One of obesity medicine’s stranger puzzles is that drugs activating a receptor called GIPR and drugs blocking it both help people lose weight. Cambridge scientists say they now know why: the receptor works through two different parts of the brain.

Researchers at the University of Cambridge’s Institute of Metabolic Science studied GIPR in genetically engineered mice, selectively removing it from specific brain regions. The work was published August 15, 2026 in Nature Metabolism.

The paradox that needed explaining

This is not an academic curiosity. Two major pharmaceutical programmes are built on opposite assumptions about the same receptor, and both appear to work.

Eli Lilly’s Mounjaro/Zepbound (tirzepatide) is a GIPR agonist combined with GLP-1, and it produces some of the largest weight losses achieved pharmacologically. Amgen’s MariTide pairs a GIPR antagonist with GLP-1 and is in Phase 3.

Normally, opposite pharmacology producing the same clinical effect indicates that one of the explanations is wrong. Here both approaches generated data, which meant something more interesting was happening.

The two circuits

The resolution is anatomical. Activating GIPR in the brainstem directly reduces appetite. Blocking GIPR in the hypothalamus releases a brake, letting the body’s fullness signals work more powerfully.

Two opposite drug actions, two brain circuits, the same result.

The elegance is that neither explanation was wrong — each was describing a different site. A drug given systemically reaches both regions, so which effect dominates depends on the balance between them, and the net outcome can be weight loss either way.

Why selective deletion was the right method

Removing the receptor from specific brain regions in genetically engineered mice is what makes this resolvable, and it explains why the question stayed open for years.

A drug given to a whole animal acts everywhere the receptor is expressed, so the observed effect is a sum of all sites. No dosing regimen separates them.

Deleting the receptor in one region while leaving it intact elsewhere isolates each contribution — converting a question about net pharmacology into a question about anatomy, which can be answered directly.

What GIP does elsewhere

Some context on why GIPR was contentious in the first place. GIP is an incretin hormone released from the gut after eating, and its classical role is stimulating insulin secretion — alongside GLP-1, the other incretin.

Its relationship with body weight was murky and appeared unfavourable. GIP promotes fat storage in adipose tissue, and evidence suggested blocking GIPR might protect against obesity. When tirzepatide showed that adding GIP agonism to GLP-1 produced greater weight loss than GLP-1 alone, it contradicted a substantial body of prior reasoning.

The brain circuits offer a resolution: GIP’s peripheral effects on fat storage and its central effects on appetite are different things, and the central effects appear to dominate clinically.

Why mapping circuits matters for the next generation

“Understanding which brain circuits respond to these medications…could help us design better drugs that produce more weight loss with fewer side effects,” said first author Jo Lewis.

That is the practical payoff. Current obesity drugs act broadly, and their side effects — nausea, vomiting, and the discontinuation those drive — come from actions at sites other than the ones producing benefit.

A drug engineered to reach one brain region and not another, or to act on one circuit preferentially, could in principle deliver the weight loss without the tolerability cost. Knowing which circuit does what is the precondition for attempting that.

The commercial implication

The finding also suggests the agonist-versus-antagonist question may be less binary than the competing programmes imply. If both approaches work through different circuits, the optimal drug might not be either extreme but something engineered to hit both favourably — or a combination.

It also means the two programmes are not straightforwardly comparable, and outcomes may differ in ways beyond total weight lost — in tolerability, in effects on body composition, or in how well the effect is maintained.

The caveats

This is mechanistic mouse research, not a new treatment. Mouse and human appetite circuitry are broadly similar in architecture and differ in detail, and mouse feeding behaviour translates imperfectly to human eating.

Why the brainstem keeps appearing in obesity research

The involvement of the brainstem rather than only the hypothalamus reflects a broader shift in how appetite control is understood.

The hypothalamus was long treated as the appetite centre, and it does integrate long-term signals about energy stores — leptin from fat tissue, insulin, nutrient availability. That framing shaped decades of obesity research.

The brainstem handles something different: immediate, meal-related signals arriving from the gut via the vagus nerve and the circulation, in a region positioned outside the blood-brain barrier so it can sample circulating hormones directly. It governs when a meal ends rather than how much fat the body carries.

That distinction turns out to matter therapeutically. Drugs producing large weight loss appear to act substantially on meal-termination circuits rather than on long-term energy sensing — which is also why nausea features so prominently among their side effects, since the same brainstem region mediates that.

Separating the circuits that reduce appetite from those that produce nausea is arguably the central engineering problem in the next generation of these drugs.

Genetic deletion also produces a permanent absence from development onward, which is not equivalent to a drug blocking a receptor in an adult — compensatory changes can occur during development that would not occur with acute pharmacological blockade. Mechanistic research; not medical advice.