Machine Learning Uncovers Food- and Excipient-Drug Interactions

Area of Science:

• Pharmacology
• Drug Metabolism
• Computational Chemistry

Background:

• Inactive ingredients, considered safe, may possess unknown biological effects impacting drug efficacy.
• Proteins like P-glycoprotein (P-gp) and UGT2B7 are crucial for drug pharmacokinetics, affecting approximately 20% of FDA-approved drugs.
• Understanding these interactions is vital for drug safety and formulation development.

Purpose of the Study:

• To develop a machine learning platform for predicting unknown biological effects of inactive ingredients.
• To identify specific inactive ingredients that interact with P-glycoprotein (P-gp) and Uridine diphosphate-glucuronosyltransferase-2B7 (UGT2B7).

Main Methods:

• Application of state-of-the-art machine learning algorithms.
• In silico, in vitro, ex vivo, and in vivo validation studies.
• Focus on predicting interactions with P-glycoprotein (P-gp) and Uridine diphosphate-glucuronosyltransferase-2B7 (UGT2B7).

Main Results:

• Identification of vitamin A palmitate as a P-glycoprotein (P-gp) inhibitor.
• Identification of abietic acid as a Uridine diphosphate-glucuronosyltransferase-2B7 (UGT2B7) inhibitor.
• Experimental validations confirmed the predicted interactions.

Conclusions:

• The predictive framework successfully elucidates biological effects of commonly consumed inactive ingredients.
• Findings have significant implications for understanding food- and excipient-drug interactions.
• The platform supports the development of functional drug formulations.