AI Unleashes Antibiotic to Wipe Away Drug-Resistant Bacteria

MIT and McMaster University researchers have utilized an artificial intelligence algorithm to identify a new antibiotic that has the potential to eliminate drug-resistant bacteria responsible for various diseases. This breakthrough could be particularly beneficial in combating Acinetobacter baumannii, a type of bacteria commonly found in hospitals and known to cause pneumonia, meningitis, and other dangerous infections. In regions like Iraq and Afghanistan, it is a major source of infection among wounded soldiers (1^✔ ✔Trusted Source
Acinetobacter baumannii: emergence of a successful pathogen

Go to source).

Jonathan Stokes, an assistant professor of biochemistry and biomedical sciences at McMaster University, explained that Acinetobacter can survive on hospital surfaces for extended periods and acquire antibiotic-resistance genes from its surroundings.

He emphasized that it is increasingly common to encounter A. baumannii isolates resistant to almost all available antibiotics.

Researchers Uncover Promising Antibiotic to Combat Drug-Resistant

The researchers employed machine learning and a library of approximately 7,000 potential medicinal compounds to select the new antibiotic.

They trained the model to determine whether a chemical molecule could suppress the spread of A. baumannii. The development of new antibiotics has lagged behind the growing resistance of harmful bacteria to existing medicines in recent decades.

Collins, Stokes, and MIT Professor Regina Barzilay began addressing this challenge by utilizing machine learning to identify new antibiotics with distinct chemical structures.

They successfully used a machine-learning system to discover a molecule named halicin, which demonstrated effectiveness against not only E. coli but also several other drug-resistant bacterial species.

Advertisement

Following this achievement, the researchers shifted their focus to combatting Acinetobacter, which they considered a top priority in the fight against multidrug-resistant bacterial infections. They trained their computer model by subjecting A. baumannii to over 7,500 chemicals to identify those that hindered its growth.

The model learned the chemical characteristics associated with growth inhibition based on the arrangements of each molecule and whether it limited bacterial growth.

Trained Algorithm Helps in Antibiotic Search

The trained algorithm was then used to analyze 6,680 chemicals from the Broad Institute's Drug Repurposing Hub, resulting in a selection of several promising antibiotics. Out of these, nine antibiotics were identified, one of which exhibited significant potency.

Advertisement

Initially explored as a potential treatment for diabetes, this molecule, named abaucin, demonstrated exceptional effectiveness against A. baumannii while leaving other bacteria species unaffected.

The researchers highlighted the importance of abaucin's "narrow spectrum" killing power, as it reduces the likelihood of bacteria developing resistance and preserves beneficial microbes in the human gut while suppressing opportunistic infections like Clostridium difficile.

In mouse experiments, abaucin proved effective in treating A. baumannii-induced wound infections, while lab experiments demonstrated its efficacy against drug-resistant A. baumannii strains obtained from human patients.

The medication was found to interact with lipoprotein trafficking, a mechanism involved in protein movement within cells. It specifically blocked the protein LolE, which plays a role in this process.

The researchers were surprised by abaucin's specificity in targeting A. baumannii while being present in all Gram-negative bacteria. They speculate that A. baumannii may have slight variations in lipoprotein trafficking, leading to the narrow spectrum activity observed.

Currently, Stokes' lab is collaborating with other McMaster researchers to optimize the therapeutic characteristics of abaucin, with the ultimate goal of making it available for patient use.

Moving forward, the researchers plan to apply their modeling approach to identify potential treatments for other drug-resistant infections caused by bacteria such as Staphylococcus aureus and Pseudomonas aeruginosa.

Reference:

1. Acinetobacter baumannii: emergence of a successful pathogen - (https://pubmed.ncbi.nlm.nih.gov/18625687/)

Source-Medindia