University of Pennsylvania AI System Finds 53 Effective Antibiotics from Venom Peptides

University of Pennsylvania researchers have achieved a significant breakthrough in the fight against antibiotic resistance, utilizing an artificial intelligence (AI) system named APEX to identify and validate potent new antibiotic candidates derived from animal venoms. The groundbreaking study, published in Nature Communications, details how this innovative approach rapidly screened millions of peptides, offering a promising new avenue for drug discovery. This development comes at a critical time, as drug-resistant infections contribute to approximately 5 million deaths globally each year.

The AI system, APEX, was deployed to analyze over 40 million venom encrypted peptides (VEPs), which are proteins evolved over millions of years for attack and defense. These VEPs were sourced from a comprehensive database of 16,123 venom proteins from various animals, including snakes, spiders, and scorpions. According to the researchers, APEX was able to process this immense chemical space and identify potential candidates within a matter of hours.

From this vast pool, APEX pinpointed 386 peptides exhibiting the molecular signatures of next-generation antibiotics. A selection of 58 of these identified peptides were then synthesized for laboratory testing. Remarkably, 53 of the synthesized peptides, representing 91.4% of those tested, successfully eliminated drug-resistant bacteria such as E. coli and Staphylococcus aureus without causing harm to human red blood cells.

The platform also unveiled more than 2,000 entirely new antibacterial motifs—short, specific sequences of amino acids responsible for their ability to kill or inhibit bacterial growth. > "Venoms are evolutionary masterpieces, yet their antimicrobial potential has barely been explored," stated senior author César de la Fuente, PhD. Co-author Changge Guan, PhD, noted, > "The platform mapped more than 2,000 entirely new antibacterial motifs - short, specific sequences of amino acids within a protein or peptide responsible for their ability to kill or inhibit bacterial growth." The study indicated that these venom-derived peptides primarily act by disrupting bacterial cytoplasmic membranes, a mechanism distinct from many traditional antibiotics.

Further validating their findings, lead peptide candidates demonstrated anti-infective activity in preclinical mouse models of Acinetobacter baumannii skin infection, showing significant reduction in bacterial loads with minimal observed toxicity. > "By pairing computational triage with traditional lab experimentation, we delivered one of the most comprehensive investigations of venom derived antibiotics to date," added co-author Marcelo Torres, PhD. The University of Pennsylvania team plans to continue optimizing these promising peptide candidates through medicinal-chemistry adjustments to enhance their therapeutic potential.