Stanford Scientists Use AI to Discover a “Natural Ozempic” Free of the Usual Side Effects
Key Takeaways:
- Stanford Medicine researchers used a purpose-built AI algorithm to identify BRP, a 12-amino-acid peptide that reduced food intake in mice and minipigs.
- BRP appears to act specifically in the hypothalamus, and in animal testing did not produce the nausea, constipation or major muscle loss linked to semaglutide.
- The findings remain preclinical, with receptor identification and peptide durability still to be resolved before human trials can begin.
A naturally occurring molecule with semaglutide-like effects
Researchers at Stanford Medicine have identified a naturally occurring molecule that may suppress appetite and reduce body weight in a manner resembling semaglutide, the active ingredient in Ozempic. In animal studies, the molecule also appeared to sidestep several of the problems associated with the drug, including nausea, constipation and substantial muscle loss.
The molecule, known as BRP, works through a different but related metabolic pathway and activates a separate group of neurons in the brain. That distinction could make it a more precise tool for controlling appetite and body weight.
Why a hypothalamus-focused signal matters
The appeal of BRP lies in where it appears to act. Semaglutide’s receptor targets are distributed widely across the body, which helps explain the breadth of its effects – and the breadth of its unwanted ones.
“The receptors targeted by semaglutide are found in the brain but also in the gut, pancreas and other tissues,” said assistant professor of pathology Katrin Svensson, PhD. “That’s why Ozempic has widespread effects including slowing the movement of food through the digestive tract and lowering blood sugar levels. In contrast, BRP appears to act specifically in the hypothalamus, which controls appetite and metabolism.”
The hypothalamus is a small region deep within the brain that helps regulate hunger, body temperature, hormone activity and energy use. Because BRP appears to act mainly in this area, it may influence appetite without producing as many effects elsewhere in the body.
Svensson has co-founded a company that plans to begin clinical trials of the molecule in humans in the near future.
Svensson is the senior author of the research, which was published on 5 March in Nature. Senior research scientist Laetitia Coassolo, PhD, is the lead author of the study.
Hunting for hidden peptides with artificial intelligence
The discovery depended heavily on artificial intelligence, which allowed the researchers to search through proteins belonging to a group known as prohormones.
Prohormones are inactive precursor molecules. They do not perform their final biological function until enzymes cut them into smaller fragments called peptides. Some of these peptides then act as hormones, carrying signals that influence metabolism, appetite and other complex processes in the brain and throughout the body.
A single prohormone can be cut in several different ways, producing many possible peptides. Finding the biologically important ones is difficult, because genuine peptide hormones are relatively rare and can be buried among large numbers of ordinary fragments created during normal protein processing and breakdown.
Traditional laboratory methods can isolate and identify peptides, but the process can generate enormous quantities of data. Researchers may need to sort through hundreds of thousands of molecules to find the few that have meaningful effects.
Narrowing the search to prohormone convertase 1/3
The team concentrated on an enzyme called prohormone convertase 1/3. This enzyme cuts prohormones at specific amino acid sequences and has previously been linked to obesity in humans.
One of the peptides produced through this process is glucagon-like peptide 1, or GLP-1. GLP-1 helps regulate hunger and blood sugar, and semaglutide works by copying its effects in the body.
The researchers reasoned that the same enzyme might produce other peptides that influence energy balance and appetite. To find them, they turned to artificial intelligence.
Peptide Predictor: from 20,000 genes to 373 prohormones
Rather than manually extracting proteins and peptides from tissues and then using methods such as mass spectrometry to identify huge numbers of molecules, the researchers created a computer algorithm called Peptide Predictor.
The program searched all 20,000 human protein-coding genes for the types of sites where prohormone convertases typically cut proteins. The researchers then narrowed the search to genes that produce proteins secreted outside the cell – a common feature of hormones – and that contained at least four possible cleavage sites.
That process reduced the field to 373 prohormones, giving the team a far more manageable group to investigate.
“The algorithm was absolutely key to our findings,” Svensson said.
Peptide Predictor estimated that prohormone convertase 1/3 could produce 2,683 distinct peptides from those 373 proteins. Coassolo and Svensson then focused on the sequences that seemed most likely to affect the brain.
