The race to develop safer, more effective obesity medications has taken a computational turn. Researchers at Stanford Medicine have used an artificial intelligence algorithm to identify a naturally occurring peptide that suppresses appetite and promotes fat loss in animal studies—without triggering the gastrointestinal side effects that plague widely prescribed GLP-1 receptor agonists like semaglutide (Ozempic).

The discovery, published in the journal Nature, centers on a 12-amino-acid peptide called BRP, derived from a protein known as BRINP2. Unlike semaglutide and similar drugs that activate GLP-1 receptors distributed throughout the brain, gut, pancreas, and other tissues—causing widespread effects alongside therapeutic benefits—BRP appears to act selectively on the hypothalamus, the brain region governing hunger and metabolism.

AI Narrows 20,000 Genes to One Breakthrough Molecule

The Stanford team, led by senior author Katrin Svensson, PhD, assistant professor of pathology, and lead author Laetitia Coassolo, PhD, a senior research scientist, deployed an AI tool called Peptide Predictor to analyze approximately 20,000 human genes. The algorithm identified 373 prohormones and predicted 2,683 distinct peptides that could potentially influence metabolic pathways.

From this computationally filtered pool, researchers tested 100 peptides in laboratory nerve cells. BRP emerged as the standout candidate, increasing neuronal activity tenfold compared to controls—more than three times the response generated by GLP-1 itself.

The discovery validates a scalable approach to hormone therapeutics: using machine learning to reduce millions of potential candidate molecules to hundreds suitable for practical experimental testing, dramatically accelerating the drug discovery timeline.

Animal Studies Show Dramatic Appetite Suppression and Fat-Selective Weight Loss

In subsequent animal experiments, BRP reduced food intake by up to 50% in lean mice and minipigs within one hour of injection. Over a 14-day study period, obese mice treated with BRP lost an average of 3 grams, while control animals gained approximately 3 grams.

Critically, the weight loss came primarily from body fat rather than muscle tissue—a significant advantage over current GLP-1 agonists, which have raised concerns about muscle mass depletion during rapid weight reduction. Treated animals also demonstrated improved glucose and insulin tolerance, markers of metabolic health.

The peptide's mechanism appears to involve the prohormone convertase 1/3 enzyme pathway and targets the hypothalamus specifically, rather than the distributed receptor network engaged by GLP-1 drugs. This selectivity may explain why BRP produced appetite suppression without the nausea, constipation, and other digestive complications that cause many patients to discontinue semaglutide and related medications.

From Algorithm to Clinical Pipeline

The Stanford researchers have cofounded a company to advance BRP toward human clinical trials, though years of development remain before any potential regulatory approval. If human studies confirm the animal findings, BRP could offer physicians an alternative obesity treatment with a differentiated side effect profile in a market currently dominated by GLP-1 agonists.

For the AI-driven drug discovery field, the work demonstrates how computational approaches can identify novel therapeutic peptides hidden within the human genome—a proof of concept that may accelerate development of hormone-based treatments across multiple disease areas beyond obesity.

Sources

  1. AI helps Stanford scientists discover a "natural Ozempic" with a surprising advantage — ScienceDaily