MIT researchers have developed an AI-based model to design nanoparticles for more efficient RNA vaccine delivery, promising faster development of critical treatments for metabolic and viral diseases.
Researchers at MIT have pioneered a new approach using artificial intelligence to design more efficient nanoparticles for delivering RNA vaccines and therapies. This groundbreaking method could significantly expedite the creation of new RNA-based treatments for a variety of diseases, including obesity and diabetes.
By training a machine-learning model to analyze thousands of existing delivery particles, the team at MIT has developed a predictive system to identify and create nanoparticles that perform better than current ones.
These nanoparticles can potentially enhance the efficacy of RNA vaccines and other RNA therapies by ensuring better delivery to cells and protecting the RNA from degradation.
“What we did was apply machine-learning tools to help accelerate the identification of optimal ingredient mixtures in lipid nanoparticles to help target a different cell type or help incorporate different materials, much faster than previously was possible,” senior author Giovanni Traverso, an associate professor of mechanical engineering at MIT and a gastroenterologist at Brigham and Women’s Hospital, said in a news release.
The study, published in Nature Nanotechnology, was spearheaded by Alvin Chan, a former MIT postdoc and now an assistant professor at Nanyang Technological University, and Ameya Kirtane, a former MIT postdoc and now an assistant professor at the University of Minnesota.
This innovative research has the potential to revolutionize how RNA therapies are developed and delivered.
Advancing RNA Delivery with Machine Learning
RNA vaccines, like those developed for SARS-CoV-2, are typically delivered using lipid nanoparticles (LNPs). These particles are essential for protecting mRNA from being broken down in the body and facilitating its entry into cells.
Traverso’s team developed a novel AI model named COMET to enhance the efficiency of these delivery systems.
“Most AI models in drug discovery focus on optimizing a single compound at a time, but that approach doesn’t work for lipid nanoparticles, which are made of multiple interacting components,” added Chan. “To tackle this, we developed a new model called COMET, inspired by the same transformer architecture that powers large language models like ChatGPT. Just as those models understand how words combine to form meaning, COMET learns how different chemical components come together in a nanoparticle to influence its properties — like how well it can deliver RNA into cells.”
To create the model, the researchers generated a library of approximately 3,000 LNP formulations. They tested these formulations in the lab to determine their efficiency in delivering RNA to cells and used the data to train the machine-learning model.
When they tested the AI’s predicted formulations, they found that they outperformed existing LNPs, including commercially used ones.
Broadening the Horizons
The researchers also experimented with adding new materials, such as branched poly-beta-amino esters (PBAEs), into the LNPs.
They found that these polymers can successfully deliver nucleic acids on their own.
Moreover, the AI model was able to predict which nanoparticles would best deliver RNA to different types of cells, including colorectal cancer cells, and which would withstand lyophilization, a freeze-drying process used to extend the shelf-life of medicines.
“This is a tool that allows us to adapt it to a whole different set of questions and help accelerate development,” Traverso added. “We did a large training set that went into the model, but then you can do much more focused experiments and get outputs that are helpful on very different kinds of questions.”
The implications of this research are significant. By streamlining the process of developing RNA therapies, this technology could pave the way for new treatments for various diseases, with enhanced precision and speed.
Future Directions
The team continues to work on integrating these advancements into potential treatments for diabetes and obesity as part of a project funded by the U.S. Advanced Research Projects Agency for Health (ARPA-H).
This approach could be particularly transformative for developing therapeutics such as GLP-1 mimics, akin to those used in treatments like Ozempic.