Researchers from The University of Texas at Austin have made a groundbreaking advancement in combating antibiotic resistance by developing a new antibiotic with the help of artificial intelligence.
Published in Nature Biomedical Engineering, this research represents a major step forward in creating safer and more effective antibiotics.
The team used a large language model (LLM), similar to the technology behind ChatGPT, to re-engineer Protegrin-1, a potent antibiotic naturally produced by pigs. While Protegrin-1 was effective against bacteria, it was previously too toxic for human use. By modifying Protegrin-1, the researchers aimed to retain its antibacterial properties while reducing its toxicity to human cells.
The researchers generated over 7,000 versions of Protegrin-1 using a high-throughput technique, which allowed them to rapidly identify modifications that could improve safety. The LLM was then used to assess these variations for their effectiveness in targeting bacterial membranes, killing bacteria, and avoiding harm to human red blood cells. This AI-guided approach resulted in the creation of a refined antibiotic known as bacterially selective Protegrin-1.2 (bsPG-1.2).
Preliminary trials in mice showed that bsPG-1.2 significantly reduced bacterial levels in their organs within six hours, even in cases of multidrug-resistant infections. These promising results suggest that bsPG-1.2 may soon be ready for human trials.
Claus Wilke, a professor of integrative biology and co-senior author of the study, emphasized the transformative role of AI in drug development.
“Large language models are revolutionizing protein and peptide engineering, making it possible to develop new drugs and enhance existing ones more efficiently. This technology not only identifies new treatment options but also accelerates their progression to clinical use,” Wilke stated.
This breakthrough highlights the growing role of AI in addressing significant health challenges.