AI Discovers Powerful Antibiotic in Old Diabetes Drug

A drug once intended to treat diabetes, Halicin, is showing new promise as a potent weapon against multidrug-resistant (MDR) bacteria, thanks to the capabilities of artificial intelligence (AI).

A recent study published in Antibiotics reveals that Halicin, rediscovered using machine learning techniques, is capable of inhibiting 17 out of 18 MDR bacterial strains tested, positioning it as a potential broad-spectrum antibiotic in the ongoing battle against superbugs.

The Superbug Crisis and AI’s Role

“Superbugs” bacteria resistant to multiple antibiotics pose a growing threat to global health. The ESKAPE pathogens (Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumannii, Pseudomonas aeruginosa, and Enterobacter spp.) have been labeled by the World Health Organization (WHO) as top priority threats due to their resistance to nearly all existing treatments.

Traditional antibiotic development is a slow, expensive process. But AI is transforming this paradigm by rapidly scanning existing drug libraries to uncover hidden therapeutic potentials.

This modern approach has now resurrected Halicin, a compound originally developed to inhibit diabetes-related enzymes, revealing it to have a novel antibacterial mechanism: it disrupts the bacterial proton-motive force, which is critical to energy production. Moreover, it is a method that bypasses most known resistance strategies.

First-of-Its-Kind Study in Morocco

The new study, the first of its kind in Morocco, evaluated Halicin’s effectiveness against 18 clinically validated MDR strains gathered from Moroccan hospitals. Researchers used minimum inhibitory concentration (MIC) assays to determine the lowest dose of Halicin required to stop bacterial growth.

Standard strains of Staphylococcus aureus and Escherichia coli served as control benchmarks. The team followed established EUCAST and CLSI guidelines for bacterial susceptibility testing, and utilized scanning electron microscopy (SEM) to observe how Halicin affected bacterial structure, particularly in E. coli.

Key Findings

• Halicin inhibited 94% of tested MDR strains, including standard S. aureus and E. coli.
• MIC values for clinical MDR isolates (excluding one) ranged between 32 and 64 µg/mL.
• The control strains showed MICs of 16 µg/mL (E. coli) and 32 µg/mL (S. aureus).
• One major outlier was Pseudomonas aeruginosa, which showed complete resistance to Halicin. Researchers attributed this to its strong outer membrane, which likely blocks Halicin penetration.

What sets Halicin apart is its unique method of bacterial attack. Rather than targeting the cell wall or protein synthesis, as most antibiotics do, Halicin interferes with bacterial energy production, potentially making it harder for bacteria to evolve resistance. “You can make really good stuff – fast,” says researcher Samir Mallal, emphasizing AI’s ability to speed up innovation without sacrificing quality.

Future Outlook

While Halicin’s results are promising, researchers caution that further studies are needed, especially regarding its pharmacokinetics, toxicity, and performance in living organisms. The paper calls for continued monitoring to track any emergence of resistance as development advances.

This breakthrough also underscores the power of AI to revitalize dormant drugs and fill critical gaps in modern medicine. With antibiotic resistance on the rise and the traditional drug pipeline struggling, Halicin represents a new frontier in the AI-led search for novel treatments.

By reimagining an old anti-diabetic drug with the help of cutting-edge AI, scientists may have found an effective new tool in the global fight against superbugs. This study demonstrates that AI-powered drug discovery is not just a theoretical concept but a practical solution to one of modern medicine’s most urgent challenges.