GIP: the incretin people forgot, then remembered

GIP is glucose-dependent insulinotropic polypeptide, the hormone K cells in the upper small intestine release when a meal brings fat and carbohydrate. Its job at the pancreas is simple to say and easy to underestimate. If blood glucose is high, GIP tells the beta cell to release more insulin. If glucose is not high, it does not. GLP-1, from L cells a little further down the gut, does that same glucose-dependent job and adds two more: it slows the stomach, and it turns appetite down in the brainstem. Mounjaro is tirzepatide, one chain built to press both receptors. That second receptor is the whole argument.

In short. After you eat, the gut releases GIP and GLP-1. Both tell the pancreas to release insulin when sugar is high. GLP-1 also slows the stomach and quiets appetite. Mounjaro copies both.

The older name is gastric inhibitory polypeptide. John Brown's group isolated it as something that could quiet stomach acid. Dupré and colleagues then showed, in people, that it amplified insulin when glucose was given. For a while GIP looked like the incretin, the gut hormone that explains why oral glucose pulls more insulin than the same glucose in a vein. Then GLP-1 was identified, it was more obviously druggable, and the first medicines — exenatide, liraglutide, later semaglutide — were GLP-1 receptor agonists and nothing else. GIP looked weak in type 2 diabetes. Infusing it did not restore insulin the way GLP-1 did. The tidy conclusion was that the GIP receptor had gone deaf, so there was no point building a drug for it.

In short. GIP was noticed first, then sidelined, because in type 2 diabetes an infusion of GIP did not work as clearly as GLP-1. The medicines that followed only copied GLP-1.

The revival is a chemistry result, not a slogan. Unimolecular dual agonists — one chain that occupies GIPR and GLP-1R — did things in animals, and then in people, that GLP-1 alone had not done at the same tolerability. Coskun, Sloop and colleagues published tirzepatide, LY3298176, in Cell Metabolism in 2018. It is a 39-residue chain, biased toward the GIP receptor, still active at GLP-1, with a C20 fatty diacid so a fraction rides on albumin in the blood. Native GLP-1 lasts about two minutes, because DPP-4 clips it at the second residue. The engineered chain is built so that clip misses. Albumin binding is why a dose is discussed in weeks, not in minutes.

In short. Tirzepatide is one 39-letter chain that prefers the GIP receptor and still presses GLP-1. A fatty tail lets it stick to a blood protein, so it lasts about a week. Your own GLP-1 lasts about two minutes.

Both receptors are class B G-protein-coupled receptors. On the beta cell they work mainly through Gs and cyclic AMP, and the insulin they release is tied to glucose being high. That is the incretin effect, and it is why these molecules are not insulin. Insulin lowers glucose whether the meal was there or not. GIP and GLP-1 ask for insulin when the sugar is already up. GLP-1's extra jobs sit elsewhere. On the stomach it slows emptying, so the same meal arrives in the blood more slowly and you stay full longer. In the brainstem it reduces appetite, so the next meal is smaller. Nausea, when it happens, is that same slowing and that same fullness signal turned up. It is not a separate mysterious side effect.

In short. They ask the pancreas for insulin only when sugar is high. GLP-1 also slows the stomach and reduces hunger. Feeling sick is usually that same fullness, not a different drug effect.

GIP is not a copy of GLP-1 with a different badge. It is the receptor tirzepatide was biased toward. Adipose tissue also carries GIP receptors, and the papers argue about lipid handling there, not only about insulin. You do not need that argument to understand the trial results. You need it to understand why a dual agonist is not 'GLP-1, but more'. Semaglutide occupies GLP-1 only. Tirzepatide occupies both. Retatrutide, LY3437943, occupies both and adds the glucagon receptor, which is a liver-and-energy-expenditure question rather than a third copy of fullness.

In short. Semaglutide is the fullness hormone only. Mounjaro is fullness plus GIP. Retatrutide adds a third, the one that tells the liver to spend fuel.

SURPASS-2, Frias and colleagues, New England Journal of Medicine, 2021, put tirzepatide 5, 10 and 15 mg against semaglutide 1 mg in people with type 2 diabetes. The dual agonist lowered HbA1c further at the higher doses. That is a blood-sugar trial, in people who already had diabetes. SURMOUNT-1, Jastreboff, Aronne, Ahmad and colleagues, New England Journal of Medicine, 2022, is the weight trial in people with obesity and without diabetes. At 72 weeks the mean weight change was about 15.0, 19.5 and 20.9 percent at 5, 10 and 15 mg, against about 3.1 percent on placebo. STEP-1, the semaglutide 2.4 mg weight trial, had landed near 15 percent. Different trials, different people, not a head-to-head at the weight doses. The honest sentence is that adding the GIP receptor, in the trial that was actually run, moved the average further than GLP-1 alone had in its own trial, and that gastrointestinal events remained the class effect.

In short. In the diabetes trial, tirzepatide beat semaglutide 1 mg on blood sugar. In the weight trial, 15 mg for 72 weeks meant about 21 percent average loss, against about 3 percent on placebo. Those are trial averages, not a promise to one person.

Stopping the drug hands the appetite signal back. The trials that followed people off treatment watched weight return toward where it started, which is what you would expect if the effect was occupancy of receptors rather than a rewritten set-point. That is physiology, and it is also why a vial is not a plan. The pen in those papers is a medicine with a dose, a titration and a licence. This page is the mechanism. The Mounjaro listing on this site is 30 mg of that chain, lyophilised, made in America, HPLC-MS on identity, for research use only and not for human consumption.

In short. When the drug stops, appetite comes back, because the effect was the receptors being occupied. The papers used a licensed pen. The listing here is the research vial of the same chain.