BRP, an AI-discovered peptide, suppressed appetite in mice and pigs without the nausea common to GLP-1 drugs. The Nature study is 16 months old, and no human trial has begun.
Stanford Medicine researchers built an algorithm, Peptide Predictor, to scan the human genome for prohormones — inert proteins that only become active once they are cut into smaller peptide hormones — and used it to map more than 2,600 candidate fragments from 373 proteins.1 Of 100 synthesised and tested on lab-grown neurons, one 12-amino-acid fragment stood out: it raised neuronal activity roughly tenfold over untreated cells, against a threefold rise for GLP-1 itself, the hormone semaglutide mimics.2 The team named the fragment BRP, after its parent protein BRINP2, and traced its action to the hypothalamus, working independently of the leptin, GLP-1-receptor and MC4R pathways most existing appetite drugs rely on.1
In lean mice and minipigs — chosen because their eating patterns and metabolism sit closer to a human's than a mouse's alone — a single injection of BRP before feeding cut food intake by up to 50% within the next hour.1 Obese mice given daily doses for two weeks lost about 3g, almost entirely fat, while untreated mice gained roughly the same amount over the same period; the treated mice also showed better glucose and insulin tolerance.2 Researchers reported no difference in the animals' movement, water intake or fecal output — their proxies for the nausea and food aversion that limit GLP-1 drugs.1
BRP's animal numbers are real. So is its age. The strip below runs from the day Stanford submitted the study to the day this piece was filed — and the drug has not moved off it.1
The study took nearly a year and a half in peer review — submitted in August 2023, accepted in January 2025 — and has now sat published for about as long again without a single reported human dose.1 Katrin Svensson, the paper's senior author, has co-founded a company, Merrifield Therapeutics, that said in 2025 it planned to launch human trials “in the near future”; no trial has since been registered or dated.2 Svensson and lead author Laetitia Coassolo are named inventors on patents covering BRP, a financial interest the paper discloses alongside its results.1
The side-effect comparison is the commercial hook, and it has a real base rate behind it: in a 2025 RAND survey of nearly 8,800 US adults, about half of people who had used a GLP-1 drug reported nausea and a third reported diarrhoea.3 But the published BRP experiments were run against native GLP-1 peptide and an older GLP-1 drug, liraglutide — not against semaglutide itself — so the “no nausea” finding is one step removed from the branded drug it is being measured against in headlines, including this one.1
Three things aren't yet established. Whether BRP is safe or effective in a human body — every published result is from mice and minipigs.1 What receptor it binds to: the pathway is real, but the target protein is still unidentified.1 And when, or whether, a human trial will start — Merrifield Therapeutics has not announced a filing or a date since the 2025 paper.2