of BEX-pyrazine, BEX-2,5-dimethylpyrazine, BEX-methyl isonicotinate, and BEX-ethyl isonicotinate were successfully obtained. The crystal structures were determined by single-crystal X-ray diffraction. Powder X-ray diffraction, differential scanning calorimetry, and thermogravimetric analysis were utilized to characterize these multi-component forms. All cocrystals present superior solubility and dissolution over the parent drug. The pharmacokinetic studies show that the plasma exposures (AUC) of BEX-pyrazine and BEX-2,5-dimethylpyrazine are 1.7 and 1.8 times that of the commercially available BEX powder, respectively. This work sets a good example for integrating virtual prediction and experimental screening to discover the new cocrystals of water-insoluble drugs.

Overview

  • The study aims to develop a GCN-based deep learning model (CocrystalGCN) for in-silico screening of the cocrystals of Bexarotene (BEX).
  • The methodology used for the experiment includes developing a GCN-based deep learning model, screening 109 coformer candidates, and validating the top 30 candidates experimentally. The crystal structures were determined by single-crystal X-ray diffraction, and pharmacokinetic studies were conducted to evaluate the plasma exposures of the new cocrystals. The study aims to achieve the primary objective of discovering new cocrystals of BEX with superior solubility and dissolution over the parent drug.

Comparative Analysis & Findings

  • The results show that the GCN-based deep learning model (CocrystalGCN) obtained high performance relative to baseline models. The top 30 of 109 coformer candidates were scored by CocrystalGCN and then validated experimentally. Cocrystals of BEX-pyrazine, BEX-2,5-dimethylpyrazine, BEX-methyl isonicotinate, and BEX-ethyl isonicotinate were successfully obtained. The crystal structures were determined by single-crystal X-ray diffraction. Powder X-ray diffraction, differential scanning calorimetry, and thermogravimetric analysis were utilized to characterize these multi-component forms. All cocrystals present superior solubility and dissolution over the parent drug. The pharmacokinetic studies show that the plasma exposures (AUC) of BEX-pyrazine and BEX-2,5-dimethylpyrazine are 1.7 and 1.8 times that of the commercially available BEX powder, respectively.

Implications and Future Directions

  • The study's findings demonstrate the potential of GCN-based deep learning models for in-silico screening of cocrystals of water-insoluble drugs. The results highlight the importance of integrating virtual prediction and experimental screening to discover new cocrystals with improved solubility and dissolution. The study suggests future research directions to explore the use of GCN-based deep learning models for the prediction of other drug-drug interactions and the optimization of drug formulations. The study also emphasizes the need for further research to evaluate the clinical efficacy and safety of the new cocrystals.
Read Full Article