We've updated our Privacy Policy to make it clearer how we use your personal data. We use cookies to provide you with a better experience. You can read our Cookie Policy here.


Model Identifies Potential Drug Combinations To Avoid

A person pouring a bottle of pills into their hand.
Credit: Towfiqu barbhuiya/Unsplash
Listen with
Register for free to listen to this article
Thank you. Listen to this article using the player above.

Want to listen to this article for FREE?

Complete the form below to unlock access to ALL audio articles.

Read time: 2 minutes

A new model has been designed to identify drugs that should not be prescribed together. It has already identified common drugs – an antibiotic and a blood thinner – that may interfere with each other.

Avoiding adverse drug interactions

The drugs we take are mostly absorbed into the body through the digestive system. Here, drugs bind to transporter proteins found on the outside of cells – but in most cases, we don’t know the specific transporters that drugs use to pass through the lining of the digestive tract.

If two drugs share the same transporter, they may interfere with each other and therefore shouldn’t be prescribed together. Trying to unravel the exact transporter each drug uses could help to inform clinicians as to which drugs can be used in combination safely ‒ and which should be avoided. This information may also help to improve the absorption of new drugs by influencing interactions with their transporters.

Now, a new model has combined lab-grown tissue and machine learning to screen and predict potentially toxic interactions, potentially helping to make drugs safer and more effective.

The study – a collaboration between MIT, Brigham and Women’s Hospital and Duke University – is published in Nature Biomedical Engineering.

Porcine tissue model and machine learning combined

“One of the challenges in modeling absorption is that drugs are subject to different transporters. This study is all about how we can model those interactions, which could help us make drugs safer and more efficacious, and predict potential toxicities that may have been difficult to predict until now,” said Giovanni Traverso, an associate professor of mechanical engineering at MIT, gastroenterologist at Brigham and Women’s Hospital, and the study’s senior author.

The researchers focused on a set of common drug transporters called BCRP, MRP2 and PgP. To study their roles in drug absorption, each was removed – or “knocked down” – in an experimental tissue model.

Their tissue model, developed in 2020, is used to measure drug absorption. The researchers expose lab-grown pig tissue to different drug formulations to measure how well they are absorbed.

The expression of each was knocked down using strands of short interfering RNA that prevent proteins from being produced from specific genes. Different combinations of genes were knocked out in various areas of the tissue model to study how each transporter interacted with different drugs.

“There are a few roads that drugs can take through tissue, but you don't know which road. We can close the roads separately to figure out, if we close this road, does the drug still go through? If the answer is yes, then it’s not using that road,” Traverso said.

The model enabled them to screen 23 existing drugs as well as over 1,500 experimental drugs, uncovering nearly 2 million predicted drug interactions. For example, the antibiotic doxycycline was predicted to interfere with the blood thinner warfarin, the heart failure drug digoxin, the anti-seizure drug levetiracetam and the immunosuppressant tacrolimus.

Next, the researchers tested these predictions by examining data from 50 patients taking one of these four drugs who then were prescribed doxycycline. For warfarin, blood warfarin levels increased in patients while they took doxycycline, but then decreased again after they stopped. Patient data also confirmed the model’s predictions for digoxin, levetiracetam and tacrolimus while taking doxycycline – despite suspicions only being raised previously for tacrolimus.

Applications for new and existing drugs

“These are drugs that are commonly used, and we are the first to predict this interaction using this accelerated in silico and in vitro model,” Traverso said. “This kind of approach gives you the ability to understand the potential safety implications of giving these drugs together.”

This method could be applied to drugs currently in development as well as existing drugs to identify potential interactions. Drug developers could therefore fine-tune new drugs to avoid interactions with other drugs or improve absorption.

Reference: Shi Y, Reker D, Byrne JD, et al. Screening oral drugs for their interactions with the intestinal transportome via porcine tissue explants and machine learning. Nat Biomed Eng. 2024:1-13. doi: 10.1038/s41551-023-01128-9

This article is a rework of a press release issued by MIT. Material has been edited for length and content.