BRP: The AI-Discovered 'Natural Ozempic' Peptide From Stanford (Nature 2026)

Important Notice: This article is intended exclusively for scientific information and research purposes. All substances mentioned are not intended for human consumption. Always consult qualified professionals before using peptides.

Introduction: A New Endogenous Appetite Hormone

On 12 April 2026, the laboratory of Katrin Svensson at Stanford University published in Nature the discovery of an entirely new human peptide hormone with appetite-suppressing properties. The compound, named BRP (BRINP2-related peptide), is a short 12-amino-acid fragment cleaved from the prohormone BRINP2 that already exists naturally in the human body. In preclinical studies in mice and Yucatán minipigs, a single dose reduced food intake by roughly half within an hour, without the gastrointestinal side effects that typify GLP-1 agonists.

The popular framing in StatNews and other outlets was immediate and almost unavoidable: a "natural Ozempic". For the research-peptide community the story matters for two reasons. First, BRP appears to operate through a mechanism distinct from every approved obesity drug. Second, its discovery was driven by an AI model (the Peptide Predictor) that screened the entire human prohormone proteome for therapeutically relevant cleavage products. Both elements signal a structural shift in how new peptide drugs may emerge over the next decade.

What Is BRP?

BRP stands for BRINP2-related peptide. It is a 12-amino-acid fragment derived by proteolytic cleavage from the precursor protein BRINP2, a member of the BMP (bone morphogenetic protein) and retinoic-acid-inducible neural-specific protein family. Three points are worth holding in mind:

Endogenous origin. BRP is not a synthetic analogue of an existing drug like semaglutide, but a peptide that the human body itself appears to produce in detectable quantities.
Short length. At 12 residues, BRP sits in the same size class as many bioactive endogenous peptides such as oxytocin, somatostatin or alpha-MSH. This length is therapeutically attractive: short enough for solid-phase synthesis, long enough to encode receptor specificity.
Distinct receptor pathway. The Svensson group reports that BRP does not act through the GLP-1 receptor. The exact receptor remains the subject of follow-up work, but the downstream signature is clearly hypothalamic.

BRP also expresses tissue-specifically: brain expression dominates, with limited peripheral signal. This peripheral silence is one of the reasons the initial side-effect profile in animals looks unusually clean.

How Was It Discovered?

The discovery story is what makes BRP genuinely novel. The Svensson lab built a neural network they call the Peptide Predictor and trained it to recognise sequence motifs in human prohormones that are likely to be cleaved into bioactive peptides. They then ran the model across roughly 20,000 human proteins, ranking candidate fragments by predicted bioactivity.

From that screen, around 100 candidate peptides emerged. The team then synthesised the top hits and tested them in cellular assays for activation of pathways relevant to metabolism and feeding. BRP rose to the top of the list because it produced strong, dose-dependent signalling in hypothalamic neurons without simultaneously activating peripheral feeding-related receptors.

Two methodological points deserve emphasis:

AI as a hypothesis generator. The neural network did not "design" BRP. It identified BRP's presence inside an existing human protein as a likely overlooked endogenous signalling molecule. This is a different paradigm from de novo design tools such as RFdiffusion.
Reverse pharmacology. Because the peptide came first and the receptor is still being characterised, BRP represents a classical reverse-pharmacology problem of the kind that originally produced opioid and cannabinoid receptor research.

How Is BRP Different From GLP-1?

The contrast with the GLP-1 class is the headline of the story. GLP-1 agonists such as semaglutide and tirzepatide work largely through the gut-brain axis: they slow gastric emptying, modulate vagal afferents and act on hypothalamic neurons indirectly. The well-known side effects (nausea, vomiting, constipation, occasional gastroparesis) are a direct consequence of that peripheral activity.

BRP, in the published animal data, appears to reduce appetite primarily through a central hypothalamic route, with very little gastric or intestinal action. The Svensson group reports activation of neurons distinct from the GLP-1-responsive POMC population, and they observe no measurable delay in gastric emptying in treated animals. The practical translation: in mice and minipigs there is no nausea phenotype, no vomiting, and no obvious GI stress.

This does not yet mean that BRP is "GLP-1 without side effects" in humans. Animal models, especially of nausea, are notoriously poor predictors. But the mechanism is genuinely different, which is the more important fact for researchers comparing these molecules. For broader context on how the GLP-1 class itself works, see our overview of GLP-1 agonists and the science of semaglutide in 2026.

