A novel approach developed by researchers at MIT, Brigham and Women’s Hospital, and Duke University helps identify the transporters used by various drugs to pass through the digestive tract, thus enhancing patient treatment. The method uses both tissue models and machine-learning algorithms. This can play an instrumental role in mitigating possible drug interference that occurs when two drugs rely on the same transporter.
Developing an understanding of transporters involved in the journey of drugs through the digestive tract could benefit drug developers, as they could improve the absorption of new drugs by incorporating ingredients that bolster their interaction with transporters. The recently conducted study revealed interference between a popular antibiotic and a blood thinner. The team adopted a tissue model they created earlier to measure a drug’s absorbability. They systematically exposed the laboratory-grown pig intestinal tissue to different drug formulations and determined their absorption.
The study involved 23 frequently used drugs, helping researchers determine the transporters each of these drugs used. Following this, a machine-learning model was trained on that data to predict which drugs would interact with which transporters based on the chemical structures of the drugs. The researchers analyzed 28 currently used drugs and 1,595 experimental drugs, yielding nearly two million predictions of potential drug interactions.
The predictions were tested using data from about 50 patients taking one of three drugs when prescribed doxycycline, an antibiotic. The data confirmed the model’s predictions. Findings showed that when doxycycline was given to patients already taking warfarin, a common blood-thinner, the level of warfarin in the patient’s bloodstream increased. The data also confirmed that the absorption of doxycycline is affected when paired with three other specific medications. The researchers asserted that this technology could predict the safety implications of administering these drugs together. The same approach could also be applied to drugs currently in development.
