Redazione RHC : 4 October 2025 21:29
Antibiotics for inflammatory bowel disease can be a double-edged sword. While they suppress inflammation, they also kill beneficial bacteria, not just harmful ones. This often worsens symptoms. In this situation, generic drugs prove to be a less effective tool.
Scientists at the Massachusetts Institute of Technology (MIT) and McMaster University have reported the discovery of a new molecule called enterololin . It works selectively, suppressing bacteria associated with Crohn’s disease (IBD) flare-ups, while leaving the rest of the gut microbiome largely unaffected.
To understand its mechanism of action, the researchers used the generative AI model DiffDock , which reduced the time from research to analysis to a few months, instead of the usual years.
In experiments on mice, the drug selectively suppressed Escherichia coli, which increases inflammation, and helped the animals recover more quickly. Furthermore, the microbiome remained significantly healthier than treatment with the standard antibiotic vancomycin.
DiffDock predicted that the molecule binds to the LolCDE protein complex, responsible for lipoprotein transport in bacteria . This clue allowed the researchers to test their hypothesis in the laboratory: they bred enterololin-resistant E. coli strains, conducted RNA sequencing, and used CRISPR to confirm the mechanism of action.
All experiments demonstrated that the drug effectively disrupts the pathways associated with lipoprotein transport, as predicted by the AI.
The authors note that traditionally, research into the mechanisms of action of new antibiotics takes up to two years and costs millions of dollars. In this case, artificial intelligence has reduced the time to six months, significantly reducing costs.
Stoked Bio, founded by one of the study’s authors, is currently further developing enterololin and preparing it for human trials . Research is also underway on derivatives of the molecule against other dangerous pathogens, including Klebsiella pneumoniae .
The promise of targeted antibiotics is particularly important for patients with Crohn’s disease and other inflammatory bowel diseases: new drugs could reduce symptoms without altering the microbiome . On a larger scale, such developments could provide an answer to the growing threat of bacterial resistance to existing antibiotics.
Scientists emphasize that it’s not just a single molecule that matters. The approach itself is crucial: a combination of artificial intelligence and laboratory methods allows for rapid clarification of the workings of potential drugs. This could transform the process of discovering new drugs for many diseases.
The researchers published the sequencing data to public repositories and made the DiffDock-L code open source on GitHub.
Redazione