Researchers from the Massachusetts Institute of Technology (MIT), Brigham and Women’s Hospital, and Duke University have developed an innovative strategy to identify the transporter proteins used by different drugs in the body’s gastrointestinal (GI) tract. The method, which employs tissue models and machine-learning algorithms, aims to improve drug administration by enabling predictions of drug interactions and potential toxicity.
Transporter proteins in the GI tract are responsible for facilitating the passage of orally administered drugs. However, it is not well known which specific transporters most drugs utilize to exit the GI tract. It’s crucial to identify these interactions to avoid drug combinations that rely on the same transporters, which could lead to interference and potentially reduce treatment effectiveness.
The research team’s strategy involved examining the interaction of drugs with three key transporters, namely BCRP, MRP2, and PgP. Using an adapted tissue model based on pig intestinal tissue, the team explored the role of individual transporters by manipulating the transporters’ expression with short strands of RNA, known as siRNA.
The methods allowed for testing of 23 commonly used drugs, leading to the identification of the transporters each drug uses. Subsequently, the findings were used to train a machine-learning model, which then made predictions about which drugs would interact with specific transporters.
The researchers were able to predict potential interactions using this model, one of which suggested that an antibiotic called doxycycline could interfere with blood thinner warfarin, as well as several other drugs. Retrospective patient database data confirmed these predictions, demonstrating that warfarin levels in patients’ bloodstream increase when they are also taking doxycycline and return to normal once doxycycline treatment is ceased.
The research team asserts that understanding these interactions has significant safety implications for drug administration and provides an opportunity to develop effective drug models. Additionally, this approach could be used in the development of new drugs, allowing developers to adjust the composition of new drug molecules in order to avoid interactions or enhance absorbability.
A biotech firm named Vivtex, co-founded by researchers involved in the study, is currently exploring potential applications for drug tuning. This would make it possible to anticipate and avoid harmful drug interactions or enhance the absorption rates of new drugs during development. The study was funded in part by the U.S. National Institutes of Health, the Department of Mechanical Engineering at MIT, and the Division of Gastroenterology at Brigham and Women’s Hospital.
