The GIP Receptor: Agonist or Antagonist?

I spent part of this week reading one sentence from a scientific review the way you would read an indemnification clause, slowly and with suspicion, and I have come to think it is the best sentence in obesity drug development. It is about the GIP receptor, the receptor for glucose-dependent insulinotropic polypeptide, which is one of the two gut hormones called incretins. Like its better-known sibling, the GLP-1 receptor, it sits on the insulin-making cells of the pancreas, and a structural biology paper on incretin receptors says plainly that “incretin receptor agonists stimulate the synthesis and secretion of insulin from pancreatic β-cells.” (It shows up elsewhere too. Rat pancreatic alpha cells, for instance, turn out to carry working receptors for both hormones.)

So should a drug turn that receptor on or turn it off? The industry’s answer is yes (which is also what a toddler says when you offer a choice between two desserts). Tirzepatide is a dual agonist: it switches on the GIP receptor and the GLP-1 receptor. Maridebart cafraglutide, also called MariTide or AMG 133, is a bispecific antibody that switches on the GLP-1 receptor and blocks the GIP receptor. Both are pitched as doing better than switching on GLP-1 alone, and the honest summary of why is that nobody has settled it.

The sentence, and what the GIP receptor fight is about

Here is the sentence, from a review titled “The Premise of the Paradox: Examining the Evidence That Motivated GIPR Agonist and Antagonist Drug Development Programs”:

“The paradox that both GIPR agonism (tirzepatide) and GIPR antagonism (maridebart cafraglutide) provide additive weight loss when combined with a GLP-1 receptor agonist is a central unresolved scientific question in incretin pharmacology.”

I mean, that is a whole field in one sentence. Let’s take it clause by clause, as if somebody might sue over it.

“Both GIPR agonism and GIPR antagonism”: who is on each side

An agonist activates a receptor. An antagonist blocks it. In ordinary pharmacology these are opposites, the way the gas pedal and the brake are opposites, and you would expect them to have opposite effects on whatever the receptor controls.

The sentence names one champion for each side. In the agonist corner is tirzepatide, “a 39-amino-acid synthetic peptide” (per a cardiology review) that mimics both hormones at once. In the antagonist corner is maridebart cafraglutide, which is not a peptide at all but an antibody built to grab two different receptors and do opposite things to them. (Agonist on one, antagonist on the other, in a single molecule, which is the kind of flexibility you normally only see in consultants.)

It is worth noticing that both sides actually built drugs. Nobody spends years and a great deal of money developing a molecule on a hunch. The review’s own title tells you there was “evidence” that “motivated” both programs. Each team looked at what was known about GIP and concluded, sensibly from where it stood, that its direction was the right one. The team pushing activation had a hormone whose whole job description is releasing insulin; the team pushing blockade presumably had its own reasons to think less GIP signaling would mean less fat. Both were reading real data.

They just disagreed about the sign.

What “additive” is doing in the GIP receptor paradox

“Additive” is a precise word. It does not say that GIP agonism causes weight loss on its own, or that GIP blockade does. It says that each adds something on top of a GLP-1 drug.

The numbers that make the agonist side look additive are fairly clean. One pharmacodynamics review puts GLP-1 receptor agonists at “approximately 6-15% weight loss in clinical trials” and dual GIP/GLP-1 agonists such as tirzepatide at approximately 15-20%. In the SURMOUNT-1 phase 3 trial, tirzepatide produced mean weight reductions of 19.5% at 10mg and 20.9% at 15mg. For the antagonist side, the facts in front of me describe maridebart cafraglutide as “in clinical development” and do not give a weight-loss figure, so I won’t make one up.

But look at how “additive” gets measured. It works like splitting a dinner bill by subtraction: you know roughly what the table would have spent without your friend, you see the final check, and whatever the difference is gets assigned to the friend (who, for all the math knows, only had water). That tells you the total changed when he showed up. It does not tell you what he ordered. A GIP component that pushes the receptor and a GIP component that blocks it can both leave the check bigger, and the subtraction method will cheerfully credit each of them.

“When combined with a GLP-1 receptor agonist” is the load-bearing clause

This is the clause that keeps the paradox from being simply a contradiction. Neither drug is a pure GIP drug. Both arrive attached to a GLP-1 agonist, and GLP-1 agonists already “reduce appetite, delay gastric emptying, and enhance glucose-dependent insulin secretion,” to borrow a pediatric obesity review’s list. Whatever GIP is doing, it is doing it in a room where GLP-1 is already doing most of the talking.

And the two GLP-1 halves are not the same thing. One is part of a 39-amino-acid peptide. The other is part of an antibody. They differ in size, in how long they last, in where in the body they can reach. So when you compare “GLP-1 plus GIP-on” against “GLP-1 plus GIP-off,” you are not really holding the GLP-1 constant. It’s fine as a description of two products. It is less fine as a controlled experiment.

I picture a lab meeting somewhere: “Okay, so it works when we turn it on, and it works when we turn it off.” “Great. Does it work when we leave it alone?” “That’s the GLP-1 arm.”

It is a bit like the Seinfeld episode where George Costanza decides to do the opposite of every instinct and his life improves, except here both the instinct and the opposite improved things, which is not how that episode went.

“Unresolved” means unresolved

The last clause is the most honest one. “A central unresolved scientific question” is a review telling you, in its own words, that the field does not know. I don’t know either, and the facts I have do not rank the competing explanations, so I’ll describe the shapes an answer could take rather than pretend one has won.

One shape: the two drugs are not really doing opposite things over time. A receptor that is pushed hard for weeks might respond differently than one nudged for a minute, and the line between “activating it a lot” and “blocking it” might be blurrier than the textbook vocabulary suggests (the relationship between these drugs and this receptor is, let’s say, complicated). Another shape: the GIP receptor does different jobs in different tissues, so turning it on in one place and off in another could both point toward weight loss. A third shape: GIP is sort of a passenger, and the gap comes mostly from the different GLP-1 machinery in each molecule.

There are also mechanism papers that complicate things in their own way. One proposes that dual and triple agonists of the GLP-1, GIP and glucagon receptors “promote thermogenic-like amino acid profiles” in brown fat, with “robust heat production.” That is a story about agonists burning energy. Even if it holds, it does not obviously explain why an antagonist would help too.

None of this has stopped anyone from building more drugs. Multi-agonists keep arriving; UBT251, a triple agonist of the GLP-1, GIP and glucagon receptors, showed weight loss of -13.6% to -19.7% versus -2.0% on placebo in a 205-person, 24-week phase 2 trial. Those programs all turn the GIP receptor on. The antibody turns it off. I guess the market will run the experiment the biology hasn’t finished.

Meanwhile the agonist side already has its branding sorted out. Tirzepatide’s “unique dual mechanism,” according to that cardiology review, earned it “the nickname ‘twincretin.’”