Scientists use AI to identify new antibiotic that could fight drug-resistant infections
Researchers from MIT and McMaster University have utilized an artificial intelligence algorithm to discover a novel antibiotic capable of eliminating a strain of bacteria responsible for numerous drug-resistant infections. If further developed for patient use, this medication could prove instrumental in combating Acinetobacter baumannii, a bacterial species frequently found in hospitals and known to cause severe infections such as pneumonia and meningitis. Notably, it is also a leading cause of infections among injured soldiers in Iraq and Afghanistan.
Acinetobacter, known for its ability to persist on hospital surfaces and acquire antibiotic resistance genes from its surroundings, has become increasingly prevalent in strains that display resistance to nearly all available antibiotics. Jonathan Stokes, previously a postdoctoral researcher at MIT and now an assistant professor at McMaster University, explains the commonality of such resistant A. baumannii isolates.
Employing a machine-learning model, the scientists screened a vast library of approximately 7,000 potential drug compounds to identify this newly discovered medication. The model was trained to assess whether a given chemical compound could hinder the growth of A. baumannii.
''This finding further supports the premise that AI can significantly accelerate and expand our search for novel antibiotics. I'm excited that this work shows that we can use AI to help combat problematic pathogens such as A. baumannii."
Drug discovery
In recent decades, the rise of antibiotic-resistant bacteria has outpaced the development of new antibiotics, posing a significant challenge. To address this issue, Collins, Stokes, and MIT Professor Regina Barzilay embarked on a mission several years ago, leveraging machine learning—an artificial intelligence technique capable of recognizing patterns within vast datasets. The researchers aimed to employ this approach in identifying novel antibiotics with distinct chemical structures compared to existing drugs.
In their initial endeavor, the team trained a machine-learning algorithm to identify chemical structures that could impede the growth of E. coli. After screening over 100 million compounds, the algorithm successfully yielded a molecule named halicin, inspired by the fictional AI system in "2001: A Space Odyssey." Halicin demonstrated the ability to kill not only E. coli but also various other antibiotic-resistant bacterial species.
Encouraged by this breakthrough, the researchers turned their attention to Acinetobacter, which they consider a major threat in multidrug-resistant bacterial infections. To train their computational model, the team exposed A. baumannii, cultivated in a laboratory dish, to approximately 7,500 different chemical compounds. They assessed the inhibitory effects of these compounds on the microbe and provided the model with information regarding their chemical structures and growth inhibition capabilities. Consequently, the algorithm learned the chemical features associated with growth inhibition.
Once the model was trained, the researchers employed it to analyze a separate set of 6,680 compounds from the Drug Repurposing Hub at the Broad Institute, which took less than two hours. This analysis produced several hundred top candidate compounds. Among these, the researchers selected 240 compounds for experimental testing in the laboratory. Their focus was on compounds with chemical structures distinct from existing antibiotics or the molecules used in training the model.
The experimental tests yielded nine antibiotics, including one with remarkable potency. Originally investigated as a potential treatment for diabetes, this particular compound proved highly effective in eradicating A. baumannii while demonstrating no impact on other bacterial species such as Pseudomonas aeruginosa, Staphylococcus aureus, and carbapenem-resistant Enterobacteriaceae.
The antibiotic's "narrow spectrum" capability, limited to specific targets, is advantageous as it reduces the likelihood of bacteria rapidly developing resistance. Additionally, the drug is expected to spare the beneficial bacteria in the human gut that play a role in suppressing opportunistic infections like Clostridium difficile. This feature is crucial since antibiotics are often administered systemically, and it is undesirable to disrupt the balance of the gut microbiota in already vulnerable patients, potentially leading to secondary infections.
A novel mechanism
Through experiments conducted on mice, the researchers successfully demonstrated the effectiveness of the newly discovered drug, abaucin, in treating wound infections caused by A. baumannii. Additionally, laboratory tests exhibited its efficacy against various drug-resistant strains of A. baumannii obtained from human patients.
Further investigations into the drug's mechanism of action revealed that it disrupts a cellular process called lipoprotein trafficking, which is responsible for transporting proteins from the interior of the cell to its outer envelope. Specifically, abaucin appears to inhibit the activity of LolE, a protein crucial to this process.
Interestingly, while all Gram-negative bacteria express this enzyme, the researchers were surprised by abaucin's selectivity towards A. baumannii. They speculate that slight differences in how A. baumannii performs lipoprotein trafficking may explain this narrow spectrum activity.
Jonathan Stokes and his team are currently collaborating with McMaster researchers to optimize the compound's medicinal properties, aiming to develop it for future use in patients.
Moreover, the researchers intend to apply their modeling approach to identify potential antibiotics for other types of drug-resistant infections, including those caused by Staphylococcus aureus and Pseudomonas aeruginosa.
Funding for this research was provided by several sources, including the David Braley Center for Antibiotic Discovery, the Weston Family Foundation, the Audacious Project, the C3.ai Digital Transformation Institute, the Abdul Latif Jameel Clinic for Machine Learning in Health, the DTRA Discovery of Medical Countermeasures Against New and Emerging Threats program, the DARPA Accelerated Molecular Discovery program, the Canadian Institutes of Health Research, Genome Canada, the Faculty of Health Sciences of McMaster University, the Boris Family, a Marshall Scholarship, and the Department of Energy Biological and Environmental Research program.
Comentários