They selected 100 peptides, including GLP-1, and tested whether they could stimulate neuron-like cells grown in the laboratory.
A 12-amino-acid peptide with an outsized effect
As expected, GLP-1 strongly activated the neuronal cells, increasing their activity to three times the level seen in untreated control cells.
One much smaller peptide produced an even more dramatic response. Made from only 12 amino acids, it increased neuronal activity tenfold compared with controls.
The researchers named the peptide BRP after its parent prohormone, BPM/retinoic acid inducible neural specific 2, or BRINP2 (BRINP2-related-peptide).
Amino acids are the basic building blocks of proteins and peptides. A molecule containing only 12 of them is extremely small compared with most full-sized proteins, yet BRP produced the strongest response in the initial cell tests.
Food intake fell by up to 50 per cent
The researchers next tested BRP in lean mice and in minipigs, which more closely mirror human metabolism and eating patterns than mice do.
An intramuscular injection given before feeding reduced food intake during the following hour by as much as 50 per cent in both species.
The team also gave daily BRP injections to mice with obesity for 14 days. On average, the treated animals lost 3 grams, with nearly all of the reduction coming from body fat. Animals in the control group gained about 3 grams over the same period.
The treated animals also showed improved glucose and insulin tolerance. These measures reflect how effectively the body regulates blood sugar and responds to insulin, the hormone that helps move glucose from the bloodstream into cells.
Because incretin-based therapies are now a routine part of weight management for many people living with obesity, clinicians looking to keep pace with the underlying pharmacology and with structured patient support pathways increasingly turn to focused CPD such as the College of Contemporary Health’s GLP-1RA Complete Programme.
No clear signs of the usual side effects
Behavioural testing found no meaningful differences between treated and untreated animals in movement, water consumption, anxiety-like behaviour or faecal output.
The absence of changes in faecal output was especially notable, because semaglutide can slow digestion and cause constipation. The researchers also did not observe the nausea-related responses or major muscle loss associated with some existing weight loss treatments.
Additional measurements of brain activity and body function showed that BRP acts through metabolic and neuronal pathways that differ from those activated by GLP-1 or semaglutide.
Those findings suggest that BRP may reduce appetite through a more focused biological route, although the results remain limited to animals.
What still needs answering before human testing
The researchers are now working to identify the cell-surface receptors that attach to BRP. Receptors are molecular structures that receive signals from hormones, drugs and other chemical messengers. Determining which receptor BRP uses will help scientists understand exactly how the peptide changes appetite and metabolism.
The team also wants to map the full sequence of events that occurs after BRP binds to its target.
Another challenge is duration. Small peptides are often broken down quickly in the body, which can shorten their effects. The researchers are investigating ways to make BRP last longer so that, if it eventually works in people, it could be administered on a more practical schedule.
“The lack of effective drugs to treat obesity in humans has been a problem for decades,” Svensson said. “Nothing we’ve tested before has compared to semaglutide’s ability to decrease appetite and body weight. We are very eager to learn if it is safe and effective in humans.”
CCH insight
Discoveries like BRP sit at the frontier of a field that is already reshaping day-to-day practice. For healthcare professionals supporting people living with obesity, a working command of GLP-1 receptor agonist pharmacology, patient selection, side-effect management and long-term follow-up is now core knowledge rather than a specialism.
The College of Contemporary Health’s GLP-1RA Complete Programme brings this together across three courses and a capstone, leading to an Advanced Certificate, with seven CPD hours and seven downloadable clinical tools designed for use in practice.
Explore the GLP-1RA Complete Programme →
Source: Stanford Medicine
Read More
New AI Tool Pinpoints Genes and Drug Combinations That Restore Health in Diseased Cells
Key Takeaways:
- Harvard researchers have developed PDGrapher, an AI tool that identifies genes and drug combinations most likely to restore diseased cells to a healthy state.
- The model uses graph neural networks to map cellular relationships and predict effective single or combined drug targets, significantly reducing the need for exhaustive drug screening.
- PDGrapher demonstrated high accuracy and speed across multiple cancer datasets, offering promise for personalised medicine and drug discovery in complex diseases such as cancer, Parkinson’s, and Alzheimer’s.
