Structure-Based Prediction of Molecular Interactions for Stabilizing Volatile Drugs

Area of Science:

• Pharmaceutical Sciences
• Computational Chemistry
• Materials Science

Background:

• Volatile drugs face clinical limitations due to high evaporation rates.
• Traditional excipient screening is inefficient and resource-intensive.
• An AI-driven approach is needed to optimize drug-excipient interactions.

Purpose of the Study:

• To develop an AI strategy for efficient screening of drug-excipient interactions.
• To identify suitable excipients for stabilizing volatile drugs like d-borneol.
• To create stable nano-formulations for volatile pharmaceuticals.

Main Methods:

• Utilized Protenix structure prediction for high-throughput simulation of d-borneol and 472 excipients.
• Selected top 50 candidates based on simulation, followed by molecular docking and stability experiments.
• Validated AI predictions against experimental data.

Main Results:

• AI predictions showed high consistency with experimental validation.
• Soybean phospholipid (PC) emerged as the optimal excipient for d-borneol.
• PC-based lyophilized liposomal formulations significantly reduced d-borneol volatilization and improved stability.

Conclusions:

• This study pioneers the use of structure prediction models for volatile drug excipient screening.
• Successfully developed a stable nano-formulation for d-borneol using AI-identified excipients.
• Offers a novel paradigm for volatile pharmaceutical nano-formulation development.