Preclinical Results in Detail

The Nature paper reports a coherent set of in vivo results across two species:

Mice

A single subcutaneous injection of BRP reduced food intake by approximately 50% over the first hour after dosing in lean and diet-induced-obese mice.
Repeated dosing over 14 days produced sustained weight reduction in obese mice.
Body composition analysis showed a preferential loss of fat mass rather than lean mass, although the timeframes are short and warrant longer follow-up.
No nausea-like behaviour (pica, conditioned taste aversion) was reported in standard rodent assays.

Yucatán Minipigs

Minipigs are a translational model closer to human metabolism than rodents.
Acute dosing produced a reproducible reduction in food intake without behavioural signs of malaise.
Glucose tolerance was modestly improved, although the effect appears smaller than that seen with GLP-1 agonists.

Mechanism Hints

BRP increased c-Fos expression in hypothalamic regions associated with satiety, including parts of the arcuate and paraventricular nuclei.
Activity in the area postrema, the brainstem region implicated in GLP-1-induced nausea, was conspicuously absent.

The combined picture is consistent with a centrally acting satiety peptide that does not engage the brainstem nausea circuitry.

What's Next: Phase 1 Human Trials

Alongside the publication, Stanford and a spin-out vehicle announced plans to advance BRP into a Phase 1 human trial. The publicly available details are deliberately limited:

A first-in-human study is planned, focused initially on safety, tolerability and pharmacokinetics in healthy volunteers and people with obesity.
Dose-escalation will be subcutaneous, in line with the animal protocols.
No formal IND timeline, sponsor name or trial registration number was disclosed in either Nature or StatNews at the time of the announcement.
Specific endpoints, such as defined weight-loss thresholds or duration of dosing, have not been published.

Researchers should treat any specific number circulating online with caution. As of the writing of this article, the credible public information stops at "Phase 1 announced, design pending". Follow-on Phase 2 efficacy data, of the kind that defined semaglutide and tirzepatide, will be at minimum 18 to 24 months away under any reasonable timeline.

Why This Matters for the Research Peptide Community

For laboratories, suppliers and informed researchers, BRP carries three signals worth absorbing.

1. AI-driven peptide discovery is now producing real candidates

The Peptide Predictor approach is not the first attempt at machine-learning-guided peptide discovery, but it is one of the first to produce an in vivo-validated hit in a high-profile metabolic indication. Expect a wave of similar screens across the human secretome and prohormone landscape over the next two years. Many will fail. A meaningful number will not.

2. "Endogenous" is a marketing word and a scientific one

The framing of BRP as a "natural Ozempic" is journalistic shorthand. In the research context, "endogenous" matters because the body already has machinery to clear and regulate the peptide, which often translates into a different safety profile from xenobiotic analogues. It does not, by itself, mean safer or better.

3. Hype cycles around novel peptides will accelerate

The trajectory of BPC-157 and the recent mainstream attention to peptides documented in our March 2026 media review suggest that BRP could become a major buzz item before any human efficacy data exists. Researchers should resist that compression of timelines and continue to evaluate candidates on data, not narrative.

Conclusion and Key Takeaways

BRP is a serious scientific finding, not just a press cycle. The combination of an AI-guided discovery method, a plausible new mechanism distinct from the GLP-1 axis, and clean preclinical tolerability in two species makes it the most credible novel obesity peptide in years.

Key points:

BRP is a 12-amino-acid endogenous human peptide derived from BRINP2.
It was identified by the Peptide Predictor neural network at the Svensson lab, Stanford.
In mice and minipigs it suppresses appetite via the hypothalamus, with no GI side effects observed.
Phase 1 human trials have been announced, but design details are sparse.
Mechanistically distinct from GLP-1, GIP and amylin pathways.

Open questions:

Which receptor mediates the central effect?
Will the absence of nausea translate to humans?
What is the durability of weight loss with chronic dosing?
How does BRP compare directly to GLP-1 agonists in humans?

For the research-peptide community, BRP is worth tracking carefully and worth keeping firmly in the "early but real" category until human data arrives. The next inflection point is the first read-out from the Phase 1 trial, which on current information is at least a year away.

Sources

Svensson KJ, et al. "AI-guided discovery of a hypothalamic appetite-regulating peptide encoded in the human prohormone BRINP2." Nature, 12 April 2026. https://www.nature.com/articles/s41586-026-10330-z
StatNews. "A new approach to GLP-1-style weight loss, discovered by AI." 16 April 2026. https://www.statnews.com/2026/04/16/glp1-weight-loss-new-approach/