Since the 1970s, finding new antibiotics has been challenging. The World Health Organization now considers the antimicrobial resistance crisis as one of the top 10 global public health threats. Bacteria can become resistant to antibiotics, especially when an infection is treated repeatedly. Some bacteria become metabolically inert, avoiding detection by antibiotics that only respond to metabolic activity, and can reawaken once the threat has passed. The recurring infections are due to this bacteria dormancy.
Researchers are attempting to overcome this problem through artificial intelligence (AI). Jackie Valeri, a researcher from the Massachusetts Institute of Technology’s Abdul Latif Jameel Clinic for Machine Learning in Health, is leading this new exploration. Valeri recently published a study in Cell Chemical Biology demonstrating how machine learning could help identify compounds that can target dormant bacteria.
This research builds on discoveries of ancient bacterial strains that have survived in an energy-saving state for millions of years. One of the key challenges is finding antibiotics that can kill metabolically dormant bacteria. According to a Lancet study, in 2019, about 1.27 million deaths could potentially have been prevented if the relevant infections were susceptible to drugs.
Valeri’s team is using AI technology to speed up the identification process. They used AI to screen millions of molecules, a process that would traditionally take years. One such molecule was the compound semapimod, which they identified in just a weekend. Semapimod is an anti-inflammatory drug usually used for Crohn’s disease. Scientists found that it was also effective against dormant Escherichia coli and Acinetobacter baumannii.
Researchers also discovered semapimod’s ability to disrupt the thicker, less penetrable cell walls of “Gram-negative” bacteria, which are often resistant to antibiotics. Examples of such bacteria include E. coli, A. baumannii, Salmonella, and Pseudomonis. Semapimod’s structure allows it to target the outer membrane of these bacteria, increasing their sensitivity to drugs generally effective only against Gram-positive bacteria. This finding is significant as there is now a potential new way to treat Gram-negative infections.
This innovative approach to using AI in the discovery of new antibiotics holds potential for significantly expanding the currently available arsenal of antibiotics. The use of AI to hasten the screening process reduces time and resources spent and could potentially save countless lives by identifying new ways to combat antibiotic resistance. Valeri’s work using AI technologies is pioneering this new frontier in antibiotic discovery.