Introduction: A new era in drug discovery
Researchers at Harvard Medical School (HMS) have unveiled a powerful artificial intelligence (AI) model capable of identifying therapies that can reverse disease at the cellular level. This innovation, published in Nature Biomedical Engineering on 9 September, could reshape how new drugs are discovered, designed, and personalised for patients.
Unlike traditional approaches that examine one protein target or drug candidate at a time, this new model — named PDGrapher and freely available to researchers — analyses multiple cellular drivers of disease. It identifies the genes most likely to return diseased cells to normal function and pinpoints the most promising single or combined drug targets to correct the underlying cellular dysfunction.
“Traditional drug discovery resembles tasting hundreds of prepared dishes to find one that happens to taste perfect,” explained senior study author Marinka Zitnik, Associate Professor of Biomedical Informatics at HMS’s Blavatnik Institute. “PDGrapher works like a master chef who understands what they want the dish to be and exactly how to combine ingredients to achieve the desired flavour.”
Limitations of traditional drug discovery
Historically, drug discovery has focused on activating or inhibiting a single protein target. This approach has produced successful therapies such as kinase inhibitors, which block proteins that drive cancer cell growth. However, Zitnik emphasised that such strategies often fall short for diseases driven by multiple interacting pathways and genes.
She noted that many recent therapeutic breakthroughs — including immune checkpoint inhibitors and CAR T-cell therapies — work by targeting broader disease processes rather than single molecules. PDGrapher aims to expand this concept by identifying drug targets that can reverse signs of disease, even when the precise molecular mechanisms are not yet fully understood.
How PDGrapher works: Mapping complex cellular networks
PDGrapher is based on a graph neural network — a form of AI that analyses not only individual data points but also the relationships and interactions between them. In biological research, this means mapping how genes, proteins, and signalling pathways influence one another inside a cell.
Instead of screening thousands of compounds blindly, PDGrapher simulates what would happen if certain genes or pathways were switched off, dialled down, or targeted with a drug. It then predicts whether these interventions would shift a diseased cell towards a healthier state.
“Instead of testing every possible recipe, PDGrapher asks: ‘Which mix of ingredients will turn this bland or overly salty dish into a perfectly balanced meal?’” said Zitnik.
Testing and validation: Proving its predictive power
To train the model, researchers fed PDGrapher a dataset of diseased cells both before and after treatment, allowing it to learn which gene changes led to recovery.
They then evaluated the tool using 19 independent datasets across 11 cancer types, combining both genetic and drug-based experiments. PDGrapher was asked to propose treatment options for samples and cancer types it had never seen before.
The model accurately predicted known drug targets that had been deliberately excluded during training and identified new candidates supported by emerging evidence. Notably, it highlighted KDR (VEGFR2) as a target for non-small cell lung cancer, consistent with clinical findings, and identified TOP2A, an enzyme already targeted by chemotherapy, as a promising target for preventing metastasis in certain tumours.
PDGrapher consistently outperformed comparable AI models — ranking correct therapeutic targets up to 35 percent higher and producing results up to 25 times faster.
Implications for future drug discovery
By focusing on targets that directly reverse disease traits, PDGrapher streamlines the drug discovery process. This allows researchers to prioritise fewer, more promising interventions and to design experiments that are faster and more cost-effective.
This capability is particularly valuable for complex diseases such as cancer, where tumours often evade therapies that strike only one target. Because PDGrapher identifies multiple disease drivers, it offers a way to design combination treatments that could prevent drug resistance.
In the future, with further validation, PDGrapher could be applied to individual patients’ cellular profiles to create personalised treatment strategies.
Broader applications and ongoing research
Beyond cancer, the research team is using PDGrapher to investigate neurological conditions such as Parkinson’s disease and Alzheimer’s disease, aiming to identify genetic drivers that could restore neuronal health.
They are also collaborating with Massachusetts General Hospital’s Center for X-linked Dystonia-Parkinsonism (XDP) to map potential drug targets for this rare, inherited neurodegenerative disorder.
“Our ultimate goal is to create a clear road map of possible ways to reverse disease at the cellular level,” Zitnik stated.
Read More